Grantee Research Project Results
2001 Progress Report: Development and Evaluation of Ecosystem Indicators for Urbanizing Midwestern Watersheds
EPA Grant Number: R825871Title: Development and Evaluation of Ecosystem Indicators for Urbanizing Midwestern Watersheds
Investigators: Spacie, Anne , Hondzo, Midhat , Engel, Bernard A. , Harbor, Jonathan M.
Institution: Purdue University
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1997 through September 30, 2000
Project Period Covered by this Report: October 1, 2000 through September 30, 2001
Project Amount: $672,323
RFA: Ecosystem Indicators (1997) RFA Text | Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Aquatic Ecosystems
Objective:
Urbanization is thought to have negative impacts on stream ecosystems, and yet the actual causal relationships between land use change and stream community response have not been well studied. The focus of this project is on the development of predictive indicators of urbanization that are applicable to midwestern watersheds and stream ecosystems. Project objectives are to: (1) quantify the impacts of urbanization on hydrologic regimes, water quality, and habitat structure of stream ecosystems using paired experimental watersheds, and to develop linked models that accurately predict these impacts; (2) use the linked models as a virtual laboratory within which to generate and test indicators of urbanization and hydrologic change in terms of responses of fish and macroinvertebrate communities; and (3) use these models and indicators to assess the response of stream communities to alternative urbanization scenarios with extension to larger watersheds in the region.Progress Summary:
The research primarily examines eight third-order watersheds in central Indiana. Seven are in the greater Indianapolis area, and one is near Purdue University in West Lafayette, IN. Additional streams also have been surveyed for specific aspects of the project. All are in various stages of transition from rural to urban. We are evaluating linkages between imperviousness, increased urban runoff, altered channel morphology, water quality effects, and reduced biotic integrity for three sites on each of these streams.
Hydrologic and Water Quality Modeling
Stream Hydrodynamics. We are developing physically based flow and water quality models for urbanizing watersheds using numerical models with comparison to direct field measurements. To characterize stream flow during storm events, one-dimensional equations based on the conservation of mass and momentum have been used to describe the unsteady free-surface flow caused by a storm in an upland agricultural watershed. The model accounts for the effects of arbitrary stream geometry, variable slopes, variable flow regimes, and unsteady boundary conditions. The model was verified using the measured stream flow in Little Pine Creek, IN, at the upstream and downstream sites, from May 15 to June 9, 1998. The model predicted the diurnal variability in the stream flow well (Figure 1). The standard error of estimation was 50 L/s for the stream flow range from 400 L/s to 3,200 L/s during the simulation period. In addition, the model can simulate cross-sectional averaged velocities, shear stress velocities, and water depth variability. The hydrodynamics model will be used to predict shear stress and substrate mobility in urbanizing channels so that these physical habitat characteristics can be correlated with the responses of fish and macroinvertebrate communities.
Water Temperature Dynamics. A numerical model based on heat transport equations was formulated to predict water temperature dynamics in an upland agricultural watershed (Figure 1). The model accounts for the effects of solar radiation, air temperature, relative humidity, cloud cover, wind speed, heat conduction between water and stream bed, subsurface flow, and shading by riparian vegetation. The model was verified using the measured temperatures in Little Pine Creek, IN, from May 15 to June 9, 1998. The model predicts water temperatures at the above duration with the standard error of estimation 0.7oC for a temperature range of 12-23oC. The heat transport model requires hourly average air temperature, relative humidity, wind velocity, atmospheric pressure, cloud cover, and solar radiation, which are generally available in climatological data reports from the National Weather Service.
Figure 1. Comparison of observed and predicted streamflow and temperature for Little Pine Creek. (Younus and Hondzo, in press).
Nutrients and Dissolved Oxygen Dynamics. Six transport equations for organic nitrogen, ammonia, nitrite, nitrate, phosphorus, and dissolved oxygen were augmented with the above models. The model was verified using the measured water quality data (dissolved oxygen, nitrate, and phosphorus, as well as water temperature) in Little Pine Creek, IN. The nutrient transport model predicts nitrogen and phosphorus dynamics well. The dissolved oxygen simulations were not satisfactory. Although the average daily values of dissolved oxygen compared well with the measured data, the diurnal dynamics of dissolved oxygen was not achieved. We believe that periphyton dynamics is a significant contributor to diurnal dissolved oxygen dynamics. Additional field measurements have been conducted this summer to verify the nutrients and dissolved oxygen models. If the results are not acceptable, we will conduct additional field and possibly laboratory measurements to formulate a closure for the model. The models will be augmented with responses of fish and macroinvertebrate communities to assess the responses of stream communities to alternative urbanization scenarios with extension to larger watersheds in the region.
Hydrologic Studies. The above model was applied to a small reach of Little Pine Creek between station 800W and 825W. To extend the model for the entire watershed, it is necessary to compute the lateral flow to a stream. For this purpose, the KINEROS rainfall-runoff model was selected. Because this is a distributed model, the preparation of input files for this model is tedious and cumbersome. An Arcview GIS interface was developed to make this job easy. The interface worked very nice; however, the hydrographs obtained were not promising. It was later determined that the numerous subsurface drains in the Little Pine Creek watershed create difficulties for KINEROS in correctly simulating the runoff from such watersheds. Thus, significant effort was made to simulate runoff from such watersheds by using TOPMODEL. The AGNPS model was run to compute the nitrogen, phosphorus, and sediment loading from Green Hill, Martell, 800W, and Dairy Farm watershed. The simulated results were compared with the observed data and a reasonable agreement was obtained.
SWAT Modeling. At this point, an existing empirical nonpoint runoff model (SWAT) incorporating land use has given the best predictive results when compared to observed hydrology at our urbanizing study sites over historic periods. For the period 1993-1997, SWAT has given good results (R2=0.85) for the prediction of daily flows in Lick Creek, Indianapolis, during the summer, but less satisfactory results for spring and fall flows (R2=0.53). SWAT significantly underestimates daily flow at these times, but better results can probably be obtained with additional model calibration.
Model Linkages With Land Use. Hydrologic models are typically data intensive and require significant expertise to use. The L-THIA (Long-Term Hydrologic Impact Assessment) model was developed to overcome these limitations. L-THIA was initially developed as a spreadsheet application by Jon Harbor (co-Principal Investigator on this project). L-THIA estimates the average annual runoff for an area given land uses and hydrologic soil groups. With L-THIA, one quickly can assess the impact of land use changes on average annual runoff. L-THIA has been extended to consider nonpoint source (NPS) pollutant, interfaced with GIS, and a simple, widely accessible World Wide Web (WWW) version has been created. L-THIA is being applied to the study watersheds and is being compared to more sophisticated hydrologic models to determine how well it predicts expected hydrologic trends within watersheds as they urbanize. The Web accessible version of L-THIA can be run at http://danpatch.ecn.purdue.edu/~sprawl/LTHIA/. Export coefficients for nutrients recently have been added to the LTHIA version used to predict annual runoff.
To date, we have completed work on sensitivity analyses of key parts of this approach, including the selection of distributed versus composite modeling approaches, the sensitivity of results to the spatial resolution of the land use data, soils, and antecedent moisture conditions. We also have completed model comparison with preexisting field data from one of our study watersheds in Indianapolis. We have begun work on using L-THIA to examine differential hydrologic impacts of various land use and development patterns, and have expanded the capabilities of the L-THIA model by implementing it in ArcView GIS and adding the ability to include data on nonpoint source pollution loadings. L-THIA runoff predictions have been compared to U.S.Geological Survey (USGS) gauging data on the Indianapolis streams. At this point, the model output has not proved to be sufficiently accurate for use in predicting channel flow characteristics or biological responses.
Hydrologic and Water Quality Field Monitoring. Water quality and flow regimes in Indian and Little Pine Creeks near Purdue University were monitored in 1998-1999. Hydrology, temperature, and nutrients have been assessed over several years at several sites, providing a data set for use in model development. In summer 1999 and 2000, YSI multiparameter probes (dissolved oxygen [DO], temperature, pH, and conductivity) were deployed on Indian Creek in conjunction with ISCO 6700 series automated water samplers during two time periods. The samplers were located approximately 0.3 stream miles apart. Two Hydrolab multiprobes also were deployed on Little Eagle Creek in Indianapolis; one at 59th Street (upstream site) and one at 52nd Street (downstream site), approximately 1.34 miles apart. Data collected included DO, temperature, pH, and conductivity. They were deployed during a low flow period when high diurnal DO fluctuations were observed. The diurnal oxygen data currently is being used in the development of the predictive oxygen model. Flow data for the 52nd Street site is recorded by the USGS gauging station. Additional water quality data including conductivity, suspended solids, nitrogen and phosphorus were collected in 2000 at all Marion County study sites to provide direct comparisons among streams and their various levels of urbanization (Figure 2). These data are not yet fully analyzed, but it is already evident that nutrient concentrations generally decline in urban areas as compared to agricultural areas.
Figure 2. Land use map for Marion County and Indianapolis based on 1997 SPOT imagery.
Land Use Map Preparation. In support of the research on land use
change and its impacts on stream ecology and geomorphology, it is necessary to
have land use maps for the period of observation. The City of Indianapolis did
not have a recent land use map available (although they do have a zoning map).
Thus, we undertook to develop two land use maps using remote sensing, one for
1997 representing pre-study conditions, and one for 1999 representing study
period conditions. Two land use maps also allow us to assess land use change
over a short time period, as well as over longer time periods using previous
land use maps (1990s and 1980s). Production of the maps and data acquisition has
been supported both by this grant, and a contract from the City of Indianapolis,
because of the potential multiple use of the land use coverage.
Once all the
images were registered, they were joined to produce a mosaic of the whole area.
Individual watershed areas were later extracted from this image. The land use
map had seven classes (agricultural, low-density urban, high-density urban,
commercial, forest, and water body); the map was then refined by adding four
more classes (industrial, airport, institutional, and special uses) using GIS
and zoning data acquired from the City of Indianapolis. This composite land use
map forms the basis of the land use data used in comparing ecological and
geomorphological conditions along gradients of urbanization, and is also useful
for city planners in the Indianapolis Department of Metropolitan
Development.
Watershed Imperviousness Estimation. One measure of land use change we are exploring to relate to measures of ecological and geomorphic change is the percent impervious area in a watershed. An independent digital map of building footprints and roads provided by the City of Indianapolis was used for impervious cover calculations. This was accomplished in three steps. The first step was to convert the land use layer into a grid format so that it was available as raster. The second step was to intersect the grid map with buildings and street map. The last step was to calculate the area covered by buildings and streets according to land use. This analysis was done using the "Spatial Analyst" module and "Cell Statistics" manipulation facilities available in Arcview. The study watersheds range from 3 to 24 percent impervious by this method.
Geomorphic Stability of Urban Channels. As a part of the overall study, the impact of urbanization on channel stability is being quantified, and preliminary results from several sites are available (Table 1). As a first step in this geomorphic approach, indicators of geomorphic stability were used to assess the relative geomorphic stability between study sites of various levels of urbanization. Preliminary assessment was accomplished based on determination of dominant channel processes and the state of the channel relative to local structures (e.g., bridge pier footings, abutments, hanging culverts, etc.). In addition, we tested the ability of existing and proposed measures of channel stability to discriminate between stable and degrading channels in an urban setting. The existing measures consist of two qualitative methods (e.g., rating schemes), five existing quantitative methods (stream power, unit stream power, excess shear stress, bankfull discharge per unit watershed area, recurrence interval of bankfull discharge), and one proposed quantitative measure (recurrence interval of critical discharge). The latter measure of stability is important to test as it has been suggested as a potential tool for use in river restoration design.
Based on field work and preliminary hydraulic and sediment transport analysis of data available to date (Table 1), two sites showed obvious signs of channel instability through bed degradation and channel widening while all other sites appeared stable. In the analysis of the available data, the two qualitative stability assessment methods did not give exceptional results in distinguishing between what were considered degrading and stable sites (Table 2). Both methods seem to equate channel stability with channel uniformity. Based on our results, the two qualitative methods of stability analysis are best suited for local channel instability (i.e., road crossing scale) but not necessarily for identifying larger-scale instability such as incision of a channel resulting from urbanization. This is an important finding as recent government field manuals have suggested using these qualitative methods for identifying watershed-scale channel instabilty, which our results indicate may not be appropriate in comparison to available quantitative measures (Table 3), which were much more successful at identifying unstable channels.
For preliminary analysis of the effects of urbanization on geomorphic stability, the sites were divided into high urbanization, medium urbanization, and low urbanization (Table 4). Using both the qualitative and quantitative measures of channel stability, very few measures revealed significant differences between the groups segregated by urbanization. Those assessment methods which discriminated between groups of urbanization class (with p < 0.10) include excess shear stress, bankfull discharge recurrence interval, and critical discharge recurrence interval (Table 5). Results of comparison using all other measures of channel stability, including the qualitative methods, were insignificant (p > 0.10).
In our study of geomorphic impacts, stream channels were responding to
urbanization by enlarging (i.e., bed degradation and channel widening) (Figure
3). In general, channel stability was not directly related to levels of
urbanization as stability. We attribute this to stability also being related to
the rate of urbanization and the relative location of a channel reach within a
watershed. The geomorphic data have yet to be analyzed in tandem with the
ecological data, which is a critical goal of the project. With respect to
further geomorphic analysis, we are in the process of developing a model for the
prediction of the time needed for a channel to reestablish dynamic equilibrium
after a perturbation to the river system (e.g., increase in runoff). The data
collected for the study to date will provide some of the data needed for testing
and calibrating this model.
Figure 3. Channel enlargement occurs with increasing watershed imperviousness in the Indianapolis area.
Ecological Indicator Development
Fish Habitat and Assemblages. We are testing the hypothesis that one of the dominant impacts of urbanization on stream community structure and function results from the alteration of the natural flow regime due to increases in watershed imperviousness. By increasing runoff and decreasing infiltration, imperviousness leads to increased peak flows, more severe floods, and decreased dry weather flows. Stream channels respond by increasing channel cross-sectional area by widening, downcutting, or both, which results in increased bank erosion, siltation, and bedload movement, leading to degraded aquatic habitats. Our main focus in this portion of the project is to relate potential indicators of the effects of urbanization (% impervious area, % high density urban) to altered hydrology (e.g., frequency of peak flows, low flows) and channel geomorphology (e.g., increased width:depth ratio, frequency of bed mobilizing events). We will then relate these alterations in hydrology and geomorphology to habitat quality (e.g., loss of pool habitat) and biotic integrity (lowered Index of Biotic Integrity, loss of sensitive fish species).
Preliminary results of this portion of the project show a clear degradation
of overall fish habitat quality (QHEI) in relation to urbanization as measured
by watershed imperviousness (Figure 4). Overall quality of the fish community
(as measured by IBI) likewise declines (Figure 5), although the correlations for
biotic variables are lower than for physical variables.
Figure 4. A measure of fish habitat quality (QHEI) declines with increasing urbanization.
Macroinvertebrate and Periphyton Indicators: A total of 17 sites in six creeks were sampled during July and August 1999. These were Crooked Creek, Little Buck, Little Eagle, Pleasant Run, and Williams in Marion County, and Indian Creek in Tippecanoe County. Three sites (upper, middle, and lower) were selected on each creek that corresponded with the geomorphic surveys and fish collections. Middle sites of all streams were again sampled in 2000, with several additional highly urban sites in Lick Creek and Pogue's Run added that year. Methods and crew members were the same in both years. At each site, three riffles were sampled for benthic macroinvertebrates and periphyton. A total of 180 samples each were collected for macroinvertebrates and for periphyton. Aquatic insects were collected with a Surber sampler with a mesh size of 500 mm. Four samples were taken from each riffle, making a total of 12
Figure 5. The index of biotic integrity declines with increasing imperviousness in Indianapolis streams.
samples at each site. Sampling began at the downstream end of the riffle and proceeded upstream until the four samples were taken. Two of the Surber sites were positioned on the upstream portion of the riffle, and two of the Surber sites were positioned on the downstream portion of the riffle.
In 2000, work also was completed on the relationship of unionid mussels to
physical habitat characteristics at the study sites. Unionid mussels are among
the largest and longest-lived animals in small midwestern streams. Their
populations are generally declining in this region. Research by Myers-Kinzie et
al. (2001) shows a conspicuous absence of mussels in Marion County that is most
likely related to increased scour, downcutting, and shear stress associated with
unstable urban channels. In 2000, a series of experiments in streamside channels
colonized with periphyton was used to test specific effects of high and low
nutrient levels and shear stress on periphyton communities. Both were found to
be highly effective in determining both structure and productivity of typical
stream periphyton communities.
Future Activities:
The hydrologic and water quality models developed thus far will be extended for use with dissolved oxygen and nutrient transport, and will be linked to the SWAT and LTHIA runoff models. Hydrologic modeling is necessary so that critical flows can be calculated at each site, which can then be related to the observed substrate characteristics and benthic community composition. The risk analysis portion of the project also will be completed. Identification of macroinvertebrates will be finished in 2001, and final statistical analyses of all biological collections will be completed and compared to the hydrologic model inferences as the last step in ecological indicator development.Journal Articles on this Report : 6 Displayed | Download in RIS Format
Other project views: | All 35 publications | 9 publications in selected types | All 8 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Bhaduri B, Harbor J, Engel B, Grove M. Assessing watershed-scale, long-term hydrologic impacts of land-use change using a GIS-NPS model. Environmental Management 2000;26(6):643-658. |
R825871 (2001) R825871 (Final) |
|
|
Bhaduri B, Minner M, Tatalovich S, Harbor J. Long-term hydrologic impact of urbanization: A tale of two models. Journal of Water Resources Planning and Management 2001;127(1):13-19. |
R825871 (2001) R825871 (Final) |
Exit |
|
Doyle MW, Harbor JM, Rich CF, Spacie A. Examining the effects of urbanization on streams using indicators of geomorphic stability. Physical Geography 2000;21(2):155-181. |
R825871 (2001) R825871 (Final) |
Exit |
|
Grove M, Harbor J, Engel B. Composite vs. distributed curve numbers: effects on estimates of storm runoff depths. Journal of the American Water Resources Association 1998;34(5):1015-1023. |
R825871 (2001) R825871 (Final) |
Exit |
|
Myers-Kinzie M, Spacie A, Rich C, Doyle M. Relationship of unionid mussel occurrence to channel stability in urban streams. Verhandlungen International Vereinigung Limnologie 2002, 28: 1-5. |
R825871 (2001) R825871 (Final) |
not available |
|
Younus M, Hondzo M, Engel BA. Stream temperature dynamics in upland agricultural watersheds: measurements and modeling. Journal of Environmental Engineering 2000;126(6):518-526. |
R825871 (1999) R825871 (2001) R825871 (Final) |
Exit |
Supplemental Keywords:
integrated assessment, EPA Region V., RFA, Scientific Discipline, Geographic Area, Water, Ecosystem Protection/Environmental Exposure & Risk, Hydrology, Midwest, Water & Watershed, Ecosystem/Assessment/Indicators, Ecosystem Protection, State, Ecological Effects - Environmental Exposure & Risk, Wet Weather Flows, Environmental Monitoring, Ecological Risk Assessment, Ecology and Ecosystems, EPA Region, Watersheds, aquatic ecosystem, hydrologic dynamics, ecological exposure, anthropogenic processes, urbanization, aquatic, nutrient transport, anthropogenic stresses, ecological effects, remote sensing, risk assessment, scaling, suburban watersheds, Indiana, aquaculture, aquatic biota , ecosystem assessment, satellite images, watershed protection, ecosystem evaluation, runoff, urban runoff, stream ecosystems, urban ecosystems, urban watersheds, modeling, large-scale regional studies, integrated assessment, regional scale, water quality, stream flow, aquatic ecosystems, Indianapolis, ecosystem health, environmental stress, remotely sensed data, Region 5, IN, dissolved oxygen , ecological indicators, fish , aquatic biotaRelevant Websites:
http://www.fnr.purdue.edu/fi/spacie/spacie.htm
http://ABE.www.ecn.purdue.edu/ABE/Fac_Staff/engel
http://www.eas.purdue.edu/geomorph/
http://danpatch.ecn.purdue.edu/~sprawl/LTHIA5/
http://danpatch.ecn.purdue.edu/~tgis/cases/ipcase/background.html
http://danpatch.ecn.purdue.edu/~tgis/cases/ipcase/object.html
http://www.ce.umn.edu/people/faculty/hondzo/
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.