River Continuum Concept Ecological Limit Functions for Fish and Benthic Data in Virginia

Science Center Objects

The ecological limit functions (ELF) developed in cooperation with the Virginia Department of Environmental Quality (DEQ) are a graphical representation of the current and historical state of aquatic biota in Virginia streams.  The goal of this study was to quantify the potential species richness and habitat response to flow alteration using available long-term ecological data. Fish and benthic richness patterns were discerned from Ecological Limit Functions (ELFs) that describe the relations between streamflow and species richness. This River Continuum Concept (RCC) ELF framework, employing Virginia’s extensive fish and benthic richness database, may provide an alternative method for assessing flow depletion impacts without the need for extensive habitat characterization or in-depth flow modeling.  The ELFs were used to calculate rate of change in richness represented by the ELF slope as a function of drainage area, annual streamflow, or monthly streamflow and annual or monthly flows.  Species richness responses varied as a function of hydrologic unit classification (from larger regions HUC6 to smaller basins HUC10), ecoregion, or major drainage.

Background    Building Ecological Limit Functions    Examining Slopes of Ecological Limit Functions   

Change in Habitat and Species    Management Implications and Conclusions    References



This study sought to:

  • Examine fish and benthic richness patterns discerned from ELFs developed with automated methods (Kleiner et al., 2019) at varying spatial scales.
  • Calculate rate of change in richness represented by the ELF slope across watershed unit  scales and annual or monthly flows.
  • Calculate percent habitat change in a case study of IFIM site data and compare with corresponding percent species change from ELF data across watershed unit scales and annual or monthly flows.
  • Explore the extent to which ELF equations might be used as an estimate of absolute or relative risk of species richness loss to inform water supply management.

The River Continuum Concept (RCC) postulates that richness or density in aquatic communities will increase with stream size, up to mid-order streams (Figure 1).  The richness in a community may be predicted by stream size (which represents an integration of habitat, nutrient inputs, and energetics of the system (Vannote et al., 1980, Minshall et al., 1985, Angermeier and Schlosser, 1989, Vander Vorst et al., 2017).


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 1. Plot showing increase and decrease of species richness with stream discharge or stream order.  (Minshall et al., 1985).


The extensive biological sample database developed by DEQ in conjunction with TetraTech Inc. (2012) linked each sample location to the NHDPlus V2 COMID.  Additional attributes like drainage area, stream order, and streamflow (monthly or mean annual flow, MAF) were also linked. This attribution of the sample data created a consistent flow-regime estimate for all of the 10,000-plus sample locations across the Commonwealth, and allowed the evaluation of ecological conditions anywhere.


Virginia Ecoflows Figure 2

Figure 2. Fish and benthic sample data 1971 – 2010 across the commonwealth of Virginia. EDAS Database (Tetratech Inc. 2012) includes data from: DEQ (ProbMon) | VDGIF (VaFWIS) | US EPA (MAHA) & (MAIA) | USGS (NAWQA) | VCU (INSTAR) | TN (MARIS) |DCR (Natural Heritage)


Building the ecological limit functions

Using R-scripting the large datasets could be evaluated based on a variety of watershed scales or Hydrologic Unit Codes (HUC) and ecoregions.  Quantile regression was used to extract the upper 80% of data (upper quantile) for use in the identification of the maximum species richness across stream sizes (Figure 3).  The Ecological Limit Function was generated by fitting a linear regression model to the data in the upper quantile which could then be used to calculate the change in richness (slope) as a function of watershed size (Kleiner et al. in prep).


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 3. Diagram of the ecological limit function. 

The RCC breakpoint (Figure 3), or inflection point, represents a shift in the slope of the ELF upper quantile regression of maximum richness values from increasing to decreasing or to a flat, zero slope. RCC Breakpoints were developed for all watersheds across the commonwealth to indicate the size of rivers for which the River Continuum Concept is applicable, valid, and increasing along a stream size gradient, and how these RCC Breakpoints vary regionally and seasonally. The RCC breakpoints represent a maximum richness threshold reached between mid-order streams (5th or 6th order), and larger-order streams which typically have lower species richness than the maximum observed at the RCC breakpoint.  For example, the smallest streams in the Potomac River watershed had five to eight taxa, increasing up to 32 taxa at the RCC breakpoint of 306.8 ft3/s (Figure 4).


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 4. Potomac HUC6 ecological limit function showing the breakpoint at 306.8 ft3/s and maximum species richness of 32 taxa.

Examining the slopes of the Ecological Limit Functions

The ELF slope was a primary statistic that describes the potential change in species richness in relation to flow. Sufficient fish data coverage was present for the development of statistically significant ELFs for total richness as well as a few select families or trait groups (biometrics) for all eight HUC6 units, and for 35 of 50 HUC8 units, and 39 out of 572 HUC10 units.  ELFs with a wide range of slopes in HUC8 watersheds are dispersed throughout different major basins (Figure 4a). When there were adequate data, ELFs for HUC10 units has similar or slightly elevated slopes than the HUC8 ELFs (Figure 4a). Local patterns in species richness and habitat conditions result in more unique ELFs slopes and species profiles for HUC8 and HUC10 whereas the ecoregions and to some extent, HUC6, represent more generalized groupings. 

Virginia USGS-DEQ Ecoflows Project Results Image

Figure 5. Maps showing the range of slopes of the Ecological Limit Functions ELFs for A) HUC8, HUC10, and B) Ecoregion.

Rate of Change in Richness Represented by Slope: Slopes of the ELF can be used to calculate richness change as a constant number (#) (Figure 6) and resulted in one numeric rate for each ELF that was applicable to any size stream, but these numbers differ across watersheds in a similar way as slope varied.  With smaller watershed unit sizes (HUC10 for example) the ELF slopes and the predictions of number of species change were greater than they would be for a HUC8 watershed which contained the HUC10. To obtain specific results for individual streams, the slope, stream size (represented by Mean Annual Flow), and the initial species richness value are used to develop an understanding of %species change for that size stream (Figure 7). The percent (%) richness change resulted in incremental species richness change along the streamsize gradient within each watershed unit.


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 6. Number of taxa change for each HUC8 watershed for 10, 20, and 40% flow reductions scenarios.


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 7. Elf plot for one HUC8 watershed A) and percent species change for the same HUC8 B).

The maximum species richness at the RCC Breakpoint dictated the slope of the ELF.  Those HUCs that had the highest maximum species richness also had the greatest potential for species change with flow reduction.  ELF slopes across Virginia tended to increase from east to west, and HUC8 watershed units that are located in the mountainous ecoregions tended to have higher ELF slopes than those that are in the Piedmont or Southeastern Plains.  There appears to an Ecoregion signature affecting these patterns as the ELFs with steeper slopes are found in places with higher stream gradients, but the maximum richness was dictated by major drainage basin also noted as an important variable by Angermeier and Winston, 1999 and Hain et al., 2017.


Percent change in habitat and percent change in species

The use of Instream Flow Incremental Methodology (IFIM) studies habitat change modeling as a surrogate for aquatic organism response has been widely used for the past 60 years; however, quantitative measures of species change associated have not been as easily obtained. Using the ELF predictions for species change at the same locations where IFIM models exist in Virginia this pilot examined the similarity of predictions of percent habitat-change and percent species-change. During summer and fall months of the year when habitat is most limited, flow reductions contribute to reductions in fish habitat across all five habitat sites (Figure 8).  Stream size is a main determinant for the percent habitat-change and percent species-change at these groups of sites.


Virginia USGS-DEQ Ecoflows Project Results Image

Figure 8. August predictions for percent habitat-change and percent species-change with a 20% flow reduction at habitat modeling sites less than 500 square miles in size. (Habitat data calculated from Averett et al., 2004; Krstolic et al., 2006; Leonard et al., 1986; EA Engineering, Science, and Technology I., 2009).

This study demonstrated intersections between habitat and species richness relations with streamflow, reinforcing each approach, and also demonstrated ways that habitat is not a one-to-one reflection of richness, and suggesting the need to study each.  Both approaches show that small streams have a greater sensitivity to flow reduction, and identified a breakpoint or threshold, below which flow reductions result in decreasing habitat and richness.  However, both methods might result in differing allowable percent-reduction in flows, so further understanding the interplay between percent richness change and percent habitat change will be important in developing concise assessments of the sensitivity of streams to flow reductions.

Management implications and conclusions

The majority of water withdrawals in the Commonwealth are either direct withdrawals without storage, or are explicitly governed by percent allowable rules, so the development of datasets to support the development of ecologically protective management guidelines for the majority of stream miles in Virginia was of extreme importance to the management community.   Richness-flow relationships from the ELF were developed for all the biological data available in the commonwealth using R-scripting to describe the relation between species richness and steam size or flow volume (MAF).  The strength of these relationships suggest the slopes and rates of change for ELFs could be used as a strong indicator for quantifying and predicting the impacts of consumptive water-supply. Species richness change could be used as an indicator of negative effects due to consumptive water use and used to characterize potential risk during water-use permit development.


  • Angermeier, P.L. and Schlosser, I.J., 1989. Species-Area Relationship for Stream Fishes. Ecology, 70(5), pp.1450–1462.

  • Averett A.W., Persinger J., Chan M.D., Lozinski J.L. and Orth D.J., 2004. Stream Habitat Modeling to Support Water Management Decisions for the North Fork Shenandoah River, Virginia. Blacksburg, VA: Virginia Tech Department of Fisheries & Wildlife Sciences.
  • Commonwealth of Virginia, 2015, State Water Resources Plan, Draft.  Virginia Department of Environmental Quality.
  • EA Engineering, Science, and Technology I. 2009. “Instream Flow Incremental Methodology (IFIM) Studies on the North Anna and Pamunkey Rivers, Virginia.” Sparks, MD.
  • Karr, J.R., Fausch, K.D., Angermeier, P.L., Yant, P.R., Schlosser, I.J., 1986. Assessing biological integrity in running waters: A method and its rationale. Illinois Natural History Survey Special Publication 5, 5(August), p.28. Available at: http://limnoreferences.missouristate.edu/assets/limnoreferences/Karr_et_....
  • Krstolic, J.L., Hayes, D.C.; Ruhl, P.M., 2006. Physical habitat classification and instream flow modeling to determine habitat availability during low-flow periods, North Fork Shenandoah River, Virginia. Scientific Investigations Report 2006-5025, 55 p. https://pubs.usgs.gov/sir/2006/5025/
  • Leonard P.M., Orth D.J. and Goudreau C.J. ,1986. Development of a Method for Recommending lnstream Flows for Fishes in the Upper James River, Virginia.
  • Minshall, W.G., 1988. Stream Ecosystem Theory : A Global Perspective G . Wayne Minshall Journal of the North American Benthological Society, Vol . 7 , No . 4 , Community Structure and Function in Temperate and Tropical Streams : Proceedings of a Symposium ( Dec ., 1988 ), p. North American Benthological Society, 7(4), pp.263–288.
  • Richter, B.D., Davis, M.M., Apse,C., and Konrad, C., 2012. A presumptive standard for environmental flow protection. River Research and Applications, 28, 1312–1321
  • Rosenfeld, J.S., 2017. Developing flow – ecology relationships : Implications of nonlinear biological responses for water management. Freshwater Biology, Vol. 62, pp.1305–1324.
  • U.S. Geological Survey and the U.S. Environmental Protection Agency, 2012, National Hydrography Dataset Plus - NHDPlus Version 2, Washington D.C., available on-line:  https://www.epa.gov/waterdata/nhdplus-national-hydrography-dataset-plus.
  • Vander Vorste, R., McElmurray, P., Bell, S., Eliason, K.M., and Brown, B.L., 2017. Does stream size really explain biodiversity patterns in lotic systems? A call for mechanistic explanations. Diversity, 9(3), pp.1–21.
  • Vanote, R.L. et al., 1980. PERSPECTIVES. The River Continuum Concept, Canadian Journal of Fisheries and Aquatic Sciences, Vol. 37, No. 1, pp.130–137.