Grantee Research Project Results
Final Report: Ecological Indicators for Large River-Floodplain Landscapes
EPA Grant Number: R826600Title: Ecological Indicators for Large River-Floodplain Landscapes
Investigators: Turner, Monica G. , Stanley, Emily H.
Institution: University of Wisconsin - Madison
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1998 through September 30, 2001 (Extended to September 30, 2002)
Project Amount: $677,351
RFA: Ecological Indicators (1998) RFA Text | Recipients Lists
Research Category: Aquatic Ecosystems
Objective:
Identifying landscape indicators that are well correlated with specific aspects of ecological function is a crucial research need that requires an integrated approach combining landscape monitoring with field studies. Large river-floodplain systems are among the most diverse and dynamic landscapes in the United States, providing many important societal values. However, relatively little effort has been devoted to the development and testing of landscape indicators for these systems. The objectives of this research project were to develop and test ecological indicators for large river-floodplain landscapes along reaches of the Wisconsin River to determine which landscape metrics are the most useful for monitoring population, community, and ecosystem processes in large river-floodplain landscapes. We combined spatially extensive field sampling with landscape analysis in nine reaches of the Wisconsin River sampling to quantify the ability of landscape indicators to predict ecological variables over broad scales. We evaluated landscape indicators by their utility for detecting changes in the structure and function of the Wisconsin River floodplain landscape that were related to modification of the natural flow regime, historical land use, and current land-use patterns. Our field studies were concentrated in floodplain forest in nine 12 to 20-km reaches along the lower 400 km of the Wisconsin River.
Summary/Accomplishments (Outputs/Outcomes):
Land-Cover Patterns and Change. Midwestern river-floodplain landscapes have undergone considerable land-use change since historic times, and our research required an understanding of the current and historic distributions of cover types within the floodplain. We quantified changes in floodplain land use and land cover since the 1930s within the nine study reaches by analyzing and digitally classifying 200 historic aerial photos corrected against modern orthophotographs. The analyses revealed that floodplain deciduous forest cover is abundant (40-60 percent of each reach) and has increased (up to 51 percent) in almost every reach, even in locations adjacent to the riverbank and near rapidly developing areas. Forest cover also is relatively well connected, although there are some gaps; forest patches also have become less compact and more complex along their edges. Using scenarios based on realistic forest management agendas, we explored how various conservation options might affect the connectivity of these floodplain landscapes. Converting similar amounts of land using very different management options can produce similar increases in forest cover and connectivity metrics, depending on the initial distribution of the landscape patches. Such scenarios may be useful to future floodplain management.
The aerial photo analysis of nine reaches was complemented by studying land-cover change since the 1930s in a larger, continuous area (approximately 8,800 km2) along the lower Wisconsin River using other historical records (Bürgi and Turner, 2002). A general shift was observed from agricultural land to forest throughout the study region. Agricultural abandonment and a general decline in farming intensity were the most important processes of land-cover change among the processes considered. Areas characterized by the different processes of land-cover change differed in the abiotic and socioeconomic variables that had explanatory power and could be distinguished spatially.
Collectively, our analyses of current and historic landscape patterns in the floodplain of the Wisconsin River demonstrated that natural cover types (forest and wetland) were dominant and well connected, but that significant areas of the floodplain that currently are occupied by forest had been in nonforest cover earlier during the 20th century. The spatial data that were generated also provided the foundation for several other aspects of our study: locating study plots, characterizing the land-cover history of each study plot, and analyzing the current landscape pattern within the study reaches and surrounding individual sampling locations.
Ecological Responses. At the population and community levels, our field studies emphasized the composition and structure of the floodplain forest and the avian community. At the ecosystem level, our studies have focused on rates of denitrification and microbial activity.
We conducted sampling during three summer field seasons (1999, 2000, and 2001). We sampled vegetation, birds, and soil properties in 10 x 20 m plots within the nine study reaches. Private landowners and state and nongovernmental organizations were contacted to secure permission for sampling. Precise locations of each plot (n = 655) were obtained using a global positioning system (GPS). Within each plot, each tree was identified to species and its diameter-at-breast-height (DBH) was recorded. All shrubs and saplings were identified to species and censused, and the overall canopy cover and shrub cover were estimated. Tree seedling abundance, slope position, herbaceous cover, and leaf litter were recorded within each of three randomly located 1-m2 quadrats. Within three additional 100-cm2 randomly located quadrats, litter accumulation was assessed by collecting surface organic matter, and three soil cores were obtained to estimate potential denitrification rates and characterize soil conditions. We censused birds at a subset of the vegetation plots along transects using 8-minute point counts, during which all birds seen and heard were noted. Each transect was visited twice, with the second round commencing only after all transects had been censused once.
Forest vegetation varied with flooding and past land cover, but the current landscape pattern explained little variation. Species observed most frequently were Fraxinus pennsylvanica, Acer saccharinum, and Ulmus americana. Indicators of physiography (e.g., geographic province) and flooding regime (e.g., relative elevation and distance from main channel) were consistently important in predicting occurrence, community composition, and abundance of trees in the Wisconsin River floodplain, although the direction of effects varied among species. Correspondence analysis revealed that flood-tolerant and intolerant species segregated along the primary axis, and late-successional species tended to segregate from flood-tolerant species along the secondary axis. Current landscape configuration was generally not important for species presence or abundance, except in forests that developed during recent decades. However, land-cover history was important for tree species presence and for abundance of late-successional species. The comparison between statistical models developed with and without soils data suggested that broad-scale indicators (such as geographic province) generally performed well. Our results suggest that indicators of physiography and flood regime are particularly useful for explaining floodplain forest structure and composition, particularly in floodplains that contain a relatively high proportion of natural cover types. Floodplain forest is extensive and well connected along much of the Wisconsin River, and forest structure and composition varies with past land use and flooding, rather than with contemporary landscape structure. Interestingly, our data on tree seedling and sapling abundance within existing floodplain forest also revealed a striking absence of recruitment by Acer, suggesting that forest composition may shift in coming decades.
Analyses of the bird data have focused on the effects of landscape context of forest patches and prior land uses. Overall, typical measures of habitat fragmentation explained a relatively small portion of the variation in the structure and composition of the bird communities that we studied. In other landscapes, measures of habitat fragmentation have accounted for a large amount of the variation in habitat use by forest interior birds, but this was not the case along the Wisconsin River. These results may reflect the fact that riparian forests continue to occupy a relatively large portion of the floodplain and also tend to be naturally fragmented. Distance from the confluence with the Mississippi River showed a strong negative correlation with species richness and abundance for several avian guilds; particularly, woodpeckers. This suggests that forests along the Mississippi may serve as a population source for some species.
Patterns of habitat use by individual bird species reinforced some of the conclusions drawn from guild- and community-based studies. For instance, variables describing regional variation dominated models for most floodplain birds. The best predictor for species in our study that have been described elsewhere as having strong affinities for floodplain habitats was the distance from the confluence with the Mississippi River (e.g., the Prothonotary Warbler, the Warbling Vireo, and the Yellow-Bellied Sapsucker). These birds tended to occur more often in forest patches in the southern reaches of the Wisconsin River. Landscape variables and measures describing local habitat structure improved these models somewhat, but played a rather minor role. Overall, landscape measures assumed a greater importance in models for individual species than they did in models for guilds, but often in counterintuitive ways. The Pileated Woodpecker (the largest extant woodpecker in North America) is usually associated with large forest patches. In our study, the best model for this species showed a negative association with the overall width of floodplain forest; width was most often used as a surrogate for patch area in riparian studies. Woodpeckers also tended to be associated with southern reaches, as also was indicated in our community study. One of the more surprising results involved models for several species that have been described in upland studies as being "forest interior specialists" or "area-sensitive species." The best predictors for such species (e.g., Ovenbird, Scarlet Tanager) were indeed landscape measures, but showed negative associations with forest patch area or overall amount of forest at broad scales (i.e., within 3 km). Overall, these results suggest that at least some species respond in a fundamentally different way to landscape measures such as patch size or proximity to habitat edge in floodplains, as opposed to upland habitats. Although we included landscape measures in models for some species, the strongest gradient associated with the Wisconsin River apparently occurs at regional, rather than landscape, scales.
Denitrification in floodplain soils may serve as an important sink for excessive nitrate in river ecosystems. We used the acetylene-block technique to determine static core denitrification (SCD) and denitrifying enzyme activity (DEA) denitrification of soils within the nine study reaches to determine patterns and potential controls for nitrate reduction. Denitrification varied within and among reaches; rates and coefficients of variation ranged from 0 to 65.14 ng N g-1 h-1 and 150 to 198 percent, respectively, and differences between reaches were not significant. Strongest correlates with denitrification shifted between organic matter and moisture among reaches, and usually explained less than 30 percent of observed variation. All reaches had a small number of samples with high rates, and several samples with low or zero values. These results suggest a fruitful avenue for future research-to examine landscape attributes that may be related to the frequency or probability of occurrence of sites with high denitrification rates (hotspots), rather than trying to identify patterns that consider the full range of variation within or among reaches.
Importantly, significant effects of tree species composition on biogeochemical cycling were observed in our soil studies. We assessed potential effects of forest community composition on soil biogeochemistry by comparing litter inputs and chemical and microbial attributes of soils adjacent to five tree species that have experienced long-term shifts in relative abundance: silver maple (A. saccharinum), swamp white oak (Quercus bicolor), basswood (Tilia americana), green ash (F. pennsylvanica), and American elm (U. americana). Silver maples produced the greatest amount of litter during a year. Microbial activity and organic matter content in surrounding soils were significantly greater for silver maples than for other trees. This implies that an increase in maple may have modified biogeochemical processes and soil C cycle. Therefore, future changes in tree species abundances may have important implications for ecosystem function.
As part of our soil analyses, we developed new method for the measurement of ammonium in soils extracts or water using orthophthaldehyde (OPA) compound. This method involves less toxic chemicals than a conventional method, requires minimal equipment (fluorometer and spectrophotometer), and provides fairly high turnover rates. This method is not only sensitive for low concentration samples (e.g., submicromolar NH4), but also applicable for samples with much higher concentrations (over 100 micromolar NH4) by the changing protocols given in this paper. This method would provide a valuable tool for the measurement of ammonium from various sources of environmental samples including soil extract and litter leachates.
Influence of Levees. Levees constructed in the early 1900s in one of the reaches used in this study offered an outstanding opportunity for more intensive studies of the effects of river flow modification on ecological responses. Vegetation studies conducted on the riverine and upland sides of the levee revealed compositional differences related to the levee, with flood-intolerant tree species having greater abundance upland of the levee. However, tree growth rates varied among species but did not differ significantly with position relative to the levee, suggesting that compositional differences may be an important driver of productivity. To assess the effects of levees on soil properties, we measured microbial activities (dehydrogenase and extracellular enzyme activities), soil properties (organic matter, pH, ions, soil texture), coarse woody debris, and litter accumulations. No significant differences in the amount of organic matter or coarse woody debris were observed outside the levee compared to areas inside the levee. However, significantly higher microbial activities (dehydrogenase, ß-glucosidase, phosphatase) per gram organic matter were found for soils inside the levee. These changes were attributed to modified hydrology, which caused differences in water availability and tree species composition (e.g., higher number of oaks outside the levee). These results suggest that modification of the hydrologic regime by levees may alter not only the structure of plant communities, but also the organic matter dynamics of floodplain.
Contributions of Ecological Indicators. We reviewed and synthesized the contributions of landscape indicators (metrics that quantify the amount and spatial arrangement of cover types) to riverine studies as part of this study. Landscape indicators can complement existing approaches to quantifying human influences on rivers by considering the importance of land cover and land use in the floodplains or surrounding catchment. Many studies, for example, have identified relationships between water chemistry and surrounding land cover.
The landscape indicators tested in this study explained relatively little of the variability in the forest community structure, bird community, and soil indicators including denitrification. We hypothesize that this may be a result of the relatively high spatial connectivity of natural habitat types that characterize the Wisconsin River floodplain, combined with our focus on sampling within the forest habitat. Landscape indicators may be more important in floodplains, in which natural habitats are greatly reduced and fragmented, as has been demonstrated in other landscapes. Furthermore, ecological responses within the river channel (e.g., nitrogen concentrations) may be more sensitive to the occurrence and distribution of gaps in natural habitats that would be reflected in landscape indicators. However, landscape indicators were very useful in exploring alternative land-acquisition scenarios, in which relative increases in habitat connectivity for a given increase in habitat extent could be compared.
Our study found that indicators related to flood regime (especially the elevation of plots relative to the 100-year flood event, or position relative to levees) were important in explaining variation in most ecological responses. Thus, obtaining accurate and spatially explicit estimates of flooding likelihood is important for understanding ecological dynamics in floodplains and likely for extrapolating from one location to another. This remains a difficult task because of the importance of small differences in elevation and microtopography to flood regime variation in these low-relief landscapes. Nonetheless, our studies suggest that broad-scale indicators that can be estimated using available sources of spatial data performed reasonably well. Our results also revealed a persistent legacy of historic land cover on current vegetation patterns, suggesting the importance of knowing the past to understand the present.
The approaches currently employed by ecologists to extrapolate spatially from fine-scale field measurements to broader scales were reviewed and synthesized. Our review identified surprising similarities in the approaches being used across a wide range of ecological response variables (e.g., species occurrence, species abundance, and process rates), and an increasing need for rigorous evaluation of predicted response patterns and the spatial distribution of uncertainty. We will continue final analyses focused on extrapolation of the data collected in this study to different locations and across different scales, with additional results and publications from this project.
Journal Articles on this Report : 14 Displayed | Download in RIS Format
Other project views: | All 58 publications | 16 publications in selected types | All 16 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Burgi M, Turner MG. Factors and processes shaping land cover and land cover changes along the Wisconsin River. Ecosystems 2002;5(2):184-201. |
R826600 (2001) R826600 (Final) |
Exit Exit |
|
Dixon MD, Turner MG, Jin C. Riparian tree seedling distribution on Wisconsin River sandbars: controls at different spatial scales. Ecological Monographs 2002;72(4):465-485. |
R826600 (2001) R826600 (Final) |
Exit |
|
Dixon MD. Effects of flow pattern on riparian seedling recruitment on sandbars in the Wisconsin River, Wisconsin, USA. Wetlands 2003;23(1):125-139. |
R826600 (Final) |
Exit |
|
Forshay KJ, Stanley EH. Rapid nitrate loss and denitrification in a temperate river floodplain. Biogeochemistry 2005;75(1):43-64. |
R826600 (Final) |
Exit |
|
Freeman RE, Ray RO. Landscape ecology practice by small scale river conservation groups. Landscape and Urban Planning 2001;56(3-4):171-184. |
R826600 (2001) R826600 (Final) |
Exit Exit Exit |
|
Freeman RE, Stanley EH, Turner MG. Analysis and conservation implications of landscape change in the Wisconsin River floodplain, USA. Ecological Applications 2003;13(2):416-431. |
R826600 (1999) R826600 (2000) R826600 (2001) R826600 (Final) |
Exit Exit |
|
Gergel SE. Assessing cumulative impacts of levees and dams on floodplain ponds: a neutral-terrain model approach. Ecological Applications 2002;12(6):1740-1754. |
R826600 (2001) R826600 (Final) |
Exit |
|
Gergel SE, Dixon MD, Turner MG. Consequences of human-altered floods: levees, floods, and floodplain forests along the Wisconsin River. Ecological Applications 2002;12(6):1755-1770. |
R826600 (1999) R826600 (2000) R826600 (2001) R826600 (Final) |
Exit |
|
Gergel SE, Turner MG, Miller JR, Melack JM, Stanley EH. Landscape indicators of human impacts to riverine systems. Aquatic Sciences 2002;64(2):118-128. |
R826600 (Final) |
Exit Exit |
|
Kang H, Stanley EH, Park S-S. A sensitive method for the measurement of ammonium in soil extract and water. Communications in Soil Science and Plant Analysis 2003;34(15-16):2193-2201. |
R826600 (2001) R826600 (Final) |
Exit Exit |
|
Kang H, Stanley EH. Effects of levees on soil microbial activity in a large river floodplain. River Research and Applications 2005;21(1):19-25. |
R826600 (Final) |
Exit |
|
Miller JR, Turner MG, Smithwick EAH, Dent CL, Stanley EH. Spatial extrapolation: the science of predicting ecological patterns and processes. BioScience 2004;54(4):310-320. |
R826600 (2001) R826600 (Final) |
Exit Exit |
|
Miller JR, Dixon MD, Turner MG. Response of avian communities in large-river floodplains to environmental variation at multiple scales. Ecological Applications 2004;14(5):1394-1410. |
R826600 (Final) |
Exit |
|
Turner MG, Gergel SE, Dixon MD, Miller JR. Distribution and abundance of trees in floodplain forests of the Wisconsin River: environmental influences at different scales. Journal of Vegetation Science 2004;15(6):729-738. |
R826600 (Final) |
Exit |
Supplemental Keywords:
water, watersheds, soil, land, ecosystem, indicators, ecological effects, regionalization, scaling, habitat, ecology, remote sensing, Midwest., RFA, Scientific Discipline, Geographic Area, Ecosystem Protection/Environmental Exposure & Risk, Ecosystem/Assessment/Indicators, Ecosystem Protection, exploratory research environmental biology, State, Ecological Effects - Environmental Exposure & Risk, Ecology and Ecosystems, Ecological Risk Assessment, Ecological Indicators, ecological exposure, landscapes, anthropogenic stresses, habitat, remote sensing, landscape indicator, Midwestern U.S., large river floodplain landscapes, Wisconsin (WI)Relevant Websites:
http://brahms.zoology.wisc.edu/ Exit
http://limnology.wisc.edu/personnel/stanley/floodplain/floodplain.html Exit
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.