Environmental Flow Research in the Tennessee River Basin

Science Center Objects

The objective of this project is to improve understanding of how alteration of streamflow characteristics affects the ecological health of rivers and streams in Tennessee. Initial efforts are aimed at identifying critical streamflow characteristics and providing a set of statistical tools and analytical approaches for the prediction of these characteristics. Application of these tools will enhance understanding of how hydrologic alteration may change fish community structure in the Tennessee River basin.

This research is part of the USGS Cooperative Funding Program and is supported by the Tennessee Wildlife Resources Agency, The Nature Conservancy, Tennessee Valley Authority, and the Tennessee Department of Environment and Conservation.

The objective of this project is to improve understanding of how alteration of streamflow characteristics affects the ecological health of rivers and streams in Tennessee. Initial efforts are aimed at identifying critical streamflow characteristics and providing a set of statistical tools and analytical approaches for the prediction of these characteristics. Application of these tools will enhance understanding of how hydrologic alteration may change fish community structure in the Tennessee River basin.

This research is part of the USGS Cooperative Funding Program and is supported by the Tennessee Wildlife Resources Agency, The Nature Conservancy, Tennessee Valley Authority, and the Tennessee Department of Environment and Conservation.

 

Key Findings

Streamflow influences fish health and abundance in Tennessee

Analysis of streamflow and fish-community data across the Tennessee River Valley identified three aspects of streamflow essential to habitat suitability and food availability for insectivorous fish communities: constancy (flow stability or temporal invariance), frequency of moderate flooding (frequency of habitat disturbance), and rate of streamflow recession. Watershed management decisions that minimize change in these aspects of streamflow have the potential to increase the health of the fish community. 

Insectivorous fishes, such as Percinidae darters, Cyprinidae minnows, and Noturus madtoms, are among the most jeopardized fish in the Tennessee River Valley. Insectivorous fish represent a middle ground in the trophic structure of a stream, feeding on invertebrates while being prey for predator species. For their eggs to hatch insectivorous fish must lay them in gravel beds that are relatively clear of sediment. Insectivorous fish are sight-feeding and need clear water for feeding.

Constancy is a measure of flow stability—the consistency of streamflow from one day to the next. Constancy reflects average conditions and is most closely associated with base flows. Persistence of streamflow at base-flow levels determines the available wetted perimeter of the channel. Stability of wetted perimeter corresponds to stable and available habitat for invertebrate colonization and subsequent uptake by insectivorous fish. Equally important, stability of base-flow is critical to maintaining water-quality conditions such as dissolved oxygen, temperature, and basic water chemistry. Insectivorous fish scores increased with increasing constancy.

Frequency of moderate flooding is defined as the average number of occurrences per year of floods with magnitudes that are at least three times the median annual flow. It is speculated that the velocity and stream power associated with floods of this magnitude are sufficient to remove silt from the substrate and moderately disturb the bed material. Decreased siltation and increased water clarity from decreased silt have proven to be beneficial to sight-feeding fishes. Average stream velocities for the streamflow three times the median annual are 3.9, 1.6, and 0.65 feet per second for the Oconoluftee River, North Fork Holston River, and Wartrace Creek. Insectivorous fish scores increased with the decreasing frequency of moderate flooding. The negative correlation seen in our study indicates that insectivorous fish respond positively to decreased disturbance. Habitat becomes increasingly unstable with increasing frequency of moderate floods, resulting in the decline of insectivorous fishes and invertebrates. 

Rate of streamflow recession is a measure of how fast or slow streamflow recedes to baseflow following a flood peak. The rate of streamflow recession from runoff events can provide habitat-limiting factors for different segments of the fish community. High recession rates elicit several consequences in the stream, including stranding fish in isolated pools made available during high flows and limiting the amount of time for passage of fish between potential spawning and feeding areas. High streamflow recession rates have been associated with saturated streambank failure, potentially increasing embeddedness through higher sediment loadings and decreasing the clarity of the water. Sediment deposition from bank failure also disrupts and diminishes available spawning and feeding habitat. Diminished habitat results in crowded conditions that result in hybridization between fish species, which is an undesirable occurrence. In this study, insectivorous fish scores increased as recession rates decreased. Additional details regarding the correlation between fish community structure and streamflow characteristics can be found in the journal article titled “Relating streamflow characteristics to specialized insectivores in the Tennessee River Valley: a regional approach,” published in Ecohydrology.

 

Climate and basin factors can be used to predict ecologically-relevant aspects of streamflow

 

Throughout the US there are many more places were ecological sampling has occurred than where streamflow data is collected. Because of this disconnect, streamflow at places of ecological sampling must be estimated if any correlation between these two factors is to be determined. There are several ways to do this; we used a statistical approach. Streamflow is dictated by climate and basin influences, such as monthly mean precipitation, percent forest, and depth to bedrock. A statistical model was built to relate climate, land use, physical and regional basin factors to 19 ecologically-relevant aspects of streamflow (called streamflow characteristics). There were 231 sites in the Tennessee and Cumberland River basins available for model development. Some aspects of streamflow such as mean annual runoff, were highly predictable using statistical methods, other streamflow characteristics like annual maximum flow were less so. 

Streamflow characteristics are numerical values that describe the characteristic temporal pattern of low, high and average flows within a stream. Streamflow characteristics can be grouped in functional categories such as the magnitude, frequency, duration, timing, and rate-of-change in flow. Our statistical model is comprised of 19 separate equations, each predicting one of the 19 ecologically-relevant streamflow characteristics identified in earlier work. Basin characteristics used for in the models can be subdivided into the four following categories:

Climate—these include monthly mean precipitation, difference in January precipitation compared to average monthly precipitation, daily temperature range, and difference in minimum and maximum August temperatures compared to the year. Daily temperature range occurred in the most equations, 17 out of 19.

Land use—includes percent forest and percent agriculture. Even though percent urban isn't included as a variable, this information is still incorporated into the model structure. Percent agriculture turned out to be 1.5 to 2 times more important than percent forest.

Physical Properties—describe physical aspects of the watershed such as ability to generate runoff, mean elevation, percentage of permeable soil and depth to bedrock.  Mean elevation was very important in 11 equations.

Regional Physiography—include the number of days in a streamflow recession (connectivity to baseflow or groundwater discharge) and the percent of the watershed that within the Blue Ridge or Interior Plateau level III ecoregions. Regional factors were used in 17 of the 19 equations and were very influential in most equations.Place is more important than land use for influencing streamflowWhen predicting ecological-relevant streamflow characteristics the most influential variables relate to where the site is located (regional physiology) versus what type of land cover (forest, agriculture, or urban) is present.  In our statistical models, regional physiology was the most influential group of variables in predicting streamflow characteristics. This includes the percent of a watershed within either the Interior Plateau or Blue Ridge Level 3 ecoregions (Omernik 1987). 

Ecoregions are areas with characteristic and generally unique ecosystems. A variety of physical and biological factors such as geology, geography, vegetation, climate, soils, land use, wildlife and hydrology influence and define these broadly-defined ecosystems. Ecoregion classifications range from Level I (coarsest level with 15 ecoregions) to Level IV (finest level). We used Level III ecoregions in our study.

 

The Interior Plateau Level 3 ecoregion extends from southern Indian and Ohio to northern Alabama. Geology within the Interior Plateau consists of Mississippian to Ordovician-age sedimentary rocks such as limestone, chert, sandstone, siltstone and shale. Landforms include hills, irregular plains and tablelands; elevations in this ecoregion are lower than elevations further east in the Appalachian ecoregions (i.e. Blue Ridge). Vegetation in the area is mainly oak-hickory forest with some prairie and cedar glades. The Interior Plateau has a diversity of fish species.

The Blue Ridge extends from southern Pennsylvania to northern Georgia. The Blue Ridge is mountainous with narrow ridges, hilly plateaus and some peaks greater than 6,600 feet. Geology in the Blue Ridge consists mainly of metamorphic rocks with some small areas of igneous and sedimentary geology. Annual precipitation is highly variable, ranging from over 100 inches to 40 inches in particular areas. The southern portion of the Blue Ridge has some of the highest biodiversity in the eastern United States.

Further descriptions of Level 3 ecoregions can be found at USEPA Western Ecology Division:
http://www.epa.gov/wed/pages/ecoregions/level_iii_iv.htm

Omernik, J.M., 1987, Ecoregions of the conterminous United States. Map (scale 1:7,500,000). Annals of the Association of American Geographers, v. 77, no. 1, p. 118-125.

Altered streams have a different “hydrologic profile” than other streams

Most streams in the United States have experienced some amount of alteration from natural conditions. Alteration can be a direct or indirect result of such processes as construction of a dam, withdrawing water from the stream or converting forested land to urban land within the watershed. Streamflow characteristics for altered streams have a different “hydrologic profile” when compared to a pristine or minimally-altered stream. For our study we considered the most forested sites within each ecoregion to be minimally-altered and the reference “hydrologic profile” to be the range of calculated values for each streamflow characteristic within this group. 

For example, the 20 least-disturbed sites in the Blue Ridge all have greater than 91% forest cover and the middle 50% of streamflow characteristic values for these 20 sites define the reference “hydrologic profile” for the Blue Ridge. Some streamflow characteristics for other sites in Blue Ridge fall within this hydrologic profile, such as Mud Creek (least forested, 43%), whereas other sites, such as Nantahala River (94.3% forested) and Brier Creek (99.9%), do not. The substantial differences of streamflow characteristics for some sites (Nantahala River and Brier Creek) from the reference “hydrologic profile” are because these sites are flow-controlled rivers with upstream dams.

It was surprising that the least forested site (46%), Mud Creek, had streamflow characteristic values that lie within the reference “hydrologic profile” for the Blue Ridge. However, this supports our earlier finding that land use (percent forest or agriculture) was the least influential category when predicting streamflow characteristics. Furthermore, this finding suggests that percent forest cover, while intuitively related to pristine streamflow conditions, does not fundamentally change the hydrologic response of a stream, at least across the range of 40 to 100 percent forest cover. Additional details regarding the development of models of streamflow characteristics and discussion on hydrologic regime can be found in the journal article titled “Modelling ecological flow regime: an example from the Tennessee and Cumberland River basins” published in Ecohydrology and can be downloaded here.