There are several potential relationships between algal biomass and total organic carbon. If algal biomass is high, than particulate carbon also must be high. The ratio of chl/C (mg/mg) ranges from 0.002-0.055 and varies as a function of growth rate (Cloern et al 1995). The highest growth rates correspond to the highest ratios, so eutrophic lakes are more likely to exhibit ratios around 0.03-0.05. If there are high concentrations of phytoplankton, dissolved organic C also could be high because of lysis and leakage by phytoplankton cells.

Alternatively, dissolved organic carbon can be high, and algal biomass low (e.g. Carpenter et al. 1998, del Georgio and Peters 1994). Tannic and humic compounds color the water, absorb light, chelate nutrients, and do not allow for algal growth. Such systems are termed dystrophic. A final possibility is the presence of non-colored dissolved organic C, which may stimulate algal growth by lowering UV inputs to lakes (Goes et al. 1995). This effect is not well documented.

Carpenter, S.R., J. J. Cole, and J. F. Kitchell and M. L. Pace. 1998. Impact of dissolved organic carbon, phosphorus, and grazing on phytoplankton biomass and production in experimental lakes. Limnology and Oceanography 43: 73-80.

Cloern, J. E. C. Grenz and L. Vidergar-Lucas. 1995. An empirical model of the phytoplankton chlorophyll: carbon ratio-the conversion factor between productivity and growth rate. Limnol. Oceanogr. 40(7):1313-1321.

del Giorgio, P. A. and R. H. Peters. 1994. Patterns in planktonic P:R ratios in lakes: influence of lake trophy and dissolved organic carbon. Limnol. Oceanogr. 39(4):772-187.

Goes, J. I. N. Handa S. Taguchi T. Hama and H. Saito. 1995. Impact of UV radiation on the production patterns and composition of dissolved free and combined amino acids in marine phytoplankton. J. Plankt. Res. 17(6):1337-1362.

Response:

I agree that there is little literature on the effect of diurnal oxygen fluxes on biota, especially in lakes. An extensive look through Google and Biological Abstracts yielded little information. Hence my answer is based on a compilation of expected results. Fortunately I ran across a document by the Alberta Department of Environmental Protection (Shaw, 1997) that had done some research on the subject ( http://environment.gov.ab.ca/info/library/7254.pdf). [ ^ ] In the paper they confirm that there is limited data on the subject.

I will let the inquirer look to that document for further information, especially on fish, and instead focus on potential avenues of research that may be somewhat fruitful in establishing the effects of supra-standard DO effects.

The impact of the oxygen depression on biota will depend on a number of factors: the nature of the aquatic habitat, the presence, type, and density of macrophytes or other physical structures, velocity of flow, species of the target organism, and the severity and length of the depression. One might expect the following effects of any oxygen depression.

Avoidance: In some instances, a mobile organism may be able to move away from the severest oxygen depression, while a sessile organism cannot.

Species replacement: Low oxygen will cause the replacement of organisms by low oxygen tolerant species. The degree of replacement would apparently depend on the degree of oxygen depression relative to the tolerance of the organism. Since tolerance varies considerably, one might expect that any DO depression may influence the species composition.

Location: In lake and reservoir situations, diurnal DO fluctuations can be both in the open water and in the littoral. Summer fish kills are known in the open water of shallow lakes, presumably because of high temperatures stimulating respiration in excess of photosynthesis. If the low DO extends across the entire lake, then species shifts are possible. Hypolimnetic benthic species replacement with increased anoxia is well-documented, but it refers to long-term increases in anoxia.

In the littoral dense stands of macrophytes that reach from bottom to surface should produce the greatest DO depression because of the amount of respiration and because they physically inhibit reaeration. Caraco and Cole (Year not known) found that the water chestnut (Trapa natans) had DO values below 2.5 mg/L while DO values in eelgrass (Vallisneria americana) never fell below 5 mg/L.

In streams diurnal oxygen depression also occurs, but I am not familiar with impact. Presumably it would be affected by the amount of organic inputs and the amount of organic matter in the stream or river as well as the amount of productivity and re-aeration. Again the question of impact may depend on the relative mobility of the organism. Adult fish may move to more oxygenated waters while eggs or larval fish may be subject to or drift into regions of lowered DO.

Connolly, et al. (date unknown) found lethal effects on mayflies and chironomids to be in the range of DO saturation of 15-48% for mayflies and <6.43% for chironomids. However sublethal effects for mayflies were found at 25 to 35% saturation. They also detected the suppression of emergence by insect taxa at low DO. These studies were however conducted at continuous low levels rather that a fluctuating levels, so still leave the question of diel fluctuations open. However, a statement by Shaw (1997) suggests that "available research seems to indicate that effects of diurnally fluctuating dissolved oxygen levels could be similar to constant levels equal to the minimum of the fluctuating dissolved oxygen levels."

A secondary effect of diel oxygen changes may be that the dislocation of low oxygen intolerant species may affect their susceptibility to predation. Studies have documented that some fish and zooplankton (Burks et al, 2002) utilize the littoral as a refuge from predation. If low oxygen forces them into the open water, they may be subject to increased predation. Fortunately, the low oxygen may occur at night, when predation by sight feeders may be lower. Lowell and Culp (1999) demonstrated that Baetis tricaudatus moved towards areas of greater velocity in artificial streams with lowered DO, which, according to the authors could expose them to increased predation.
Other secondary effects may include the effects of other materials that may be released by low DO. Metals in increase in a fashion inversely mirroring the decrease in DO. Release is related to redox, presumably related to periods of anoxia. However microenvironments near the sediments may be anoxic non detected by a mid-water sensor.
Temperature may also be a factor as it affects not only metabolism but also material release rates and solubility of gasses.

Some papers to consider.

Alabaster, J. S. and K. G. Robertson. 1961. The effect of diurnal changes in temperture, dissolved
oxygen and illumination on the behaviour of roach (Rutilus rutilus (L.)), bream (Abramis
brama (L.)) and perch (Perca fluviatilis (L.)). Anim. Behav. 9:187-92.

Bouck, G. R. and R. C. Ball. 1965. Influence of a diurnal oxygen pulse on fish serum proteins.
Trans. Am. Fish. Soc. 94:363-70.

Burks, R.L., D. M. Lodge, E. Jeppesen, and T. L. Lauridsen. 2002. Diel horizontal migration of zooplankton: costs and benefits of inhabiting the littoral Freshwater Biology (2002) 47, 343-365

Carlson, A. R., J. Blocher, and L. J. Herman. 1980. Growth and survival of channel catfish and
yellow perch exposed to lowered constant and diurnally fluctuating dissolved oxygen
concentrations. Prog. Fish-Cult. 42:73-78.

Caraco, N. F. and J. J. Cole. Contrasting Impacts of a Native And Alien Macrophyte On Dissolved Oxygen In A Large River. Ecological Applications: Vol. 12, No. 5, pp. 1496-1509.

Connolly, N. M., M. R. Crossland and R. G. Pearson. Effect of low dissolved oxygen on survival, emergence, and drift of tropical stream macroinvertebrates. Journal of the North American Benthological Society: Vol. 23, No. 2, pp. 251-270.


Fisher, R. J. 1963. Influence of Oxygen Concentration and of Its Diurnal Fluctuations on the
Growth of Juvenile Coho Salmon. M.Sc. Thesis. Oregon State University . Corvallis ,
Oregon . 48p.

Lowell , R.B. and J.M. Culp. 1999. Cumulative effects of multiple effluent and low dissolved oxygen stressors on mayflies at cold temperatures. Can. J. Fish. Aquat. Sci./J. can. sci. halieut. aquat. 56(9): 1624-1630.

Nebeker, A. V. 1972: Effects of low oxygen concentrations on survival and emergence of aquatic in-sects. Transactions of the American Fisheries Society 4: 675-679.

Shaw, J. 1997. Alberta Water Quality Guideline For The Protection Of Freshwater Aquatic Life: Dissolved Oxygen. Pub. No.: T/391 Standards and Guidelines Branch
Environmental Assessment Division, Environmental Regulatory Service, Alberta Environmental Protection
 
Whitworth, W. R. 1968. Effects of diurnal fluctuations of dissolved oxygen on the growth of
brook trout. J. Fish. Res. Bd. Canada 25:579-84.

Review:

With regard to streams there is quite a bit known about long term depression in dissolved oxygen, including early work reviewed by Hynes in Hynes, H. B. N. 1960. The biology of polluted waters. Liverpool University Press, but less is known about the effects of variation.

Low dissolved oxygen has been implicated in fish kills several rivers and streams including problems in the lower Madison River in Montana, and problems in North Carolina causing fish kills in rivers in their tidal (estuarine) regions.
http://h2o.enr.state.nc.us/esb/Fishkill/2003KillReport.pdf [ ^ ]
Low oxygen impairment has also been implicated in water quality impairment in the Tualatin River in Oregon .
http://www.deq.state.or.us/WQ/TMDLs/WillametteBasin/Tualatin/TualatinAppxH.pdf [ ^ ]

Dodds, W. K. and E. Welch. 2000. Establishing nutrient criteria in streams. J. No. Am. Benthol. Soc. 19:186-196.
Provide little data on the effects of low dissolved oxygen excursions in streams, but note that problems with variation dissolved oxygen may impair biotic integrity.

Review:

The question appears directed at the effects of large fluctuations in DO that do NOT violate DO criteria - ie, the question is related to how large swings even in the well oxygenated range affect biota. As the other experts have noted, the literature is relative quiet on how the magnitude of diurnal fluctuations within safe DO concentrations affect freshwater organisms - rather it focuses on the effects of low oxygen concentrations on biota. Certainly if the DO falls below the minimum 1-day criterion at any point during a day, that is a problem. But whether a daily fluctuation of 4 degrees (say from 6 to 10 mg/L) versus that of 1.0 degree (7.5 to 8.5 mg/L) is more stressful for biota, there is very little to support any clear decision. Until more data are generated on the effects of diurnal swings within safe DO concentrations, this question will likely remain unanswered. That having been said, I imagine that large swings well within safe oxygen concentrations will not be found to be as stressful as even short excursions into low oxygen, hypoxic, or anoxic ranges.

In the interim, it is likely safe (and well supported by the literature) to assume that diurnal fluctuations that result in frequent concentrations near or below stressful concentrations (5.5 mg/L warmwater; 6.5 mg/L coldwater), will have significant sublethal effects on biota including but not limited to altered movement (drift and avoidance), reduced growth, and decreased emergence. If sustained over time, these will undoubtedly result in adverse community effects. The ambient water quality criteria for oxygen (cited below), addresses sustained concentrations at or near criteria. This document also has a section entitled "Effects of Diurnal Fluctuations" that supports everything written here.

The other obvious caveat is if high concentrations during these swings contribute to supersaturation that results in dangerously high total dissolved gas concentrations. This hyperoxia is also dangerous to biota, especially for fish.


USEPA. 1986. Ambient Water Quality Criteria for Dissolved Oxygen. EPA 440/5-86-003

Response:

The issues raised here are part of a larger subject of reconciling empirical predictions of chlorophyll a density and simulation model predictions of periphyton biomass. The larger subject is addressed in detail in a white paper submitted in response to the question regarding the Dodds empirical method & the QUAL2K model. The following summary is extracted from the linked white paper.( Periphyton Simplified Methods, 368KB, PDF )

Various simulation models such as QUAL2K predicts as a state variable benthic algal biomass as grams of ash-free dry weight (AFDW) per square meter. The bulk of the literature values on benthic algal biomass and targets - as well as most regression-type empirical models that predict benthic algal biomass - are reported as milligrams of chlorophyll a per square meter (mg/m2).

Unfortunately, the chlorophyll content of periphyton is highly variable, depending on species, number of heterotrophs present, and light conditions, rendering a conversion difficult, and many different values are reported in the literature.

We examined provisional EMAP data for California (see description in response to Question 44), which contains 173 data points benthic algal density results from sites throughout California in 2000-2002.The biomass data include both AFDW and chlorophyll a for periphyton, along with their ratio. The ratio of chlorophyll a density (mg/m2) to ash-free dry weight (g/m2) on 173 California EMAP samples ranged from 0.06 to 6.73, with an average of 1.673 and a median of 1.14. Most of the values lie between 0.6 and 2.1 (interquartile range). Numbers in this range are not representative of benthic algae. Typical stoichiometry of algae would be a carbon to AFDW ratio of about 0.45 and a carbon to chlorophyll a ratio typically in the range of 25 to 100 but possibly as high as 275 (Bowie et al., 1985). The typical range implies that the ratio of chlorophyll a density (mg/m2) to ash free dry weight (g/m2) for autotrophic benthic algae should be at least 4.5 and possibly as high as 18, while the highest reported value (for dinoflagellates) would lead to a ratio of 1.64. The low ratios at the EMAP sites suggest that the periphyton communities were likely dominated by heterotrophic fungi and bacteria whose growth is based on allochthonous carbon sources (e.g., terrestrial detritus or wastewater) rather than photosynthetic production. Heterotrophs usually dominate the shaded, fast-flowing, shallow first- to third-order streams in forests (Welch and Jacoby, 2004).

Simulation models such as QUAL2K directly predict the accumulation of phototrophic algal biomass only. However, it is the total periphyton biomass that leads to nuisance conditions and the impairment of uses. Further complications arise because (1) some algae exhibit mixotrophy, in which they are able to assimilate energy from fixed carbon compounds as well as by photosynthesis, and (2) exudates of benthic phototrophic algae may support bacterial and fungal heterotrophic populations, thus tying the heterotroph density indirectly to photosynthetic production.

For comparison of simulation model output to data and predictions for chlorophyll a, an appropriate ratio of chlorophyll a to AFDW must be selected. As noted above, a wide range of values are available, and the model can be fit with any reasonable value. Various authors (cited in Welch and Jacob, 2004) have noted that the autotrophic index (defined as the ratio of AFDW to chlorophyll a) is a useful indicator of the influence of organic wastewater on periphyton communities. Collins and Weber (1978) suggested that an autotrophic index of 400 defines the upper limit of clean water conditions, and use of this boundary has been verified by observations of Biggs (1989). Because the primary use of the model in connection with nutrient criteria is to predict periphyton and chlorophyll a density for "clean water" or supporting conditions, where appropriate nutrient limits can be determined, it is appropriate to use the autotrophic index value of 400 (which corresponds to a ratio of 2.5 mg/m2 chlorophyll a to 1 g/m2 AFDW) for comparison.

References

Biggs, B,J.F. 1989. Biomonitoring of organic pollution using periphyton, South Branch, Canterbury , New Zealand . New Zealond Journal of Marine and Freshwater Research, 23: 263-274.

Bowie, G.L., W.B. Mills, D.B. Porcella, C.L. Campbell, J.R. Pagenkopf, G.L. Rupp, K.M. Johnson, P.W.H. Chan, S.A. Gherini, and C.E. Chamberlin. 1985. Rates, Constants, and Kinetic Formulations in Surface Water Quality Modeling (Second Edition). EPA/600/3/-85/040. Environmental Research Laboratory, U.S. Environmental Protection Agency, Athens , GA.

Collins, G.B. and C.I. Weber. 1978. Phycoperiphyton (algae) as indicators of water quality. Trans. Am. Microscop. Soc., 97: 36-43.

Welch, E.B. and J.M. Jacoby. 2004. Pollutant Effects in Freshwater, Applied Limnology, 3rd Edition. Spon Press, London .

Review:

There are a few misconceptions in the initial background information that need to be commented on. The empirical models by Dodds (updated in Dodds, W.K., V.H. Smith, and K. Lohman 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 865-874) have very wide geographic distribution, as they were built across many ecoregions. They also analyzed for ecoregional bias and found little.

An additional point is that highly shaded streams are not expected to have excessive levels of chlorophyll, but have very high inputs of organic materials. Point sources of BOD will also have excessive organic material in the periphyton related to allochthonous sources (this is what the heterotrophic index is for). Therefore, if it is necessary to reconcile model predictions to AFDM data use an appropriate conversion and, if possible make certain to only choose AFDM data from areas without point sources of BOD or heavy shading.

Review:

These responses point to the difficulties of relying solely on chlorophyll and AFDM. Therefore, it is critical to use a weight-of-evidence approach for overall assessment, as proposed in the background for the question.

Response:

The best information on this question can be found in Standard Methods for the Examination of Water and Wastewater, APHA. It discusses potential interferences of the various methods. The EPA approved methods are the best to use. I will try to encapsulate below.

The first distinction is between total and dissolved nutrients. Total nutrients are just that. It includes all the particulate material and all dissolved organic or inorganic materials. These samples are not filtered or settled. Note, in all cases dissolved gasses (e.g., N2) are not included.

For dissolved nutrients, filtration is required. Depending on the method, different filters are used, but commonly a 0.7 - 1 um glass fiber filter is used because the filter can be analyzed seperately. If you know any 2, you can get the 3rd because:

Total nutrient = particulate nutrient + dissolved nutrient.

Nutrients can also be in inorganic or organic form. The organic forms and some others require digestion since analyses occur on the inorganic forms. Filtration of a sample followed by determination of total nutrients results in total dissolved nutrient

Total dissolved nutrient = organic nutrient + inorganic nutrient

Total N is commonly analyzed after digestion by Kjeldahl or persulfate digestion procedures.

Kjeldahl digests all organic N to ammonium. It does not include nitrate or nitrite. To get total N on an unfiltered sample then you need to add TKN to NO3 + NO2.

Persulfate digestion techniques digest everything to NO3, so can be used directly to indicate total N.

Total phosphorus is usually determined by persulfate digestion.

Dissolved inorganic N and P are analyzed seperately, usually on filtered samples, but sometimes when turbidity is low on unfiltered samples. In this case the results should be very similar since these assays only react with dissolved forms for the most part.

Nitrate assays often include nitrite but nitrite is usually a minor component, so they are not reported seperately, and often just as nitrate.

Ammonium assays are relatively straight forward but are susceptible to contamination problems (e.g. use of ammonium cleaners in the laboratory, smoke, a newly cut lawn nearby, small amounts of contamination in tubes).

Soluble reactive P, ortho-P, and dissolved reactive P are all the same thing. Results are not simply reported as phosphate because the assays react with inorganic or organic P forms in addition to phosphate. Exactly which forms do react to the assay is not well known.

There are many potential interferences, and each individual method should be referred to. Also precision and accuracy varies among assays, and is highly dependent upon the method used. In general it is best to choose a method with a limit of detection substantially lower than most of your samples (not always practical in low-nutrient waters). It is also advisable to purchase stock standards and run those through your assay to see how well your results match a verified quality control sample. An additional important step is to calculate recoveries in natural waters. This involves adding a known additional amount of nutrient to a natural water sample, and determining if the concentration in your amended sample - concentration in the unamended sample = the expected additional amount of nutrient added. This is a very good way to rule out interferences specific to the waters you are analyzing.

A final issue is interpreting the results in an ecologically sound fashion. There is a tendency to use the easiest measurements in the hopes that they will give ecologically relevant information. However, that is not always the most prudent approach. More on this issue can be found in:

Dodds, W. K. 2003. The misuse of inorganic N and soluble reactive P to indicate nutrient status of surface waters. Journal of the North American Benthological Society 22:171-181.

Review: 

Walter's responses is excellent - concise and complete and not overly technical. The APHA Standard Methods is the relevant technical guidance - especially for regulatory work. Readers interested in more insight than APHA offers may also be interested in the relevant sections from the following texts - the first a more recent standard in academic limnological methods, the second a classic. There are also emerging methods for some analyses - principally a fluorometric method for ammonium that has proven accurate and sensitive. I think it is mentioned as exploratory in the most recent APHA, but it may not yet be "Standard". I include its citation as well, for those interested, since ammonium can be tricky.

Wetzel, R.G. and G.E. Likens. 2000. Limnological Analyses. 3rd edition. Springer, New York .

Strickland, J.D.H. and T.R. Parsons. 1968. A practical handbook of seawater analysis. Queens'S Printer, Ottawa .

Holmes, R.M., A. Aminot, R. Kerouel, B.A. Hooker, and B.J. Peterson. 1999. A simple and precise method for measuring ammonium in marine and freshwater ecosystems. Canadian Journal of Fisheries and Aquatic Sciences 56: 1801-1808.

Review:

The responses so far are really great. I would add the following related to a part of the question: "I would appreciate any warnings about potential confounding factors that may alter measurements. For example, are there "non-target" forms of nutrients that may elevate a measurement?"

This is a particular problem for "total" nutrients, since the criteria are usually based on these. Phosphorus is especially subject to the interference problem of non-algal sediment, which, if high enough, can actually suppress algal growth. Both TP and TN are problems if there is considerable dissolved forms, suggesting non-limitation. These cases make setting nutrient criteria that are related to biological problems problematic. Please check the appropriate manual for discussions of these.

Response:

As noted in the question, relatively simple means of predicting the potential benthic algal biomass in response to nutrient loads are needed to aid in assessing potential nutrient impairment of streams and in developing preliminary management targets. Two candidate approaches are the empirical regression methods developed by Dodds et al. (1997, 2002) and the simplified simulation model approach of QUAL2K (Chapra and Pelletier, 2003).

The Dodds empirical approach is based on data collected throughout the world, and has been successfully applied to setting nutrient criteria to control benthic algal density in Clark Fork, MT. Periphyton response to nutrients is difficult to predict because of the many factors other than nutrients (including light availability, grazing, scour) that affect the observed periphyton density. In addition, measurements of ambient inorganic nutrient concentrations may not accurately reflect the bioavailable concentrations present at the substrate. As a result, Dodds et al. found that models provide a better fit to maximum potential benthic chlorophyll a than to mean chlorophyll a, and work better when based on total nitrogen (TN) and total phosphorus (TP) than on inorganic N and P. The updated equation for maximum benthic chlorophyll a given in Dodds et al. (2002) is:

log(max Chl a) = 0.714 + 0.372 log(TN) +0.223 log(TP), R2 = 0.31,

in which the logarithms are base-10. While the R2 value of this regression is relatively low, Biggs (2000) showed that the R2 value of relationships of this type could be approximately doubled by incorporating a factor to account for the frequency of scour events.

A problem with this type of empirical model, in addition to its inherent uncertainty, is that it does not provide any adjustment for changes in factors such as light availability, water depth, water temperature, or grazing pressure; nor does it account for scour effects unless the modification proposed by Biggs is used. This limits the usefulness of the approach for setting nutrient targets on specific stream segments. For instance, a well-shaded stream is likely to be able to tolerate much higher nutrient concentrations without producing nuisance levels of benthic algal growth than is a stream fully exposed to the sun.

Process-based simulation modeling approaches are attractive because they can explicitly incorporate the effects of factors such as shading and grazing. Successful applications of simulation models to predict time series of benthic algal densities are, however, rare. Most such models also have large data requirements and complex analytical formulations. They are thus of limited use for initial scoping applications with limited data.

QUAL2K provides a simplified process-based approach to the simulation of benthic algal density in response to light, nutrients, and stream characteristics. In theory, running this model to steady state ought to provide an estimate of maximum benthic algal density consistent with that of Dodds et al. (Translation must be made, however, been the biomass units of ash free dry weight generated by QUAL2K and the chlorophyll a density predictions of Dodds' model).
We undertook a comparison between the Dodds and QUAL2K approaches. Using the default kinetic parameter values recommended in the QUAL2K model results in predictions of steady-state periphyton density that are considerably higher than those predicted by the Dodds model and not linearly related to those predictions (see attached white paper for details).

With some adjustment of the default parameters (specifically increasing the inorganic nutrient half-saturation constants and the periphyton natural death rate), it is possible to bring the QUAL2K predictions into general alignment with the Dodds model. The reconciliation works, however, only up to periphyton densities of about 175 mg/m2 chlorophyll a. Beyond that point, the relationship becomes nonlinear, and the Dodds equation predicts considerably higher densities than are obtained from QUAL2K.

The relationship between the Dodds maximum density predictions and QUAL2K steady-state predictions can be made linear with the addition of an assumption that nutrient availability to periphyton is not constant, but instead decreases with increasing benthic biomass. This effect seems reasonable, because, as mat thickness increases, the availability of nutrients to cells within the periphyton mat becomes diffusion-limited. In addition, dense algal growths may deplete the inorganic nutrient fraction near the sediment-water interface. Once the relationship is linearized, the parameters can be readjusted to obtain agreement between the Dodds and QUAL2K predictions.

In sum, steady-state predictions from QUAL2K with default parameters do not agree well with the Dodds equation for maximum benthic biomass. Adjustments to the default parameters can render QUAL2K and the Dodds equation consistent, but only up to a certain periphyton density. Obtaining agreement at higher densities appears to require modification to the QUAL2K equations through the assumption of decreased availability with increasing algal density.

For further details, the attached white paper provides an analysis of the Dodds empirical method and QUAL2K simulation approach applicability to predicting maximum potential periphytic algal density in streams, focusing on California applications. This white paper first examines the empirical prediction methods developed by Dodds and their applicability to California streams. Secondly, it examines the simplified parametric approach to periphyton simulation contained in the QUAL2K model. Disagreements between the two approaches and a potential reconciliation are then evaluated. Both types of approaches appear to be useful for establishing initial, scoping-level target ranges for nutrients in streams to meet periphyton density targets, and the QUAL2K model can be adjusted to provide estimates of maximum potential periphytic algae density consistent with the Dodds approach. However, the complex nature of interaction of nutrient levels with other habitat factors (light, flow regime, etc.) indicates that more detailed, site-specific studies would likely be needed to set precise nutrient loading targets for individual streams.

References

Biggs, B.J.F. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll relationships for benthic algae. Journal of the North American Benthological Society, 19(1): 17-31.

Chapra, S. and G. Pelletier. 2003. QUAL2K: A Modeling Framework for Simulation River and Stream Water Quality: Documentation and Users Manual. Civil and Environmental Engineering Dept., Tufts University , Medford , MA .

Dodds, W.K., V.H. Smith, and B. Zander. 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: A case study of the Clark Fork River . Water Research, 31(7): 1738-1750.

Dodds, W.K., V.H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences, 59: 865-874.

Review:

I think a fundamental distinction that is missing in the initial answer to the question is that the Dodds and Biggs approaches are empirical. The Dodds et. al data are based on seasonal means of nutrients and of mean chlorophyll (or a single maximum chlorophyll value) across many temperate streams and rivers. The data of Biggs are based upon a variety of streams found in New Zealand . New Zealand streams are very variable in characteristics across short distances. Thus, both are empirical relationships based upon real data across a wide variety of stream types. If a model cannot reproduce the empirical results, at least approximately, the question needs to be asked what leads to the discrepancy? The empirical relationships cannot be changed, so the model must or be shown to apply to systems that differ in some fundamental way from those used to establish the empirical relationships.

If the model consistently gives higher chlorophyll yields per unit nutrient that the empirical relationships, then factors that may lower chlorophyll are probably not correctly accounted for (flood scour, grazing would be the primary candidates). The situations of highly turbid rivers or those with tight canopy cover are probably not as important because most investigators that reported benthic chlorophyll and nutrient levels used to create the empirical relationships probably only sampled reaches where chlorophyll was expected to be important.

The ultimate test of the utility of both the Qual2k model and the empirical relationships would be to apply both to several streams where seasonal data for both chlorophyll and total nutrients are available, but had not been used to develop either approach. If the model matches the stream data better than the empirical relationship, then the streams used in the empirical relationship are somehow different.

Note that there were some errors in the Dodds et al. equations (minor) and corrected versions for TN and TP alone (but not both together) have recently been published in Dodds. 2006. Limnol. Oceanogr. 51:671-680. A soon to be published errata in Canadian Journal of Fisheries and Aquatic Sciences will provide the equations with both TN and TP.

Review:

I think the basic issue here is whether California adopts the Dodds approach, with 30% R2, or applies QUAL 2K for site-specific criteria at considerable time and expense per application. What are the costs and benefits in increasing the precision of predictions? The cost of the simulation modeling includes data collection, and calibrating and applying the model, but more importantly, also includes an opportunity cost of increased time before implementation of any nutrient reduction scheme, as well as money spent on modeling that could have been spent on nutrient reduction.

The benefit of the modeling approach is increased precision of predicted nutrient criteria. But what do we know about the increased accuracy of models? We can infer that the Dodds empirical approach has an R2 of 30% or greater in predicting actual observed algal biomass, in many rivers. What is the comparable accuracy of QUAL2K models over many places and times, and what is the aggregate (or per-site) cost of those models?

A basic rule in information collection is to spend effort to reduce the greatest sources of uncertainty, because the largest uncertainty drives the uncertainty of the whole process. Accordingly, I believe the greatest single source of uncertainty is in nutrient reduction and management. If the model predicts, with 90% confidence, that you need to reduce nutrients by 45%, do you also have 90% confidence that you can achieve the target? I would guess not.

If I am correct that the greatest uncertainty is in nutrient reduction, then CA should be spending its effort on increasing the reliability and efficacy of BMPs so that one can be confident, for example, of hitting a target somewhere between 25% and 75% nutrient reduction, with 80% confidence.

A prominent example of the misapplication of simulation models is EPA's Chesapeake Bay Program, which has spent many millions on simulation modeling of nutrients and phytoplankton, to estimate targets for nutrient reduction. The models had many optimistic (and untested, possibly untestable) assumptions, with the result that they "reported" that the Bay program was on track to meeting nutrient reduction goals. In the meantime, actual monitoring data showed that nutrient concentrations were being held relatively constant, but that there were no substantial reductions in loading. The Bay program subsequently suffered a major embarrassment when they had to admit that no progress had been made towards nutrient reduction (Ernst 2003).

Criteria for a pollutant should be appropriate to our ability to manage the pollutant. Excessive concern over the precision of nutrient criteria when we have no clue if we can actually reduce the nutrients, seems a waste of resources.

Literature Cited:

Ernst, H.R. 2003. Chesapeake Bay Blues. Science, Politics, and the Struggle to Save the Bay. Rowman & Littlefield Publishers, Inc.

Response:

One could write an entire book chapter to answer this question (and many years ago, I did to some extent). In brief, one needs to define the desired "population" that reflects the particular issue of concern (nutrient criteria development/assessment). For example, for a lake, this means that we must decide what is of importance - the entire lake, the photic zone only, the summer growing season only,. plus have a priori understanding of space/time variability, and then concentrate sampling efforts in space/time where/when greatest variability exists (in the population of interest). See chapter 5 in Reckhow and Chapra (1983); this characterizes the role of variability, size, and cost, plus it provides equations and examples.

Reference:

Reckhow, K.H., and S.C. Chapra. 1983. Engineering Approaches for Lake Management , Volume I: Data Analysis and Empirical Modeling. Butterworths Pub. Boston . 340 p.

Review:

See also general textbooks on sampling design, e.g., Thompson, S.K. 1992. Sampling. John Wiley & Sons.

Review:

This is a large topic and Dr. Reckhow has written texts on it. His and other appropriate water quality sampling design texts are referenced in the lake nutrient criteria document (# 1 below). This and the stream nutrient criteria document (2 below) both contain chapters on data collection and study design and include a variety of references that will help guide an appropriate study design for characterizing chlorophyll a, secchi depth, turbidity, etc. in a stream or lake.

1: http://www.epa.gov/ostwater/criteria/nutrient/guidance/lakes/index.html [ ^ ] 2: http://www.epa.gov/ostwater/criteria/nutrient/guidance/rivers/index.html [ ^ ]

Response:

The first question needs to be answered with respect to what is being sampled for. In cases of sampling for stream eutrophication, the established relationships between nutrients and chlorophyll were based upon longer term monitoring of nutrients and chlorophyll, suggesting that baseflow is the appropriate condition to sample. The reason for this is that nutrient concentrations are elevated with stormflow, but the pulses are transient and are likely not available to algae. Furthermore, since algal problems are most commonly encountered during periods of baseflow, eutrophication control schemes are often based around baseflow. For an example see: Dodds, W. K., V. H. Smith, and B. Zander. 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: A case study of the Clark Fork River . Water Res. 31:1738-1750.

Accounting for the time since flood (stormflow) can improve predictive models of chlorophyll because storm flows reset algal biomass by scouring away accrued periphyton. A good example of this can be found in: Biggs, B. J. F. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll relationships for benthic algae. J. No. Am. Benthol. Soc. 19:17-31.

The technical guidance document http://www.epa.gov/ost/criteria/nutrient/guidance/rivers/index.html [ ^ ]
states that "If only a one-time sampling is possible, then sampling between two to four (2-4) weeks after a storm or high flow event has disturbed algal assemblages is recommended."
Stevenson, R. J. and L. L. Bahls. 1999. Periphyton protocols. In: Revision to Rapid Bioassessment Protocols for Use in Streams and Rivers: Periphyton, Benthic Macroinvertebrates, and Fish. Barbour, M. T., J. Gerritsen, and B. D. Snyder (eds.). U.S. Environmental Protection Agency , Washington , DC.

If monitoring is meant to encompass the total daily loading downstream, then it is important to include stormflow sampling. In this case, flow-weighted sampling is conducted. The concentrations are weighted to the relative amount of discharge and its associated concentration to obtain the average concentration in the water relative to the entire volume of water that has flowed over some period of time. Flow-weighting obviously also requires continuous monitoring of discharge. Flow-weighting is important for assessing downstream loading, because nutrient concentrations increase as well as discharge, so storms can lead to large amounts of nutrient transport downstream. There is a nice study comparing flow weighted sampling to routine grab sampling that can be accessed at: http://www.nwqmc.org/98proceedings/Papers/24-KAMM.html [ ^ ] Kammerer, P. A., Jr., H. S. Garn, P. W. Rasmussen, J. R. Ball. A Comparison of Water-Quality Sample Collection Methods Used by the U.S. Geological Survey and the Wisconsin Department of Natural Resources.
As for the second question, how to determine "normal" flow, this could be somewhat subjective because there is no clear definition of what a flood is (bankfull is often used as a definition) and a smaller increase in flow is often called a spate. Determining what constitutes baseflow is probably an issue of best professional judgment given the region of interest. If it always rains (e.g. Pacific Northwest in the winter), then winter baseflow may be determined as the condition occurring when no heavy rains have occurred in the last few days. In drier areas, any rain may be used as the criteria for not being at baseflow. In regions influenced by mountain runoff or colder climates, the period of snowmelt may need to be avoided if baseflow sampling is required.

In streams that are normally clear flowing (lower turbidity) baseflow may be defined as occurring when there are no recent storms and there is no substantial elevated turbidity in the water. If hydrographs are available for a site, the typical time to drop down to a steady flow after a larger rainstorm can be determined more specifically. A good treatment of baseflow and hydrographs can be found in: Ward, A. D. and S. W. Trimble. 2004. Environmental Hydrology. Lewis Publishers. If you are familiar with a specific site, then assessing stormflow versus baseflow is fairly easy since you have seen the stream under conditions without substantial storms and steady flows.

Review:

Dr. Dodd's response is complete, technically accurate and well written. The citations offered will provide readers with additional sources for more detailed information on conducting nutrient monitoring at the appropriate flow condition for differing objectives.

Review:

Dr. Dodds has provided a very thorough response. Just as index periods are used to reduce natural variability associated with seasonal changes in periphyton communities, so too are recommendations made to sample at an adequate interval post-disturbance - namely, to reduce natural variability in periphyton associated with flood scours.

As he rightly states, if predictable "floods" are actually the normal hydrologic condition (e.g., certain seasons in the tropics), then sampling could occur consistent with normal flow patterns for that region (see Poff and Ward 1989 or Poff et al. 1997 for a discussion of flow predictability and its implications for stream ecology). However, if flooding is not a normal condition and is largely unpredictable, then adequate time for re-establishment is necessary. Ultimately, the concern is over the effect algal growth will have on uses and this is mostly a problem of excess growth - so it makes sense to sample when the growth is at its greatest potential. This would require allowing the flow to reset under a period of stable baseflow.

Flow separation (distinguising storm from baseflow in hydrographs) is commonly performed by hydrologists and there are a number of techniques for distinguishing stormflow from baseflow within a hydrograph (e.g. see Dunne and Leopold 1978 or Gordon et al. 1992). However, since 2-4 weeks post flood is recommended for reestablishing the periphyton community, clearly this should be sufficient for "normal" baseflow to have reestablished itself.

For the situation where samples just "have to be taken" during an index period in a particularly stormy year, I would recommend, as Dr. Dodds has, taking the samples but recording the time since the storm occurred and using this judiciously to intrepret data in the analysis.

Dunne, T. and L.B. Leopold. 1978. Water in Environmental Planning. W.H. Freeman, San Francisco .

Gordon , N.D. , T.A. McMahon, and B.L. Finlayson. 1992. Stream Hydrology. An Introduction for Ecologists. Wiley, Chichester , UK .

Poff,N.L. and J.V. Ward. 1989. Implications of streamflow variability and predictability for lotic community structure: a regional analysis of streamflow patterns. Canadian Journal of Fisheries and Aquatic Sciences 46: 1805-1818

Poff N.L. and others. 1997. The natural flow regime. BioScience 47:769-784.

Response :

This is a great graduate exam question because it has several ramifications.

1. Continuous point source loadings are typically from wastewater treatment plants or an industrial source. Often the concentrations of N and P from wastewater plant are quite high - 10-12 mg/L N, mostly as ammonium and dissolved organic nitrogen, and 4-7 mg/L P, mostly as SRP and some dissolved organic P. These are high concentrations, and available forms, of both nutrients. This point-source inflow tends to be warmer than ambient during the peak of winter (so it floats) and near ambient temperature during the rest of the year. Temperature is an important consideration because the point-source inflow will generally mix with the receiving water body thereby increasing nutrient concentrations in the photic zone. Point-source discharges to lakes are a near-constant source of nutrients. Our analysis of reservoir nutrients in Korea shows that reservoirs with point source inputs from wastewater inputs are diluted during the Asian monsoon (summer) and increase during the rest of the year when the point-source is continuous input and there is minimal surface runoff, so concentrations spike in response. Conversely, if point-source inputs are piped long distances and discharged to a reservoir the effluent cools to ground temperature (in many parts of the US this is about 10 C). A piped/cooled effluent will pop to the surface during winter because it will be warm and light relative to ambient, but during summer it will sink as an interflow or underflow and not influence the photic zone until fall mixing. Point-source discharges to streams and rivers are quantitatively more important during low flow when they constitute a greater proportion of total flow than during spates when this input is diluted by surface runoff.

2. Storm loadings are complicated by the ascending and descending limbs of the hydrograph. On the upswing their surface runoff dominates flow and in agricultural and urban areas there is a great deal of soil-bound P lost during this phase of the hydrograph. On the descending limb, nitrate from the soil profile increases in importance. Storm loads can be quantitatively important. There are examples showing 80-90+% of annual nutrient loading from a given stream is delivered by one or two events. Nutrients associated with particles will be less available and tend to sediment more readily than soluble forms of the nutrients. Particulates promote light limitation rather than nutrient limitation. Simply put, not all nutrients delivered by storm loads have an equal shot at being translated into algal biomass. This statement gets to the "equality" issue raised in the original question but does not quantify relative availability of nutrient inputs on the ascending and descending limb of the hydrograph. But, descending limb inputs are likely less often associated with particles and more likely to be available. Again, dilution and temperature are important. Storm inputs to lakes and reservoirs during summer often have minimal influence on the photic zone because the inflow is typically the result of a thunder storm and is much cooler than the average temperature of a summertime epilimnion. In the Midwest we see these inflows sink into the upper hypolimnion and form a "mud sandwich" below the photic zone. This phenomenon was driven home to me during 1993 and 1995 (the flood years) when I was collecting daily samples from a local lake. Massive rain events had virtually no influence on the epilimnion. I think this "plunging" inflow phenomenon is a major feature of reservoirs (they have one or two large inflows) and less of a consequence in glacial, kettle lakes where inflows are smaller. We did notice this phenomenon in Arctic Alaska, however. Inflows from watersheds with wetlands were warm and mixed with the epilimnion directly, and inflows from permafrost were cold and plunged to the hypolimnion. Storm flows in streams have little chance of promoting algal blooms in the short-term because of low residence time, scour and prevailing light limitation.

Review:

The response gives good detail in lakes and less on rivers, so a bit more qualification may be in order.

If one is interested in nutrient loading with respect to eutrophication in rivers and streams, then storm events elevate nutrient loading, but these events may be transient and the nutrients may be flushed through before the primary producers have a chance to get them. The same may be true for run-of river reservoirs where big flushes of nutrients flow right through. However, moderate floods or spates, or high runoff events, may not scour algae and deposit nutrients that can be taken up in subsequent time periods. The ability of algae to utilize transient pulses of phosphate through luxury consumption is particularly important. Lohman and Priscu nicely documented just such an effect in the Clark Fork River in Montana . (Lohman, K. and J. C. Priscu. 1992. Physiological indicators of nutrient deficiency in Cladophora (Chlorophyta) in the Clark Fork of the Columbia River, Montana . J. Phycol. 28:443-448.)

The considerations about floating and sinking inflow in lakes are not consequential in very shallow lakes or reservoirs that do not stratify.

A very good point is made about relative availability of nutrients in point-sources versus non-point storm flow runoff.

Review:

This answer speaks very well to the complexity of this issue. The answer to this question will (as indicated both by Jack's answer and Walter's comments) change with the waterbody type (e.g., lake, reservoir, stream or river), physical characteristics of the system (e.g., depth, size), season, and watershed characteristics as well as the characteristics (e.g., type, source, temperature) of the inputs. Determining "importance" therefore, will need consideration of all of these factors.

Response:

An answer to this question should include considerations of monitoring design. If a nutrient criterion is written and interpreted as covering all space and time in a waterbody, then a single sample that is greater than the criterion would constitute an exceedance. Since, in that case, the criterion is to be met everywhere and for all time in the waterbody, then the appropriate design is to sample at the time and place where/when an exceedance is most likely to be found (based on a priori knowledge).

If, however, the criterion is interpreted to allow a certain percentage (e.g., 10%) of violations of the criterion before an "exceedance" is declared, then another issue of interest is if the interpretation refers to the samples or a population. That is, are we assessing compliance based on 10% of the samples or 10% of the population? In either case, probability sampling (e.g., stratified random sampling) is the most defensible design. If the exceedance decision is based on 10% of the samples, then a simple ratio of the number of violations to the number of samples should be computed, with the common-sense notion that more samples results in a more confident decision on exceedances. If, the exceedance decision is based on a population assessment, then an hypothesis test and a significance level need to be selected associated with a probability of error in the inference from the sample to the population.

Review:

For general guidance on this isssue, I recommend EPA's Consolidated Assessment and Listing Methodology (the CALM document): www.epa.gov/owow/monitoring/calm.html

Chapter 4 has sections on so-called "conventional" pollutants that are applicable to nutrient criteria, and appendixes C and D explain statistical considerations and tests for determining chemical exceedances.

One issue is the appropriate null hypothesis: Is it worse to declare the waterbody in exceedance of criteria when it is not (Type I error), or to declare everything OK when it has actually exeeded the criteria (Type II error)? Appendix C of the CALM document shows two case examples illustrating this concept.

Review:

Drs Reckhow and Gerritsen have covered the salient points of this question thoroughly. The CALM guidance is an excellent resource for this issue.

Response:

This question raises an interesting aspect of classic Vollenweider/Dillon&Riger era modeling that were somewhat brushed aside by Jones/Bachmann and Canfield/Bachmann in work of around the same time period. Vollenweider drew on ideas of Edmondson to put forth steady state/mixed tank reactor models for lakes. Sakamoto gave us the Chl-P regression model and Dillon&Rigler interpreted it as a spring phosphorus value sets the stage for predicting summer Chl. This spring-to-summer couplet is a common theme in the Dillon&Rigler work of the mid-1970s. Jones&Bachmann worked up a summer-summer Chl-TP relation but the coefficients don't differ greatly from that of Dillon&Rigler because their work was drawn mostly from glacial lakes and the spring-summer P concentrations are similar during open water season. This information provides a historical perspective and background to the general topic of this question. Most management of temperate lakes is based on summer collections. Nurnberg suggests trophic state is best assigned on conditions during summer and the bulk of lake data comes from summer. Recent literature deals with conditions during summer or seasonal patterns but the "spring" emphasis, particularly for estimating chlorophyll, has taken second stage. Dynamic modeling of lake phosphorus has also contributed to this shift in viewpoint.

A key element raised by this question is that classic Vollenweider steady state models were based on the concept of a mixed tank reactor for simplicity and nutrient loads were linked to in-lake phosphorus concentrations. In glacial lakes in-lake phosphorus during spring seems like a good approximation of the average. The paragraph above points out that we've set that approach aside for estimating summer chlorophyll. But, the question remains - what happens to nutrient loading when the lake is stratified? The answer - most of it sinks below the epilimnion and has little direct influence on summer conditions in the short run. Large summertime inflows tend to be associated with thunderstorms and cool conditions this causes inflows to hit the warm epilimnion and sink into the meta- or hypolimnion of most temperate lakes (see Vincent et al. 1991, Hydrobiologia 226:51-63). The larger the inflow the more likely it will plunge and not influence the photic zone during summer. Working on reservoirs I've learned about this process the hard way. Early on comparisons were made between reservoirs and lakes (Jones and Bachmann, Canfield and Bachmann) and in each case reservoirs seemed to have lower in-lake P values for a given nutrient load than did glacial lakes. We speculated that this was because of reservoirs being constructed in valleys (drained by one or two large streams) that were still undergoing active erosion and much of the phosphorus load was associated with particles that rapidly settled (somewhat true but not the full story). Upon further study we realized that during summer stratification reservoir inflows plunge and simply aren't being measured in our summer samples of the epilimnion. In short, this tendency for cool summer inflows to plunge effectively "blunts" the potential influence of external loading during summer (see Knowlton and Jones, 1995. Arch. Hydrobiol. 135:321 and Jones and Knowlton 2005, Lake and Reservoir Management 21: 361 for several examples and a discussion). These summertime inflows form a mid-layer in the water column (we call it the mud sandwich) and particles settle over time. Any influence of these summer plunging inflows are seen during fall overturn and are just part of (in addition to) the internal load that is returned at fall overturn. Simply put, the timing of the external inflow critical when inflows occur prior to stratification they mix throughout the water column and increase P concentrations (there is a spike in P with these inflows, followed by sedimentation of particles but P levels are ramped-up relative to pre-load levels). In contrast glacial lakes are located in depressions and there is comparatively less erosion (fewer particles) and the streams are typically smaller. Plunging inflows may be less consequential in these systems.

How do we deal with this? See Jones and Knowlton (2005, LRM) for some text. Because of budget problems we typically sample lakes during summer, but keep close track on weather patterns and catalog the years according to when inflow occurs. Wet springs result in higher P levels in our reservoirs during the subsequent summer. We've learned that wet spring vs. dry spring has an influence on our reservoir P values and this adds greatly to temporal variation (see Knowlton-Jones LRM paper coming in the June issue).

Review:

If I may summarize Dr. Jones's response:

1) Use summertime TP because that is most important for predicting chlorophyll and estimating load capacity
2) Use a model calibrated for summer TP, but you will find its estimate is not much different from the spring models.

Is this correct?

Review:

The response and the review of the response are correct.

An additional reason that TP in the epilimnion may reflect spring TP is that algae may luxury consume P in the spring, and keep it near the surface until after stratification. Since luxury P consumption can allow up to 20 doublings, there may be a slight mismatch between summer chl yield and TP.

If it is necessary to use spring-determined loading values, then it could be possible to correct the loading prediction with correlation between spring and summer epilimnion values.

Also, if the bulk of the loading is derived from point source output, the summer loading may be warm enough to mix with the epilimnion.

Response:

A nice, statistically significant relationship between microalgal mat thickness and nutrients has not been documented in field surveys. This technique was developed and used in a survey of Kentucky and Michigan streams and did distinguish the great effects of grazers on diatom-dominated algal assemblages in Michigan versus Kentucky streams (Stevenson et al. 2006). Microalgal mat thickness was related to chlorophyll a concentration in this study. However, diatom-dominated periphyton was not related well to nutrients in these studies. We know from other studies that growth rates and peak biomass of diatom-dominated periphyton are related to nutrients in experimental settings (Bothwell 1989; Rier and Stevenson 2006), but most of the changes occur at low nutrient concentrations. Even though these low concentrations were observed in some streams of our survey, we did not see the expected relationship between diatom dominated periphyton and nutrients. In Michigan , this was certainly due to grazer control until high nutrient concentrations. In Kentucky , flood disturbances may have created enough variation in biomass that we did not observe a relationship. In addition, high microalgal biomasses can accrue over extended periods of time in low nutrient habitats if not disturbed by floods or grazing. Plus, some diatom mats in low nutrient streams produce thick mats with long stalks. So, there are some confounding factors that may prevent detection of microalgae-nutrient relationships.

The relationships investigated in Stevenson et al. (2006) were in riffles. Another habitat in which microalgal accumulations can be important is in runs and pools. Extensive cyanobacteria mats have been observed in some of these situations. Low gradient streams can develop extensive periphyton and plankton assemblages in slow moving waters.

Shade should reduce photosynthetic rates and peak algal biomasses, however, microalgal growth can be very high in low light conditions (Rier et al. 2006). Periphyton are abundant under bridges and in many shaded habitats. Reduced photosynthetic rates may reduce carbon production and resulting microbial oxygen demand in shaded habitats, but particulate and dissolved carbon loading from upstream may create problems in these systems. In addition, downstream effects and upstream-downstream linkages should be considered. Seldom is a stream shaded along its entire length. Reach direction is an important factor affecting light availability. Even though east-west oriented streams with riparian trees on the south side probably receive little light during the day, north-south oriented streams can receive high light during mid-day hours. Plus reaches do vary in direction along the length of the entire stream.

No known direct relationship has been developed between microalgal mat thickness and biological condition of invertebrate and fish communities. However, this has been hypothesized, where thick diatom mats would interfer with grazing, alter competitive hierarchies, and provide habitat for non-native metazoan and benthic macroinvertebrates.
 
Bothwell, M. L. 1989. Phosphorus-limited growth dynamics of lotic periphytic diatom communities: areal biomass and cellular growth rate responses. Canadian Journal of Fisheries and Aquatic Sciences 46:1293-1301.

Rier, S. T. and R. J. Stevenson. 2006. Response of periphytic algae to gradients in nitrogen and phosphorus in streamside mesocosms. Hydrobiologia 561:131-147.

Rier, S.T., R.J. Stevenson, G. LaLiberte. 2006. Photo-acclimation response of benthic stream algae across experimentally manipulated light gradients: A comparison of growth rates and net primary productivity. Journal of Phycology 42:560-567.

Stevenson, R.J., S.T. Rier, C.M. Riseng, R.E. Schultz, and M.J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:149-165.

Review:

Lack of relationship between benthic algal thickness and nutrients may be due to variations of a number of other factors, such as shading, habitat type, and grazing. We conducted a study recently in channelized streams where substrate and shading is consistent throughout stream segments. The preliminary results show that nutrients are correlated with benthic algal thickness and score. Our results also show that macroinvertebrate IBI score is highest when algal thickness is at intermediate level. An interesting hump-shaped relationship is found between benthic algal cover and macroinvertebrate IBI. Too little or too much algal mass will likely impact macroinvertebrate community.

Review:

I agree that mat thickness or percentage cover are not good indicators of algal biomass. While these characteristics are easy to quantify, there are too many contingencies to use these as surrogates to estimate trophic state. Benthic chlorophyll is probably the best indicator of autotrophic state.

One reason mats may be thick is that there is a substantial input of biochemical oxygen demand, a reduced inorganic compound, or water that is prone to calcification. Downstream of sewage treatment plants dense bacterial mats ("sewage fungus") often develop. Downstream of springs that are rich in iron, sulfide or other reduced compounds, dense growths of chemoautotrophic bacteria can accumulate. These two problems can form in shaded areas. There are times in areas with limestone in the catchment where algal photosynthesis leads to deposition of carbonates so a thick crusty layer can form in relatively pristine very high quality streams.

Another example of thick mats is very obvious in the recent Didymosphenia geminate expansion into fairly oligotrophic areas such as New Zealand and the Pacific North West
http://www.biosecurity.govt.nz/didymo. [ ^ ] These algae grow very dense mats (that may interfere with biotic integrity) but are not any more productive than in a stream that looks fairly clean.

Another issue the question brings to mind is the idea of trophic state. In shaded regions, nutrients may alter heterotrophic state. Therefore, there may be a reason to protect shaded low algae areas from nutrients as well as open canopy areas. See

Dodds, W. K. 2006. Eutrophication and trophic state in rivers and streams. Limnology and Oceanography. 51:671-680.

Probably the best approach is to use invertebrate communities as the biotic indicator of nutrient effects in shaded areas (establish relationships with existing data). Another approach is to assume that most shaded areas eventually broaden into areas where algal growth may be a problem, and nutrients are transported downstream. Therefore, criteria that project open areas are also relevant for shaded areas.

Response:

It is possible that P associates with organic matter to become less available. However, most organic P compounds are thought to be bioavailable. Fe and P associate with organic material and light can lead to the disassociation. It is not clear how available the material is. This is actually a very complex question because Fe chelation may lower P precipitation rates and make P more available. The chemistry is not simple given the variety of humic compounds that can occur, interactions with light, concentration, and microbial degradation.

Most of the P precipitation that is reported leading to non-available forms happens in high pH with calcium or magnesium, or at lower pH with iron. Adsorption should not alter long-term bioavailability, thus total P should be a good indicator of available P. One way to think about this is that internal loading of P from sediments in lakes is generally high once they become anoxic even with very high organic C contents. Thus, it is not likely that irreversible binding of P occurs with organic materials. A final note: because high organic material is known to chelate Fe, with high degrees of chelation, Fe limitation is a possibility.

In conclusion, there are few solid data to suggest that nutrient criteria should be modified for dystrophic waters, other than the fact that high light attenuation lowers algal productivity and nutrient limitation may be unimportant in this situation. The question of what happens downstream when organic material is processed by heterotrophic organisms is less clear, but there is no persuasive case to be made that nutrient criteria should be more or less restrictive based on organic content of streams or lakes.

For further reading, see: Wetzel. R. G. Limnology: Lake and River Ecosystems, third edition. 2001. San Diego , Academic Press.

Review:

Two reasons could be used to justify different P criteria for "tea-colored" waters versus others. First, there is some indication that a greater proportion of P in "tea-colored" waters versus others is not bioavailable. As Dr. Dodds' indicates, P can associate with organic matter. In addition, "tea-colored" waters from wetlands tend to have more suspended particulate P than clearer waters. Algal response to nutrients seems to occur at slightly higher P concentrations in "tea-colored" waters than clear, harder waters. This may explain the response of Cladophora to lower TP concentration in Kentucky streams than Michigan streams, because Michigan streams have higher dissolved organic carbon from wetlands than Kentucky streams (Stevenson et al. 2004). Diatom indicators as well as macroalgal response seems to reflect this (unpublished data). SRP:TP ratios were lower in Michigan than Kentucky streams (Stevenson et al. 2004). Both of these effects are modest, but could be taken into account for different nutrient criteria in these 2 naturally different classes or types of streams.

Second, P concentrations may be naturally higher in "tea-colored" waters than clear waters. This would also justify different nutrient criteria for streams in different landscape settings. In a predictive model of TP in Michigan streams, we found that wetlands throughout a watershed have a positive effect on TP concentration that we would naturally expect in streams. Interestingly, riparian wetlands have a negative effect on expected natural TP concentration. We've interpreted these relationships to indicate that throughout the watershed contribute dissolved organic P in the recalcitrant organic forms and riparian wetlands trap particulate P. Unfortunately, we do not have DOC measurements and characterizations of DOC in these watersheds to support or refute our interpretations of these patterns.

Downstream, the P bound to DOC could become available through the processes that Dr. Dodds describes. But instream, it is reasonable to suspect that streams "tea-colored" waters and clear waters could have different expected natural P conditions and responses to P.
 
Stevenson, R.J., S.T. Rier, C.M. Riseng, R.E. Schultz, and M.J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:149-165.

Review:

Both the current responses are excellent. Some organic compounds will have P bound within them. For example, P has been shown to bind strongly to humic acids, particularly when, as Dr. Dodds mentioned, ferric iron is present (e.g., De Haan et al. 1990). However, microbes and algae have enzymatic mechanisms to get at bound P under conditions when P is limiting (e.g., alkaline phosphatase) and assays of some of these enzymes have been used as indicators of P limitation and could be examined here as well. In addition, UV light can liberate P from bound forms, primarily by converting iron into its ferrous (soluble) form. But I agree with Dr. Dodds that organically bound forms are unlikely to be "unavailable" - there are few completely unavailable bound P forms and these are bound primarily to inorganic compounds.

In colored Florida lakes, we noticed higher current background total P concentrations than in clear lakes and we noticed higher concentrations associated with chlorophyll response - likely due to light limitation. Paleolimnological evidence, however, suggested lower historical (pre-settlement) concentrations of P for a subset of these same lakes, though, suggesting that this current higher P is not a natural condition.

An interesting note to mention here, is that the same situation will apply, if not even more, to N. N forms are also tightly bound in organic forms and to organic compounds and, essentially, unavailable. Again, bacterial degradation and UV light essentially liberates mineral N from bound humic forms, increasing mineral N forms in blackwater systems, especially ammonium (Bushaw et al. 1996). This could be a problem under N limitation.

Bushaw, K.L. et al. 1996. Photochemical release of biologically available nitrogen from dissolved organic matter. Nature 381:404-407

De Haan, H. et al. 1990. Abiotic transformations of iron and phosphate in humic water revealed by double isotope labeling and gel filtration. Limnology and Oceanography 35:491-497

Response:

There are several major adverse effects of nitrogen in freshwaters. The first is Methemoglobinemia -- it is a public health issue and EPA, WHO and others recommend limits of 10 mg/L nitrate-nitrogen in potable water. In babies the nitrates are reduced to nitrite and bind to hemoglobin resulting in oxygen depletion in the tissues. This condition is widely known, but rare. There only have been about 2000 cases in the US and Europe in the past 50 years. Nitrates have been a problem with hogs and chicken production in the midwest, but only at levels above 10 mg/L. Second, ammonia can be toxic to fish above pH 7 (where un-ionized ammonia dominates) and varies with temperature. EPA criterion for ammonia for freshwater life was updated in 1999 and should be reviewed if this is a concern. Third, nitrification occurs in fresh water wherein ammonium is oxidized to nitrite and then nitrate. It takes two moles of oxygen to oxidize one mole of ammonium. This oxidation process adds to the classic oxygen sag below wastewater discharges in streams and it adds the the rate of oxgen depletion in the hypolimnia of stratified lakes where ammonium is released from organic matter by decomposers. Nitrification also can contribute to low oxygen in lakes during overturn. Fourth, is eutrophication. In many temperate systems phosphorus is considered the limiting element. But studies over the past few decades have identified co-limitation with nitrogen is commonplace and there are many examples of nitrogen limitation. Nitrogen has a 'gas phase' and it is common for total nitrogen concentrations to decline in the epilimnia of stratified lakes. Nitrogen is taken up and released by phytoplankton but it is not as tightly cycled as phosphorus and some nitrogen is lost to the atmosphere as gas. This loss of nitrogen gas takes place in the shallows and at the sediment surface associated with the oxygenated/mixed surface layer. By late summer inorganic forms of nitrogen (ammonium, nitrite and nitrate) are not measureable in the surface waters of many lakes. These lakes show a decline in the ratio of TN:TP during summer stratification and by late summer it is possible that nitrogen and not phosphorus is the limiting element. There are several studies showing this seasonal importance of nitrogen. Nitrogen is typically replenished internally during fall turnover, causing the return to phosphorus (or light) limitation during winter to mid-summer. Stream studies have often shown that attached algae are stimulated by nitrogen and not phosphorus. Nitrogen is the limiting element in some regions of sub-tropical Asia , where high temperatures and agriculture (rice) favor the uptake of nitrogen or its loss by denitrophication. Another important feature of nitrogen in lake eutrophication is that algal chlorophyll values tend to be higher per unit of phosphourus when TN:TP ratios are high. Simply put, chlorophyll is greater at a given phosphorus value when there is lots of nitrogen in the system. This patten has been demonstrated in empirical chlorophyll-nutrient regressions but is also seen in some lake management efforts. The James River Arm of Table Rock Lake recently benefited from P-load reduction resulting from P-removal at the wastewater treatment plant in Springfield , MO. Phosphorus declined in the lake but chlorophyll declined by a smaller percentage than phosphorus. There was no apparent change in nitrogen because the plant did not address nitrogen removal. The current thought is that chlorophyll levels are higher per unit of P in the system because of the "seemingly" abundance of nitrogen.

An important paper addressing nitrogen limitation is that of--

Elser, J.J., E.R. Marzolf and C.R. Goldman. 1990. Phosphorus and nitrogen limitation of phytoplankton growth in the freshwaters of North America : A review and critique of experimental enrichments. Can. J. Fish. Aquat. Sci. 47: 1468-1477.

Review:

Nitrogen has direct health effects on humans and animals, can shift nutrient ratios, and stimulate algal growth, especially in systems that are N-limited or maginally N or P limited. There are some voices cautioning against P reduction alone because of (1) the excess N that would be transported downstream and (2) the possibility of greater algal reduction if both are removed. Also N transport downstream does not terminate in the next reservoir or stream. Potentially it will find its way to an estuary and marine environment, where concerns regarding N-limitation is even greater than in freshwater environments (eg. the Gulf of Mexico controversy).

Review:

1. Adverse effects of nitrogen on lakes and reservoirs;
Well answered

2. Potential for downstream transport of nitrogen from lakes and reservoirs (generally thought to be a sink/transformer of N) to other waterbodies;
This is somewhat true, but variable. Some lakes with low N:P could have cyanobacterial blooms and actually be a source of N (Dodds, W. K. 2002 Freshwater Ecology: Concepts and Environmental Applications. Academic Press).

3. Potential for adverse effects in the rivers/streams receiving N loadings from lakes/reservoirs; and 4. What those adverse effects might be in the receiving river/stream, if realized;

This has been reviewed by
Dodds, W.K. and E. B. Welch 2000. Establishing nutrient criteria in streams. Journal of the North American Benthological Society 19: 186-196.

Many of these effects or streams are what were in the original general answer. The original answer did not directly address biotic integrity. Nutrient pollution can lead to species shifts in streams and stimulate excess algal production, leading to DO violations in some cases. There is even the potential for stimulation of toxic cyanobacteria in streams by high nutrients

Aboal, M., M. A. Puig, and A. D. Asencio 2005. Production of microcystins in calcareous mediterranean streams: The Alharabe River, Segura River basin in south-east Spain . Journal of Applied Phycology 17: 231-243.

Bowling, L. C. and P. D. Baker. 1996. Major cyanobacterial bloom in the Barwon-Darling River , Australia , in 1991, and underlying limnological conditions. Mar. Freshwater Res. 47:643-657.

Sabater, S., E. Vilalta, A. Gaudes, h. Guasch, I. Muñoz, and A. Romaní 2003. Ecological implications of mass growth of benthic cyanobacteria in rivers. Aquatic Microbial Ecology 32: 175-184.

5. A list of the scientific literature that has addressed any and/or all of these points, some included above, and citations therein.

6. Information (with sources cited) on the effects of phosphorus controls for the protection of one waterbody (e.g., reducing algal blooms in a lake) on transport of nitrogen to the receiving waters (does reducing P effectively elevate N where waters are near saturation with respect to nutrients?).

Little is published on this to my knowledge. Many of the non-point control methods will reduce N and P both, but point source reduction of P could lead to more downstream N transport. The degree of N limitation is related to the residence time of N in biota.

Dodds, W.K., E. Martí, J. L. Tank, J. Pontius, S. K. Hamilton, N. B. Grimm, W. B. Bowden, W. H. McDowell, B. J. Peterson, H. M. Valett, J. R. Webster, and S. Gregory 2004. Carbon and nitrogen stoichiometry and nitrogen cycling rates in streams. Oecologia 140: 458-467.

Response:

I have some experience with this in smaller streams, and also talked to Andrew C. Ziegler United States Geological Survey, 4821 Quail Crest Place , Lawrence , Kansas 66049 about their experiences with extended deployments in the Kansas River .
In general it is best to use a locked cable attachment on the downstream side of the bridge. In smaller streams we prefer a cable attached to a larger tree near the bank and obscured by vegetation above water with the probe obscured by an overhanging bank in the water. The flexible connection offers the best protection from debris in all situations. A solid structure will be taken away more easily with significant flooding. The major equipment losses tend to come from floods, and also, if ice is involved major problems can occur. If the cable is very long, an ATV winch attached to a vehicle battery can be used to raise and lower the equipment on a cable.

As for mounting, the best would be something that could be locked and not seen well from the road. But it seems bridges vary so widely that a single solution is not very easy to recommend. If you could couple with USGS monitoring sites they may have more protected equipment you could piggy back onto, with permission of the USGS of course.

If you are interested in dissolved oxygen, there are some calibration issues to be aware of with the sondes. In general air calibration is not very accurate, and it is better to get the entire probe close to the temperature of measurement and use a bucket of air-saturated (bubbled for several hours) water to calibrate the probe. Even better if you can do on-site Winklers they provide better data. It is import that the entire body of the probe comes to temperature equilibrium before calibration. There is a new optical probe for DO. This seems good, but in my experience there is a signal deflection with the default logging setting because the cleaning causes a depressed dissolved oxygen signal. This can be taken care of with a 30 sec lag between cleaning and logging with some sacrifice of battery time. If you are just looking for low DO excursion events, the standard calibration may be good enough. If you want to calculate system metabolism then better calibration procedures are in order.

Another issue I have come up against is the sonde housing serves as a great place for fish and crayfish to hide in. They can lower dissolved oxygen with respiration. If you also have a turbidity probe this can be seen as a drastic increase in turbidity. You may want to modify the size of the slots in the cover with a finer mesh realizing that the mesh will foul more easily and more impede water flow the finer the openings.

Review:

Once again, Dr. Dodds has provided an excellent response. I would encourage the questioner to review the response related to the use of DO diel sags in nutrient criteria development (Q7A page on N-STEPS website under Indicators, question 1) as it may relate to this question.

That being said, the question appears to be the best way to deploy. My experience has been primarily using sondes for whole system metabolism measurements (diel curves) in urban and agricultural systems. I had the best luck locking the sonde using thin wire cable (very strong) attached to the sonde and to trees on the bank or onto a camouflaged post firmly/permanently buried in the bank. I also placed the sondes in camouflaged PVC pipe that had been perforated with holes to allow thorough exchange. These were very easy to retrieve and well protected. I never lost a sonde although I did have one thrown onto the bank and left high and dry during a flood. It was not permanently damaged. I suppose you could anchor the sonde end in the stream as well, but that decreases deployment ease and increases damage risk from having a fixed object in the stream (as Dr. Dodds said, any permanent object will attract detritus - some of it large and damaging). Of the hundreds of measures, having only one sonde leave the channel seems pretty good. I was careful to select very well mixed positions in the channel (undercut or outer banks are promising, because they indicate areas of concentrated channel flow) - geomorphology is a good guide here.

I am less concerned about the DO accuracy issues raised by the first respondent, as air calibration for this application would seem to be sufficient and can be checked by taking Winkler samples from river water that the calibrated sonde is then placed into and comparing measures. However, it is always wise to occasionally calibrate in air saturated water in the lab and to corroborate with Winklers. This could be done on a periodic basis to reinforce air calibrations. You could decide on a data quality objective beforehand and decide what accuracy is required. It would take a day of lab work to see what this would require in terms of calibration effort.

Sonde housings can also collect organic matter, which has more respiration associated with it. We seemed to have less of a problem with this using the perforated (small holes) PVC pipe housing we used. But make sure you get adequate exchange in the housing with the sonde. You can use a dye test to check this easily. We left the bottom of this housing open and covered with coarse mesh to keep fish and crayfish out.

Review:

With all the excellent suggestions above, and precautions, I can recommend 2 other techniques. Their applications depend upon how deep the river is and how frequently it's used. For shallow waters, Stephen Porter and the NAWQA crew used black plastic garbage bags to camouflage anything that they left on the bank. For both shallow and deep waters, I like the idea of a camouflaged, well perforated PVC pipe for deploying sondes. For deeper waters where security may be an issue, we've used subsurface bouys to mark sonde locations and to add retrieval. You can make submersible bouys out of anything that floats, just put an anchor on it that is substantially heavier than the water displaced by the bouy. Connect the 2 bouys with a cable that is well connected to the sonde. You can then easily find the hidden sonde with a grappling hook. We've used this idea in ponds and lakes for a variety of underwater devices and have only had problems where there's lots trolling by fishermen.

Response:

What kind of nutrient diffusing substrates should I use?
What are the strengths and weaknesses of the various options (e.g., clay flower pot saucers, cups with permeable tops)?

The question recognizes that there are two general approaches. Both are well described in the recent book METHODS IN STREAM ECOLOGY (2nd ed) 2006 F. Richard Hauer and Gary A. Lamberti (eds.) which has two excellent articles on this issue. Chapter 10 by Tank et al. describes the cups, and Chapter 32 by Pringle and Triska the clay pots. In general I prefer the cups because: 1) they are easier to construct, 2) you can run more replicates, 3) they are easier to sample (the glass frit can be removed into a test tube and the whole thing can be analyzed for chlorophyll), and 4) they are easy to deploy in small streams and it takes a fairly large flood to wash them away (the individual rails can be roped in to resist all but the largest floods).

The number of replicates is very important.

Francoeur, S.N. 2001. Meta-analysis of lotic nutrient amendment experiments: detecting and quantifying subtle responses. Journal of the North American Benthological Society 20: 358-368.
Suggests that 8-10 replicates per treatment are required to obtain a 95% chance of detecting a doubling of algal biomass. 36-40 clay pots would take up a tremendous amount of stream bottom. The one exception is when the algal taxa, not just biomass are being studied. The cup method is not likely to accrue enough cells per treatment to analyze and the rough frits may hold many cells and not allow for scraping.

It is also important to run enough replicates to get a full statistical design. For N and P you need Control, N, P and N+P. This is analyzed with a two-way ANOVA to allow detection of co-limitation. This is discussed more fully in
Tank, J.L. and W.K. Dodds 2003. Nutrient limitation of epilithic and epixylic biofilms in 10 North American streams. Freshwater Biology 48:1031-1049.

What kind of agar and what forms of N and P should I use in the mixture?

The least expensive agar can be used, although very impure types may be checked for nutrient leaching before use.

Pringle and Triska (cited above) suggest 0.5 M Na NO3 and 0.1M NaH2PO4. These concentrations should be adequate and are typical of this type of experiment. Concentrations in these ranges gave linear release rates with the cups for about 17 days (Tank et al. chapter cited above).
I also prefer a hot ethanol extraction technique as it more completely extracts chlorophyll (Cladophora can cause problems), and acetone gives some people headaches.

Dodds, W. K., B. J. F. Biggs and R. L. Lowe. 1999. Photosynthesis-irradiance patterns in benthic algae: Variations as a function of assemblage thickness and community structure. J. Phycol. 35:42-53.

Sartory, D. P. & Grobbelaar, J. E. 1984. Extraction of chlorophyll a from freshwater phytoplankton for spectrophotometric analysis. Hydrobiologia 114:177-87.

Welschmeyer, N. A.1995 Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and pheopigments. Limnol. Oceanogr. 39(8):1985-1992
gives a nice method that does not require acidification for phaeophytin correction, saving considerable time.

Review:

I agree with the first reviewer's recommendations, but suggest that you consider the following for ease of construction, sampling, and flow design characteristics.

I would not use flower pots in current. Weird flow paths create patchy colonization patterns around the clay pots. For that reason, a flat colonization surface on the top will reduce spatial variability and among replicates and enable more precise assessment of treatment effects.

My favorite NDS are made from snap cap vials and fine Nitex® (maybe about 40-50 um pore size) screen. Punch a whole in the snap caps of the vials, put the agar in the vial, put a spare of nitex over the vial, and hold it on by snapping the top back on. Sharpen a pipe or get a cork borer to put holes in snap cap vials. You want Nitex® screen or even nylon stockings with a pore size that is sufficiently large to allow nutrients and water flow through the screen, but small enough for algae to colonize the pores over the holes. Nitex® with a pore size of 40-50 um seems to work well. I tried this and had great success. Fritted glass filters should work as well.

Insert the vials through holes drilled in a board that is fastened to a brick. See the attached illustrations. Put the vials as flush as possible with the surface of the board and at the downstream end of the board to reduce patchy currents over the colonization surface. I even beveled the edge of the holes to get the snap cap vials flush with the surface of the board. I don't like my substrata up in the front eddy as flow separates at the upstream edge of the brick and creates an eddy. Putting the vials at the back allows for currents to rebound and flow more smoothly and consistently across the substrata than if they are at the upstream edge, and reduces settling in that upstream eddy; theoretically making growth a more important process in community development than immigration. I randomly arrange 4 vials through the boards on each brick, one for each C, N, P, and NP treatment. This can, of course, be modified with different experimental designs.

Review:

These answers are pretty thorough and I have little to add. I have used clay pots and cups with glass fiber filters instead of fritted glass. I am not a big fan of the glass fiber filter method. I have also seen people use really squat pots or the bottom plate that fits under a pot to collect the water rather than the pot itself to increase the amount of flat area exposed (in relation to the second reviewer's concerns). The glass fritted approach with vials appears to work well.

An important thing to consider is that ALL approaches are an abstraction of real limitation - which is governed by a myriad factors, including activity within the biofilm as well as in the water column - and, therefore, are biased. That being the case, either method, as long is it is consistently applied and replicates are similarly distributed in space and time (including flow, light, etc.) and analyzed comparably, should answer your question. I would aim for placement that will maximize the response - likely slower moving, fully lit reaches - but flow and light should be comparable and both data should be collected for each replicate set to help interpret results. I would encourage a randomized placement of treatment sets (ctrl, N, P, and NP) within specific habitat types (slower moving, well lit) that are randomly selected for each treatment set and plenty of replicates (The Francouer et al paper is an excellent resource on NDS studies). Don't forget water column chemical sampling during the experiment either.

Good luck. I am not sure I have seen this method applied for a TMDL. I am interested in how well it will work to identify relative limitation under enriched conditions. Relative limitation being key here - too much algae + more algae is still too much algae. So if you are getting nuisance algal amounts on your controls - it would argue that both nutrients are probably provided in excess, even though one may be, relatively, limiting.

Response:

The compiled response indicates one important part of the answer. It is true that P uptake will lower P transport by rivers and streams to downstream areas. The problem with the answer is that it is based on P uptake, not long-term transport. This is a much more difficult question.

Take the following situation as an example. A sewage treatment plant discharges P into a river and it is assimilated into biota or precipitated with inorganic materials (Ca, Fe, etc.) and incorporated into the sediments to a point where there is negligible transport to a lake 20 km downstream. If there is a flood, and all the sediment is mobilized, it could be transported downstream and into the lake. If it settles into an anoxic hypolimnion, the conditions are ripe for internal loading processes to allow the P to eventually enter the lake and stimulate eutrophication. The question then becomes if there is a flood, how much of the sediment is transported to the lake, and what proportion of that settles before it is flushed through the lake? The hydraulic retention of the lake can be important. Also, if the lake is eutrophic enough to have an anoxic hypolimnion figures in as does the time of year the transport occurs.

An example of this may be Flathead Lake , where spring snowmelt flooding brings P-rich sediments into the lake and stimulates summer productivity. The city of Kalispell's release of P-rich sewage could have little effect most of the year, but each spring the flooding could ultimately transport much of the P.

What makes the effect of transient P inputs worse is that algae have a high capacity for luxury consumption so transient events that cause high P availability can be taken advantage of. Luxury consumption can lead to 10-20 doublings worth of P supply so a pulse of P in the spring can lead to extended growth through the first part of the summer and fuel a bloom later in the year.

Review:

I agree with the expert - the answer to the question is YES. Streams are not just pipes, but transformers, however, they transform in both directions from available (mineral) to particulate (organic) and back again - so what is taken up as dissolved is remineralized eventually. Most uptake studies (including those cited in the document) only measure the uptake length not the spiraling length - that is, they measure transport in the dissolved phase - not how far the nutrient then travels as a particle - so these numbers can be deceiving. Spiraling length is far harder to measure. Yet very important for P transport as the expert noted. P has essentially no gaseous phase (phosphine plays a negligible role), therefore there is little or no long-term net removal of P, especially since we know that long-term particulate storage in streams is pretty small. The take home message is that with P inputs, the piper eventually must be paid (double entendre intended) and any P inputs will eventually end up downstream, constantly moving from available to organic forms. Any lake TMDL will have to consider even the most remote sources of P as substantial contributors - no matter what a simple grab sample of dissolved P may tell you downstream of a WWTP during baseflow.

Important to note in the Ozark WWTP work is that they were 1) dealing with a particular geology/hydrology that absorbed and stored a lot of P during baseflow, 2) did not consider stormflow budgets - when most of that P from the WWTP was likely reentrained, 3) only measured soluble P fractions, and 4) did not report the significance of uptake length regressions. All of these will underestimate the P uptake length and greatly underestimate the total spiraling length.

Review:

I agree with the others. The issue here is long-term P transport, not uptake length. A stream transports virtually 100% because streams have no permanent sinks.

Review:

A stream does not necessarily transport 100% if it is depositional. Sediments may be buried and stored for a very long time. However, how much is stored for a very long period of time is difficult to predict, so it may be prudent to assume that eventually all the P is transported downstream.

Related to Data Gathering and Assessment #9

Related to Data Gathering and Assessment #6

Response:

There are few criteria I am aware of that are compared against an "average" condition. All the ones I am aware of are assessed against a single, fixed value. There are two main approaches for criteria attainment - single exceedance criteria (where any one exceedance is considered a violation) and percent exceedance criteria (where, for example, 10% of the samples cannot exceed the criterion). One could argue that the latter approach considers variability in the pollutant of interest more than the former. As far as criteria that require comparing a sample value to a "population" of reference values, I am unaware of any such approach. There are no statistical tests that can tell you whether one value is different from a population of values. However, you could test whether several values from a lake are significantly different from a distribution of reference lake values. Realize, however, that the power of such a test will be very low if the sample size is low and even lower if the values are highly variable. But you can conduct such means comparisons test, they are described in any statistical textbook and, more specifically for this application, in the Consolidated Assessment and Listing Methodology - an invaluable resource for these exact types of questions (www.epa.gov/owow/monitoring/calm.html). In classical statistical terms, the Type II error rate (of concluding the average test lake nutrient concentration is the same as reference when it is indeed different) would be very high with this type of test and these conditions (small sample size of highly variable data). As far as "certainty" goes, this same analysis provides statistics that indicate the degree of confidence with which the sample is different (or not) from the criterion. This has been a long and convoluted, perhaps, way of saying it will be impossible to express the certainty with which a waterbody exceeds a criterion with only one sample. With few samples, there will be very low certainty UNLESS those samples are a LOT higher than the criterion. The more samples you have, the smaller difference you can detect with high certainty. This is the classic statistical power issue. Given this statistical reality, it is probably best if a state or tribe set the confidence (certainty) desired to detect a set difference from a criterion. As part of the implementation procedures, the state could define the necessary statistical assurance - for example, acceptable type I and type II error rates and the minimum effect size required. Given observed historical variability, this would essentially establish the required sample size for such an approach.

An interesting topic for discussion is whether the hypothesis tests for regulatory compliance should not be switched for nutrient and other pollutant criteria (e.g., test of non-inferiority) as they do in human health drug testing. In non-inferiority or bioequivalence testing the null hypothesis is that the pollutant exceeds the criterion and error types are switched - Type II error in this case is the risk of concluding that the pollutant exceeds the criterion when it really does not. This would greatly increase the power of regulatory sampling, since it would be of great interest to collect sufficient samples so that the type II risk is low (and power is high). Type I error (the risk of concluding a waterbody does not exceed a criterion when it does) would be low because it would remain fixed at the traditional 0.05 or 0.1 alpha level. This is discussed in detail on pp 54-59 of Appendix C of the CALM guidance cited earlier.

N.B. This question is similar in context to two that have been posted to T-REQS. They have been answered and are in the Q&A page under Data Analysis (question 2) and Study Design (Question 5).

Review:

As has been mentioned, there is no question that variability is a problem. Other than the statistics discussed, one approach would be to minimize variability, such as limit the window of time used in criteria setting. The larger the window, the greater the variability possible.

The second point is that you have actually 4 values at any sampling, not one (CHL, TP , TN , and Secchi Depth). If the criteria have been set in such a manner so that there is a known correlation between the values, exceedance of one should result in the exceedance of others as well. Concern can then be related to the number of variables that exceed the criterion rather than just the exceedance of a single variable. In Michigan , where chlorophyll is often nutrient limited and transparency is related to chlorophyll, this should be possible, especially. if color is used in a correlation of CHL to SD.

Review:

See also questions Implementation #6, Implementation #7 and Study Design #9 on number of samples and index period (There may also be others I'm not aware of).

Criteria -- The question on averages here is not quite clear: average of what: multiple observations on a single lake; multiple lakes of a reference population; all lakes in the population? The reference condition approach uses a percentile of all reference lakes (within a class) as the criterion; the 75th percentile was proposed (but not required) in the nutrient criteria guidance. You may have valid reasons for selecting a higher or lower percentile than the 75th.

For a single lake, we reduce the observations (from the index period) to a single number: mean, geometric mean, median; and compare to the criterion. If the criterion was developed from a percentile of the reference distribution, then I think it is inappropriate to add a CI based on the multiple observations of the lake: the percentile you selected is your alpha-value that the lake is or is not similar to reference.

If the criterion is based on algal or whole lake response to nutrients, then it is OK to use a CI or hypothesis test. The basic problem with too few observations is low power: a hypothesis test will conclude "cannot reject null hypothesis of no effect" even though there may be nutrient impairment occurring. The non-inferiority approach suggested by the first response deals with this problem. Another way might be to require that significance and (1-power) be specified to the same level; for small samples you may need to increase alpha to 0.25 or higher.

The bottom line is that there is no way to have a great deal of confidence in a small number of nutrient observations, although it can be increased with a good index period (see other questions). The purpose of this question is to assist in 305(b) and 303(d) listing. The consequence of a 303(d) listing is a TMDL, and TMDLs are expensive! In the long run, increased sampling effort and confidence is much cheaper than unnecessary TMDLs and undetected impairment (which will still require TMDL in the future).

In the short run, set some minimum sample size for a hypothesis test or CI, and don't try to estimate a CI for those lakes with too few samples. Instead, just use the point estimate that you have, with the understanding that lakes that fail must be resampled and reevaluated prior to embarking on a TMDL.

Related to Implementation #2.

Response:

Nutrient criteria can recognize the seasonal nature of algal growth and associated impacts by restricting application of criteria to index periods. Index periods are one more specific times of the year when effects may occur.

Restricting application of nutrient criteria to index periods should be justified by showing that nutrients have no in-stream or downstream effects during the period when criteria would be suspended (non-index periods).

Application of a so-called index period for nutrient criteria requires relationships between nutrients and potential effects be clearly understood and predicted and that lack of relationships and effects be thoroughly considered and investigated.

Relationships and lack of relationships between ecological condition and nutrients in waterbodies (streams, rivers, lakes, wetlands, estuaries, and coastal waters) should be evaluated using stressor-response approaches, versus simple frequency distribution approaches. Nutrient criteria based on frequency distributions alone are not based on relationships between nutrients and valued ecological attributes (e.g. low algal biomass, high biological condition, and water clarity). Covariation, preferably through some kind of regression analysis, should be used to test relationships between nutrient concentrations and valued ecological attributes.

Application of index periods would differ among waterbody type. For one reason, we expect different nutrient effects and timing of effects in different waterbody types. In addition, nutrient retention varies greatly among waterbody types. For example, in streams, nutrient retention is short and effects are limited to specific seasonal periods. Nutrient concentrations measured in during summer low flow periods may not indicate nutrient concentrations that caused nuisance algal blooms during some other time of year. Whereas in lakes, nutrients are retained for much longer periods, so measurements in one season indicate conditions throughout the rest of the year. Most index periods for lakes are probably during summer, however spring and summer assessments may provide more accurate and precise characterizations of lake conditions than simple summer assessments for relating nutrients and effects. Most index periods in streams would be when nuisance algal growths occur and during warm, low-flow periods during summer and early fall. Of course, these are examples for more northern US systems with great seasonal variability.

Again, restricting application of nutrient criteria to index periods should be justified by showing that nutrients have no in-stream (in-waterbody) or downstream effects during the period when criteria would be suspended (non-index periods). These will vary among waterbody types, with regional climatic and geologic conditions, and with site-specific hydrogeomorphic conditions.

Review:

Agree. There is no recommended a priori index period for nutrient criteria. For rivers and streams, it may be best to use growing-season averages, or annual averages of nutrient concentrations as well as response attributes, estimated from 6 or more samples. When you have enough data, you may be able to demonstrate that an index period would yield sufficient information for both stressor levels and overall responses.

Review:

The original answer is correct and to the point. It depends upon the response variable. Chlorophyll may be particularly problematic for phosphorus because the propensity of algae for luxury uptake can decouple water column nutrients from the response variable (algal biomass).

Most agencies use seasonal or monthly means because it is not practical to sample for nutrients more frequently if many sites are of concern. In some cases shorter periods are of interest. For example, when the state of Montana was concerned about low-flow conditions in the Clark Fork River , they sampled for a week or few week periods to examine response variable of algal biomass. Functional relationships between algal biomass and total water column nutrients are usually constructed on seasonal or at least monthly means because variance, particularly in streams, is so great that single sampling events are difficult to use. The regulated community is generally in favor of longer periods of averaging because this allows for temporary excursions above criteria. This should be balanced against using very long averages (say annual) that have little to do with many of the response variables.

EPA Response

The answers are correct, but the EPA guidance on nutrient criteria development discusses index periods for a very specific reason. Nutrient cycling and effects of nutrient pollution (algal growth, algal turbidity, and diurnal swings in oxygen (high in daytime surface waters, low at night) is substantially different when waters are warmer. As the experts said, index periods vary for different system types, and in different climates (northern versus southern US). The Guidance documents recommend that samples be collected during the growing season (index period).

In rivers and streams, the index period should include high and low flow periods (rain event sampling to identify particulate transport is often useful for determining downstream effects for implementation purposes) during the growing season; that generally means spring and summer, with fall included in more southern climates.

In lakes the index period varies, but should include spring summer and the season near fall overturn for those systems that stratify to determine P regeneration from the sediments.

Response:

Nitrogen is available to plants primarily in the form of nitrate (NO3) and ammonium (NH4). Certain plants that contain symbiotic nitrogen-fixing bacteria (Rhizobium) in their roots are able to utilize gaseous nitrogen (N2) directly. Red alder (Alnus rubra) is one such plant in the Northwest that is particularly important. Nitrogen-fixing bacteria also have been identified in association with decaying wood in streams (Buckley and Triska 1978). Blue-green algae can also use atmospheric nitrogen but a large portion of the peripyton community must obtain their nutrients (nitrogen and phosphorus) from the stream water and sediments.

In oligotrophic waters primary production can result in large diel shifts in nutrient concentration under low-flow conditions. The magnitude of diel shifts depends on stream order because channel width determines canopy cover, and heavy shading inhibits photosynthesis. Gregory (1979) reported virtually no NO3 uptake on a first-order stream in the Oregon Cascades but a greater than 80% NO3 decrease from midnight concentration in a fifth-order stream. Thus primary production by benthic algae can be a significant mechanism for dissolved nutrient retention. Triska et al. (1989) measured the uptake of NO3 injected into the stream channel of Little Lost Man Creek in northern California . Over the course of their 10-day injection experiment, approximately 19% of injected NO3 was taken up by the stream's biota, which amounted to an average uptake rate of 26mg/m2/d. Uptake rates were greatest during daylight hours but continued through the night. Grimm and coworkers (Grimm et al. 1981) found that nitrogen to phosphorus ratios and concentrations of nitrate and soluble reactive phosphate (SRP) associated with a large number of Southwestern streams suggested that nitrogen was the limiting nutrient where such limitation occurs. Nitrate uptake in sections of intermittent streams was attributable to autotrophic activity, which again would be greatest during the light period and lowest in the dark portion of the diel cycle.

Phosphorus is available to plants in the form of orthophosphate (PO3), which is also the only inorganic form in which phosphorus occurs in appreciable amounts. Gregory (1978) examined isotopically labeled PO3 uptake by primary producers in Mack Creek, Oregon, and found epilithic algae to exhibit higher PO3 uptake rates than either filamentous algae or riparian vegetation. Gregory did not discuss diel flux in PO3 in this study. However, Grimm (Grimm et al. 1981) did find that the uptake of SRP in Southwestern streams was unrelated to any indicator of autotrophic activity. Grimm and coworkers concluded that SRP concentrations in desert and semi-desert stream waters may be controlled by other factors.

Citations

Buckley, B.M. and F.J. Triska. 1978. Presence and ecological role of nitrogen-fixing bacteria associated with wood decay in streams. Verhandlungen Internationale Vereinigung fiir Theoretische und Angewandte Limnologie 20:1333-1339.

Gregory, S.V. 1978. Phosphorus dynamics on organic and inorganic substrates in streams. Verhandlungen Internationale Vereinigung fur Theoretische und Angewandte Limnologie 20:1340-1346.

Gregory, S.V. 1979. Primary production in Pacific Northwest streams. Ph.D. dissertation. Oregon State University , Corvallis , Oregon . USA .

Grimm, N.B., S.G. Fisher and W.L. Minckley. 1981. Nitrogen and phosphorus dynamics in hot desert streams of Southwestern U.S.A. Hydrobiologia 83: 303-312.

Triska, F.J., V.C. Kennedy, R.J. Avanzino, G.W. Zellweger, and K.E. Bencala. 1989. Retention and transport of nutrients in a third-order stream in northwestern California : Channel processes. Ecology 70:1877-1892.

Review:

Although patterns of nutrient concentrations in streams are generally linked to photosynthesis and respiration of plant and microbes as reviewed, studies (Duff & Triska 2000) also indicated that exchanges of nutrients between sediment and surface water could be an important source of diel fluctuation of nitrate or SRP concentration in streams. In a diel study of hydrologic exchange between the hyporheic zone and surface environment in a desert stream in Arizona , Valett (1993) found that concentration and flux of NO3-N was greatest in surface water overlying upwelling regions at night when upwelling heads were greatest. Valett also found downstream decreaseed in NO3-N concentration, which reflected loss of groundwater discharge into surface water and strong assimilatory demand by surface algal communities. Diel fluctuation of biotic activity in sediments seems to be an important factor in changing of nutrient concentration (Claret and Boulton 2003, Hatch et al. 1999).

Claret C. and A. J Boulton. 2003. DIel variation in surface and subsurface microbial activity along a gradient of drying in an Australian san-bed stream. Freshwater Biologoy 48:1739-1755.
 
Duff, J. H. & Triska, F. J. 2000. Nitrogen biogeochemistry and surface-subsurface exchange in streams. In: Streams and Ground Waters. J. B. Jones and P. J Mulholland (eds), pp 197-220. Academic Press, San Diego .

Hatch L. K., Reuter J. E. , and Goldman C.R.1999. Daily phosphorus variation in a mountain stream. Water Resources Research 35: 3783-3791.

Valett, H. M. 1993. Surface-hyporheic interactions in a Sonoran Desert stream: hydrologic exchange and diel periodicity. Hydrobiologia. 259:133-144.

Review:

We have some unpublished data from 9 Kansas streams, with samples taken at 1:00 am and 11:00 am. Nine replicates each sample site and time. In all but one case the nitrate levels were lower in the day than the night. In 7 of those cases the difference was only slight (< 10%). In one, a concrete lined drainage ditch with good growth of epilithic algae and high nitrate, daytime concentrations were ½ those found at night (359 and 192 ug/L night and day respectively). These streams ranged from pristine with barely detectable nitrate to 20mg/L nitrate streams. Based on these data, I would suggest the diurnal range found in the models is a bit wide.

It is important to keep in mind that nitrate concentrations are a balance between uptake and nitrification, so concentration is not the same as demand. As the other reviewers indicated other factors such as groundwater influence can also be important. Finally, if denitrification causes major nitrate demand, rates may be higher at night, and nitrification lower, as dissolved oxygen concentrations decrease and denitrification is stimulated and nitrification inhibited.

Related to Implementation #1

Related to Implementation #2

Response: 

High flushing rate/low residence time reservoirs in the plains will always be problematic for nutrient criteria, because these lakes respond to elevated nutrients only under certain conditions: primarily times of low flow, when residence time is longer and the inflowing water has less suspended sediment and turbidity than in times of high flow. Using Secchi depth as the primary response is problematic in these lakes because of the contribution of suspended sediment to Secchi depth measurements.

Algae in these lakes only grow to excess when conditions are just right: high TP, long residence time, and sufficient light. Thus, they do not respond only to nutrients and a regression on TP doesn't work well. The trick to defining nutrient criteria will be to find the TP (and TN?) concentrations required when all the other conditions are also "just right".

Do the lakes with high flushing rates ever turn unacceptably green? That is, does the chlorophyll of those lakes exceed an acceptability threshold? If there is a recognized problem, but it is not predictable based solely on TP, you may want to consider an alternative analyses: conditional probability analysis (Paul and McDonald, 2005) which estimates the probability of impairment (excessive chlorophyll) for all stressor levels (say, TP) in your data set. It is similar to logistic regression, but easier to use. If you wish to explore multiple factors contributing to blooms, you could use multiple logistic regression, where the response is chlorophyll exceeding your threshold or not, and independent variables could include TP, TN, residence time, TSS, non-algal turbidity, and others. Rather than instantaneous measures of the explanatory variables, they may need to be averaged over a period of time that contributes to the bloom. See topics Implementation #6 and Implementation #7 for discussions on averaging and index periods.

If conditional probability and logistic regression fail, another approach would be to define reservoirs with low residence times as pools in the rivers, instead of as lakes, and apply TP criteria for streams. These criteria could be based on a reference percentile of streams (per EPA guidance for stream nutrient criteria), instead of on the response of chlorophyll to elevated nutrients. For reservoirs that do not respond to nutrient enrichment at all, it seems that river criteria, or criteria based on downstream receiving waters would be appropriate.

Reference:

Paul, John F. and Michael E. McDonald, 2005. Development of Empirical, Geographically Specific Water Quality Criteria: A Conditional Probability Analysis Approach. Journal of the American Water Resources Association (JAWRA) 41(5):1211-1223.

Review:

Low residence time waterbodies are problematic and the reply does address the problem.
Fast flushing rates will mean suppression of CHL because of the flushing rate itself and, as the response indicates, the high probability of non-algal turbidity. Secchi depth will probably correlate better with TP than CHL because these non-algal materials may/will contain phosphorus. It would be useful to differentiate between the effect of flushing and that on non-algal material. If the residence time in August or September is, say, 30 days or more, then there should be adequate time to develop near-maximum chlorophyll.

Setting criteria for the period of maximum chlorophyll makes sense. Rather than generate new CHL-TP relationships, it would be best to use existing equations, many of which have been constructed for Missouri lakes. The problem with erecting new equations is that they will be specific for the existing conditions and will be made obsolete as the conditions change.

The approaches to criteria setting in the response are quite appropriate. I would expand on one. If, by any method, you were to set the criteria first on chlorophyll levels, then you can use existing CHL-TP relationships to determine the TP criteria. If the problem is non-algal TP which are messing up the relationships, then assume that there is no light limitation by non-algal particulates and that the NAP does not contribute phosphorus to algal growth (both of which are probably to some extent in error). You can then use a model approach to develop TMDLs based on the phosphorus contributing to chlorophyll.

Review:

Nutrient levels in Missouri reservoirs are closely tied to the proportion of cropland in the watershed, reservoir depth and flushing rate. All factors equal, the greater the flushing rate (water exchange per year) the greater the nutrient value. So, it would be possible to have two reservoirs in Missouri with equal proportions of cropland in the basin and equal depth, but if one of the two water bodies has a larger catchment (more runoff) it will have a greater flushing rate and higher P levels. Simply put, reservoirs located high in the valley have lower nutrients than reservoirs located lower in an otherwise similar valley.

This description of nutrients and flushing rate is consistent with the Vollenweider approach to modeling lakes. In reservoirs with flushing rate greater than 4-times per year the lake TP value will be about 70% of inflow TP concentration. But, if the flushing rate is only 1-time per year the in-lake TP drops to about 45% of inflow concentration. If flushing rate is about 0.1-times per year then in-lake TP drops to about 25% of inflow TP. This relationship between in-lake TP and flushing rate is underappreciated in lake management.

As far as Chl-TP relationships go in MO reservoirs the highly flushed systems fit the pattern when the long-term mean condition is considered. These highly flushed systems are right on the regression line when averages are used. The tendancy is for rapidly flushed reservoirs to show a spike of turbid, high-TP and low Chl water in early spring in response to rain and inflow. Once these systems stratify the turbidity sediments and the Chl-TP relation matches the norm. This seasonal feature of the Chl-TP relation in MO reservoirs was described in a 2005 paper of Lake and Reservoir Management (see full citation below).

Rapidly flushed reservoirs are a special case and in may circumstances it will be impossible to greatly reduce in-lake TP if the reservoir has an agricultural watershed and the flushing rate is greater than 2-times per year.

Jones, J.R. and M.F. Knowlton. 2005. Chlorophyll response to nutrients and non-algal seston in Missouri reservoirs and oxbow lakes. Lake and Reserv. Manage. Vol 21(3): 361-371.

Response:

A number of factors can affect the basic relationships between nutrients and algal responses in streams. However, these relationships are not easily defined nor can many general statements be made about how individual or groups of similar streams will respond to factors such as light attenuation, stream velocity, scour and substrate size and stability. This is reflected in the weak relationships most researchers report when looking at N or P and benthic chlorophyll a (Lohman, Jones and Perkins 1992, Briggs 2000, Dodds, Smith and Loman 2002 and Bernhardt and Likens 2004). Benthic algal biomass accrual has been shown to be related to both drying events and floods of sufficient power as to cause streambed scouring (Biggs 1996 and 2000). However, we have found that measurable benthic accrual can take place in as little as 6 days in fourth and fifth order streams (80 stream sites measured 10 times/each) in Kansas, Nebraska and Iowa (mean, median and range of benthic chlorophyll a was 1.79, 0.72 and 0 to 14.9 mg/ M 2, respectively). Scour does reduce the accrual of benthic algal biomass but excessive algal growth can still occur during the critical low or normal flow periods of the year when high stream temperatures and low flows can contribute to lower DO values that might be acerbated by high algal productivity and respiration. Some authors (Findlay, Howe and Fontvieille 1993, Rosemond, Mulholland and Elwood 1993, Hill 1996) have noted that light and herbivores can limit algal biomass even in the presence of high nutrients. The total concentrations of benthic chlorophyll a from benthic algae taken from glass slides allowed to colonize in the small streams mentioned above, showed no relationship with canopy cover (percent shade). Using robust regression, a non-significant relationship (p=0.077) was found between 554 benthic chlorophyll samples and average % canopy (measured twice during the colonization period). Percent canopy values ranged from zero to 100% and average TP values were 0.117 mg/L. These low gradient streams flow through fairly open agricultural landscapes and riparian boundaries are typically narrow so even though canopy cover was 100% in many areas there was still a lot of incidental light reaching the streams. We feel that measures of canopy cover reflect the amount of direct light that reaches a stream but not the amount of total light as reflected by the lack of a relationship between periphyton accrual and % canopy. Based on our findings, it cannot be assumed that riparian cover can affectively reduce or limit benthic algae or for that matter sestonic chlorophyll a. Looking at our robust regression analysis for sestonic chlorophyll a concentrations and % canopy we did note that there was a highly significant, negative relationship (n=521, p=0.00004) but the r 2 value was only 0.032 suggesting that light was not a controlling factor in these streams regardless of the amount of local riparian cover.

Certainly the best relationships we see between nutrients and chlorophyll are for sestonic chlorophyll. The table below is from the EPA Region 7 RTAG findings and suggests that excessive algal biomass related to nutrients are more likely a problem in larger streams where light attenuation is not a possibility and flow regimes may be more conducive to the production and maintenance of sestonic algae.

The results in this table are from robust regressions run for selected stressors, classification and response variables using all sites having one or more values for both the dependant and independent variables (i.e. no data filters used). All analysis preformed using the NCSS ® database entitled Chem2005Sep.SO. This database is currently maintained by the Central Plains Center for BioAssessment and is available on request (www.cpcb@ku.edu). NCSS ® is a statistics software package that we use in most of our statistical testing.

The short answer to part of your question is that riparian cover should be maintained or replaced where appropriate but is not a substitute or reliable modifier for acceptable nutrient levels. However, I feel that stream classifications (e.g. watershed size as an indicator if stream order or size) might be appropriate in establishing criteria. Also we didn’t find much of a relationship between chlorophyll a and substrate sizes. Chlorophyll concentrations in and on sand and silt can get nearly as high as concentrations found on large gravel and cobble but the instability of small particles leads to more frequent and perhaps more severe scour losses. Smaller, more frequent runoff events often possess enough energy to move and entrain smaller particles and thus disrupt algae grown and accrued.

Bernhardt, E.S. and G.E. Likens, Controls on algal biomass in heterotrophic streams, Freshwater Biology, vol. 49 (2004), pp. 14-27.

Bernhardt, E.S. and G.E. Likens. 2004. Controls on periphyton biomass in heterotrophic streams. Freshwater Biology 49: 14-27.

Biggs, B.J.F. 1996. Patterns in benthic algae of streams. Pages 31–56 in R. J. Stevenson, M. L. Bothwell, and R. L. Lowe (editors). Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.

Biggs, B.J.F. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll relationships for benthic algae. J. North American Benthological Society 19:17-31.

Dodds, W.K., V.H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences. 59:865-874.

Findlay , S., K. Howe, D. Fontvieille. 1993. Bacterial-algal relationships in streams of the Hubbard Brook experimental forest. Ecology 74: 2326 2336.

H ill , W.R. 1996. Effects of light. Pages 121–148 in R.J. Stevenson, M. L. Bothwell, and R. L. Lowe (editors). Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.

Lohman K., J.R. Jones & B.D Perkins. 1992. Effects of nutrient enrichment and flood frequency on periphyton biomass in Northern Ozark streams. Canadian Journal of Fisheries and Aquatic Sciences 49, 1198–1205.

Rosemond, A.D., P.J. Mulholland, J.W. Elwood 1993. Top-down and bottom-up control of stream periphyton. Ecology 74: 1264-1280.

Review:

Along with field survey, manipulated experiments might better demonstrate the true relationships among light, nutrients, and benthic algal biomass (Taulbee et al. 2005). In a meta-analysis of experimental manipulation, Hillebrand (2005) found that both grazer removal and light enhancement had positive effects on algal biomass. Light effects became more important with increased trophic state of the habitat and with algal biomass. In a field experimental study comparing algal biomass and light availability, Roberts et al.(2004) found that although the initial slopes of photosynthesis versus irradiance relationships were similar in a light saturated creek and a partially shaded creek during the initial stages of colonization, chl concentration per surface area content of chl per algal cell in the light saturated creek in later stages of colonization were higher than in the partly shaded creek. Another study in trophic streams (Larned and Santos 2000) found levels of chlorophyll a in partially-shaded stream pools were significantly greater than in heavily-shaded pools, and nutrient-enrichment increased the level of chlorophyll a in partially-shaded pools but not in heavily-shaded pools, indicating light limitation inhibits algal biomass. It seemed that total algal biomass could be limited by light availability, and theoretically, regulating riparian cover might limit algal biomass to some extent. However, light availability to algae can not be easily regulated since algae can utilize the incidental light to reach a high biomass level even if the canopy cover is 100%. In addition, riparian cover in temperate streams varies seasonally which makes light control difficult in spring and fall seasons.

Hillebrand, H. 2005. Light regime and consumer control of autotrophic biomass. Journal of Ecology 93:758-769.

Larned, S. T., and S. R. Santos. 2000. Light- and nutrient-limited periphyton in low order streams of Oahu, Hawaii. Hydrobiologia 432:101-111.

Roberts, S., S. Sabater, and J. Beardall. 2004. Benthic microalgal colonization in streams of differing riparian cover and light availability. Journal of Phycology 40:1004-1012.

Taulbee, W. K., S. D. Cooper, and J. M. Melack. 2005. Effects of nutrient enrichment on algal biomass across a natural light gradient. Archiv Fur Hydrobiologie 164:449-464.

Review:

Do exogenous factors affect nutrient criteria and assessment?

In this answer, I try to put the question in a broader context of how exogenous factors should be considered in nutrient criteria development. I focus on shading and scour as two examples of exogenous factors, but also include others in examples of how exogenous factors should be considered for nutrient criteria and assessment. I’ll address how exogenous factors affect criteria and then assessment.

A number of questions are related to importance of exogenous factors for criteria development and assessment. What are the management goals? Which exogenous factors? Are they natural or anthropogenic? Are interactive effects with nutrients positive or negative - simple or complex? Are both instream and downstream effects considered? Do effects considered represent all effects that may be important? And, as Dr. Huggins points out, do they have measurable and predictable effects?

Conceptually, exogenous factors should be considered in criteria development and assessment. Exogenous factors could originate from 2 sources, natural and anthropogenic, that would have very different effects on how they should be considered in nutrient criteria and assessment (Stevenson et al. 2004a, b). In the following discussion, I assume that management goals are related to the Clean Water Act and support of aquatic life use. Natural exogenous factors, such as climate, geology, hydrogeomorphology (stream slope, hydrologic variability, etc), substrate size, and species membership based on natural biogeographic patterns should be considered in predicting the “natural” expected biological and physicochemical condition of an ecosystem or the “reference” condition (see Stoddard et al. in press). Anthropogenic factors, such as the contaminants and habitat alterations that are regulated by human activities, should be evaluated in assessment and the subject of criteria development. Some factors, such as shading, scour, and even nutrients, are regulated by both natural and anthropogenic processes; so the natural and anthropogenic portions must be considered separately.

Natural exogenous factors should be considered in nutrient criteria if they affect nutrient concentrations or they affect the response of valued ecological attributes (e.g. measures of aquatic life or other stream uses) to nutrient criteria. If they affect nutrient concentrations directly, then they become incorporated into the characterization of reference condition. The differences in nutrient concentrations among nutrient ecoregions (Omernik 2000, Omernik et al. 2000) are partially the result of natural exogenous factors.1 Different regions would have different geology, soil types, and landscapes that could affect nutrient concentrations in streams. For example, natural phosphorus concentrations probably would be higher in streams draining clay soils and landscapes with extensive wetlands (wetland-derived dissolved organic P) versus streams draining thin glacial soils. Dodds and Oakes (2004) and the discussions associated with the documentation for the proposed EPA nutrient criteria for nutrient ecoregions (http://www.epa.gov/waterscience/criteria/nutrient/ecoregions/rivers/index.html) provide additional details and perspectives for refining nutrient criteria based on regional attributes.

In the case where natural variability in exogenous factors among stream segments regulates responses of valued ecological attributes, then those exogenous factors can be used to classify streams and establish different criteria for different classes of streams. Let’s develop an example for which different nutrient criteria could be justified based on effects of nutrients on valued attributes for two classes of streams. We need three assumptions (Figure 1): 1) we have two classes of streams, shaded headwater streams and less-shaded mid-sized streams and for sake of the argument, natural nutrient concentrations in both classes of streams are the same; 2) Cladophora growth rates and % cover of a reach increase with nutrients; and 3) light differences between streams affect Cladophora growth and % cover. If we were to develop nutrient criteria based, for example, on the TP concentration causing Cladophora cover of 25%, then we would use different TP concentrations for nutrient criteria. We’d need lower criteria in the naturally unshaded midreach streams than in the shaded headwater streams to prevent average Cladophora cover from exceeding 25%. In this case, if Cladophora were the only thing we were concerned about that was affected by nutrients, then different nutrient criteria would be justified. But other instream and downstream effects of nutrients should be considered to ensure that allowing nutrients to exceed reference condition does not allow unforeseen problems to develop.

If lack of shading along the riparian corridor is due to anthropogenic causes, this introduces issues associated with multistressor interactions. Multistressor interactions may increase or decrease effects of the nutrients, and that will guide how interactive effects of those stressors (contaminant or habitat alteration) should be incorporated into development of nutrient criteria. In general, if non-nutrient anthropogenic stressors exacerbate nutrient effects, then lower nutrient criteria would be justified to prevent nutrient effects. If non-nutrient anthropogenic stressors ameliorate nutrient effects, then different (especially higher) nutrient criteria would not be justified because when those stressors are remedied and during periods when those stressors are not a problem, then nutrients will be a problem. I’ll develop a couple examples with shading and silt below, where we’d expect that light would increase effects of nutrients and silt would decrease them.

In the case of shading, shading should decrease the potential for problems to develop as it reduces the positive effects of light. Therefore, anthropogenic reductions in shading increase problems associated with nutrients. Solving nutrient related problems could be solved by lowering nutrients, and thereby justify lower nutrient criteria, or by reestablishing shading or having lower nutrient criteria. Of course that does not solve problems associated with higher water temperatures and lower detrital loading without riparian trees. Establishing lower nutrient criteria for streams affected by humans clearing riparian corridors could provide an incentive for restoration of those corridors.

In the case of silt, low silt is the natural condition and increasing siltation decreases algal accrual and the potential for nutrient effects to be manifested. Silt causes other problems with aquatic life use support, which should be corrected. This is a common problem in many regions where nutrients and silt are co-occurring stressors. The amelioration of nutrient effects by silt from anthropogenic sources would not justify higher nutrient criteria for silt-impacted streams for two reasons. First, silt problems should be corrected; and when they are, nutrients will be a problem. Second, anthropogenically loaded silt is not always a problem in most streams, because silt effects vary with time and space within the stream; so when and where silt is not a problem, nutrients will be a problem.

From these examples, we can develop a general rule. Effects-based nutrient criteria (those based on response of valued ecological attributes to nutrients) should be established with models that have other anthropogenic factors set to natural or reference condition. Then, in the case of non-nutrient stressors that have positive interactive effects with nutrients (e.g. light, temperature related to loss of shading), lower nutrient criteria can be established to reduce effects of those stressors than would be required if those stressors were not present. The difference in nutrient criteria with and without these stressors would provide an incentive for their restoration. In the case of non-nutrient stressors that have negative interactive effects with nutrients, nutrient criteria should be based on expectations when these stressors are near natural or reference levels.

Scouring is a complex factor that interacts with nutrients in streams because it has non-linear effects on algal biomass in streams and has both anthropogenic and natural causes. Plus, scouring seems to have different effects on different kinds of algae. Let’s simplify this discussion by assuming that we’re dealing with natural causes of scour and want to consider natural variability in hydrologic disturbance and scouring as factors for classifying streams and having different nutrient criteria for different streams. These issues are covered in a recent paper that I wrote (Stevenson et al. 2006, Figure 2). In this paper, I describe a model where climate and geology are ultimate factors that regulate hydrologic variability and natural scour disturbance regimes. This is an elaboration and clarification of models presented a number of years ago, and which may be useful for understanding the broader predictive models for algal-nutrient interactions (Biggs et al. 1990, Biggs 1996, Stevenson 1997, Biggs et al. 1998) . Hydrologic disturbance affects invertebrate grazer density, which affects top-down or bottom-up regulation of diatoms, but has less effect on filamentous green algae. Filamentous green algae are less affected by invertebrate grazers than diatoms. Grazers are negatively affected by hydrologic disturbance, so diatom accrual should be greater in systems with higher levels of hydrologic disturbance, until disturbance frequency is so high that even diatoms can not recolonize substrata. Filamentous green algae grow well in low and intermediate disturbance conditions, but are strongly and negative affected by scouring in high disturbance conditions. Nutrients have a greater effect on filamentous green algae than diatoms, because diatom growth rates saturate at lower nutrient concentrations than filamentous green algae. In addition, higher nutrients have more effect on filamentous green algae (FGA) in streams than diatoms because FGA grow to thicker density and higher nutrients concentrations are needed to counter the negative density effect on nutrient supply.

Based on this model, we could justify higher nutrient criteria for high disturbance streams if problems did not develop downstream. We’d be tempted to have higher nutrient criteria in low versus intermediate disturbance streams because peak biomass in high nutrients is lower in low disturbance due to diatom grazing by herbivores. However, if most impacts are caused by FGA, then FGA are as much of a problem at high nutrients in low as intermediate disturbance streams.

The initial reviewer brought up another very important issue, which is the effects of many of the non-nutrient factors are not well understood or documented at scales that are important for resource management. We know from experiments that light and grazers are important interactive factors with nutrients. However, seldom have these interactions been documented outside experimental scales. Results of experiments should be applied in the development of nutrient criteria very carefully (Pan et al. 2000, Stevenson et al. 2002). In the field, correlations between light and algal biomass are not clear. This may be due, in part, to diatom growth rates having lower light requirements than photosynthetic rates (Rier et al. 2006) or long accrual periods compensating for lower algal growth rates in low light. We must be certain that we have a good understanding of the interactive effects of exogenous factors on both algae and non-algal valued attributes. The reference condition approach for establishing nutrient criteria does help solve some of these problems by establishing the nutrient conditions needed to protect the reference condition, versus effects-based criteria approaches for which good quantitative models for many relationships have not been developed thoroughly.

Finally, exogenous factors affect both criteria and assessment. Criteria have been discussed above. Assessment will be affected differently depending upon how the assessment is conducted. Two interesting contrasts are when assessment is defined as a deviation from criteria versus from our best estimate of natural. If assessment is deviation from preset criteria, then exogenous factors can be taken into account by correcting criteria based on exogenous factors. For example, different classes of streams (either based on ecoregion or hydromorphic characteristics) would have different expected conditions and criteria for those expected conditions. These may simply be pass/fail assessments or assessments where assessment grades of excellent, good, fair, and poor are assigned. The TP concentrations, metrics, or IBI scores for these different grades could vary among stream classes. The tiered uses and different regional classes of waters in Ohio provide a good example of this kind of system. Alternatively, assessments may be deviations from natural and a fixed deviation from natural is acceptable. Wiley et al. (1997) apply this concept with Michigan stream fish, where different numbers of fish species are expected in different valley segment based on water temperature and stream hydrogeomorphology. The number of fish species for a specific valley segment is predicted with a regression model using natural features as independent variables. If observed conditions exceed predicted, then waters are deemed excellent. If they deviate less than predicted by 0.5 standard errors from the expected condition, then they are deemed good, fair if between 0.5 and 1 standard errors, and poor if greater than 1.0 standard error less than predicted. In general, however, the same basic considerations should be used to account for exogenous factors for nutrient criteria and assessment.

The effects of exogenous factors on criteria and assessment have many determinants, and each must be considered within the context of the management program and region in which it occurs.

The guidelines and examples above should cover many examples of how scour, shading, and many other factors would affect nutrient criteria and assessment. Many of these guidelines should be transferable to other non-nutrient factors.

Literature Cited

Biggs, B. J. F. 1996. Patterns in benthic algae of streams. Pages 31-56 in R. J. Stevenson, M. L. Bothwell, and R. L. Lowe, editors. Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.

Biggs, B. J. F., M. J. Duncan, I. G. Jowett, J. M. Quinn, C. W. Hickey, R. J. Davies-Colley, and M. E. Close. 1990. Ecological charactersation, classification, and modeling of New Zealand rivers: an introduction and synthesis. New Zealand Journal of Marine and Freshwater Research 24:277-304.

Biggs, B. J. F., R. J. Stevenson, and R. L. Lowe. 1998. A habitat matrix conceptual model for stream periphyton. Archiv für Hydrobiologie 143:21-56.

Dodds, W. K., and R. M. Oakes. 2004. A technique for establishing reference nutrient concentrations across watersheds affected by humans. Limnology and Oceanography: Methods 2:331-341.

Omernik, J. M. 2000. Draft Aggregations of Level III Ecoregions for the National Nutrient Strategy. http://www.epa.gov/waterscience/criteria/nutrient/ecomap.html.

Omernik, J. M., S. S. Chapman, R. A. Lillie, and R. T. Dumke. 2000. Ecoregions of Wisconsin. Transactions 88:77-103.

Pan, Y., R. J. Stevenson, P. Vaithiyanathan, J. Slate, and C. J. Richardson. 2000. Changes in algal assemblages along observed and experimental phosphorus gradients in a subtropical wetland, U.S.A. Freshwater Biology 43:1-15.

Rier, S. T., R. J. Stevenson, and G. LaLiberte. 2006. Photo-acclimation response of benthic stream algae across experimentally manipulated light gradients: A comparison of growth rates and net primary productivity. Journal of Phycology 42:560-567.

Stevenson, R., S. Rier, C. Riseng, R. Schultz, and M. Wiley. 2006. Comparing Effects of Nutrients on Algal Biomass in Streams in Two Regions with Different Disturbance Regimes and with Applications for Developing Nutrient Criteria. Hydrobiologia 561:149-165.

Stevenson, R. J. 1997. Scale-dependent causal frameworks and the consequences of benthic algal heterogeneity. Journal of the North American Benthological Society 16:248-262.

Stevenson, R. J., R. C. Bailey, M. C. Harass, C. P. Hawkins, J. Alba-Tercedor, C. Couch, S. Dyer, F. A. Fulk, J. M. Harrington, C. T. Hunsaker, and R. K. Johnson. 2004a. Designing data collection for ecological assessments. Pages 55-84 in M. T. Barbour, S. B. Norton, H. R. Preston, and K. W. Thornton, editors. Ecological Assessment of Aquatic Resources: Linking Science to Decision-Making. Society of Environmental Toxicology and Contamination Publication, Pensacola, Florida.

Stevenson, R. J., R. C. Bailey, M. C. Harass, C. P. Hawkins, J. Alba-Tercedor, C. Couch, S. Dyer, F. A. Fulk, J. M. Harrington, C. T. Hunsaker, and R. K. Johnson. 2004b. Interpreting results of ecological assessments. Pages 85-111 in M. T. Barbour, S. B. Norton, H. R. Preston, and K. W. Thornton, editors. Ecological Assessment of Aquatic Resources: Linking Science to Decision-Making. Society of Environmental Toxicology and Contamination Publication, Pensacola, Florida.

Stevenson, R. J., Y. Pan, and P. Vaithiyanathan. 2002. Ecological assessment and indicator development in wetlands: the case of algae in the Everglades, USA. Verhandlungen Internationale Vereinigung für Theoretische und Andgewandte Limnologie 28:1248-1252.

Wiley, M. J., S. L. Kohler, and P. W. Seelbach. 1997. Reconciling landscape and local views of aquatic communities: lessons from Michigan trout streams. Freshwater Biology 37:133-148.

1 Anthropogenic exogenous factors also vary regionally andare incorporated into nutrient ecoregions. This is evident when comparing natural and best attainable characterizations of nutrient reference conditions and in EPA’s proposed nutrient criteria based on 25 th percentiles of conditions in different nutrient ecoregions http://www.epa.gov/waterscience/criteria/nutrient/ecoregions/rivers/index.html. Arguably, issues related to different levels of protection in different regions of the country are an issue when extent of landscape altertation by humans differs and affects definitions of reference condition. Solutions to these issues are being addressed with concepts elaborated in the USEPAs Tiered Aquatic Life Use concepts and how a common scale of comparison can be applied across the country.

Response:

Secondary sewage effluent is dominated by inorganic N (nitrate for most plants because of laws against ammonia release leading to aerobic digestion reactors). Tertiary treatment in an advanced waste water plant can use a denitrification step (most nitrogen lost as a gas) or biological removal with harvesting of the biomass (e.g. algae in raceways or on screens or membranes that are subsequently harvested). I can find no evidence for highly refractory N compounds being the majority of the N left after such processes. If this were the case then highly refractory N would dominate all aquatic ecosystems since biological removal differs little from the processes that occur in natural biofilms of rivers, streams, lakes, and groundwaters. I would expect the organisms to leak a little organic matter into the system, and leave behind some of the organic matter that subsequently leaves the plant. This material should not be very refractory since it would be just cellular material.

One paper (Patel et al. 2005) used a membrane bioreactor system and it removed 99% COD, 98.4% NH4, and 77.5% total N. This pilot system produced little biological oxygen demand although it did release 5.8 mg/L nitrate.

There is also research published that the higher the N content of humic material, the more bioavailable it is (Hunt et al. 2000).

Hunt A.P., J.D. Parry, J. Hamilton-Taylor. 2000. Further evidence of elemental composition as an indicator of the bioavailability of humic substances to bacteria. LIMNOLOGY AND OCEANOGRAPHY 45 (1): 237-241. January.

Patel, J., G. Nakhla, A. Margaritis. 2005. Optimization of biological nutrient removal in a membrane bioreactor system. Journal of Environmental Engineering. 131 (7): 1021-1029. July.

Review:

The Patel paper cited by the first reviewer used synthetic wastewater and therefore may not be particularly applicable in this situation. Secondary sewage effluent is dominated by inorganic N (ammonia), but also carries a small fraction of organic nitrogen. An advanced wastewater plant with BNR uses a denitrification step (most nitrogen lost as a gas) but some refractory organic N is reported at some locations. This refractory nitrogen ranges between 0 and 1.5 mg/L.

The answer to the question, therefore, is yes, the final effluent could have refractory nitrogen but it does not exert much direct BOD. At an advanced treatment facility with BNR, the effluent would carry nitrate in the range of 3 to 10 mg/L or less, depending upon the process selected and the wastewater characteristics at the location.

Review:

I think of direct BOD from nitrogen being related to ammonia, nitrification, and NBOD. If that is low as a result of repeated biological transformation (the N is never degraded, it is transformed) and transformation into nitrate or organic N, then direct BOD would be low. Unless the N is removed by denitrification or harvesting, N doesn’t disappear from the waste stream. Since the N being discussed is being discharged, it is part of the waste stream. So, now one question could be, “When will it become bioavailable again and create an oxygen demand?” Direct nitrification is a direct BOD. Nitrate will cause most of its BOD “indirectly” from respiration of algae and plants that accrue as a result of the N enrichment and by the bacteria decomposing the algae and plants. Organic N can also be decomposed in-stream by bacteria, creating both a “direct BOD” and indirect BOD as the resulting NH3 and NO3 are later utilized by primary producers.

The concept of refractory is relative. The effect of N loading into streams, no matter what the form, will cause a BOD at some time and some place. This may be as far away as coastal estuaries, but may be local during low flow conditions. The potential for particulate N and P to be transformed to bioavailable forms was one rationale for TN and TP to be used as stream criteria. Dissolved inorganic forms of N and P describe the bioavailable forms of nutrients being carried downstream and immediately available for algae and plants, but not the potential for internal loading of transformed N and P. Although internal loading is certainly more of a local problem in lakes and wetlands, it is also a problem, at least to some extent, in streams. The problem in lotic systems increases with water residence time and settling of particulate N and P in pools. These problems may also emerge with entrainment in benthic algal and plant assemblages where decomposition of detrital N and P may increase their local availability.

Response:

There is no one assessment method that fits all situations. There is, however, a reasonable set of questions to answer to narrow the options down to a few assessment methods. These decisions deal with 2 main issues, objectives of the assessment and type of stream.

The preferred or recommended periphyton assessment method depends upon the endpoint that is a concern. For example, if concern is limited to the biological condition of invertebrates and fish, then the endpoint would probably be an indicator of algal biomass or algal biomass potential. The latter could actually be a diatom indicator of the algal biomass that could accumulate at that site if given the chance without flood disturbance. Biomass is also an important aesthetics endpoint. Alternatively, if the concern is periphyton condition, which some states measure with species composition of diatoms assemblages, then species composition of diatoms should be assessed. Finally, you may just want to use species composition of periphyton to help infer nutrient concentrations. The many parameters for which algae can be assessed in streams assessments have been covered in many reviews (e.g. Lowe and Pan 1996, Stevenson and Pan 1999, Stevenson and Smol 2002) – but these 3 parameters are a good start: biomass, biological (taxonomic) condition, and inferences of nutrient conditions.

Once you know the parameter to assess, then the sampling and assay methods can be chosen. Choosing a method depends upon budget for assays, time in the field, and time in the lab and in-house expertise. It also depends on the habitats present in the stream. Sampling methods vary most among stream types, such as those that have rocks in riffles and those that don’t.

Biomass can be measured visually in the field and after sampling followed by analysis of chlorophyll a, ash-free dry mass, cell biovolume, or cell density. Visual assessments are conducted to estimate cover of the stream bottom by macroalgae and thickness of both microalgae and macroalgae. These methods are described in the Rapid BioAssessment Protocols (Stevenson and Bahls 1999)and in Stevenson et al. (2006). Visual assessments of benthic algal biomass in streams (and probably lakes as well) are more rapid, more spatially extensive, more related to aesthetics, and likely more precise for estimating macroalgae than parameters (chlorophyll a, ash-free dry mass, cell biovolume, or cell density) derived from most substrate sampling protocols (e.g. 10 rocks from the riffle). For example, % Cladophora cover of habitat may be a better estimate than chlorophyll a of habitat alteration and cause of invertebrate and fish aquatic life use impairment. Visual assessments of microalgae cover are well related to measured chlorophyll a on rocks. Secchi disk transparency is the analogous visual assessment method for water column algal biomass, although this measure is more affected by suspended sediment and dissolved organic carbon than visual assessments of benthic algal biomass.

Assays of chlorophyll a, ash-free dry mass, cell biovolume, or cell density derived from substrate (or water column) are more quantitative and more broadly accepted as a methodologically rigorous approach than visual assessments for assay of algal biomass. Sampling should be as extensive as possible throughout the “defined habitat”, discussed later. Each of these parameters has strengths and weaknesses, which are described in table in Stevenson (1996) and probably elsewhere. Basically, chlorophyll a varies in cells with light, nutrients, and type of algae and it’s most difficult to store for assay when in remote field settings. Ash-free dry mass is usually contaminated with detritus and non-algal biomass. Cell biovolume and cell density are most expensive to assay because they require microscopic examination. Cell density is problematic because cells vary greatly in size. Cell biovolume is a pretty variable measure of algal biomass because large cells tend to be relatively rare in samples compared to small cells and they have disproportionately greater effects on biovolume estimates and have proportionally larger vacuoles in cells. Given these issues, chlorophyll a is the recommended assay of algal biomass.

The recommended method for assaying chlorophyll a is not well agreed upon. The main differences in methods is whether pigments are extracted in acetone, methanol, or ethanol, whether samples are ground or sonicated or not to facilitate cells breakage and extraction extraction, and whether a spectrophotometer or a fluorometer is used to assay the extracted pigments. Acetone is historically the standard method of pigment extraction and grinding becomes more important with the greater proportion of green algae in samples with their tough cell walls. The published benchmarks for characterizing stream trophic status (e.g. Welch et al. 1988, Dodds et al. 1998) were mostly based on acetone extraction. Although some scientists continue to use acetone extraction with grinding because it has been the standard method, many scientists are changing protocols from acetone to methanol or ethanol because chlorophyll extraction is much more efficient. Calibration of results is problematic because there are differences in extraction efficiency between solvents. Acetone estimates of chlorophyll a are usually lower than methanol and ethanol. However, these differences can be accounted for with cross-method comparison and developing a calibration factor. See the comparison of these methods in Paptisa et al. (2002). They found that methanol is the most efficient, but safety issues with methanol versus ethanol and acetone are a concern that should be considered.

Some states use species composition of benthic algae to assess aquatic life use and some are planning to (e.g. Kentucky, Idaho, Maine). Aquatic life use of periphyton is usually based on diatom species composition versus all algae, but all algae are being considered more and more. Species composition of benthic algae can be determined without quantitative sampling of substrata, which saves some time in the field. If field sampling time limits the effort that can be allocated to quantitative periphyton sampling, then consider visual assessments of biomass with random sampling of periphyton for assessing species composition and resulting metrics of biological condition (e.g. Fore and Grafe 2002, Wang et al. 2005).

Although species composition of diatoms in a stream will not tell you much about aesthetics, it can help infer nutrient concentrations and whether nuisance algal blooms are possible. Species composition of diatoms very accurately infers nutrient concentrations in streams and other habitats (Kelly and Whitton 1995, Potapova et al. 2004).

Nutrient concentrations in algae can also be used estimate nutrient limitation and relative importance of N:P ratios. Little evidence supports the use of this technique in larger scale surveys. Problems seem to be important with contamination by detritus and variation in ratios with algal density.

The scale of assessment (“defined habitat”) should also match the specific objectives of the study or program. The reach scale assessments of EMAP with random sampling along 10 or more randomly placed transects are appropriate for assessing the whole reach and obtaining an unbiased estimate of reach conditions. The advantages of this method are the unbiased estimate of reach-scale conditions, i.e. the whole stream, and the relative certainty that major problems will be observed. One disadvantage of the reach scale EMAP methods may be the relative insensitivity to changes in biological condition in specific habitats that are most vulnerable to environmental changes. This is most problematic in low gradient streams where most samples will be the species-poor bottom habitat, which does represent most of the stream, versus the more species-rich snag and root-mass habitats. This problem could be solved by separating samples from different habitats within the stream, which minimally would reflect original Mid-Atlantic EMAP and NAWQA sampling protocols with separate samples from erosional and depositional habitats. Such targeted habitat sampling has problems too. You have more samples to analyze and a bit more field time. Selecting one targeted habitat solves the too many samples issue, but could miss loss of biological condition in non-target habitats. For example, most sampling is done in stream riffles, so we know little about what is happening in pools, which may be very important habitat for many invertebrates and fish, certainly during low flow periods. Polls also may be important locations of high algal accrual during low water periods. Region 3 and Great Plains sampling protocols should be constraining interpretation of their results to what is actually measured in riffles and shallow water areas, respectively, and what those results can infer about other habitats. But they are not, of course, direct measures of whole-reach conditions.

All streams do not have to be sampled the same. Different types of streams could have different sampling methods, even within a region. This is covered in the Nutrient Criteria Guidance document under “classification.” One protocol for doing this for periphyton is to sample hard substrata if they exist and soft substrata if hard substrata are not present. For example, sample rocks, wood, and plants in streams that have these habitats, in proportions that reflect their relative abundance in streams. If the so-called hard substrata are not present, sample soft substrata in slow current, stable habitats with light (as the Great Plains protocol indicates). Different expectations should be developed for streams with hard and soft substrata.

Consider sampling plankton as well as or in place of periphyton as streams get larger and the importance of plankton in the ecosystem increases. Suspended algae can even be important in small, low gradient streams with high residence times (e.g. wetland streams and ditches), not just in rivers. Lower tier assessment efforts can accommodate the importance of sampling this habitat by assessing turbidity or chlorophyll a. Periphyton are the broadly accepted standard for assessment of flowing waters, but the objectives of the program, associated with the likely causes and threats to valued ecological attributes, should always be kept in mind.

Dodds, W. K., J. R. Jones, and E. B. Welch. 1998. Suggested classification of stream trophic state: Distributions of temperate stream types by chlorophyll, total nitrogen, and phosphorus. Water Research 32:1455-1462.
Fore, L. S., and C. Grafe. 2002. Using diatoms to assess the biological condition of large rivers in Idaho (USA). Freshwater Biology 47:2015-2037.
Kelly, M. G., and B. A. Whitton. 1995. The trophic diatom index: a new index for monitoring eutrophication in rivers. Journal of Applied Phycology 7:433-444.
Lowe, R. L., and Y. Pan. 1996. Benthic algal communities and biological monitors. Pages 705-739 in R. J. Stevenson, M. Bothwell, and R. L. Lowe, editors. Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, California, USA.
Papista, E., E. Acs, and B. Boddi. 2002. Chlorophyll-alpha determination with ethanol - a critical test. Hydrobiologia 485:191-198.
Potapova, M. G., D. F. Charles, K. C. Ponader, and D. M. Winter. 2004. Quatifying species indicator values for trophic diatom indices: a comparison of approaches. Hydrobiologia 517:25-41.
Stevenson, R. J. 1996. An introduction to algal ecology in freshwater benthic habitats. Pages 3-30 in R. J. Stevenson, M. Bothwell, and R. L. Lowe, editors. Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, California, USA.
Stevenson, R. J., and L. L. Bahls. 1999. Periphyton protocols. Pages 6-1 through 6-22 in M. T. Barbour, J. Gerritsen, and B. D. Snyder, editors. Bioassessment Protocols for Use in Wadeable Streams and Rivers: Periphyton, Benthic Macroinvertebrates, and Fish, Second Edition. U. S. Environmental Protection Agency, Washington, D.C.
Stevenson, R. J., and Y. Pan. 1999. Assessing environmental conditions in rivers and streams with diatoms. Pages 11-40 in E. F. Stoermer and J. P. Smol, editors. The diatoms: applications for the environmental and earth sciences. Cambridge University Press, Cambridge, UK.
Stevenson, R. J., S. T. Rier, C. M. Riseng, R. E. Schultz, and M. J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:140-165.
Stevenson, R. J., and J. P. Smol. 2002. Use of algae in environmental assessments. Pages 775-804 in J. D. Wehr and R. G. Sheath, editors. Freshwater Algae in North America: Classification and Ecology. Academic Press, San Diego.
Wang, Y. K., R. J. Stevenson, and L. Metzmeier. 2005. Development and evaluation of a diatom-based index of biotic integrity for the Interior Plateau Ecoregion. Journal of the North American Benthological Society 24:990-1008.
Welch, E. B., J. M. Jacoby, R. R. Horner, and M. R. Seeley. 1988. Nuisance biomass levels of periphytic algae in streams. Hydrobiologia 157:161-168.

Review:

Jan gives a thorough review of current periphyton methods. As a measure of aesthetics in a stream, algal biomass receives more attention than algal species composition. Most states are interested in establishing a threshold/criterion for periphyton impairment. While level of algal impairment (biomass) has been proposed by several authors (Welch et al. 1988, Dodds et al. 1997, Biggs 2000, Dodds et al. 2002), it is also important to realize that level of impairment could vary among different regions and stream systems. It should be cautious to adopt numbers from other ecoregions or even from New Zealand streams. I also recommend sampling algal biomass in a stream reach in combining with RBP habitat assessment and pebble count so that the proportion of periphyton biomass in different habitat can be estimated. At least for the purpose of measuring algal biomass, a consistent method should be used in one study.

Biggs, B.J. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll relationships for benthic algae. J.N. Am. Benthol. Soc. 19: 17-31.
Welch, E.B., J.M. Jacoby, R.R. Horner, and M.R. Seeley. 1988. Nuisance biomass levels of periphytic algae in streams. Hydrobiologia 157: 161-168.
Dodds, W.K., V.H. Smith, and B. Zander. 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: A case study of the Clark Fork River. Wat. Res. 31: 1738-1750.
Dodds, W.K., V.H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relatiosnships to benthic algal biomass in temperate streams. Can. J. Fish. Aquat. Sci. 59: 865-874.

Review:

What is the preferred or recommended periphyton assessment method?
What is the comparability between the methods? What is most effective for the unit of effort?

I am unaware of any literature article that attempts to provide a comprehensive comparison of assessment methodologies for periphyton. I am assuming that the assessment methodology that is the basis of this request focuses on the overall approach and not field and lab methods associated with collection, preservation and analysis.
One state approach that I feel has real merit and has been in use long enough to receive some positive feedback is Montana ’s use of periphyton in their bioassessment of streams (Bahls 1993). Bahls covers site selection and indicates that only riffles are used as collection sites and states that reference comparisons must use only data from sites of located on the same stream order as defined by Strahler (1957). Montana bioassessment data is based on material collected only on natural substrates and lists a number of reasons why artificial substrates should not be used in bioassessment efforts. I would add that the use of the predominant substrate in a stream accounts for the natural colonization and accrual limitations associated with differing substrates and their venerability to scour. Bahls also covers the rationale for their selection of an indexing period (June 21 to September 21) stating that periphyton diversity peaks in summer and early fall, stream flows are most stable then and fieldwork is more easily accomplished.
Stevenson and Bahls (1999) provide rapid bioassessment protocols for periphyton collection from natural substrate single habitat sampling but this article falls short of defining a methodology beyond collection periphyton from large substrate and preparing sample for taxonomic identification. Taxonomic identifications are necessary to characterize the periphyton community and quantify algal community metrics that can be related to water quality and other waterbody conditions. It should be noted that most bioassessment programs and assessment studies use community metrics and metrics based on tolerance values for certain taxa to evaluate aquatic conditions (e.g. Descy 1979, Bahls 1993, Kentucky Division of Water 1993, Oklahoma Conservation Commission 1993, Dixit et al. 1992, Winter and Duthie 2000, Fore and Grafe 2002). Some use chlorophyll concentrations as a estimate of biomass and as a potential “effects” variable indicative of eutrophic conditions that might lead to unacceptable change in ecosystem structure or function, and loss (impairment) of designated uses (Bahls 1993).
Lastly, I would strongly recommend reading the “New Zealand Periphyton Guideline: Detecting, monitoring and managing enrichment in streams” prepared by Barry Biggs (Biggs 2000). This publication addresses all the essential aspects of a periphyton based assessment program from their importance in stream ecosystems, controlling factors, effects on stream uses, periphyton as environmental indicators and guidelines for protection of uses.
    
Is one method more appropriate for determining impacts for aquatic life versus aesthetics?

I don’t have a good answer for this except to say that macro-colonies (e.g. amts and treads) and macroalgae (e.g. Chara) can be an encumbrance to recreation uses and can detract from more aesthetical displeasing stream conditions. I would suspect that large visible growns of algae also could contribute to imparted aquatic life.

Which stream substrate (riffles, soft sediments, or a composite of the two) is preferable for sampling?

The type of substrate that is sampled for periphyton varies greatly with most assessment efforts utilizing large, inorganic substrates (e.g. cobble) when present. Our work with co-sampled substrate types (sand and silt verses cobble) indicated that there was little relationship between the two chlorophyll concentrations (see following scatter plot). Region 7 RTAG felt that the substrate sampled show be reflective of the prominent substrate which in tern would reflect actual stream conditions.

What is the preferred or recommended extraction method?
According to Sartory and Grobbelaar (1984) the alcoholic solvents, methanol and ethanol, proved to be superior to acetone and acetone with DMSO in extraction of chlorophyll pigments. Homogenization and sonication did not improve the extraction in the alcoholic solvents. Boiling at 100°C had an adverse effect whereas complete extraction of the pigments was obtained at the solvents boiling point and allowing the samples to stand for 24 h in the dark. Our experience with methanol extraction of algal samples after freezing of the filtered sample has been that this method allows for consistent and almost complete extraction of pigments from our algal samples. We filter the sample on to a G/F filter, place it in a labeled scintillation vial and freeze the samples for at least 48 hours.
    

Turner Designs ( http://www.turnerdesigns.com/t2/esci/chlqa.html) [^] state the following “The most commonly used extraction solvent is a 90% acetone 10% DI water solution. Other solvents, such as methanol, ethanol and acetone/DMSO mixtures are also commonly used and can improve extraction efficiency with specific phytoplankton or may be found useful for the extraction of sediment samples. There is no 'best' solvent or procedure for chlorophyll extraction. Several work well and have their own pros and cons. The EPA Method 445.0 describes the recommended step-by-step process for analysis using 90% acetone. There are many factors in the extraction process that can lead to different results.”

Which method best corresponds to the method used by Welch 1988 when he posed the 100, 150 mg/ m^2 thresholds for nuisance algae?

The method used by Welch 1988 is that of Carl Lorenzen (Lorenzen 1967) who investigated the use of a spectrophotometric method that would not be susceptible to the typically small error that can be introduced due to the presence of moderate to high amounts of chlorophyll b, which shows an increase in fluorescence when the sample is acidified. This issue was that algae containing chlorophyll b (mostly we are interested in measuring chlorophyll a and phaeophytin a) generally have an acid factor slightly less than those without chlorophyll b, this can lead to a slight underestimate of chlorophyll a and a slight overestimate of phaeophytin pigments. This is usually not a problem in taxonomically diverse samples. In summary Lorenzen recommended filtering with a GF/C glass filter (we recommend a GF/F filter), extraction with acetone during maceration of the cells in a Teflon tissue grinder, final volume adjusted to 10 ml from washing and then centrifuged after 30-60 minutes. He then recommends the specific wavelengths (750 and 665 m&956;) to use in his equations. A method very similar to Lorenzen’s method appears in Standard Methods (APHA 1995) and is described in Section 10200 H.

Citations

American Public Health Association, American Water Works Association, and Water Environment Federation, Washington, D. C.

APHA. 1995. Standard Methods for the Examination of Water and Wastewater, 19th Ed.

Bahls, L.L. 1993. Periphyton bioassessment methods for Montana streams. Revised edition, January 1993. Water Quality Bureau, Montana Dept. Environmental Quality, A-206 Cogswell Bldg., 1400 Broadway, Helena, Montana 59620. 23 pp + figures, tables and appendices.

Descy J-P. 1979. A new approach to water quality estimation using diatoms. Nova Hedwigia 64: 305-323.

Dixit, S.S., J.P. Smol, J. C. Kingston, and D.F. Charles. 1992. Diatoms: Powerful indicators of environmental change. Environmental Science and Technology 26:23-33.

Fore, L.S. and C. Grafe. 2002. Using diatoms to assess the biological condition of large rivers in Idaho (U.S.A.) Freshwater Biology (2002) 47, 2015–2037

Kentucky Division of Water. 1993. Methods for assessing biological integrity of surface waters. Kentucky Dept. of Environ. Protection, Frankfort, KY.

Lorenzen, C.J. 1967. Determination of chlorophyll and pheo-pigments: spectrophotometric equations. Limnology and Oceanography. 12:343-346

Oklahoma Conservation Commission. 1993. Development of rapid bioassessment protocols for Oklahoma utilizing characteristics of the diatom community. Oklahoma Conservation Commission.

Sartory, D.P. and J.U. Grobbelaar. 1984. Extraction of chlorophyll a from freshwater phytoplankton for spectrophotometric analysis. Hydrobiologia 114: 177-187.

Stevenson, R. J. and L. Bahls. 1999. Chapter six: periphyton protocols. Rapid bioassessment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates, and fish, 2nd edition. (eds. Barbour, M.T., J. Gerritsen, B.D. Snyder, and J.B. Stribling). EPA 841-B-99-002. U.S. Environmental Protection Agency, Office of Water, Washington, DC.

Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. American Geophysical Union Transactions. 38: 913-920.

Winter, J. G. and H. C. Duthie. 2000. Epilithic diatoms as indicators of stream Total N and Total P concentration. Journal of the North American Benthological Society 19:32-49.

Text added from Word file by KPavlik. If you cannot see the figure, click on the Word attachment above.

Review:

The reply and reviews are quite thorough. Periphyton can be measured largely based on some estimator of biomass, usually chlorophyll, or by examining the community for indicator species or communities. Each technique has its advantages.

Chlorophyll techniques are varied in both phytoplankton and periphyton studies. Until recently acetone was the recommended solvent and it is still recommended in Standard Methods. There appears to be a gradual shift to alcohols since the 1980’s because the grinding step is eliminated assuring less time spent per sample and eliminating the potential exposure to acetone. Methanol has safety issues and problems with the determination of phaeophytin a. We use hot ethanol extraction in our laboratory.

Measurement of chlorophyll and phaeophytin using acidification and the Lorenzen equations does not necessarily produce an accurate estimation of these pigments because of interfering pigments such as chlorophyllides and chlorophylls b, and to some extent, c. Trichromatic techniques do not address phaeophytin interference. Only HPLC techniques can deliver an accurate chlorophyll measurement, but that accuracy is usually not necessary for bioassessment. However, because of the mixed algal community and detrital rain on that community, chlorophyll estimates should be adjusted by removal of phaeopigment interference.

A key criterion to method may be techniques previously used. If periphyton chlorophyll has been previously measured, some caution should be exercised in switching techniques. Differences in results will invariably be found and for reasons such as differential extraction, cross calibration, though necessary, may not result in a 1:1 correlation of technique.

Although it is probably too obvious to mention, the place of sampling can have an effect on periphyton biomass. The technique would be useless in a shaded area, and may be compromised in partially shaded locations. Water velocity will affect growth and biomass, as will the species of algae.

It may be advantageous to consider qualitative measures of algal extent as was suggested earlier.

Response:

This is a difficult question, and I am aware of little information on this point.

There are data that demonstrate a low number of species occur at elevated pH at the time of sampling (see Dodds 2002 Freshwater Ecology).

There has been a good bit of work on short-term acidification to simulate acid snowmelt events. The main effect of these acidification episodes seems to be increased invertebrate drift.

Short-term high pH is unlikely to have much influence on photosynthetic organisms since they should be adapted to high pH locally when active.

As the question mentions, high ammonia can result from elevated pH so this is one area of concern.

Review:

This is not too surprising because there is not even much information on effects of short-term exposure of low DO (a really common stressor) on invertebrates and fish.

Review:

Although not as well-researched as acidity, there is a body of information documenting that pH > 9.0 is stressful to many common aquatic organisms (not those specifically adapted to high pH, of course). I searched Google scholar and found the abstracts attached. Some observations:

High pH alone can be stressful to some organisms.

Most metals are soluble (and toxic) under acidic conditions and precipitate out of solution at neutrality; however, a smaller subset are also re-dissolved at high pH, notably aluminum, and possibly arsenic and selenium. Thus, high pH could induce metal toxicity under certain conditions.

If pH excursions above 9 due to photosynthesis are common, then it would seem that the waterbody is highly eutrophic with a host of potentially undesirable outcomes: pea-soup conditions, cyanobacteria blooms, severe DO fluctuations, potential ammonia toxicity, as well as pH stress.

EPA Note:

The response to this question is a technical one. Policy implications - such as allowable duration and frequency are beyond the scope of T-REQs but will be dealt with by EPA's Implementation Workgroup. Answers to the Workgroup questions will be posted on N-STEPS as they become available.

Response:

Response variables such as benthic chlorophyll and dissolved oxygen should certainly be considered when evaluating nutrient criteria. Attached ash free dry mass is not as good of an indicator because it is so variable and can be related to carbon inputs from outside the stream.

Exactly how to use the response variables is somewhat situationally dependent. Dissolved oxygen is relatively easy because most states have levels of dissolved oxygen in which a violation has occurred if it drops below some critical level. There are not existing and generally accepted values for what is too much benthic chlorophyll.

Low DO excursions are most likely to happen with high algal biomass, but here are not good direct relationships published for instream nutrients and stream metabolism. There are relationships published for instream nutrients and benthic chlorophyll concentrations. Thus we try to set endpoints for algal biomass and nutrients based on some level related to the historic or reference condition, usually with some allowance for increases over pristine conditions.

The really clear cases of low DO occur in very slow rivers where planktonic chlorophyll values are very high and aeration rates are low.

DO monitoring in particular should focus on times with low stable discharge and higher temperatures because this is when there is the greatest probability for low DP excursion.

Establishing an extensive database of reference values for DO, benthic chlorophyll and nutrient levels is the first step in establishing the nutrient criteria. This way it is possible to know roughly the conditions that support biotic integrity. A related database should be established for reference sites. Mean total N, total P, and benthic and planktonic chlorophyll, as well as minimum and maximum DO, and maximum chlorophyll values are desirable.

The entire database can be used to determine the relationships between water column total N and P and the response variables of interest. If there are no reference sites available, then there are several methods possible to extrapolate to reference conditions:
Dodds, W. K. and R. M. Oakes. 2004. A technique for establishing reference nutrient concentrations across watersheds impacted by humans. Limnology and Oceanography Methods 2:333-341.

Once nutrient criteria are set, the driver and response variables should still be monitored to assess the progress related to nutrient control.

Review:

Responses of algal biomass and dissolved oxygen should be considered in the development of nutrient criteria and as variables for criteria because they are or are more closely relate do “problems” than nutrients. Why regulate nutrients? Who cares about nutrients unless they cause a problem or are at sufficient concentrations to cause a problem? We care about nutrients because they stimulate algal and plant growth, which can cause low dissolved oxygen and reduced aquatic life support, both respect to biodiversity and productivity when. In addition, algae alter habitats with may also affect aquatic life. Nutrients also stimulate algal growths that cause reduced water clarity, aesthetically unpleasing and nuisance growths in both water column and on the bottom substrata, impair taste and increase odor in drinking water, and toxins in drinking water and food supplies. Thus algal biomass and dissolved oxygen, as well as other parameters that can be related to valued ecological attributes (VEAs), should be considered in the development of nutrient criteria.

How use response variables in criteria development? Their response to nutrients should be quantified and thresholds in responses should be determined to help justify specific nutrient concentrations as criteria. Water bodies with “acceptable” levels of algal biomass, dissolved oxygen, aquatic life use, and other VEA parameters can be selected to define as reference condition and develop nutrient criteria with the frequency distribution approach. Here “acceptable” should be related to natural, minimally disturbed, best attainable, or best available conditions, depending upon management goals.

Algal biomass and dissolved oxygen (as well as other VEAs) should be included as parameters having criteria in water quality standards because they are explicitly related to “uses”; plus having criteria for multiple variables provides a more certain assessment of condition in the habitat than just one or two variables (e.g. total phosphorus). In addition, including criteria for algal biomass and dissolved oxygen in water quality standards communicates resource management goals more directly to the public than standards with just nutrients. Although gathering data for additional analyses and these analyses may seem like more trouble in the short term, they will provide a more complete set of water quality standards in the long term that generate better public support for the standards and more certain assessment of whether standards to support uses are being met.

Stevenson, R. J. and J. P. Smol. 2003. Use of algae in environmental assessments. In: J. D. Wehr and R. G. Sheath, eds. Freshwater Algae in North America: Classification and Ecology. Pp. 775-804. Academic Press, San Diego.

Stevenson, R. J., B. C. Bailey, M. C. Harass, C. P. Hawkins, J. Alba-Tercedor, C. Couch, S. Dyer, F. A. Fulk, J. M. Harrington, C. T. Hunsaker, and R. K. Johnson. 2004. Designing data collection for ecological assessments. In: M. T. Barbour, S. B. Norton, H. R. Preston, and K. W. Thornton, eds. Ecological Assessment of Aquatic Resources: Linking Science to Decision-Making. Pgs 55-84. Society of Environmental Toxicology and Chemistry, Pensacola, Florida. ISBN 1-880611-56-2.

Stevenson, R. J., B. C. Bailey, M. C. Harass, C. P. Hawkins, J. Alba-Tercedor, C. Couch, S. Dyer, F. A. Fulk, J. M. Harrington, C. T. Hunsaker, and R. K. Johnson. 2004. Interpreting results of ecological assessments. In: M. T. Barbour, S. B. Norton, H. R. Preston, and K. W. Thornton, eds. Ecological Assessment of Aquatic Resources: Linking Science to Decision-Making. Pgs 85-111. Society of Environmental Toxicology and Chemistry, Pensacola, Florida. ISBN 1-880611-56-2.

Stevenson, R.J., S.T. Rier, C.M. Riseng, R.E. Schultz, and M.J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:149-165.

Stevenson, R.J. 2006. Refining diatom indicators for valued ecological attributes and development of water quality criteria. In: Ognjanova-Rumenova, N. And K. Manoylov, eds. Advances in Phycological Studies. Pp. 365-383. Pensoft Publishers. Moscow, Russia.

Review:

I think the initial response and first review really hit the salient points. The critical issue is response variables can and should have criteria developed for them as well. One of the key points from the reviewer is that due to the many indirect ways by which nutrients impact aquatic systems, the response criteria are often more clearly related to use impairment than are nutrients themselves. For this reason, these variables are used in effects-based approaches for deriving protective nutrient concentrations. That is the first way response variables are incorporated into nutrient criteria – they are used to derive appropriate values. But, as clearly recommended in the guidance documents, criteria for response variables should also be developed or at least considered (e.g., benthic chlorophyll and turbidity in streams and planktonic chlorophyll and secchi depth in lakes). The use of response variables has many advantages in addition to their direct linkage to use impairment. Nutrients are also hard to sample at the temporal and spatial scales necessary to detect problems. Response variables, however, integrate the episodic and dynamic nature of nutrient behavior and often give a clearer picture of when problems exist. This is akin to the advantages of using biocriteria in addition to chemical water quality criteria. The second way response variables can be incorporated into nutrient criteria is that they can have criteria developed for themselves as well.

How can nutrients and response criteria be integrated? Different approaches can be taken. For example, one could require that all criteria (nutrient and response) be met for a waterbody to meet its use. Certainly, this would be most protective. However, there are situations, even natural ones, where this condition may not be attainable and, yet, uses are protected. Another option, therefore, is to identify that when either nutrient or response criteria are not met, a site is triggered for closer scrutiny and further sampling to assess whether uses are being maintained, and then a decision made re: nutrient criteria attainment. If response variable criteria are not being met but nutrient criteria are, this process could also include an analysis of whether nutrients are the cause of the failure to attain the criteria for the response variable. A third option may be to use two tiers of nutrient criteria, low and high, for example. Below the low concentrations, all uses are known to be protected; above the low criterion but below the high nutrient criterion, response criteria are triggered to assess whether a nutrient related problem exists that threatens uses; and, lastly, the high criterion is set so that any value above it would be known to have an unacceptably high risk to causing problems either in situ or downstream, and therefore, a site is listed.

Response:

A number of factors can affect the basic relationships between nutrients and algal responses in streams. However, these relationships are not easily defined nor can many general statements be made about how individual or groups of similar streams will respond to factors such as light attenuation, non-algal turbidity, stream velocity, scour and substrate size and stability. While it is appropriate to consider these in the evaluation of factors to be incorporated into nutrient criteria, it will not be easy to deal with all of these factors and arrive at a single value for regional streams. Several examples will be provided to illustrate the difficulty in using multiple factors to derive criteria.

Variance of concentrations and processes are remarkable in streams, over time. For example, monitoring of a local stream daily for nearly a year demonstrated that total phosphorus values varied between 16 ug/L and 1597 ug/L -- that's two orders of magnitude in a single system.  Total nitrogen ranged between 150 and 2700 ug/L. Suspended chlorophyll ranged between 0.2 and 58 ug/L.

Empirical, cross-system comparisons also show that the planktonic chlorophyll content per unit of phosphorus in stream water increases with the residence time in the stream (large streams at a given phosphorus concentration have more chloropyhll than small streams with the same phosphorus content). The process that accounts for this difference is "time" in the system. It allows for algae to replicate within the stream water and for cells to move from the sediments into the water column.  Further, many researchers report weak relationships when looking at N or P and benthic chlorophyll a (Lohman, Jones and Perkins 1992, Biggs 2000, Dodds, Smith and Lohman 2002 and Bernhardt and Likens 2004).
Some authors (Findlay, Howe and Fontvieille 1993, Rosemond, Mulholland and Elwood 1993, Hill 1996) have noted that light and herbivores can limit algal biomass even in the presence of high nutrients. The total concentrations of benthic chlorophyll a from benthic algae taken from glass slides allowed to colonize in the small streams mentioned above, showed no relationship with canopy cover (percent shade). Using robust regression, a non-significant relationship (p=0.077) was found between 554 benthic chlorophyll samples and average % canopy (measured twice during the colonization period). Percent canopy values ranged from zero to 100% and average TP values were 0.117 mg/L. These low gradient streams flow through fairly open agricultural landscapes and riparian boundaries are typically narrow so even though canopy cover was 100% in many areas there was still a lot of incidental light reaching the streams. Measures of canopy cover may reflect the amount of direct light that reaches a stream but not the amount of total light as reflected by the lack of a relationship between periphyton accrual and % canopy. Available data do not support the idea that riparian cover can affectively reduce or limit benthic algae or for that matter sestonic chlorophyll a. Looking at a robust regression analysis for sestonic chlorophyll a concentrations and % canopy, there was a highly significant, negative relationship (n=521, p=0.00004) but the r 2 value was only 0.032 suggesting that light was not a controlling factor in these streams regardless of the amount of local riparian cover.

The table below is from the EPA Region 7 RTAG findings and suggests that excessive algal biomass related to nutrients are more likely a problem in larger streams where light attenuation is not a possibility and flow regimes may be more conducive to the production and maintenance of sestonic algae.

The results in this table are from robust regressions run for selected stressors, classification and response variables using all sites having one or more values for both the dependant and independent variables (i.e. no data filters used). All analysis preformed using the NCSS ® database entitled Chem2005Sep.SO. This database is currently maintained by the Central Plains Center for BioAssessment and is available on request (www.cpcb@ku.edu). NCSS ® is a statistics software package that we use in most of our statistical testing.

Riparian cover should be maintained or replaced where appropriate but is not a substitute or reliable modifier for acceptable nutrient levels. However, stream classifications (e.g. watershed size as an indicator if stream order or size) might be appropriate in establishing criteria.

Also, a significant relationship was not found between chlorophyll a and substrate sizes. Chlorophyll concentrations in and on sand and silt can get nearly as high as concentrations found on large gravel and cobble but the instability of small particles leads to more frequent and perhaps more severe scour losses. Smaller, more frequent runoff events often possess enough energy to move and entrain smaller particles and thus disrupt algae grown and accrued.

Finally, benthic algal biomass accrual has been shown to be related to both drying events and floods of sufficient power as to cause streambed scouring (Biggs 1996 and 2000). However, measurable benthic accrual can take place in as little as 6 days in fourth and fifth order streams (80 stream sites measured 10 times/each) in Kansas, Nebraska and Iowa (mean, median and range of benthic chlorophyll a was 1.79, 0.72 and 0 to 14.9 mg/ M 2, respectively). Scour does reduce the accrual of benthic algal biomass but excessive algal growth can still occur during the critical low or normal flow periods of the year when high stream temperatures and low flows can contribute to lower DO values that might be acerbated by high algal productivity and respiration.

Given the difficulties with accounting for light, non-algal turbidity, temperature, stream velocity, stream width, depth, ph, temperature, or hardness in setting nutrient criteria across systems, it is probably better to establish mean criteria for nutrients without considering these factors.  If consistent relationships can be found with these variables as independent drivers of benthic or sestonic chlorophyll using statistical models within a region, then they could be taken into account.  However, our experience suggests that statistically significant relationships will not be established, but that a surrogate for variation in these and other variables, ecoregion, may yield results that can be used to set specific nutrient criteria for some regions.

This answer was drawn heavily from a response to a similar question through EPA’s nutrient technical assistance website.   Please see that response for further information and examples. (http://n-steps.tetratech-ffx.com/Q&A-StudyDesign.cfm). 
Bernhardt, E.S. and G.E. Likens, Controls on algal biomass in heterotrophic streams, Freshwater Biology, vol. 49 (2004), pp. 14-27.
Bernhardt, E.S. and G.E. Likens. 2004. Controls on periphyton biomass in heterotrophic streams. Freshwater Biology 49: 14-27.
Biggs, B.J.F. 1996. Patterns in benthic algae of streams. Pages 31–56 in R. J. Stevenson, M. L. Bothwell, and R. L. Lowe (editors). Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.
Biggs, B.J.F. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll relationships for benthic algae. J. North American Benthological Society 19:17-31.
Dodds, W.K., V.H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences. 59:865-874.
Findlay , S., K. Howe, D. Fontvieille. 1993. Bacterial-algal relationships in streams of the Hubbard Brook experimental forest. Ecology 74: 2326 2336.
H ill , W.R. 1996. Effects of light. Pages 121–148 in R.J. Stevenson, M. L. Bothwell, and R. L. Lowe (editors). Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.
Lohman K., J.R. Jones & B.D Perkins. 1992. Effects of nutrient enrichment and flood frequency on periphyton biomass in Northern Ozark streams. Canadian Journal of Fisheries and Aquatic Sciences 49, 1198–1205.
Rosemond, A.D., P.J. Mulholland, J.W. Elwood 1993. Top-down and bottom-up control of stream periphyton. Ecology 74: 1264-1280.

Review:

The above response is excellent. I would add that criteria are being developed, in one sense, to control risk to beneficial uses associated with eutrophication to waterbodies. In that context, it is unlikely that one number will be perfect for every site – but, properly developed, it will be generally protective of most sites. Clearly, site specific factors will contribute to the behavior of chlorophyll and nutrients for any one site, creating the variability and explaining that residual variance referref to. One could argue that distribution derived approaches incorporate this site-specific variability into the value selected, thus its value in the weight of evidence approach for criteria development. In addition, measuring parameters during specific index periods (for example, late summer low flow periods) may reduce some of this variability.

Developing rules for the variety of controlling factors would be difficult. The rules for pH- or hardness-dependent criteria for some chemicals are based on well established laboratory toxicity experiments and chemical laws. Such clearly defined relationships have not yet been defined. Some preliminary work incorporating the effects of time of accrual (cited by Dodds above), grazers, and available light have been developed but hardly to the level one would consider sufficient for making general rules for nutrient-chlorophyll relationships nationally.

The downstream effects issue cannot be overlooked as an important consideration. Clearly, shade, turbidity, etc. will lower the chlorophyll response to nutrients at a given point. However, these nutrients will be conveyed downstream, the distance depending on flow and instream processing. They will eventually reach a segment or waterbody with increased clarity and residence time – where the risk associated with these nutrients will be realized. Criteria must consider this effect and developing exceptions for each reach based on site-specific factors loses sight of the watershed/landscape context within which nutrients affect aquatic systems.

Review:

Stream criteria are going to be difficult to frame. Stream flow and the size of the stream/watershed must be considered. The various responses and comments on this topic reflect the state of our understanding of stream processes. I suggest all parties read the paper by Van Nieuwenhuyse and Jones (1996) for an understanding of how suspended chlorophyll levels in streams are influenced not only by phosphorus by stream catchment area. The material covered is not going to clarify how to set stream criteria but it illustrates some major aspects that deserve consideration.

see-
Van Nieuwenhuyse and Jones. 1996. Phosphorus-chlorophyll relationship in temperate streams and its variation with stream catchment area. Can. J. Fish. Aquat. Sci. 53:99-105.

Response:

This is a difficult question to answer given the amount of information provided. We know that in most cases if there are ample nutrients, stable hydrologic conditions, lack of scour, a stable attachment surface, low grazing pressure, and a good amount of light, a sizable algal biomass will develop. If all these conditions are present and there is still relatively small attached algal biomass, then other explanations would need to be sought. I can think of several potential explanations, and these can be related to work published by Tank and Dodds (2003). In some experiments, high levels of phosphate actually inhibit algal accrual. Mechanisms for this are not well established. When ample carbon is supplied, bioassay experiments reported in Tank and Dodds (2003) demonstrate that heterotrophic stimulation by nutrients (accrual of fungal biomass) can apparently inhibit an algal response. This observation is consistent with the consultant’s hypothesis. A final possibility is that some sort of toxic effect is keeping algal biomass low. If, for example, a herbicide such as atrazine is in the water, then photosynthetic organisms would be inhibited. High or low pH, ammonium, sulfide or some other condition could inhibit algal growth.

I suspect that it is more likely that conditions at the proposed site (e.g., frequent floods, shading, excessive turbidity, high concentrations of grazers, or fine sediment or sand bottom) lead to lower mean algal biomass than expected by general models such as those presented by Dodds et al. (2002, 2006). It is important to keep in mind that those models only explain slightly more than a third of the variance in nutrient-chlorophyll relationships. Mean values at baseflow provide the best overall results, and any sampling strategy should sample at least 3 times across a season, and avoid sampling for a few weeks after each scouring flood. Sampling should also include lighted areas and take into account that the plume will move downstream a considerable distance if there is not much algal activity to incorporate it.

Dodds, W. K., V. H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Can. J. Fish. Aquat. Sci. 59:865–874.

Dodds, W. K., V. H. Smith and K. Lohman. 2006. Erratum: Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Can. J. Fish. Aquat. Sci. 63:1190-1191.

Tank, J. and W. K. Dodds. 2003. Responses of heterotrophic and autotrophic biofilms to nutrients in ten streams. Freshwater Biology 48:1031-1049.

If the sampling was limited, the low chlorophyll to nutrient ratio could be related to error from too small of a number of sample events, or if the individual number of rocks sampled was low per sampling event, choosing some low chlorophyll rocks.

Review:

Low biomass in streams may be caused by high grazer densities, low light, and lack of large stable substrata, as well as heterotrophic bacteria and toxins of many forms in urban settings. Given those alternative hypotheses, yes, the hypothesis that heterotrophic bacteria and fungi may be inhibiting algal accrual on substrata is realistic. This may occur by at least three mechanisms. Competition for nutrient resources, allelopathic interactions, and bioturbation (physical disturbance) is possible. All have been hypothesized in association with observations of various types, such as those described by the first reviewer and in Rier and Stevenson (2002). Bacteria may outcompete algae for nutrients and secrete allelopathic substances that inhibit algal growth. In addition, filamentous bacteria like Sphaerotilus can grow to substantial lengths and prevent alga from accumulating by brushing algae of substrata as its filaments wave back and forth in the current.

My personal observations indicate that bacterial or fungal interactions with algae may be greatest during high temperature, summer conditions. Bacteria tend to have higher tolerance to high temperatures (e.g. 30° C) than eukaryotic algae. In Beargrass Creek, an urban stream passing through Louisville, Kentucky, benthic algae do not accumulate during low flow conditions during summer even though nutrients are abundant and grazers are rare. The algae looked like they were sick. These observations lead to including bacteria in a survey of algae and bacteria in different nutrient conditions (Rier and Stevenson 2001), and an experiment on algal bacterial interactions (Rier and Stevenson 2002). More recent work by graduate students in Dr. Leff’s lab at Kent State University and Dr. Feminella’s lab at Auburn University indicate algal-bacterial interactions may regulate algal performance.

 So yes, algal-bacterial interactions may regulate algal density in streams. But a simple ratio of AFDM-chl a is not sufficient evidence to “conclude” that algae are regulated by other bacteria.

Rier, S. T. and R. J. Stevenson. 2001. Relation of environmental factors to density of epilithic lotic bacteria in 2 ecoregions. Journal of the North American Benthological Society 20: 588-600.
Rier, S. T. and R. J. Stevenson. 2002. Effects of light, dissolved organic carbon, and inorganic nutrients on the relationship between algae and heterotrophic bacteria in stream periphyton. Hydrobiologia 489:179-184.

Review:

The previous respondents provide interesting views, most of which I concur with. I must admit there is little in the literature to recommend complete reach or stream scale exploitative competition of bacteria over algae. Bacteria have been found to have pretty competitive nutrient uptake rates, but to exclude algae as a result, is hard to imagine, without other extenuating factors and certainly not any natural extenuating factors other than the ones mentioned by the previous respondents (light, substrate, flow, grazers, etc.) But if none of those factors are limiting in a natural system, algae and bacteria would co-exist and algae would increase with nutrients. If none of the above factors are limiting and there are no algae in a stream, I would be concerned about other factors of great concern (pesticides were mentioned, high organic matter loads leading to excess Sphaerotilus - which is why they are called sewage fungi (even though they are not fungi but filamentous bacteria), low dissolved oxygen, etc.). I have never heard of bacterial competition explaining lower than expected algal concentrations, since every stream contains abundant bacteria, including the ones used to generate nutrient-algal relationships. I am hedging towards one the above limiting factors or something unknown.

I also agree that the AFDM:chlorophyll ratio is an unreliable way to deduce this observation as it can result from any of a number of reasons. I would recommend much more targeted sampling of the algal community, investigation of the level of natural factors known to limit algae mentioned in the responses (light, substrate, grazers, etc), and if these show nothing, looking towards other potential contaminants (pesticides, excess OM loading, low DO, metals).

Response:

There is no literature that I am aware of specifically targeted toward setting lower nutrient criteria for filamentous algae in cold water trout streams. However, there is some justification for setting such criteria.

One of the first rivers where nutrient criteria became an issue was the Clark Fork River in Montana (Dodds et al. 1997), where excessive Cladophora growth interfered with aesthetic uses and also was unpopular with anglers. The filaments can clog lures and trout ingest filaments with invertebrates and acquire an off taste from the algae in their guts (Dodds and Welch 2000). In my experience this problem was even worse in the Madison River, an even more eutrophic Montana river. In the summer at low flow this river has fish kills related to DO sag and high temperatures.

I am not aware of any data indicating that community structure tilts toward or away from filamentous algae with greater nutrient concentrations, so criteria should be set at an attainable chlorophyll level (total algal biomass). Equations for these relationships are described in Dodds et al. (2002, 2006).

A side note, filaments take a while to develop and seem to be worst at extended periods of low flow (cold or hot, some impressive filamentous accumulations occur in winter ice-free low flow conditions). If there is control on hydrology, an occasional spate my help keep down filamentous algae, and maintaining natural flow regime may be as important as nutrients. The research by Biggs (2000) substantiates this view.

Biggs, B.J.F. 2000. Eutrophication of streams and rivers: dissolved nutrient-chlorophyll. Journal of the North American Benthological Society 19: 17-31.

Dodds, W. K., V. H. Smith and B. Zander. 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: A case study of the Clark Fork River. Wat. Res. 31:1738-1750.

Dodds, W. K. and E. Welch. 2000. Establishing nutrient criteria in streams. J. No. Am. Benthol. Soc. 19:186-196.
Dodds, W. K., V. H. Smith, and K. Lohman. 2002 Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Can. J. Fish. Aquat. Sci. 59: 865–874.

Dodds, W. K., V H. Smith and K. Lohman. 2006. Erratum: Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams Canadian Journal of Fisheries and Aquatic Sciences 63:1190-1191.

Review:

Algae as a group are relatively tolerant to temperature conditions, with some kind of algae or another are able to tolerate the wide range of temperature conditions that occur in streams. In general, diatoms tolerate the coolest temperatures and cyanobacteria tolerate the highest. The cold temperatures of cold water streams are often optimal for filamentous algal growth during the early summer of the northern US climate. In addition, the optimal temperatures for Cladophora (a common algal bloom taxon that outcompetes other taxa under high nutrient conditions) as well as diatom growth occur in both cold and warm water streams.

Covarying factors other than temperature may be more important determinants of algal biomass in cold water versus warm water streams. Cold water streams are probably groundwater dominated streams, thus their hydrologic stability is likely greater than warm water streams. In a comparison of cold and warm water streams of Michigan and Kentucky, hydrologic stability of the cold water Michigan streams was an important determinant of algal biomass (Riseng et al. 2004, Stevenson et al. 2006). Algal biomass was usually higher in hydrologically stable versus flashy streams because grazers were able to accumulate and control diatoms accrual in hydrologically stable streams. However, filamentous algae such as Cladophora require longer times to accumulate, so the likelihood of having high Cladophora biomass at a given nutrient concentration increased with hydrologic stability (Stevenson et al. 2006). Spates were more frequent in the hydrologically variable warm water streams, and controlled the Cladophora growth. However, high biomasses of Cladophora could accumulate in warm water streams if timing of spates allowed accumulation.

Management endpoints of cold water and warm water streams may also vary. In addition to trout taste, low oxygen effects on trout may be greater than on typical warm water fish taxa. Alteration of habitat by filamentous algae seems to facilitate hydropsychids (a type of caddisfly), so EPT (Ephemeroptera/Plecoptera/Trichoptera - mayflies/stoneflies/caddisflies; trio of sensitive benthic macroinvertebrates often used in biological assessment) metrics often do not detect changes in biological condition unless finer taxonomic levels of resolution are included in metrics. This may stimulate growth of trout at low levels of enrichment. Higher levels of protection may be warranted in cold water streams than warm water streams if management endpoints are more sensitive to nutrients or more important to stakeholders in cold water streams.

Riseng, C. M., M. J. Wiley, and R. J. Stevenson. 2004. Hydrologic disturbance and nutrient effects on benthic community structure in midwestern US streams: a covariance structure analysis. Journal of the North American Benthological Society 23:309-326.

Stevenson, R.J., S.T. Rier, C.M. Riseng, R.E. Schultz, and M.J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:149-165.

Review:

Cold water streams are less frequently reported to have nuisance algal problems due to a number of factors. First, most cold water streams occur in headwaters and lower order stream segments. Water forming these streams are mainly from groundwater and less likely to be affected by human land use. Lower temperature, better canopy cover, and less human land use (nutrients) all contribute to low algal productivity in these headwater streams. Little relationship between nutrients and algal biomass found in wadeable headwater streams, which has been attributed to the influence of canopy shading (Lowe et al. 1986), the frequency of flood disturbance (Lohman et al. 1992), and macroinvertebrate and fish grazing (Bourassa and Cattaneo 1998).

Second, thermal loading into cold water streams is often considered the strongest stressor of cold water streams and it is associated with other stressors (e.g. nutrients). Beneficial use of cold water streams is often designated for game fishing, namely the presence of trout population in streams. Trout are very sensitive to rising temperatures. According to Torgersen et al. (1999), cold water stream temperature increases with distance from headwaters. Increasing stream order is often accompanied by increasing impervious surface which leads to higher urban runoff, including thermal loading, nutrient runoff and sedimentation in streams. Wang and Kanehl (2003) suggested that stream temperature might be one of the major factors through which human activities degrade cold water streams.

Third, the effect of nutrient loading into cold water streams is often masked by other collinear stressors and is overlooked. Other factors, such as habitat degradation, sedimentation, and thermal loading often outweigh the effect of nutrients on trout population. In a recent publication, Wang L. et al. (2007) studied 240 streams (not all cold water) in Wisconsin and found that selected environmental factors explained 54% of the variation in the fish assemblages and 53% of the variation in macroinvertebrate assemblages. Of this explained variance, the majority (>42%) was attributed to catchment and instream habitat, 15-22% to nutrients, 3-5% to other water quality measures, and 32%-36% to the interactions among all the environmental variables.

Finally, global warming may also contribute and confound the effect of nutrients on algal biomass and biological integrity. USEPA is holding a workshop in March to identify how climate change will likely affect bioassessment programs in the United States. According to a recent reference search, global warming and climate change are predicted to raise stream temperatures by 5 to 11° for different regions in the U.S. from 1961 to 2010. A preliminary analysis on cold water reference streams indicated that biological integrity of cold water streams in Maryland tended to decline with rising temperatures, even though the total fish taxa was likely to increase.

Therefore, to develop nutrient and algal biomass criteria, it is important to understand the covariables and collinearity among different environmental stressors. Diagnosis of impairment in cold water streams is far more complicated since rising temperatures could stimulate algal growth and affect trout populations. It is difficult to tease out nutrient and algal effects on biological integrity solely based on large scale field surveys. Future study should focus on experimental manipulation in cold water systems.

Currently, a nutrient TMDL has been developed for cold water streams as well. The nutrient TMDL is generally accompanied by a thermal and sediment TMDL due to co-occurrence. One example is the Snake River watershed in Idaho. The state of Oregon uses a threshold level of less than 0.015 mg chlorophyll a/L (floating algae or phytoplankton?) which when exceeded should trigger a Department determination of impairment of the designated use. The state of Idaho uses a narrative standard for floating algae. The TMDL document is available online at: http://deq.idaho.gov/water/data_reports/surface_water
/tmdls/snake_river_hells_canyon/snake_river_hells_canyon.cfm#SBA [^]

Bourassa N., A. Cattaneo. 1998. Control of periphyton biomass in Laurentian streams. J North Am Benthol Soc 17:420–429

Lohman K., J. R. Jones, B. D. Perkins. 1992. Effects of nutrient enrichment and flood frequency on periphyton biomass in Northern Ozark streams. Can J Fisheries Aquat Sci 49:1198–1205

Lowe R. L., S. W. Golladay, J. R. Weber. 1986. Periphyton response to nutrient manipulation in streams draining clearcut and forested watersheds. J North Am Benthol Soc 5:221–229.

McCormick P. V. 1994. Evaluating the multiple mechanisms underlying herbivore-algal interactions in streams. Hydrobiologia 291:47–59

Torgersen, C.E., D.M. Price, H.W. Li and B.A. McIntosh. 1999. Multiscale thermal refugia and stream habitat associations of Chinook Salmon in northeastern Oregon. Ecological Applications 9: 301-319.

Wang L., J. Lyons, P. Kanehl. 2003. Impacts of urban land cover on trout streams in Wisconsin and Minnesota. Transactions Am Fisheries Soc 132:825–839.

Wang L., D. M. Robertson, P. Garrison. 2007. Linkages Between Nutrients and Assemblages of Macroinvertebrates and Fish in Wadeable Streams: Implication
to Nutrient Criteria Development. Environ Manage (2007) 39:194–212

Response:

Odor and taste in water can be produced by a number of sources (a good introduction is available online by Trojan Technologies, Inc. 2005). The main causes of odor and taste in reservoirs are from chemical release by algae and bacteria. Cyanobacteria are the most notorious algal group that causes odor, taste, and toxic problems in lakes and reservoirs. However, not all cyanobacteria produce odor and taste in water. Microcystis, Anabaena, Aphanizomenon, Pseudanabaena, and Oscillatoria are main cyanobacteria genera causing taste and odor in water supplies. Geosmin (trans-1, 10-dimethyl-trans-9-decalol) and MIB (2-methylisoborneol) are common odorous chemicals that are produced by these cyanobacteria. Simpson and MacLeod (1991) reported that when MIB concentrations are greater than 12 ng/L and geosmin concentrations are greater than 7 ng/L, it is likely to have taste and odor in water. Another group of algae that cause odor and taste problems are the golden flagellates, such as Synura, Mallomonas, and Dinobryon (Lembi 2002). Odor and taste can be quantified using threshold odor number (TON) and flavor threshold test (FTT) (Eaton et al. 1995). A testing method for MIB and geosmin concentrations is provided by Zimmerman et al.(2002).

Although it is generally recognized that eutrophication leads to harmful algal blooms (HAB) , which subsequently cause odor and taste and human health related (e.g. toxins) problems in lakes and reservoirs, few studies have documented the direct association between algal biomass and concentrations of taste and odor compounds and toxins in cyanobacterial blooms in the United States. A study in a Kansas reservoir (Wang et al. 2005) found that nutrients (in particular TN, NO3–N concentrations), turbidity, and hydrologic regime all played potentially important roles in regulating cyanobacterial production. Low levels of nitrogen coupled with the internal release of phosphorus from the lake sediment promoted algal blooms, geosmin production, and most taste and odor events in the lake. In addition to nutrients, other environmental factors may also contribute to odor and taste in reservoirs. A USGS study (Christensen et al. 2006) in Cheney Reservoir, Kansas continuously monitored environmental variables, such as light, temperature, conductivity, and turbidity and modeled real-time estimation of water-quality constituent concentrations and in-reservoir conditions that might result in cyanobacterial production of taste and odor compounds. They used the model to successfully predict when geosmin concentrations would exceed the human detection limit of 10 ng/L. It is expected that the ongoing studies at Cheney Reservoir will link biological, physicochemical, hydrological, and meteorological processes to refine relations to estimate taste and odor occurrences and develop new relations with other variables of concern, such as cyanotoxins.

Another study of interest is one that relates microcystin concentrations (MC) to environmental factors (Graham et al 2004) which may have implications to relations between odor and taste and environmental factors in lakes. Graham et al. investigated 241 lakes in the Midwestern USA. They found that MC-TN and MC-TP relationships were unimodal, but the MC values exponentially declined with N:P ratio. It is not surprising that rising N:P ratio is related to declined cyanobacteria production due to the fact that many toxin and odor producing cyanobacteria prefer nitrogen limited environments so their heterocysts (in Anabaena etc.) can fix nitrogen and reproduce themselves.

A number of factors may confound the relationship between odor and taste and Chl a concentrations. Taxonomic composition of phytoplankton and their spatial and temporal patterns in reservoirs are among the main factors. Phytoplankton communities can shift from cyanobacteria to diatoms and dinoflagellates early spring to late summer without changing in biomass. If none of the odor producing algae dominate in the phytoplankton community, it is not likely to have a sensitive smell and taste in the water. Therefore, lake and reservoir monitoring have to consider algal species compositions along with algal biomass in order to correctly capture the actual causes of water impairment.
  
Eaton, A. D., L.S. Clesceri, A. E. Greenberg (Eds). 1995. Standard Methods for the examination of water and wastewater. 19th edition, published by APHA, AWWA, WEF. Part 2150, 2160.

Christensen, V.G., Graham, J.L., Milligan, C.R., Pope, L.M., and Ziegler, A.C., 2006, Water quality and relation to taste-and-odor compounds in the North Fork Ninnescah River and Cheney Reservoir, south-central Kansas, 1997-2003: U.S. Geological Survey Scientific Investigations Report 2006-5095, 43 p.

Graham, J.L., Jones, J.R., Jones, S.B., Downing, J.A., Clevenger, T.E., 2004, Environmental Factors Influencing Microcystin Distribution and Concentration in the Midwestern United States, Water Research, v. 38, p.4395-4404.

Lembi, C. A. 2002. Control of Nuisance Algae. In Wehr, J. D. and R. G. Sheath (Eds). Freshwater algae of North America. Ecology and Classification. Academic Press. Pp805-835.

Simpson, M. R. and B. W. Maclead 1991. Comparison of various powdered activated carbons for the removal of geosmin and 2-methylisoborneol in selected water conditions, in : 1991 Proceedings of the American Water Works Association Annual Conference.

Trojan Technologies Inc 2005.. TASTE-AND ODOR-CAUSING COMPOUNDS IN DRINKING WATER. Printed in Canada. London, Ontario, Canada (available online

Wang, S.H., A.R. Dzialowski, J. O. Meyer, F. deNoyelles Jr., N. Lim, W. W. Spotts, and D. G. Huggins. 2005. Relationships between cyanobacterial production and the physical and chemical properties of a Midwestern Reservoir, USA Hydrobiologia 541(1): 29-43

Zimmerman, L.R., Ziegler, A.C., and E.M. Thurman, 2002. Method of Analysis and Quality-Assurance Practices by U.S. Geological Survey Organic Geochemistry Research Group--Determination of Geosmin and Methylisoborneol in Water Using Solid-Phase Microextraction and Gas Chromatography/Mass Spectrometry, , U.S. Geological Survey Open File Report 02-337, 12 p.

Review:

In principal, I would agree that it is difficult to predict health-related problems based on chlorophyll. However, I believe that the difficulty (low level predictability) in relating the two is because other known factors and non-chlorophyll producing organisms (e.g., actinomycetes) also contribute to taste and odor events. These “other” causes and factors only lessen our ability to predict specific cause/effect relations between chlorophyll or cyanobacteria and specific odor causing compounds or taste-and-odor events. The fact remains that many taste and odor events have been linked to algal metabolites and high algal biomass often generates high levels of these compounds.

There exists a body of literature that consistently links increased nutrients to increases in algal biomass. At least one study has shown that increases in algal biomass (e.g., chlorophyll a) often result from increases in cyanobacteria biomass (Downing et al. 2001). They show that the risk of water quality degradation by cyanobacteria blooms is more strongly correlated with variation in total P, total N, or standing algae biomass than the ratio of N:P. Risks associated with cyanobacteria are therefore less associated with N:P ratios than a simple increase in nutrient concentrations and algal biomass. Cyanobacteria cause a multitude of water-quality concerns, including the potential to produce taste-and-odor causing compounds and toxins that are potent enough to poison animals and humans. The cyanobacterial compounds most commonly associated with taste-and-odor episodes are geosmin and 2-methylisoborneol (MIB)
All of the above relationships indicate that increases of nutrients can promote increases in algal biomass that often leads to increases in cyanobacteria that is a source of geosmin and MIB. However, these two compounds can also be produced by certain actinomycetes that are known to be the cause of taste and odor in municipal water supplies and wastewater treatment facilities (see 20th edition of Standard Methods, section 9250). The fact that other organisms can produce taste and odor occurrences in waterbodies does not less negate the fact that nutrient increase often lead to increases in cyanobacteria that can be a major contributor to taste and odor occurrences. Not all taste and odor problems or occurrences in inland waters are relatable to cyanobacteria and eutrophication but in many cases they are and we need to acknowledge that nutrient enrichment can contribute to taste and odor conditions. Several regional studies support the occurrence of these relationships. In a USGS study of Cheney Reservoir in south central Kansas, researchers indicated that the taste-and-odor compound geosmin was probably produced by the cyanobacterial genera Anabaena and thus was the likely cause of taste-and-odor episodes in this reservoir (Christian et al. 2006). Prior to this study of Cheney Reservoir, Smith and coworkers (Smith et al. 2001) had found a strong relationship between chlorophyll a concentrations and geosmin suggestive of a causal effect relation. Another Kansas study, USGS researchers found that taste-and-odor episodes in Lake Olathe (Johnson Co. KS) were likely linked to both cyanobacterial bloom formation as well as actinomycetes bacteria (a group of taste-and-odor producing bacteria that live in soil) washed into the lake during runoff events (Mau et al. 2004). Our studies, both published (Wang et al. 2005) and ongoing continue to support the occurrence of relationships between nutrients, algal biomass (or cyanobacteria biovolume) and levels of geosmin. Current study of five eastern Kansas reservoirs has shown that four of the five have noticeable concentrations of geosmin but relationship with chlorophyll levels is proved to be weak although much of the data remain to be analyzed. In conclusion, I believe that eutrophication can often lead to conditions that promote the occurrences of taste and odor events through the stimulation of algal biomass increases and shifts in lake conditions that favor cyanobacterial growth. While cyanobacteria are not the single source of taste and odor compounds, it is often related to most individual events in our region, enough so that control and reduction of nutrients in lakes is warranted.

References

American Public Health Association (APHA). American Water Works Association
(AWWA) and Water Environment Federation (WEF). 1998. Standard methods for the examination of water and wastewater, 20th Ed. Washington D.C.

Christensen, V.G., Graham, J.L., Milligan, C.R., Pope, L.M., and A.C. Ziegler. 2006. Water quality and relation to taste and odor compounds in the North Fork Ninnescah River and Cheney Reservoir, South-Central Kansas, 1997-2003. U.S. Geological Survey Scientific Investigations Report 2006-5095. 49 p.

Downing, J.A, Susan B. Watson, and Edward McCauley. 2001. Predicting Cyanobacteria dominance in lakes. Can. J. Fish. Aquat. Sci. 58(10): 1905-1908.

Mau, D.P., Ziegler, A.C., Porter, S.D., and L.M. Pope. 2004. Surface-Water-Quality Conditions and Relation to Taste-and-Odor Occurrences in the Lake Olathe Watershed, Northeast Kansas, 2000–02. U.S. Geological Survey Scientific Investigations Report 2004-5047. 95 p.

Smith, V.H., deNoyelles, Frank, Jr., Graham, D.W. and S.J. Randtke. 2001. A comparative water quality study of Cheney Reservoir, Kansas – final report to the city of Wichita Water and Sewer Department. University of Kansas, Dept. of Ecology and Evol. Biology, Lawrence, KS. 28 p.

Wang, S.H., A.R. Dzialowski, J.O. Meyer, deNoyelles, F. Jr., N.C. Lim, W.W. Spotts and D.G. Huggins. 2005. Relationships between cyanobacteria, production and the physical and chemical properties of a Midwestern Reservoir, USA. Hydrobiologia 541(1): 29-43.

Review:

As the second reviewer points out, the problem of taste and odor is not always one associated with eutrophication. There are confounding factors with regards to taste and odor problems (like actinomycetes)in eutrophic systems, and that organisms present in less eutrophic waters (such as golden flagellates) can also cause taste and odor issues. Taste and odor issues generally occur in relation to cyanobacteria. In general, the growth characteristics of cyanobacteria allow them to dominate in eutrophic warm water systems, there are few absolutes in ecology, however, and other than the toxins associated with Microcystis, there are few predictors of taste and odor problems from eutrophic systems.

Ok, so while the answers all pinged on taste and odor problems, there are two real human health problems that are not well discussed that do occur in association with treatment of eutrophic waters for drinking water. Chlorination associated with drinking water treatment can create trihalomethanes from cyanobacteria. Trihalomethanes present a human health problem. The other big problem was touched on but not directly addressed--the toxics associated with Microcystis. Microcystis spp. habor toxins call microcystins that can kill mammals if enough is ingested. I don't think there have been many (if any) documented human deaths associated, however, it is common that cows and dogs are killed by these toxins every summer after drinking from farm ponds that have high concentrations of Microcystis spp. Drinking water plants that pull from eutrophic reservoirs often test for Microcystis or microcystins on a routine basis.

Blue baby syndrome (methhemoglobin) is also associated with drinking water high in nitrates (greater than 10 mg/L). There is no rule on drinking water high in nitrates for folks who aren't pregnant, but I wouldn't think that it would be recommended for small children. High nitrates are a frequent problem in the mid-west in the spring. Several cities that pull there drinking water from rivers routinely distribute bottled water to the public in spring to prevent blue baby syndrome.

That is the sum of the issues for drinking waters that I am aware. There are also contact recreation issues associated with many blue-greens. WHO has a document out about it.

Backer, L.C., D.L. Ashley, M.A. Bonin, F.L. Cardinali, S.M. Kieszak, J.V. Wooten. 2000. Household exposures to drinking water disinfection by-products: whole blood trihalomethane levels. JOURNAL OF EXPOSURE ANALYSIS AND ENVIRONMENTAL EPIDEMIOLOGY, 10 (4): 321-326.

Carmichael, WW. 1992. CYANOBACTERIA SECONDARY METABOLITES - THE CYANOTOXINS. JOURNAL OF APPLIED BACTERIOLOGY 72(6): 445-459 JUN 1992

Codd, GA; Edwards, C; Beattie, KA; Barr, WM; Gunn, GJ. FATAL ATTRACTION TO CYANOBACTERIA. NATURE 359(6391): 110-111 SEP 10 1992

Dodds, L., W. King, C. Woolcott, J. Pole. 1999. Trihalomethanes in public water supplies and adverse birth outcomes. EPIDEMIOLOGY. 10 (3): 233-237.

Frazier, K; Colvin, B; Styer, E; Hullinger, G; Garcia, R. 1998. Microcystin toxicosis in cattle due to overgrowth of blue-green algae. VETERINARY AND HUMAN TOXICOLOGY 40(1): 23-24. FEB 1998

Hallegraeff, G.M. 1992.
Harmful algal blooms in the Australian region. MARINE POLLUTION BULLETIN, 25 (5-8): 186-190.

Harding, WR, N. Rowe, J.C. Wessels, K.A. Beattie, & G.A. Codd. 1995. Death of a dog attributed to the cyanobacterial (blue-green algal) hepatotoxin nodularin in South Africa. JOURNAL OF THE SOUTH AFRICAN VETERINARY ASSOCIATION-TYDSKRIF VAN DIE SUID-AFRIKAANSE VETERINERE VERENIGING, 66 (4): 256-259.

Landsberg, J.H. 2002. The effects of harmful algal blooms on aquatic organisms. REVIEWS IN FISHERIES SCIENCE, 10 (2): 113-390.

McDonald A.T., D. Kay. 1988. Water resources issues and strategies. UK: Longman Scientific & Technical, p.164-148.

Miles, A.M., P.C. Singer, D.L. Ashley, M.C. Lynberg, P. Mendola, P.H. Langlois, & J.R. Nuckols. 2002. Comparison of trihalomethanes in tap water and blood. ENVIRONMENTAL SCIENCE & TECHNOLOGY, 36 (8): 1692-1698.

Mez, K., K.A. Beattie, G.A Codd, K. Hanselmann, B. Hauser, H. Naegeli, & H.R. Preisig. 1997. Identification of a microcystin in benthic cyanobacteria linked to cattle deaths on alpine pastures in Switzerland. EUROPEAN JOURNAL OF PHYCOLOGY, 32 (2): 111-117.

Super, M., H.V. Heese, D. Mackenie et al. 1981. An epidemiological study of well water nitrates in a group of South West African/Namibian infants. Water Research, 15: 1265 - 1270.

Trubetskova, IL; Haney, JF. 2006. Effects of differing concentrations of microcystin-producing Micropystis aeruginosa on growth, reproduction, survivorship and offspring of Daphnia magna. ARCHIV FUR HYDROBIOLOGIE 167(1-4): 533-546. SEP 2006

Tencalla, F.G., D.R. Dietrich, C. Schlatter. 1994. Toxicity Of Microcystis-Aeruginosa Peptide Toxin To Yearling Rainbow-Trout (Oncorhynchus-Mykiss). AQUATIC TOXICOLOGY, 30 (3): 215-224.

vanHalderen, A., W.R. Harding, J.C. Wessels, D.J. Schneider, E.W.P. Heine, J. vanderMerwe, J.M Fourie. 1995. Cyanobacterial (blue-green algae) poisoning of livestock in the Western Cape Province of South Africa. JOURNAL OF THE SOUTH AFRICAN VETERINARY ASSOCIATION-TYDSKRIF VAN DIE SUID-AFRIKAANSE VETERINERE VERENIGING, 66 (4): 260-264.

World Health Organization. 1988. WHO Guidelines for drinking water quality. 2nd Edition, Addendum to Volume 1: Recommendations. Geneva: WHO, p. 8-10.; and addendum to Volume 2: Health Criteria and other Supporting Information.

Response:

Optimality is in the eyes of the beholder - so it depends on what your questions, desires, and constraints are.

First, some rough "rules of thumb" based on looking at a lot of state stream macroinvertebrate datasets:
• Sample around 1 to 2 m2 of stream bottom, but spread out over your habitats (whether targeted or proportional)
• My experience suggests an absolute minimum subsample is 100 organisms, 200 is much better, and I don't think you gain (proportional to effort) by going over 500 (Barbour and Gerritsen 1996). However, Cao and Hawkins (2005) proposed that minimum sample size should be 600 organisms, based on simulations of taxa loss with increased stress. Their results require the assumption that the most sensitive organisms in unstressed sites, the first to disappear, are also the most abundant taxa at those sites, thereby increasing evenness in stressed sites compared to unstressed. They identified a potential problem, but I do not know if their simulated results have been confirmed with real data.

• As mentioned above, “optimal” depends on the questions. If the objective is a statewide survey of biological condition (as in EMAP), effort is optimally spent in sampling more sites in more strata and more regions, because those are likely to be the largest sources of variability. A true effort vs. cost analysis for this question would tell you that optimal subsample is 100 organisms or fewer, if that allows you to sample more sites.
• If the objective is to decide whether a borderline site meets or fails state biocriteria, then you may want to have a larger subsample, as well as several seasonal samples over 1-2 years, to decide whether to spend the TMDL money.
• If the waterbody is already on the TMDL list, and sampling is to help identify stressors, then you would also want a substantial effort to sample physical and chemical water quality on monthly intervals for a year or more, as well as detailed mapping and source identification in the watershed.

The bottom line will be to have a single sampling and subsampling protocol that meets all your needs, which may be gold-plated for survey needs, yet bare-bones for detailed needs of TMDL, UAA, or nutrient criteria development.

If you have a set of samples that were completely counted and identified, then it’s fairly simple to subsample each one randomly to a desired target, whether a constant number of organisms (200, 300, etc.), or a specified fraction (e.g., 1/2, ¼ of the total). Examine the effect of subsampling protocol on values and variability, both within site (measurement error) and among sites (total sampling error). Taxa richness is dependent on sampling and subsampling effort, such that small subsamples tend to have larger coefficients of variation (CV). If you don’t have repeat measurements at each site, they can be simulated with bootstrapping subsamples from the original complete sample (sampling with replacement).

I think the cost issue breaks down into 2 parts:

1) For a survey, how many new sites can I sample by using a smaller subsample (what does subsampling buy in terms of more representative coverage);
2) For detailed analysis, what is the cost of revisiting a site when a small subsample was insufficient (what does a small subsample cost in terms of having to resample the same site if the subsample is too small)?

For benthic macroinvertebrates, a target subsample in the range of 200-500 organisms (pick a fixed target within the range, not the range as the target), or a fraction of the total that would yield that range, seems to be an adequate compromise for most circumstances. For fish, the key is more likely to be the total bottom area or the reach length sampled (usually in the range 10 to 40 times wetted width). Periphyton investigations often subsample to 500 cells, but filamentous is more variable.

Barbour, M.T., and J. Gerritsen 1996. Subsampling of benthic samples: a defense of the fixed-count method. J. N. Am. Benthol. Soc. 15:386-391.

Cao, Y., and C.P. Hawkins. 2005. Simulating biological impairment to evaluate the accuracy of ecological indicators. J. Appl. Ecol. 42: 954-965

Review:

The answer looks good to me. More of my experience is with algal cells, and 500 is a good target number, but a fairly large sample (e.g. 6 scrapings of 50-100 cm2 over a variety of habitats) should be homogenized and subsampled. Often times transects are used to estimate percentage cover of filaments and macrophytes. In this case 100 point measurements per reach sampled are generally adequate.

I also suggest not using artificial substrata for algal sampling; use what is naturally there.

With algae there is also the problem of living versus dead diatoms. It can be difficult to get identity to species with dead diatoms (uncleared frustules). However, if you digest all the samples you may include frustules that were not abundant in the live community at the time of sampling. It is possible to learn the species well enough to be fairly accurate on uncleared samples, but this takes some time. If diatoms are a major component, it may make sense to concentrate on erosional habitat (solid substrata in substantial water velocity) to make certain that few dead frustules are present in the sample.

Kelly and Whitton (1995) scraped 5 boulders and homogenized. They counted at least 200 frustules and found that this was adequate for determination of trophic state.
Kelly, M. G. and B. A. Whitton. 1995. The trophic diatoms index: a new index for monitoring eutrophication in rivers. J. Appl. Phycol. 7:433-444.

Lowe and LaLiberte 2006. (chapter in Methods in Stream Eco )

Review:

I agree with the first response that the objectives of the study or program must be considered in determining sufficiency of subsample size, number of samples, taxonomic resolution, habitats sampled (single vs multiple) and sampling gear.

I have attached a paper that address some of the subsampling issues but it doesn't answer the question of how big a subsample do you need. It does frame the questions that need to be considered in determining sufficiency (what is necessary to answer your question). I don't think there is a right number but we use a 500 organism subsample size in all projects as we are confident that this size will allow us to meet 99% of our objectives in most projects. Mostly the question is - is site A different than site B, but we have to ask how small of a difference am we are interested in.

Response:

We have not been able to find criteria based on nuisance values of duckweed or algal scums, or their percent cover. In general, I would suspect that visible scums that yield a measurable perception of percent cover would be unacceptable to many users. In other words, anything greater than a few individuals (for duckweed) or small scums on the lee shore of a lake would be unacceptable. There have been some user surveys of transparency:

In the late 1980s, the states of Vermont and Minnesota developed a brief lake user perception survey to be completed by volunteer monitors at the time of their routine sampling. This survey asked the monitors their opinion of the physical condition and recreational/aesthetic enjoyment of the water. When linked to water chemistry data collected concurrently, these surveys were used along with other types of data to develop ecoregion-based numeric phosphorus criteria for segments of Lake Champlain and portions of Minnesota. Since this time, many other states (Wisconsin, Ohio, New Hampshire, New York, Maine, and Indiana) have used this same survey to identify the total phosphorus, chlorophyll a, or Secchi disk transparency values at which algal nuisances and recreational impairments are perceived by the public. New York adapted the Vermont-Minnesota survey to include questions that ask users to characterize macrophyte coverage and identify the cause of any perceived recreational use impairment. It is important to note that user expectations of water condition and use impairment vary regionally.

Minnesota recognized that primary contact recreation (i.e., swimming) is not the principal use for all lakes. Therefore, they proposed eutrophication criteria (total phosphorus, chlorophyll a, Secchi disk transparency) to support specific lake types including deeper lakes for primary contact, shallow lakes for healthy aquatic plant and wildlife communities, and trout lakes. For instance, in the same ecoregion, the criteria for shallow lakes would focus on supporting macrophytes and minimizing algal dominance, while greater transparency would be important for deeper lakes.

Heiskary, S.A. and W.W. Walker. 1995. Establishing a chlorophyll a goal for a run-of-the-river reservoir. Lake and Reservoir Management 11(1):67-76.

Heiskary, S.A. and W.W. Walker. 1988. Developing phosphorus criteria for Minnesota
lakes. Lake and Reservoir Management 4:1-10.

Hoyer, M.V., C.D. Brown and D.E. Canfield Jr. 2004. Relations between water chemistry and water quality as defined by lake users in Florida. Lake and Reservoir Management 20(3):240-248.

MPCA. 2005. Minnesota lake water quality assessment report: developing nutrient criteria. Third edition. Prepared by Minnesota Pollution Control Agency, St. Paul, MN.

NALMS. 1992. Developing eutrophication standards for lakes and reservoirs. Prepared by Lake Standards Subcommittee. North American Lake Management Society, Alachua, FL. 51p.

NYFLA and NYDEC. 2003. Evaluating lake perception data as a means to identify reference nutrient conditions. Report to USEPA. New York State Federation of Lake Associations and New York State Department of Environmental Conservation.

Smeltzer, E. 1992. Developing eutrophication standards for Lake Champlain from user
survey data. Vermont Department of Environmental Conservation. Waterbury,
Vermont.

Smeltzer, E. and S.A. Heiskary. 1990. Analysis and application of lake user survey
data. Lake and Reservoir Management 6(1):109-118.

Review:

The answer gives good information on how perceived water quality is related to transparency. In my experience, any strongly noticeable level of algal bloom or macrophyte leads individuals to attempt control.

I am not aware of information functionally linking duck weed to nutrient content of water except, A model of ditch vegetation in relation to eutrophication
Janse J.H Water Science and Technology, Volume 37, Number 3, 1998, pp. 139-149(11)
Does provide some information on this.

 The work of Downing indicates a sharp increase in the probability of cyanobacterial blooms at 0.03 mg/L TP.
Carpenter and his coworkers have done some work on economic analyses and eutrophication.

Regardless of perception, the sharp increase in cyanobacterial bloom probability should be a top priority because of the potential for toxic strains to dominate the water column.

Carpenter, S.R., D. Ludwig, and W. A. Brock 1999. Management of eutrophication for lakes subject to potentially irreversible change. Ecological Applications 9: 751-771.

Downing, J.A., S. B. Watson, and E. McCauley 2001. Predicting cyanobacteria dominance in lakes. Canadian Journal of Fisheries and Aquatic Sciences 58: 1905-1908.

Review:

The reviewer gives a good summary of historical development on this issue. I dont have much to add at this time. To set criteria for nuisance aquatic plants, uses of the lakes or ponds have to be considered. To protect aquatic life (fishes), many consider that 20% or more floating aquatic plant cover is potentially causing problems (DO declines) but I am not aware of any scientific literatures to support that.

Response:

The question asks about separating DO trends to those caused by nutrient enrichment from those caused by other sources.

The first issue is what can you tell from diurnal DO swings in streams? They can be used to calculate both respiration and photosynthesis if a complete time series is available for at least a 24 hour period. For more information on this consult Bott. T. 2006. Primary Productivity and Community Respiration. In FR Hauer and GA Lamberti (eds.) Methods in Stream Ecology. If there is no DO swing, and DO is below saturation, then respiration drives it all. This method is less common in lakes, but could be applied in the same fashion.

On a cloudy day, there may be little fluctuation and the system may have low DO for the 24 hour period. The amount of light is a consideration.

If a stream substantially exceeds saturating DO (i.e. by 50%) I would suspect a very high algal or macrophyte biomass, but there is little published literature on this. Part of the reason for no absolute numbers is that the degree of saturation exceedence or deficit is not only a function of metabolic rate, but also on aeration rate of the water column to the atmosphere.

The question implies that sediment respiration is separate from algal respiration. This is a difficult separation to make because algae inhabit the surface of the sediments, and algal biomass can sink into the sediments and cause subsequent respiration. It makes more sense to think about factors that influence whole system metabolism. It is possible this question is being asked this way, but I am not quite sure.

There is a recent publication that argues that heterotrophic processes can be limited by nutrients, so the effect of nutrients on heterotrophic processes should also be considered for nutrient criteria. Nutrient pollution could contribute to chronic low DO problems.
Dodds, W. K. 2006. Eutrophication and trophic state in rivers and streams. Limnology and Oceanography 51:671-680. Think of a situation where there is heavy leaf input in the fall, but the heterotrophic community is very nutrient limited. Nutrient pollution could substantially stimulate respiration in this condition.

Nutrient pollution may harm biotic integrity (e.g. macroinvertebrates and fish), so it may not be necessary to use DO swings to prove an adverse effect of nutrients.
Wang, L. et al. (2007) Linkages between nutrients and assemblages of macroinvertebrates and fish in wadeable streams: Implication to nutrient criteria development. Environ. Manage. 39, 194-212. Our research indicates this effect is due to alteration of food sources for invertebrates (enrichment of algae and heterotrophic microbes leads to lower diversity of consumers).

There is substantial literature on low pH, but less on higher pH induced toxicity from photosynthetic processes. Most of the pH literature is on acid rain, and little that I am aware of has to do with rapid fluctuations in pH. One exception to this is the rapid decreases in pH associated with acid snow melt.

WELCH, E. B. 1992. Ecological effects of wastewater. 2nd edition. Chapman and Hall, London, UK.

Mentions negative influence of high pH, and general effects of high and low pH on invertebrates are mentioned in

Dodds, W. K. 2002. Freshwater Ecology: Concepts and Environmental Applications. Academic Press. 569 pp.

In summary, both maximum and minimum DO can be influenced by nutrients. Probably maximum DO is influenced more heavily if there is adequate light and low inorganic turbidity. There is little published literature on negative effects of the DO swings, but much on low DO and some on elevated DO (to refer back to the original question).

Review:

Although no direct evidence shows fish and macroinvertebrates are related to DO fluctuation alone, people generally assume that biological community is adversely impaired by DO fluctuations since relatively fewer field DO measurements are taken before dawn when DO depletion occurs. I am aware of very few studies that specifically address DO fluctuation (rather than low DO) and aquatic health from direct experimental manipulation. There are a number of reasons. First, DO fluctuation is usually accompanied by nutrient enrichment, and low DO below acceptable criteria most of the time. Designing studies focusing on DO fluctuation alone is difficult. DO fluctuation is also often accompanied by other stressors such as organic enrichment, TSS, habitat alteration, or sedimentation. Therefore, diagnosis of impairment has to be implemented by a stressor identification process and fluctuating DO may not be the primary stressor. Collier and Willcock (1998) were able to link biological impairment to DO fluctuation using multiple regression. However, DO fluctuation only explains a small amount of variance. Finally, DO fluctuation is a complex process involving both autotrophic production and heterotrophic respiration, as pointed out by the primary reviewer. In systems where autotrophic process predominates and temperature remains high, it is possible that DO fluctuates a lot but does not drop below a standard (<5 mg/L). (DO dropped from 17mg/L to 9 mg/L in an ongoing diurnal study in a concrete channel in Southern California. However, these systems are also affected by some of the stressors listed above).
 
High aquatic plant reproduction tends to increase both DO and pH level in streams. Instead of looking at DO decline, a study by Serafy and Harrell (1993) examined the effect of increased DO and pH values on fish from both field and laboratory observation. They found that increased DO does not affect fish but pH fluctuation alone (when pH>9.5) does affect fish behavior. However, when high pH levels (9.5–10.0) are accompanied by high DO levels (200–260% saturation), high pH level could be alleviated by high DO. It is an interesting reference for a DO fluctuation study.

Another interesting study in Iowa examines correlation between Fish IBI and fluctuation of diurnal DO
(http://www.iowadnr.com/water/watershed/tmdl/files/draft/nforkmaqsi.pdf [^] Appendix Fig. 2-2,3,4,5). They illustrate four different relations. First, there is a correlation between minimum DO and O fluctuation. Second, there is no relationship between Fish IBI and DO fluctuation. However, high DO fluctuation is also accompanied by low DO and consequently, low Fish IBI. Third, average daily respiration is negatively correlated to minimum DO; Forth, Fish IBI is negatively correlated with average daily respiration.

Collier K. J. and R. J. Willcock. 1998. Influence of substrate type and physico-chemical conditions on macroinvertebrate faunas and biotic indices of some lowland Waikoto, New Zealand, Streams. New Zealand Journal of Marine and Freshwater research 32:1-19.

SERAFY, J.E. and R.M. HARRELL. 1993. Behavioural response of fishes to increasing pH and dissolved oxygen: field and laboratory observations. Freshwater Biology 30 (1), 53–61.

Review:

There is really very little to be added to the first two expert responses. The literature is still fairly silent on the effects of DO fluctuation or even on the response of DO fluctuation to nutrients, especially for freshwater biota. This is likely due to 1) the relatively recent availability of recording DO probes and 2) the experimental designs required to study such things. Especially for biota, it still appears that DO minima are far more important that the magnitude of fluctuation. For example, a recent experimental study examined 5 DO treatments – 4 fixed concentrations (1.5, 2, 4 and 6 mg/L) and 1 fluctuating (2 to 10 mg/L) DO concentration on estuarine fish growth (McNatt and Rice 2004). Only the 1.5 mg/L treatment resulted in reduced growth – all the others were equivalent, again supporting the idea that it is DO minima alone that are critical and not fluctuations. A similar study found no difference in the growth of Atlantic salmon between fluctuating DO (4 to 13 mg/L) or constant DO (7.3 mg/L; Forsberg and Bergheim 1996). A slightly different result was obtained for young-of-the-year flounder, which grew slower under fluctuating DO (2.5 to 6.4 mg/L) than under high fixed DO (6.7 mg/L), but still faster than under a low fixed DO concentration (2.2 mg/L; Bejda et al. 1992). One report noted DO fluctuation had been correlated with fish anomalies in the Little Miami River in Ohio, but the same correlation was not observed in a follow-up study and may have been the result of colinearities with other factors (such as DO minima) that were the real cause (Evans and Miller 2006).

DO fluctuations may respond to nutrients, especially where light is not limiting. One would expect greater fluctuations in the presence of nutrients as a result of increased daytime photosynthesis (nutrients enrich plants) and nighttime respiration (more plant respiration, more plant biomass to decompose and enrich microbes). Again, however, this will depend on oxygen reaeration rates. Reaeration is controlled by a number of factors including channel dimensions, flow, temperature, barometric pressure, dissolved oxygen saturation. All of these will affect fluctuations as well. So, it will likely be hard to generalize. Not surprisingly, there is far more published on oxygen and oxygen fluctuation in response to organic effluent than to nutrients, per se. I could find little additional literature linking nutrient concentrations to diel oxygen fluctuations.

Bejda, A.J., B.A. Phelan, and A.L. Studholme. 1992. The effect of dissolved-oxygen on the growth of young-of-the-year winter flounder, Pseudopleuronectes-americanus. Environmental Biology of Fishes 34:321-327.

Evans, R.L. and M.C. Miller. 2006. Nutrients, eutrophic response, and fish anomalies in the Little Miami River, Ohio. Ohio Journal of Science 106: 146-155.

Forsberg, O.I. and A. Bergheim. 1996. The impact of constant and fluctuating oxygen concentrations and two water consumption rates on post-smolt Atlantic salmon production parameters. Aquacultural Engineering 15(5): 327-347.

McNatt, R.A. and J.A. Rice. 2004. Hypoxia-induced growth rate reduction in two juvenile estuary-dependent fishes. Journal of Experimental Marine Biology and Ecology 311(1) 147-156.

Response:

I do not know that anyone has addressed optimizing the number of samples that should be taken or the index period over which samples should be taken to precisely characterize the nutrient concentrations at a site.  However, understanding the issues can help you justify the number of nutrient samples that you need to take at a site to correlate with biological responses to nutrients.  Below I present results of some analyses that I conducted with a data set where I have repeated measurements of nutrients from each site over a 2 month period.

Before starting, I should emphasize that I think this is an important issue.  I usually take multiple measures of nutrients at sites over time or multiple indicators of nutrient concentrations, such as measured nutrient concentrations and diatom inferred nutrient concentrations.  I don’t have space to present analyses here for the value of diatom-inferred nutrient concentrations, but they are valuable complements for measured nutrient concentrations.  In Kentucky, diatom inferred TP concentration more precisely measured nutrient concentrations than one-time sampling of nutrients and better predicted algal biomass than measured nutrient concentrations (Stevenson, unpublished results).

First, the objective of the data collection should be clear and subsequent use of the data should be restricted by the original use.  For example, if the goal is to quantify a relationship between nutrient concentrations in streams and biological attributes and develop a nutrient criterion, the number of observations used to characterize a site during the development of the “nutrient-response” relationship and implementation of the nutrient criterion should be considered.  Error variances may differ during the two processes.  If you are going to use the relationship to develop nutrient criteria, you should follow the same procedure regarding when you sample after storm events, time of year when the model was developed, and same number of repeated measures of nutrient concentrations for each site.  In the following answer to this question, I assume the goal is just to quantify or establish a relationship between biological attributes and nutrient concentrations.

The next question is over what period of time would the biological attribute respond to nutrient concentrations?  More measurements should be taken if you are trying to assess the indirect effects of nutrients on invertebrates and fish than if you were trying to assess the direct effects of nutrients on Cladophora accrual in a stream.  In the first case, you would consider measuring nutrients throughout the year to characterize seasonal variability or target specific seasons when stresses and subsidies caused by nutrient enrichment may be affecting dissolved oxygen or productivity of larval forms of invertebrates. (i.e. consider the processes that link nutrients and the attribute(s) of interest and then measure nutrient concentrations over that “index” period).

What other factors affect nutrient variability in the stream?  Variation in loading due to point and non-point sources would be one consideration.  Another is the hydrologic variability and routes of transport of nutrients to the streams.  Nutrient concentrations are more variable in hydrologically variable streams than hydrologically stable streams with storm events.  Nutrient concentrations in runoff-dominated waters after storms differ greatly from groundwater-dominated waters in flow (baseflow) conditions.  In some regions, like the glaciated region of Michigan with poor sandy soils and high soil permeability, runoff is not an important route of nutrient transport, so nutrient concentrations vary less with time.

Now that we’ve dealt with some conceptual issues, let’s deal with a couple statistical issues and examples of what has worked.  From a statistical perspective, if we are trying to quantify effects of nutrients on biota and calculate slopes of lines or determine threshold responses, the precision of characterizing nutrient concentrations (the independent variable) is a problem for regression analysis if it is not characterized with much greater precision than the biological attribute (dependent variable).  This issue is addressed in Model I and Model II Regression.  In Model I, the independent variable is precisely estimated (without much error).  In Model II, the independent variable has natural error, which characteristically results in underestimation of the magnitude (slope) of the regression model quantifying the relationship.  The solutions are to increase precision of the estimation of nutrient concentrations in streams, simply use correlations, or use Model II approaches for analyzing regression models (check web for more details or Sokal and Rohlf 1995).  This statistical issue is seldom addressed in ecological problems where variability in independent variables is common.

With fixed budgets, or levels of effort, the number of repeated visits must be balanced with the number of sites that will be visited.  Thus a tradeoff probably will exist between number of sites visited and the number of times each will be visited.  Simple linear relationships can usually be resolved with 20-30 sites, if they exist and with good measures of nutrients and biological attributes; but quantifying the relationship well and detection of scientifically defensible non-linear trends probably requires on the order of 40, 50 or more sites.  Sites should be selected so they reflect a wide range of nutrient conditions.  I usually sort candidate sites into 5 or more groups by land use indicators of nutrient concentrations and then randomly select the appropriate number of sites evenly from each of the land use indicator bins.

Remember, in many cases, good relationships between algal biomass and nutrients were not resolved in a number of studies. In the 90’s, a few studies showed relationships in limited areas.  Then Biggs (2000) characterized dissolved nutrient (PO4 relationships) with algal biomass in streams with nutrient concentrations and algal biomass collected over a year.  Dodds et al. (2002) characterized nutrient relationships with algal biomass in hundreds of streams and usually used annual means.  Later, Stevenson et al. (2006) characterized relationships between TP concentrations and algal biomass as well as Cladophora cover using between 3 and 8 characterizations of nutrient concentrations in the same stream during a 2 month spring period when biomass was assessed.  Fewer samples were used to characterize nutrient conditions in hydrologically stable streams of Michigan versus the hydrologically variable streams in Kentucky.  Analysis of algal biomass-nutrient relationships with EMAP data from the Mid-Atlantic region showed no statistically significant relationship when only one measure of nutrients and algal biomass was used in the test (Stevenson, unpublished analysis).

Now, how many observations am I comfortable with?  More than one and probably around 5.  As indicated at the beginning, I usually take multiple measures of nutrients at sites over time or multiple indicators of nutrient concentrations, such as measured nutrient concentrations and diatom inferred nutrient concentrations.

Taking multiple nutrient measures per site does increase the effort, however it is probably necessary.  One way to economize the effort is to limit detailed biological sampling and assessment to one time during the index period and measure only nutrients repeatedly and perhaps include very rapid in-stream biological assessments.  This is the approached used in Stevenson et al. 2006 and Riseng et al. 2004.

You should also be aware that nutrients vary with time of day, but I know of no study in which this source of variation was controlled.  There are reasons that both soluble and insoluble fractions would vary with time of day.  Soluble fractions vary with diel uptake cycles, and are thus lower during the day.  Particulate nutrients vary with diel patterns in algal drift and sediment disturbance by animals (Stevenson and Peterson ).
 
The following analysis may help you evaluate how many samples to collect.  I conducted this analysis with data from the EPA-STAR funded project for Michigan and Kentucky streams.  In the hydrologically stable streams of Michigan water chemistry was usually sampled 3 or 4 times during a 2 month index period, where as in the hydrologically variable Kentucky streams water chemistry was sampled between 8 times one year and 3 times the next during the 2 month index period.  Standard deviations in nutrient concentrations were remarkably similar between regions, but the number of repeated measures in KY was twice as high as in Michigan (Figures 1 and 2).  Standard deviations in nutrient concentrations increase with nutrient concentrations and range from 1/3 – 1/10 of the mean nutrient concentrations, depending upon whether they were PO4, TP, NO3, or TN.  For Kentucky only, I compared the correlation between log-transformed benthic chlorophyll a and average TP concentration at a site, with the number of repeated measures increasing from 1 to 8.  R2 increased from 0.077 to 0.066, 0.082, 0.091, 0.136, 0.157 and 0.157 with one, two, three, four, five, six and eight samples per site included in the average, respectively.  This illustrates that precision in relationships increases with increasing numbers of repeated nutrient samples at sites and with just one measure of algal biomass.  Is 5 or 6 the magic number?  That’s difficult to determine for more than this one case.  But more than one measurement is certainly important.

Biggs, B. J. F. 2000. Eutrophication of streams and rivers: dissolved nutrient- chlorophyll relationships for benthic algae. Journal of the North American Benthological Society 19:17-31.

Dodds, W. K., V. H. Smith, and K. Lohman. 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences 59:865-874.

Riseng, C. M., M. J. Wiley, and R. J. Stevenson. 2004. Hydrologic disturbance and nutrient effects on benthic community structure in midwestern US streams:  a covariance structure analysis. Journal of the North American Benthological Society 23:309-326.

Sokal, R. R., and F. J. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research, 3 edition. W.H. Freeman and Company, New York.
Stevenson, R. J., and C. G. Peterson. 1991. Emigration and immigration can be important determinants of benthic diatom assemblages in streams. Freshwater Biology 26:295-306.

Stevenson, R. J., S. T. Rier, C. M. Riseng, R. E. Schultz, and M. J. Wiley. 2006. Comparing effects of nutrients on algal biomass in streams in 2 regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561:140-165.

Review:

The number of water samples needed depends on:

Spatial Variability. If one is sampling a complex system, then additional samples will be required to capture the longitudinal or cross-sectional variability. For example, in my paper at www.hydrology.uga.edu/rasmussen/pubs/JEQ2005.pdf we show that original Lake Lanier monitoring network can be simplified into a more parsimonious network because parts of the lake and tributary subwatersheds behave very similarly. Thus, a single station can be used to represent general areas upstream and within the lake. Knowing the complexity ahead of time may be difficult, however, because one needs to do a preliminary analysis to know what this complexity is. That is why we show that using a surrogate, such as turbidity for nonpoint sources and total dissolved solids or specific conductance for point sources, would be very helpful, in that it is quick and easy way to characterize the spatial variation within the system. If the locations of the nutrient sources are known, then sampling with this information in mind makes characterization more robust. Sampling above vs. below point sources can be confusing if the source has not been identified ahead of time.

Temporal Variability. The response to this is again contingent on the complexity of the system, and needs to account for both point and nonpoint sources:

• For point sources, the temporal variation may not be an issue, in that permitted discharges don’t usually vary substantially (except when a failure occurs . . . ). For normal operations, monthly samples would likely be sufficient, and can be obtained either from the state or the point source directly.
• For nonpoint sources, the nutrient loads change rapidly with time during storm events, so capturing the storm-related variability is the challenge. There are multiple strategies for this in the literature, including the list below provided by Josh Romeis, a doctoral candidate in our program at the University of Georgia. Note that there is a wide body of literature on this.
  For the Etowah River in north Georgia, baseflows are sampled using routine grab samples, while stormflows are characterized using ISCO samplers collected during the course of the storm. The total nutrient load is mainly a function of land use, so that nutrient yields (i.e., loads per unit area) can be used as an index to determine total watershed loads and concentrations.

Aulenbach, B.T.a.R.P.Hooper., 2006. The composite method: an improved method for stream water solute load estimation. Journal of Hydrology 29:3029-3047.

Bradu, D.a.Y.M., 1970. Estimation in lognormal linear models. Journal of the American Statistical Association, 65(329): 198-211.

Coats, R., F. Liu, and C.R. Goldman, 2002. A Monte Carlo test of load calculation methods, Lake Tahoe Basin, California-Nevada. Journal of the American Water Resources Association, 38(3): 719-730.

Cohn, T.A., 1995. Recent advances in statistical methods for the estimation of sediment and nutrient transport in rivers. Reviews of Geophysics, Supplement: 1117-1123.

Cohn, T.a., Caulder, D.L., E.J. Gilroy, L.D. Zynjuk, and R.B. Summers, 1992. The validity of a simple statistical model for estimating fluvial constituent loads: an empirical study involving nutrient loads entering Chesapeake Bay. Water Resources Research, 28(No. 9): 2353-2363.

Cohn, T.A., L.L. DeLong, E.J. Gilroy, R.M. Hirsch, and D.K. Wells, 1989. Estimating constituent loads. Water Resources Research, 25(No. 5): 937-942.

Cooper, D.M., 2004. Some effects of sampling design on water quality estimation in streams. Hydrological Sciences Journal, 49(6): 1055-1080.

Crawford, C.G., 1991. Estimation of suspended-sediment rating curves and mean suspended-sediment loads. Journal of Hydrology, 129(331-348).

Dann, M.S., J.A. Lynch, and E.S. Corbett, 1986. Comparison of methods for estimating sulfate export from a forested watershed. Journal of Environmental Quality, 15: 140-145.

Dolan, D.M., A.K. Yui, and R.D. Geist, 1981. Evaluation of river load estimation methods for total phosphorus. Journal of Great Lakes Research, 7(3).

Ferguson, 1986. River loads underestimated by rating curves. Water Resources Research, 22(1): 74-76.

Ferguson, R.I., 1987. Accuracy and precision of methods for estimating river loads. Earth Surface Processes and Landforms, 12(95-104).

Gilroy, E.J., R.M. Hirsch, and T.A. Cohn, 1990. Mean square error of regression-based constituent transport estimates. Water Resources Research, 26(No. 9): 2069-2077.

Guo, Y., M. Markus, and M. Demissie, 2002. Uncertainty of nitrate-N load computations for agricultural watersheds. Water Resources Research, 38(No. 10, 1185, doi:10.1029/2001WR001149): 3-1 - 3-12.

Haggard, B.E., T.S. Soerens, W.R. Green, and R.P. Richards, 2003. Using regression methods to estimate stream phosphorus loads at the Illinois River, Arkansas. Applied Engineering in Agriculture, 19(2): 187-194.

Johnes , P.J. 2007. Uncertainties in annual riverive phosphorus load estimation: impact of load estimation methodology, sampling frequency, baseflow index, and catchment population density. Journal of Hydrology 332: 241-258.

Koch, R.W.a.G.M.S., 1986. Bias in hydrologic prediction using log-transformed regression models. Water Resources Bulletin, 22(5).

Kronvang, B.a.A.J.B., 1996. Choice of sampling strategy and estimation method for calculating nitrogen and phosphorus transport in small lowland streams. Hydrological Processes, 10: 1483-1501.

Moatar, F.a.M.M., 2005. Compared performances of different algorithms for estimating annual nutrient loads discharged by the eutrophic River Loire. Hydrological Processes, 19: 429-444.

Moosmann, L., B. Müller, R. Gächter, and A. Wüest, 2005. Trend-oriented sampling strategy and estimation of soluble reactive phosphorus loads in streams. Water Resources Research, 41(W01020, doi:10.1029/2004WR003539): W01020 (10 pp.).

Norick, N.X., 1969. SALT (selection at list time) Sampling: A new method and its application in forestry. M.A. Thesis Thesis, University of California, Berkeley, 50 pp. pp.

Porterfield, G., 1972. Computation of Fluvial Sediment Discharge.

Preston, S.D., V.J. Bierman, Jr., and S.E. Silliman, 1989. An evaluation of methods for the estimation of tributary mass loads. Water Resources Research, 25(6): 1379-1389.

Rasmussen, T.C., 2001. Determination of average concentrations and loads. Unpublished.  Warnell School of Forest Resources.  The University of Georgia, Athens.

Rekolainen, S., M. Posch, J. Kamar, and P. Ekholm, 1991. Evaluation of the accuracy and precision of annual phosphorus load estimates from two agricultural basins in Finland. Journal of Hydrology, 128: 237-255.

Richards, P.R.a.J.H., 1987. Monte Carlo studies of sampling strategies for estimating tributary loads. Water Resources Research, 23(No. 10): 1939-1948.

Robertson, D.M., 2003. Influence of different temporal sampling strategies on estimating total phosphorus and suspended sediment concentration and transport in small streams. Journal of the American Water Resources Association, 39(No. 5): 1281-1308.

Robertson, D.M., and E.D. Roerish, 1999. Influence of various water quality sampling strategies on load estimates for small streams. Water Resources Research, 35 (12): 3747-3759.

Rose, P., 1999. Lake Allatoona phase I diagnostic-feasibility study report for 1992-1997., A.L. Burruss Institute of Public Service. Kennesaw State University. Kennesaw, Georgia.

Schwartz, S.S.a.D.Q.N., 1999. Bias and variance of planning level estimates of pollutant loads. Water Resources Research, 35(11): 3475-3487.

Thomas, R.B., 1985. Estimating total suspended sediment yield with probability sampling. Water Resources Research, 21(9): 1381-1388.

Thomas, R.B., 1988. Monitoring baseline suspended sediment in forested basins: the effects of sampling on suspended sediment rating curves. Hydrological Sciences Journal, 33(5)(10): 499-514.

Thomas, R.B.a.J.L., 1993. A comparison of selection at list time and time-stratified sampling fore estimating suspended sediment loads. Water Resources Research, 29(No. 4): 1247-1256.

Thomas, T.S.a.M.A.N., 2001. Designing stream sampling schemes for load determination, AWRA Annual Spring Specialty Conference, pp. 71-76.

Walling, D.E., 1977. Assessing the accuracy of suspended sediment rating curves for a small basin. Water Resources Research, 13(3): 531-538.

Webb, B.W., J.M. Phillips, and D.E. Walling, 2000. A new approach to deriving 'best-estimate' chemical fluxes for rivers draining the LOIS study area. Science of the Total Environment, 251/252: 45-54.

Review:

I agree that 5-6 nutrient measures per response measurement is a good place to start. Biggs (2000; cited in the first response) did monthly measurements over a year. The Biggs study is an important paper and shows interesting relationships; I recommend studying that paper, even though you will probably not do the same thing. Biggs showed a strong relationship betyween mean chlorophyll (measured monthly) and mean accrual time, or the time between scouring flood events (requires gauged streams).

One strategy is to start out oversampling (say, monthly) and then reevaluate your sampling strategy after the first year to see if you can cut back the frequency without compromising your results. Nutrient concentrations, periphyton abundance, and chlorophyll concentration will usually require log transformation. Check the mean to variance ratio.

Response:

The EFDC model only accounts for phytoplankton so cannot be used for periphyton in streams or shallow lakes or wetlands. The EFDC water quality model accounts for the fact that nitrogen or phosphorus may limit growth. However, it is based on the concept of Leibig’s Law of the Minimum (only one nutrient can limit at a time, the one in shortest relative supply) so it does not account for co-limitation and the possibility that a secondary limitation can control biomass production by the primary limiting nutrient.

The WASP model considers both nitrogen and phosphorus, and it is capable of assuming multiple nutrient limitations for phytoplankton, but the user’s manual recommends against using the multiplicative limitation capability. The WASP model also can include a periphyton module, and this model considers both nitrogen and phosphorus, but assumes only a single nutrient limits.

In conclusion, at least both models assume that nitrogen or phosphorus can be the limiting factor, but they do not deal with the concept of co-limitation very well. The WASP model accounts for periphyton, so may be more generally useful. Probably both models should be used when possible, as well as sensitivity analyses, to compare the rigor of the model results.

Review:

With version 7.0, the WASP water quality model incorporated periphyton routines1. These were implemented by Dr. James Martin of Mississippi State University, and were initially based on the simple periphyton routines contained in the QUAL2K model version 12. In that version, nutrient limitation, as implemented for periphyton in WASP, is represented via Michaelis-Menten equations representing growth limitation due to inorganic N and P. The minimum value is then used to compute the nutrient attenuation coefficient, as is done for phytoplankton. The WASP routines were subsequently updated to be compatiblel with version 2 of QUAL2K (v. 2.0.4) 3, which uses a more complex formulation.

In the current version of QUAL2K and WASP, rather than being dependent on external nutrient concentrations, the effect of nutrient limitation on periphyton photosynthesis is modeled as dependent on intracellular nutrients4.. QUAL2K, but not WASP, includes carbon limitation in addition to N and P The uptake rates depend on both the external and intracellular nutrient concentrations5, and the total nutrient limitation is dependent on the minimum factor based on intracellular N, and P. In theory, this allows for simulation of luxury uptake and storage, thus getting at one important aspect of co-limitation. Provision is also made for substrate (space) limitation on periphyton growth. Due to their recent implementation, it does not appear that the performance of the QUAL2K/WASP7 formulation has yet been thoroughly tested and validated in freshwater streams.

The EFDC water quality simulation code is primarily adapted from the Chesapeake Bay three-dimensional water quality model CE-QUAL-ICM. For phytoplankton, EFDC uses a “most limiting” concept with Michaelis-Menten kinetics, but considers potential limitation by silica in addition to inorganic N and P. It also readily simulates multiple algal groups, unlike WASP. EFDC has been adapted for periphyton simulation; for example in the Christina River TMDL6. This implementation uses Michaelis-Menten nutrient kinetics, but also includes semiempirical representation of limitations on algal growth due to velocity effects on nutrient availability as well as limitations due to substrate availability/benthic algal density. The implementation for EFDC is largely based on similar modifications made to WASP for an application to the Carson River in Nevada7. Tetra Tech has also developed modifications to EFDC that account for fixing of nitrogen by cyanobacteria and luxury uptake of phosphorus8. These modifications have not yet been incorporated in the public release versions of EFDC.

Ambrose, R.B., J.L. Martin, and T.A. Wool. 2006. WASP7 Benthic Algae – Model Theory and User’s Guide; Supplement to Water Quality Analysis Simulation Program (WASP) User Documetnation. EPA 600/R-06/106. Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC.

Chapra, S. and G. Pelletier. 2003. QUAL2K: A Modeling Framework for Simulating River and Stream Water Quality: Documentation and Users Manual. Civil and Environmental Engineering Dept., Tufts University, Medford, MA.

Chapra, S., G. Pelletier, and H. Tao. 2006. QUAL2K: A Modeling Framework for Simulating River and Stream Water Quality, Version 2.04: Documentation and Users Manual. Civil and Environmental Engineering Dept., Tufts University, Medford, MA.

Droop, M.R. 1974. The nutrient status of algal cells in continuous culture. J. Mar. Biol. Assoc. UK, 54:825-855.

Rhee, G.-Y. 1973. A continuous culture study of phosphate uptake, growth rate and polyphosphate in Scenedesmus sp. J. Phycol. 9:495-506.

USEPA Region 3. 2006. Updated Model Report for Christina River Basin, Pennsylvania, Delaware, and Maryland, High-Flow Nutrient and DO TMDL Development. http://www.epa.gov/reg3wapd/tmdl/pa_tmdl/ChristinaMeetingTMDL/.

Warwick, J.J., D. Cockrum, and M. Horvath. 1997. Estimating non-point-source loads and associated water quality impacts. Journal of Water Resources Planning and Management, 123(5): 302-310.

Tetra Tech. 2004. Total Maximum Daily Load for Nutrients in Clear Lake, Lake County, California: Technical Report. Central Valley Regional Water Quality Control Board, Rancho Cordova, CA.
http://www.swrcb.ca.gov/rwqcb5/programs/tmdl/ClearLake/cl-final-tmdl.pdf

Review:

First, I’ve not seen Dr. Dodd’s webcast, and so I’m unable to tailor my response to the comments that were made. In addition, I can’t directly address the first part of the question either (i.e., how do water quality models incorporate new understanding), yet I can partially respond to the second two parts.

With respect to the second part (i.e., ability of models to understand nutrient and algal growth dynamics), there is a wide body of literature that can be used to address these types of problems. While I doubt that any science is ever complete (in the sense that we know everything), we have a solid foundation upon which to base our models. There may be uncertainties in some areas that require additional evaluation, but the knowledge gained from previous experiences is constantly being incorporated into our understanding of the relationships between nutrients and algal growth. We do know that reducing nutrient inputs and removing nutrient accumulations in sediments are clearly beneficial with respect to reducing eutrophication.

With respect to the last part (i.e., is there any current research), the answer is that there are many federal and locally funded research efforts to identify the extent, magnitude, and mechanisms of algal growth. Many of these are evaluating alternative management procedures that may result in cost-effective reductions in nutrient loading and/or sequestration. The biological impacts of nutrient loading are also being evaluated in many ecoregions of the country. If you are interested, a few of the journals that could be examined to see what kinds of studies being performed include:

• Biogeochemistry
• Bioscience
• Environmental Management
• Journal of the American Water Resources Association
• Journal of Environmental Quality
• Journal of Water Pollution Control Federation
• Limnology and Oceanography
• Soil Science Society of America Journal
• Water Environment and Technology
• Wetlands

Although this is an extremely abbreviated list, many of these journals are helpful in trying to convey the latest research results to the scientific community, as well as to the public at large, and to regulatory and management agencies. Keeping abreast of the science, while challenging at times, means knowing the literature and having access to available studies and data sources.

Dzialowski, A.R., S.H. Wang, N.C. Lim, W.W. Spotts, and D.G. Huggins. 2005. Nutrient limitation of phytoplankton growth in central plains reservoirs, USA. Journal of Plankton Research 6(27): 587-595.

Erturk, A., A. Ekdal, M. Gurel, K. Yuceil and A. Tanik,. 2004. Use of mathematical models to estimate the effect of nutrient loadings on small streams. Fresenius Environmental Bulletin 11B(13): 1350-1359.

Gusewell, S., K. M. Bailey, W. J. Roem and B. L. Bedford,. 2005. Nutrient limitation and botanical diversity in wetlands: Can fertilisation raise species richness?. Oikos 1(109): 71-80.

Ivanikova, N.V., R. M. L. McKay and G. S. Bullerjahn,. 2005. Construction and characterization of a cyanobacterial bioreporter capable of assessing nitrate assimilatory capacity in freshwaters. Limnology and Oceanography-Methods (3): 86-93.

Maberly, S.C., L. King, M. M. Dent, R. I. Jones and C. E. Gibson,. 2002. Nutrient limitation of phytoplankton and periphyton growth in upland lakes. Freshwater Biology 11(47): 2136-2152.

Matlock, M.D., D. E. Storm, M. D. Smolen, M. E. Matlock, A. M. S. McFarland and L. M. Hauck,. 1999. Development and application of a lotic ecosystem trophic status index. Transactions of the Asae 3(42): 651-656.

Matthews, R., M. Hilles and G. Pelletier,. 2002. Determining trophic state in Lake Whatcom, Washington (USA), a soft water lake exhibiting seasonal nitrogen limitation. Hydrobiologia 1-3(468): 107-121.

Scrimgeour, G.J., W. M. Tonn, C. A. Paszkowski and C. Goater,. 2001. Benthic macroinvertebrate biomass and wildfires: evidence for enrichment of boreal subarctic lakes. Freshwater Biology 3(46): 367-378.

Squires, M.M and L. F. W. Lesack,. 2002. Water transparency and nutrients as controls on phytoplankton along a flood-frequency gradient among lakes of the Mackenzie Delta, western Canadian Arctic. Canadian Journal of Fisheries and Aquatic Sciences 8(59): 1339-1349.

Stephens, S.L., T. Meixner, M. Poth, B. McGurk and D. Payne,. 2004. Prescribed fire, soils, and stream water chemistry in a watershed in the Lake Tahoe Basin, California. International Journal of Wildland Fire 1(13): 27-35.

Stockner, J.G. and K. S. Shortreed,. 1994. Autotrophic Picoplankton Community Dynamics in a Pre-Alpine Lake in British-Columbia, Canada. Hydrobiologia 1-3(274): 133-142.

Toetz, D.W. 1999. Multiple limiting nutrients in a subalpine stream, Colorado front range. Journal of Freshwater Ecology 3(14): 349-355.

Tufford, D.L. and H. N. McKellar,. 1999. Spatial and temporal hydrodynamic and water quality modeling analysis of a large reservoir on the South Carolina (USA) coastal plain. Ecological Modelling 2-3(114): 137-173.

van der Peijl, M, J., M. M. P. van Oorschot and J. T. A. Verhoeven,. 2000. Simulation of the effects of nutrient enrichment on nutrient and carbon dynamics in a river marginal wetland. Ecological Modelling 2-3(134): 169-184.

Venterink, H.O., R. E. van der Vliet and M. J. Wassen,. 2001. Nutrient limitation along a productivity gradient in wet meadows. Plant and Soil 2(234): 171-179.

Warwick, J.J., D. Cockrum and A. McKay,. 1999. Modeling the impact of subsurface nutrient flux on water quality in the Lower Truckee River, Nevada. Journal of the American Water Resources Association 4(35): 837-851.
Wold, A.P. and A. E. Hershey. 1999. Spatial and temporal variability of nutrient limitation in 6 North Shore tributaries to Lake Superior. Journal of the North American Benthological Society 1(18): 2-14.

Zheng, L.Y., C. S. Chen and F. Y. Zhang,. 2004. Development of water quality model in the Satilla River Estuary, Georgia. Ecological Modelling 3-4(178): 457-482.

Response:

One of the reviewers stated that an afternoon DO saturation of 150% or greater may be an indication of high algal biomass, however they do state there is little literature on this. Part of the problem is that factors other than algae (such as reaeration) affect DO, making it difficult to come up with an afternoon DO saturation indicator of possible problems.
 
Just to clarify, I'm interested in using mid-afternoon DO saturation values as possible indicators of high algal biomass and early morning depressed DO levels (below 5 mg/). It is typically easier for our field people to get to monitoring sites in the mid-afternoon rather than in the early morning. If we found elevated DO saturation levels, we would go out and do more involved investigations such as early morning DO monitoring, qualitative algal assessment, etc.

Review:

A good place to look for guidance would be Professor Claude E. Boyd’s work:

1. Boyd CE, 1979, Water Quality in Warmwater Fish Ponds, Auburn University [Out of print, but used copies are available on the web . . . ]
2. Boyd CE, 2000, Water Quality - An Introduction, Kluwer Academic Publishers

While these materials primarily focus on aquaculture (fish ponds), there is much to be gained by applying this knowledge to water quality in wetlands, ponds, and lakes, as well as in slow-moving rivers and streams. In the first reference (Fig 2.6, pg. 27) he shows that a heavy phytoplankton bloom has a DO range at the surface of the 3-m deep pond that varies from approximately 3 mg/L at 6 am to about 18 mg/L in the afternoon. DO at 1 m depth and below, however, was always near zero. For a moderate phytoplankton bloom, the DO range was lower, from 6 to 12 mg/L, with the anoxic zone starting at about 2 m depth. For a sparse bloom, the DO varied between 6 and 8 mg/L, with the DO never falling below 4 mg/L at the maximum depth of 3 m.

Thus, for shallow ponds, finding the maximum DO at the surface (in the afternoon) and the minimum DO at bottom of the pond (anytime) is probably a good indicator of eutrophication. This is also likely to be true for deeper systems as well (e.g., lakes and deeper ponds), with the benthic water quality showing a more stable (anoxic for eutrophic and hypereutropic) concentration. [As an aside, we also see significant pH effects - the maximum pH of 9.5 to 11 during the day in eutrophic and hypereutrophic conditions corresponds to elevated oxygen during the afternoon. This is shown in Fig. 8.9 (pg 150) in the second reference.]

An additional point to be made is that some hypereutrophic systems have low dissolved oxygen concentrations at the surface due to mixing (wind or current-driven) that averages the high oxygen at the surface with low oxygen at the bottom. Thus, it may be that daytime DO was OK, but an early morning sample would be very low. The first reference (Fig 13.7, pg 277) also shows that the Secchi Disk Visibility is a good indicator of
Cha (r = -0.79).

Cha = 19.14 SDV ^−1.976        (1)

which would be appropriate if phytoplankton are the dominant cause of water visibility reduction. This relationship would probably not be appropriate for systems were sediment or decaying vegetation reduce the clarity, but this could be checked by noting the water color (i.e., green instead of red, orange, yellow, or black).

This analysis should be appropriate for lentic (slow-moving) rivers and streams. Yet, I am unsure how eutrophication in lotic (fast-moving) rivers and streams would be manifested, unless the source of eutrophication lies upstream and the discharge from a eutrophic, lentic waterbody is translated down- stream. In that case, the peak oxygen demand may be offset by the travel time in the lotic system. Hence, I would have to know more about the physical system to be able to advise. While total phosphorus is likely to be a good indicator of the potential for eutrophication in lotic systems, the bioavailability of the phosphorus is a key unknown. For sediment-sorbed phosphorus, the release is mediated by pH and by dissolved oxygen concentrations. An indication of the relative bioavail-ability would be a peak afternoon pH exceeding 8.5, or a benthic dissolved oxygen concentration less than 1.0. Low DO concentrations reduce the benthic iron oxyhydroxides from Fe⁺³ to Fe⁺², thus releasing the PO⁻³₄ associated with the ferric-phosphate ligand. High pH (greater than 8.5) also induce phosphate release due to alkaline desorption. For organic-bound phosphorus, the decay rate (also a function of oxygen concentration, or other electron acceptors, as well as temperature) will limit the bioavailability. Thus, the phosphorus bioavailability, and resulting phytoplankton response, is strongly dependent on ambient conditions that may or may not be reflected in the total phosphorus concentration.

Review:

This is a difficult question to answer because little is known about high DO toxicity effects in general. Question 97 has some responses to the idea of toxicity of super-saturation.

My previous suggestions on question 98 partially answer this question, but to recap:

Often the maximum and the minimum are correlated (they should be because primary producers respire at night). However, if there is extensive carbon input from outside the system, then the minimum may be lower than expected given the maximum. Once a large enough data set is established for a particular stream, pond, or general type of habitat, noon sampling may be adequate to monitor most systems. Perhaps judicious use of sondes is warranted to establish a baseline followed by spot sampling.

Review:

DO is difficult. Any DO "grab" is affected by so many factors that it is hard to generalize. I would say that any high DO saturation value would be cause for suspicion, but that this should simply trigger deploying 24 hr sondes for several days. There is little literature to suggest that a single grab has much predictive value for indicating trophic state - the higher the saturation, certainly the more suspect that there is a problem, but as for a specific DO saturation threshold that can be used to split oligo-, meso-, and eutrophic streams, I do not know of one. I do know many highly productive stream systems that are undersaturated and many very low productivity stream systems that are supersaturated.

Response:

A good place to look for guidance would be Professor Claude E. Boyd’s work:

1. Boyd CE, 1979, Water Quality in Warmwater Fish Ponds, Auburn University [Out of print, but used copies are available on the web . . . ]
2. Boyd CE, 2000, Water Quality - An Introduction, Kluwer Academic Publishers

While these materials primarily focus on aquaculture (fish ponds), there is much to be gained by applying this knowledge to water quality in wetlands, ponds, and lakes, as well as in slow-moving rivers and streams. In the first reference (Fig 2.6, pg. 27) he shows that a heavy phytoplankton bloom has a DO range at the surface of the 3-m deep pond that varies from approximately 3 mg/L at 6 am to about 18 mg/L in the afternoon. DO at 1 m depth and below, however, was always near zero. For a moderate phytoplankton bloom, the DO range was lower, from 6 to 12 mg/L, with the anoxic zone starting at about 2 m depth. For a sparse bloom, the DO varied between 6 and 8 mg/L, with the DO never falling below 4 mg/L at the maximum depth of 3 m.

Thus, for shallow ponds, finding the maximum DO at the surface (in the afternoon) and the minimum DO at bottom of the pond (anytime) is probably a good indicator of eutrophication. This is also likely to be true for deeper systems as well (e.g., lakes and deeper ponds), with the benthic water quality showing a more stable (anoxic for eutrophic and hypereutropic) concentration. [As an aside, we also see significant pH effects - the maximum pH of 9.5 to 11 during the day in eutrophic and hypereutrophic conditions corresponds to elevated oxygen during the afternoon. This is shown in Fig. 8.9 (pg 150) in the second reference.]

An additional point to be made is that some hypereutrophic systems have low dissolved oxygen concentrations at the surface due to mixing (wind or current-driven) that averages the high oxygen at the surface with low oxygen at the bottom. Thus, it may be that daytime DO was OK, but an early morning sample would be very low. The first reference (Fig 13.7, pg 277) also shows that the Secchi Disk Visibility is a good indicator of
Cha (r = -0.79).

Cha = 19.14 SDV ^−1.976        (1)

which would be appropriate if phytoplankton are the dominant cause of water visibility reduction. This relationship would probably not be appropriate for systems were sediment or decaying vegetation reduce the clarity, but this could be checked by noting the water color (i.e., green instead of red, orange, yellow, or black).

This analysis should be appropriate for lentic (slow-moving) rivers and streams. Yet, I am unsure how eutrophication in lotic (fast-moving) rivers and streams would be manifested, unless the source of eutrophication lies upstream and the discharge from a eutrophic, lentic waterbody is translated down- stream. In that case, the peak oxygen demand may be offset by the travel time in the lotic system. Hence, I would have to know more about the physical system to be able to advise. While total phosphorus is likely to be a good indicator of the potential for eutrophication in lotic systems, the bioavailability of the phosphorus is a key unknown. For sediment-sorbed phosphorus, the release is mediated by pH and by dissolved oxygen concentrations. An indication of the relative bioavail-ability would be a peak afternoon pH exceeding 8.5, or a benthic dissolved oxygen concentration less than 1.0. Low DO concentrations reduce the benthic iron oxyhydroxides from Fe⁺³ to Fe⁺², thus releasing the PO⁻³₄ associated with the ferric-phosphate ligand. High pH (greater than 8.5) also induce phosphate release due to alkaline desorption. For organic-bound phosphorus, the decay rate (also a function of oxygen concentration, or other electron acceptors, as well as temperature) will limit the bioavailability. Thus, the phosphorus bioavailability, and resulting phytoplankton response, is strongly dependent on ambient conditions that may or may not be reflected in the total phosphorus concentration.

Review:

The answer is correct, but some additional suggestions might be useful.

Since there are two major issues, high DO and low DO, these are the values necessary to find. A diurnal curve using sondes can give these data, and much more. It can also be used to estimate system metabolism.

The minimum DO should occur at night, so a sampling right at dawn can be used to determine the minimum. Sometimes minima occur at other times of night (particularly where the water is cooling substantially over the night), but this should be a good indicator. I know it is difficult to get personnel to sample early in the morning in the dark, but it is less expensive than lots of sondes at $7000 each. For maximum DO, usually it should be sampled near or immediately after noon (solar noon, so account for daylight saving time) on a sunny day. If a stream is to be sampled, then sample below an extended reach with open canopy and exposure to light.

Often the maximum and the minimum are correlated (they should be because primary producers respire at night). However, if there is extensive carbon input from outside the system, then the minimum may be lower than expected given the maximum. Once a large enough data set is established for a particular stream, pond, or general type of habitat, noon sampling may be adequate to monitor most systems. Perhaps judicious use of sondes is warranted to establish a baseline followed by spot sampling.

The problem with the Welch system is that it monitors mean DO and does not get the swings. The extremes are what we are most interested in with regard to nutrient criteria. It will be more sensitive to whether a system is net autotrophic or heterotrophic.

Review:

I have little to add to the two previous comments. I agree with the first reviewer that if one was to develop a standard protocol that was most likely to catch the minima, that early morning sampling would be the most cost efficient approach for streams and rivers. This would be an easy and less expensive option. More involved methods would have to involve developing "rating curves" between surrogate measures and DO minima - which would mean measuring diel DO in a number of sites and trying to find a good predictor (e.g., benthic organic matter, TOC, pre-sunrise DO, etc.).

This is one problem with relying on DO rather than other measures for criteria.

Response:

Yes, it is reasonable to use a measure of biological condition as a response variable, i.e. a valued ecological attribute, in nutrient criteria development. The streams and wetlands guidance documents both emphasize the use of measures of biological condition and other valued ecological attributes as parts of the suite of parameters used in development and implementation of nutrient criteria.

Biological condition, such as measured by a floristic quality index, is an important ecological attribute for which we manage waters. The aquatic plant flora of waterbodies responds directly to nutrient enrichment, as algal biomass or algal species composition would. Biological condition of aquatic plants, as measured by an FQI, seems particularly appropriate in shallow, dystrophic lakes because algae accrual may be constrained by shading or high grazing activity. In addition, biological condition of aquatic plants is an appropriate response variable in wetlands, which may be a better model for small, shallow dystrophic lakes than the suite of criteria proposed (TN, TP, chl a, secchi depth) for deeper and larger lakes.

Review:

EPA Technical Guidance Manual for Lakes & Reservoirs recommens measuring TN, TP, TOC, algal biomass (phytoplankton), and water clarity. Periphyton biomass is not recommended in the guidance documents due to relatively small amount of periphyton biomass and productivity in deep water lakes.

It is always a good idea to find out what uses you want to protect in your systems before making a decision on what indicators you want to choose to protect these uses. If aquatic life uses are the valued ecological attributes to be protected, then it is valuable to use either a plant or a macroinvertebrate index to assess these lakes. Minnesota PCA has developed both macroinvertebrate and plant IBIs for depressional wetlands in the state(Gernes and Helgen 2002). These IBIs may be adapted and re-calibrated to suit lakes in Northern Minnesota. In the shallow wetland-type lakes, aquatic vegetation seems dominant, whereas macroinverbrate indices are questionable due to unstable habitat provided by these types of vegetation (Wilcox 2002). Therefore, a plant IBI may be preferred. References for FQI/IBI in wetlands can be found in the list of papers attached. In addition, phytoplankton sampling may also be considered in the deep water area of lakes.

Batzer, D. P., B. J. Palik, and R. Buech. 2004. Relationships between environmental characteristics and macroinvertebrate communities in seasonal woodland ponds of Minnesota. Journal of the North American Benthological Society 23:50-68.

Gernes, M. C. and J. C. Helgen. 2002. Indexes of Biological Integrity (IBI) for Large Depressional Wetlands in Minnesota. Minnesota Pollution Control Agency, Final Report to U.S. Environmental Protection Agency. (http://www.pca.state.mn.us/water/biomonitoring/wet-report-largewetland.pdf [^] ).

Godefroid, S., and N. Koedam. 2003. Identifying indicator plant species of habitat quality and invasibility as a guide for peri-urban forest management. Biodiversity and Conservation 12:1699-1713.

Guntenspergen, G. R., S. A. Peterson, S. G. Leibowitz, and L. M. Cowardin. 2002. Indicators of wetland condition for the Prairie Pothole Region of the United States. Environmental Monitoring and Assessment 78:229-252.

Houlahan, J. E., P. A. Keddy, K. Makkay,and C. S. Findlay. 2006. The Effects Of Adjacent Land Use On Wetland Species Richness And Community Composition. Wetlands 26(1) 79-96.

Johnson, S., and E. Rejmankova. 2005. Impacts of land use on nutrient distribution and vegetation composition of freshwater wetlands in Northern Belize. Wetlands 25:89-100.

Lopez, R. D., and M. S. Fennessy. 2002. Testing the floristic quality assessment index as an indicator of wetland condition. Ecological Applications 12:487-497.

Matthews, J. W., P. A. Tessene, S. M. Wiesbrook, and B. W. Zercher. 2005. Effect of area and isolation on species richness and indices of floristic quality in Illinois, USA wetlands. Wetlands 25:607-615.

Miller, S. J., and D. H. Wardrop. 2006. Adapting the floristic quality assessment index to indicate anthropogenic disturbance in central Pennsylvania wetlands. Ecological Indicators 6:313-326.

Miller, S. J., D. H. Wardrop, W. M. Mahaney, and R. R. Brooks. 2006. A plant-based index of biological integrity (IBI) for headwater wetlands in central Pennsylvania. Ecological Indicators 6:290-312.
  
Simon, T.P. P.M. Stewart, and P.E. Rothrock. 2001. Development of multimetric indices of biotic integrity for riverine and palustrine wetland plant communities along Southern Lake Michigan. Aquatic Ecosys. Health and Manage. 4: 293-309.

Tangen, B. A., M. G. Butler, and J. E. Michael. 2003. Weak correspondence between macroinvertebrate assemblages and land use in Prairie Pothole Region wetlands, USA. Wetlands 23:104-115.

Wilcox, D. A. and J. E. Meeker. 1991. Disturbance effects on aquatic vegetation in regulated and unregulated lakes in northern Minnesota. Canadian Journal of Botany 69:1542–1551.

Wilcox, D. A., J. E. Meeker, P. L. Hudson, B. J. Armitage, M. G. Black, and D. G. Uzarski. 2002. Hydrologic variability and the application of index of biotic integrity metrics to wetlands: A Great Lakes evaluation. Wetlands 22:588-615.

Review:

Both the answers are very good. This approach seems aimed at approximating the native condition, so would be protective of biotic integrity.

Response:

In my understanding, states and tribes can propose to classify streams based on scientifically valid differences among regions.

Very turbid streams would be expected to have low levels of periphyton even if nutrients are high, so this would be a poor response variable for these particular streams. Nutrients could still link to biotic integrity in these streams because nutrients can alter the degree that carbon enters the food web through heterotrophic processes.

It is not necessarily true that silty or sandy substrata have low biomass of periphyton, so this alone should not be criteria for eliminating a stream or group of streams from nutrient control. However, if little light reaches the substrate because of high values for suspended solids in the water column, substantial buildup of periphyton is not expected except in very shallow habitats.

Finally, nutrient criteria may be important even in silty streams if the receiving waters downstream allow settling of the sediments and the dissolved nutrients cause eutrophication in those waters.

Review:

This is an excellent question with relevance to many areas of the country. There are several ways to
approach this issue:

1. There is good reason to relate periphyton to the substrate in that the nutrient content of sediments can be related to the particle size of the sediments - finer sediments can sorb or contain more nutrients than coarser sediments. Clays may have a high Cation Exchange Capacity which allows them to store ammonium, and iron coatings and minerals can store phosphorus. In addition, coarse and fine organic matter can decompose, releasing nutrients. A cobble or sand bottom may not have the same inventory of nutrients as does a finer (silty, clayey, or mucky) bottom.

2. There may be a relationship between the particle size distribution of sediments within the water column during storm events with those on the bed. Normally, larger sediments require greater flows (energy) to be mobilized than do smaller sediments. Stream beds dominated by larger sediments may have high flows that remove the smaller sediments, or they may exist in environments where finer sediment inputs to streams are absent. Streams that are dominated by finer sediments may exist in calmer environments, or where there is so much available fine sediment that storms are unable to export the total fine sediment volume.

3. Yet, it may be that nutrient inputs are unrelated to sediments, in that a point-source (wastewater treatment facility) or nonpoint source (nitrate releases from agriculture, groundwater inputs from septic systems or soil drainage) stimulate periphyton production.

4. Also, the type of flows may act to physically restrict periphyton abundance. High velocities may scour channel beds, while lower velocities may allow growth. In this case, it is likely to be dependent on the type of periphyton present, and its ability to withstand flow variability.

My recommendation would be to establish a classification system based on substrate, with the further documentation of:

1. The primary source of nutrients to the stream system - either from (mineral or organic) sediments or dissolved sources. This would require a baseline inventory of the nutrient content of the dissolved and sediment (both suspended and substrate) fractions.

2. The effect of (seasonal) flow variability on substrate, suspended sediments, and periphyton abun-dance. This would require seasonal sampling for a range of flows to determine the effects of discharge on sediment composition, substrate alteration, and periphyton abundance. The intent of this documentation is to provide a robust physical understanding of the relationship between nutrients and the proposed response variable.

Waters, Thomas F., 1995, "Sediment in Streams: Sources, Biological Effects and Controls", Monograph 7, American Fisheries Society, Bethesda MD.

Wetzel, Robert G., 2001, "Limnology: Lake and River Ecosystems", Third Edition, Academic Press.

Gordon, Nancy D., Thomas A. McMahon, Brian L. Finlayson, Christopher J. Gippel, Rory J. Nathan, 2004, "Stream Hydrology: An Introduction for Ecologists", 2nd Edition, Wiley Press.

Review:

I agree with the respondent. Sure you can classify if it can be defended. And defense means that use of substrate as a classification tool can be used to reduce variability in natural clarity. I am skeptical of being able to use substrate as a classification element because 1) it is affected by many anthropogenic factors, including historical ones and 2) we have limited data on subtrate. If a subecoregion has naturally high sediment and different substrata that would affect clarity naturally, than this should be reflected in looking at nutrient and chlorophyll concentrations of REFERENCE sites among subecoregions - that ecoregion would stand out as unique during the classification process. Then you could justify it without having to rely on substrate measures - just use the subecoregion map.

Response:

Three approaches come to mind. 1) Use a light meter (quantum sensor)and measure the underwater light field and attenuation of PAR. This is text book limnology but requires spendy equipment, some care in the field to get the information and a bit of math to get attenuation values (nothing difficult). 2) Davis-Colley 1988 (Limnology and Oceanography, 33:616) promotes the use of measuring water clarity with a horizontal black disk. He has several subsequent papers and this approach seems ideal for this system. It deserves serious consideration. I believe it would be much less complex than a standard light meter and therefore conducive to routine monitoring with field crews. 3) a standard appoach would be to measure absorption in a spec at 440 nm, but in these clear waters it would require a 10 cm cell for reliable readings.

Review:

Water clarity in rivers and streams commonly use turbidity and total suspended solids. While related,
they actually measure slightly different things:

1. The common turbidimeter usually takes a few seconds to take a reading, allowing the sand size
fraction to settle out. In this case the turbidity would be too low by the amount of large particulates
that have settled. Yet in a clear lake, this would not be a problem, because the large sand and silt
sized particles would likely settle very quickly.

2. The total suspended solids measurement usually requires that the sample be oven-dried. Thus, any
phytoplankton present in the sample would be reduced in mass by the amount of tissue dessication.
- the turbidity may actually be higher in this case than the total suspended solids measurement.

3. Dissolved organic matter may (or may not, depending upon the type of turbidimeter) yield a
measurable turbidity. (Filtered tea and coffee, for example, may induce a turbidity.) The total
suspended solids would not detect the cloudiness due to ability of the organics to pass through the
filter.

For these reasons, I suggest that total suspended solids be avoided, and that turbidity be preferred. Yet
there are still several reasons why turbidity may not be adequate:

1. The turbidimeter may not have sufficient resolution to detect very low values. This could be a
problem in very clear lakes.

2. There may be a narrow layer within the lake that is cloudy, while other layers above and below
are clear. In this case, sampling at specified depths may show low turbidity, and completely miss a
turbid layer that was not sampled. The advantage of the Secchi disk lies in the integration of light
absorption over the entire depth.

Review:

If you must have a number, the approaches suggested by JJones are worth trying, but remember that each one (even the black disk method) will require extensive cross-calibration with your existing Secchi and turbidimeter methods to yield readings that will be meaningful in the context of the other data. If comparison to a baseline for the particular waterbody is sufficient, then it may not be as necessary to cross calibrate. I also agree with TRasmussen that TSS is not a good substitute.

What you have is a detection limit issue. The depth of a lake or bay determines the detection limit of Secchi transparency. If there is a decrease in transparency (increased turbidity), yet you can still see the Secchi disk on the bottom, can you regulate it? Will the lake or bay go on a TMDL list? Would you institute a watershed plan? If yes, then it would be worthwhile to document the changes.

Another way to document changes in transparency would a be a photographic record of Secchi disks on the bottom of a lake or bay. If turbidity increases, the Secchi disk will apear to fade, but may not disappear, in images taken over time. All you would need is a relatively inexpensive waterproof camera and a standardized way of shooting the picture (bright, high sun, midday, late spring-early summer, no recent wind or rain storms, same location in lake). While not quantitative, changes that are obvious and consistent in a photographic record are sufficient for the public and for judges if your regulatory decisions are challenged.

Context:

We are struggling coming up with a water column chl a number that will protect our Class B aquatic life communities because we do not have a dataset of paired water column chl a and biomonitoring results. Class B waters must support aquatic life with "no detrimental change" from refence conditions. In practice, this is the point where some of the most sensitive species may have been lost but the community still has a pretty high richness and relative abundance of senstive species.

Response:

This is a question many people are interested in but it is difficult to get a precise answer. In order to answer this question, we have to understand the relationships among nutrient concentrations, water column Chl-a level, and biological integrity (e.g. macroinvertebrates) in lotic streams.

Measurement of Chl-a is considered an appropriate surrogate for primary production in water column because researchers have consistently found strong correlations between Chl-a values and algal biomass. Water column Chl-a is mainly composed of phytoplankton, with a portion of dislodged periphyton and metaphyton. Both phytoplankton diversity and biomass in low order streams are very low but increase in higher-order large rivers. The productivity of phytoplankton in lotic streams is also low compared with other primary producers in most streams. As a result, macroinvertebrates and fish feed mainly on other sources. Therefore, the increase in phytoplankton abundance is less likely affecting macroinvertebrate and fish communities directly. Rather, an increase in phytoplankton abundance often indicates nutrient enrichment in the system (e.g., van Nieuwenhuyse and Jones 1996), when other ecosystem functions, such as periphyton productivity, bacteria productivity, and biogeochemical processes have also been altered. The consequence of changes in other ecological attributes, (i.e., pH, dissolved oxygen, and habitat structure) could directly impact biological integrity.

One of the consequences of nutrient enrichment may be loss of sensitive macroinvertebrate and fish taxa and increases in tolerant taxa. Mayflies, a group of invertebrates that are considered sensitive to environmental pollutants, show the highest relative abundance when algal biomass is at intermediate levels (Miltner and Rankin 1998). The abundance of scrapers, a functional group that is closely related to grazers, is highest when nutrient levels are elevated, indicating positive effects from increased algal availability (Miltner and Rankin 1998). The strong correlation between fish metrics and nutrient pollution (Miltner and Rankin 1998, Stockner et al. 2000, Chambers et al. 2006, Wang et al. 2007) indicates that nutrient enrichment has contributed to changes in the structure of fish assemblages. The obvious impact at high nutrient loads is the reduction in DO, which would exclude many sensitive taxa. In addition, excess algal growth (periphyton) would eliminate important feeding and respiration habitat, further reducing survivorship.

There are sporadic reports showing correlation between water column Chl-a and biological metrics, but I am not aware of evidence of a strong relationship in northern U.S. lotic systems. Other alternate approaches have to be sought to develop a Chl-a criterion which make ecological sense. This has to be done at a local and regional scale where data is available. A recent paper by Royer et al (2008) set a good example of this approach. In this study, they found that watershed area is a good predictor of sestonic Chl-a and suggested that Chl-a may be an appropriate criterion for the large rivers in Illinois but inappropriate for small rivers and streams.
  
Other studies have developed for nutrient endpoints for their regions. For example, van Nieuwenhuyse and Jones (1996) consider oligotrophic-mesotrophic boundary as 10µg/L and mesotrophic-eutrophic boundary as 30µg/L. Oregon has a Chl-a criterion for phytoplankton of 15µg/L for rivers (OAR 340-041-0019). Reckhow et al. (2005) estimated that the 40µg/L Chl-a criterion has a 60% probability of supporting the designated uses of the Neuse Estuary and recommended that a 10µg/L criterion would be more suitable from a use-protection perspective. The US EPA Chl-a criterion for nutrient ecoregion VIII is 0.63µg/L based on 25th percentile data distribution.

Chambers, P. A., R. Meissner, F. J. Wrona, H. Rupp, H. Guhr, J. Seeger, J. M. Culp, and R. B. Brua. 2006. Changes in nutrient loading in an agricultural watershed and its effects on water quality and stream biota. Hydrobiologia 556:399-415.
 
Miltner, R.J. and E.T. Rankin. 1998. Primary nutrients and the biotic integrity of rivers and streams. Freshwater Biology 40: 145-158.

Morgan, R., K. Kline and S. Cushman. 2007. Relationships among nutrients, chloride,
and biological indices in urban Maryland streams. Urban Ecosystems 10:153-166.

Reckhow, K.H., G.B. Arhonditsis, M.A. Kenny, L. Hauser, J. Tribo, C. Wu, K.J. Elcock, L.J. Steinberg, C.A. Stow, and S.J. Mc Bride. 2005. A predictive approach to nutrient criteria. Environmental Science & Technology 39(9): 2913-2919.

Royer et al. 2008. Assessment of Chlorophyll-a as a Criterion for Establishing Nutrient Standards in the Streams and Rivers of Illinois. J Environ Qual. 37: 437-447.

Stockner, J. G., E. Rydin, and P. Hyenstrand. 2000. Cultural oligotrophication: Causes and consequences for fisheries resources. Fisheries 25:7-14.

van Nieuwenhuyse, E.E. and J.R. Jones. 1996. Phosphorus-chlorophyll relationship in temperate streams and its variation with stream catchment area. Can. J. Fish. Aquat. Sci. 53: 99-105.

Wang, L., D. Robertson and P. Garrison. 2006. Linkages between nutrients and
assemblages of macroinvertebrates and fish in wadeable streams: implications to nutrient criteria development. Environ. Management online 10.1007.