Grantee Research Project Results
Final Report: Developing Effective Ecological Indicators for Watershed Analysis
EPA Grant Number: R827638Title: Developing Effective Ecological Indicators for Watershed Analysis
Investigators: Patten, Duncan T. , Marcus, Andrew , Lawrence, Rick , Minshall, Wayne
Institution: Idaho State University , University of Oregon , Montana State University - Bozeman
Current Institution: Idaho State University , Montana State University - Bozeman , University of Oregon
EPA Project Officer: Packard, Benjamin H
Project Period: July 1, 1999 through June 30, 2002
Project Amount: $868,242
RFA: Ecological Indicators (1999) RFA Text | Recipients Lists
Research Category: Aquatic Ecosystems , Ecological Indicators/Assessment/Restoration
Objective:
Alteration of watersheds by human and natural perturbations is expected to result in changes in characteristics and functions of associated rivers and riparian areas. The objective of this research project was to show how attributes of riverine systems might be used as indicators of watershed condition and whether these relationships are spatially scaleable. The study was not intended to show how changing watersheds influence riverine systems. To achieve this objective, two approaches were used—one that looked at the influence of watershed characteristics on river and riparian functions and one that related river and riparian attributes to upstream watershed conditions. An additional objective of the study was to evaluate the potential use of remote sensing (e.g., hyperspectral imagery) for measuring river and riparian attributes identified as possible indicators of watershed condition. Emphasis was placed on remote sensing of stream characteristics because attributes of these systems were more "visible" than most riparian attributes, which often overlaid or shaded each other.
Summary/Accomplishments (Outputs/Outcomes):
Study sites were established in several watersheds within the northern Greater Yellowstone Area, Cache Creek, Soda Butte Creek, Pebble Creek, and Tom Miner Creek. Watersheds (and subwatersheds) were selected for their pristineness, accessibility, and modification by human (e.g., grazing) and natural (e.g., 1988 Yellowstone National Park fire) perturbation. Watersheds with mining and timber harvest also were surveyed, but were dropped from final analyses because preliminary fluvial geomorphologic analysis showed no definitive signal of an altered system, a possible consequence of low magnitude disturbance. Rivers within the watersheds were sampled every 100 m for channel structure and river form and functions (e.g., depth, river type [run, ripple, pool, etc.]). Riparian areas were sampled within the watersheds below and above confluences of rivers from subwatersheds. Characteristics (i.e., primarily land cover) of upstream watersheds for river and riparian study sites were determined by Landsat thematic mapper analysis. Water quality assessment (i.e., macroinvertebrates and chemistry) sampling sites, used to assess watershed qualities, were near watershed “outflow” locations, many sites closely aligned with riparian sites. Using traditional stream ecological measurements, this assessment examined the difference between natural and anthropogenic disturbances to stream ecosystems by comparing stream ecosystem recovery after a natural disturbance with stream ecosystem recovery after anthropogenic disturbances. This study also examined the cumulative effects of a natural disturbance, fire, on an entire watershed by assessing subwatersheds with increasing burned area.
River Indicators
River attributes related to watershed characteristics included factors associated with stream power (e.g., channel shape, hydraulics, and sediment). Location of eroded banks and deposition of bedload sediments to form ripples, pools, etc. could be related to watershed condition. We determined that the primary watershed condition related to stream power was the percentage of watershed burned by the 1988 Yellowstone fire because fire is capable of inducing complex fluvial responses. Regression models suggested that 12-13 years post fire, channels with a greater percentage of burned watershed were associated with higher cross-sectional stream power, low width/depth ratios, and lower bank failure rates. This study identified the need to characterize the spatial distribution of stream power. Because stream power integrates several elements of fluvial processes and is associated with percent area burned, it may serve as a good indicator of watershed response to changing fire management procedures and watershed and stream restoration activities.
Riparian Indicators
Riparian attributes related to watershed characteristics included factors associated with vegetation community structure and composition. Fluvial processes, a consequence of climate and watershed organization and land cover types, form different floodplain topography and soils. The spatial arrangement of these attributes, in turn, produces heterogeneous conditions for riparian vegetation establishment and maintenance, creating several vegetation patch types that were grouped into three primary patch types: herb, deciduous, and coniferous. Riparian vegetation attributes associated with establishment conditions were significantly different between these riparian patch types. Watershed characteristics, including elevation, subwatershed size, and proportion of barren and forest land cover types, accurately classified coniferous, herbaceous, and deciduous riparian patch types more than 70 percent of the time. In basins burned in the 1988 Yellowstone fires, the proportion of grasslike upland land cover classified riparian patch types with 80 percent accuracy. Canonical correlation analyses of relationships between physical and biological properties in 2-, 5-, 10-, and 100-year floodplains revealed basin-specific patterns. In most watersheds, patch type heterogeneity on the floodplain resulted from different magnitudes of stream power across the floodplain related to flood intervals. Regression models of detrended correspondence analysis axis 1 scores of herbaceous composition indicated vegetation responses to the hydrogeomorphic environment. These findings support the concept that riparian vegetation integrates watershed, subwatershed, reach, and patch level processes, as well as outputs of watershed alteration.
Validation of Watershed Condition
Levels of watershed "pristineness," a general evaluation of watershed condition that can be related to watershed characteristics, were validated through water quality assessment. Water quality results show that all three watersheds grouped independently, indicating watershed differences. The best predictors for each watershed were: for Cache Creek (percent upstream basin burned 1988, percent upstream in forest class, strand one bankfull depth); for Soda Butte Creek (percent upstream in water class, percent upstream in barren class, patch perimeter of barren class, percent south aspect); for Tom Miner Creek (percent north aspect, mean patch area of range class, percent north west aspect, patch perimeter of deciduous trees). Among watersheds, macroinvertebrate presence/absence was the best biological predictor. Within watersheds, macroinvertebrate presence/absence data and relative abundance data predicted equally well. Raw abundance was inferior at all scales.
Remote Sensing
Hyperspectral imagery was evaluated relative to its applicability to measuring riverine attributes. Several components of the study demonstrated that hyperspectral imagery could "identify" several features within and adjacent to streams (i.e., depth, river habitats, and woody debris). In addition, spectral response of fluvial characteristics (e.g., location, depth, turbidity, surface roughness, cobble size, bottom color, and stream habitats of run, ripple, pool, and glide) on Soda Butte Creek were evaluated using 34-band airborne hyperspectral imagery with 2.5-m resolution to understand how imagery allows the identification of riverine attributes. Principal components from spectral data revealed that C1 is related to visible reflectance, and PC2 is associated with the near infrared range; PC3 represents a distinct combination of visible reflectance bands. Regression results indicated that all stream characteristics, except cobble size, were significantly related to spectral response. Stream depth showed the best relationship, with surface roughness second. These results demonstrate that with further development, specific river indicators may be identified and quantified using hyperspectral imagery. Other studies have demonstrated the use of hyperspectral imagery to identify different riparian types, but these demonstrations usually were accurate only when riparian vegetation was clearly separated into individual species groups, not a common phenomenon in most western riparian areas.
Integration of Indicators
Each component of this study identified several attributes that were related to watershed conditions. Remote sensing data also showed spectral responses of fluvial attributes. Some response attributes of river and riparian systems closely overlap, and can be used for integration of indicators for watershed assessment. The initial approach to integration of indicators was to assess the overlap of watershed attributes acting as drivers of individual river and riparian indicators. The most altered watershed, Cache Creek, was used because there was more overlap in watershed drivers, and there also was a gradient among subwatersheds in percent burn from the 1988 Yellowstone fires. Several simulation models were developed with watershed drivers as independent variables and riparian patch types and stream power as dependent variables. From these models, percent burn was selected as the primary driver, and regression equations were developed relating this to stream power and riparian vegetation patch cover, both of these variables have been shown to be good indicators of watershed condition. A simple predictive model then was developed tying the two indicators together using only the herb patch type and relating them to watershed condition (i.e., percent burned). This demonstrated the potential for using these river and riparian parameters in an integrated fashion when the watershed has been altered severely.
Scalability of Indicators
One goal of this study was to develop indicators that were scalable across the landscape and/or region. Assessment of several significant watershed drivers of river and riparian indicators, and also of macroinvertebrate communities, showed that each watershed had its own set of significant attributes that significantly influenced river and riparian processes. Consequently, with watersheds and rivers being relatively unique, river and riparian indicators of watershed attributes or conditions usually will be different between and among watersheds, which reduces the possibility of cross-regional scalability of indicators. Remote-sensing techniques developed for river and riparian attributes, however, may be scalable across watersheds.
Reference:
Minshall GW, Robinson CT, Royer TV. Stream ecosystem responses to the 1988 wildfires. Yellowstone Science 1998;6(3):15-22.
Journal Articles on this Report : 8 Displayed | Download in RIS Format
Other project views: | All 14 publications | 8 publications in selected types | All 8 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Fonstad M, Marcus WA. Self-organized criticality in riverbank systems. Annals of the Association of American Geographers 2003;93(2):281-296. |
R827638 (Final) |
Exit |
|
Lawrence R, Bunn A, Powell S, Zambon M. Classification of remotely sensed imagery using stochastic gradient boosting as a refinement of classification tree analysis. Remote Sensing of Environment 2004;90(3):331-336. |
R827638 (Final) |
Exit Exit Exit |
|
Legleiter CJ, Marcus WA, Lawrence RL. Effects of sensor resolution on mapping in-stream habitats. Photogrammetric Engineering & Remote Sensing 2002;68(8):801-807. |
R827638 (Final) |
Exit Exit |
|
Legleiter CJ, Lawrence RL, Fonstad MA, Marcus WA, Aspinall R. Fluvial response a decade after wildfire in the northern Yellowstone ecosystem: a spatially explicit analysis. Geomorphology 2003;54(3-4):119-136. |
R827638 (Final) |
Exit Exit |
|
Marcus WA. Mapping of stream microhabitats with high spatial resolution hyperspectral imagery. Journal of Geographical Systems 2002;4(1):113-126. |
R827638 (Final) |
Exit |
|
Marcus WA, Legleiter CJ, Aspinall RJ, Boardman JW, Crabtree RL. High spatial resolution hyperspectral mapping of in-stream habitats, depths, and woody debris in mountain streams. Geomorphology 2003;55(1-4):363-380. |
R827638 (Final) |
Exit Exit |
|
Minshall GW, Royer TV, Robinson CT. Response of the Cache Creek macroinvertebrates during the first 10 years following disturbance by the 1988 Yellowstone wildfires. Canadian Journal of Fisheries and Aquatic Sciences 2001;58(6):1077-1088. |
R827638 (Final) |
Exit |
|
Minshall GW. Responses of stream benthic macroinvertebrates to fire. Forest Ecology and Management 2003;178(1-2):155-161. |
R827638 (Final) |
Exit Exit Exit |
Supplemental Keywords:
ecological indicators, watersheds, watershed condition, stream geomorphology, riparian vegetation, riparian/floodplain connectivity, stream power, stream macroinvertebrates, aquatic habitat, remote sensing, hyperspectral imagery, scaling, Upper Yellowstone River, Yellowstone National Park, land use, fire, woody debris., RFA, Scientific Discipline, Geographic Area, Ecosystem Protection/Environmental Exposure & Risk, Hydrology, Ecosystem/Assessment/Indicators, Ecosystem Protection, Ecological Effects - Environmental Exposure & Risk, Northwest, Environmental Monitoring, Ecology and Ecosystems, Ecological Risk Assessment, Ecological Indicators, ecological exposure, anthropogenic stresses, biological activity, remote sensing, scaling, logging, watersheds, sediment, stream ecosystems, survey, ecosystem indicators, recreational home development, multiscale assessmentProgress 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.