2001 Progress Report: Development And Testing Of A Multi-Resource Landscape-Scale Ecological Indicator: Forest Fragmentation, Structure, and Distribution Relative to Topography

EPA Grant Number: R826598
Title: Development And Testing Of A Multi-Resource Landscape-Scale Ecological Indicator: Forest Fragmentation, Structure, and Distribution Relative to Topography
Investigators: Seagle, Steven W. , Townsend, Philip A.
Institution: University of Maryland Center for Environmental Science
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1998 through September 30, 2003
Project Period Covered by this Report: October 1, 2001 through September 30, 2002
Project Amount: $683,374
RFA: Ecological Indicators (1998) RFA Text |  Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Ecosystems


The objective of this research project is to understand how water quality and avian habitat quality vary across landscapes as a function of forest fragmentation and forest topographic position. Our primary hypothesis is that avian habitat quality and water surface water quality covary within the Mid-Atlantic Highlands because of their dependence on topography. We should be able to develop indicators, which simultaneously reflect both resources from databases of topography (digital elevation models) and land use. Because vertical structure is also notably important for avian diversity and habitat suitability, we also are developing the use of synthetic aperture radar remote sensing of forest structure in conjunction with topography and land use. The primary field sites being employed are located on the Appalachian Plateau (within Savage River State Forest, Garrett County, MD) and in the Ridge-and-Valley physiographic province (Green Ridge State Forest, Allegany County, MD).

Progress Summary:

During this funding period, our work shifted toward statistical analysis of previously collected data and remote sensing data. Specifically, we focused on: (1) statistical analyses of the relationships among forest-interior bird reproductive success, forest productivity, and topography; (2) a focused analysis of the relationships among ovenbird (Seiurus aurocapillus) reproductive success, forest productivity, and invertebrate biomass; (3) collecting data on the distribution of invertebrate biomass relative to topography and bird territories; (4) development of forest structure measurements from synthetic aperture radar images; and (5) continued reclassification of land use for the Mid-Atlantic Highlands from Landsat images. Each of these topics is addressed separately.

Bird Reproduction, Forest Productivity, and Topography. These analyses evaluated the hypothesis that forest interior songbird reproduction is significantly correlated with topographic variation in forest productivity, and differs between the Appalachian Plateau and Ridge-and-Valley physiographic provinces. These provinces differ in their annual precipitation amounts (Appalachian Plateau: 114-144 cm/yr; Ridge-and-Valley: 76-88 cm/yr), and also in geology. Site index was chosen as a long-term integrative measure of forest productivity and an indicator of soil moisture conditions. Site index was measured for our 22 field study sites, which range in age from 74-134 years in tree age and are composed of mixed oaks. Previously published site index equations were used to calculate the site index for each of our study areas from field data on tree age and height. Site index proved to be a reliable index of forest productivity across the range of site conditions we measured. Analysis of covariance indicated that forest interior bird density was greater on the Appalachian Plateau than in the Ridge-and-Valley, and that this density was positively correlated with site index in both provinces. The higher density of interior species on the Appalachian Plateau was attributed primarily to its greater diversity of forest interior species. Relative fledging success, determined for each of our study areas using behavior mapping, did not differ between physiographic provinces. Contrary to our hypothesis, fledging success was not correlated with site index for fledging data collected during 1999. In contrast, fledging success was very positively correlated with site index for data collected in 2000, which was a year of typical rainfall.

Thus, our hypothesis appears correct for both provinces during normal weather years, but the direct and indirect effects of drought apparently homogenize spatial variability in reproductive success by overriding topographic effects on soil moisture. These results were incorporated into a cartographic model of potential habitat quality for forest interior birds using digital elevation models and land cover data; the model was applied to a 3,000 km2 area that included both the Appalachian Plateau and Ridge-and-Valley provinces covering western Maryland, and parts of adjacent West Virginia and Pennsylvania. Analysis of variance indicated that land use change had a greater impact on forest interior birds than expected from forest pattern alone, because development is biased toward more topographic positions of higher productivity (both forest and bird). Overall, this mapping exercise produced a map of not only where forest interior bird habitat occurs (land use effect), but also relative ranking of habitat quality (productivity) with remaining forest interior. Because of its basis in bottom-up effects (from topography to forest productivity to avian productivity) and our subsequent ability to map over large areas based on widely available databases (i.e., digital elevation models and land use maps) we feel that the relationship demonstrated in this study carries great potential as a regional indicator of both avian habitat quality and forest productivity.

Ovenbird Reproduction and Litter Invertebrate Biomass. Having established that forest productivity is predictive of avian productivity across a heterogeneous landscape, we thought that it was important to establish the causal link between forest and bird productivity. Thus, we examined avian food resources as a link between the two, focusing on one of the most common forest interior bird species. Forest-floor detrital food webs have classically been described as donor-controlled systems sustained by annual inputs of carbon and nutrients from leaf fall. Little evidence exists, however, to document whether the bottom-up control from detrital resources influences the productivity of the vertebrate species feeding on the detrital food web. We hypothesized that the reproductive success of ovenbirds, a forest-interior neotropical migrant that feeds on invertebrates of the forest-floor, is a function of invertebrate biomass within the detrital food web, and that both invertebrate biomass and ovenbird reproductive success can be predicted by an integrative measure of forest productivity (site index; see description of site index calculations above). Ovenbird nesting success was measured for 8 of our 22 10-ha forest-interior study areas, with 4 plots located on the Appalachian Plateau and 4 in the Ridge-and-Valley physiographic province. Within each province, study sites were stratified between dry upper and wetter lower slope positions. Forest litter invertebrate biomass and forest site index were determined for each study area. Invertebrate biomass also was measured within and outside bird territories for these study areas. Our results indicate that invertebrate biomass is predictable from forest site index.

In addition, ovenbird reproductive success is even more strongly related to both site index and invertebrate biomass. Even though ovenbirds apparently impact the size distribution of invertebrates within their territories, this top-down effect does not appear to be strong. Thus, we conclude that bottom-up effects of forest productivity propagate though the detrital food web to influence productivity of secondary/tertiary vertebrate predators. Stronger bottom-up effects were evident for study sites within the drier Ridge-and-Valley province; greater precipitation appears to partially ameliorate topographic effects on the wetter Appalachian Plateau. Tree species composition, insect frass fall from the forest canopy, and microtopography also were quantified. These factors vary in their influence on specific invertebrate taxa and elicit within-stand spatial variation in bottom-up effects. For example, overstory tree composition influences the quality of leaf litter entering the detrital food web. The density of invertebrate predators is clearly correlated with forest study areas dominated by tree species that produce more readily decomposable leaf litter (e.g., maple spp.). Even with a single growing season, springtails (Collembola spp.), which form an important base to the detrital food chain, are positively correlated with frass fall from the canopy. Finally, it was observed that ovenbirds (and other species of forest interior bird species) align their territories with mircotopographic features such as swales and depressions, ostensibly because microtopography influences site quality and food availability. We project that site productivity, which is predictable from topography (see above), should prove to be an effective tool for predicting landscape-scale variation in the strength of bottom-up effects in the detrital food web.

Invertebrate Biomass Relative to Bird Territories and Topography. We collected one further data set relating topography, avian distribution, and invertebrate biomass during the summer of 2000. This effort focused on ovenbirds on X of our study sits in the Ridge-and-Valley, because: (1) ovenbirds are our most common single species; (2) ovenbird densities are greatest in the Ridge-and-Valley; and (3) topographic effects on avian reproduction seemed more pronounced in the drier Ridge-and-Valley than the Appalachian Plateau. On each of the X study sites, spot mapping was used to define ovenbird territories. With these territories defined, we sampled invertebrate biomass in the forest floor litter (the ovenbird foraging substrate) within each ovenbird territory and also within at a randomly chosen set of coordinates within each study area. With each territory, multiple measures of forest canopy height (a strong correlate of site index and forest productivity) were taken. Ovenbird territories also were classified by microtopography. Complete analyses of these data are in progress with focus on comparing litter invertebrate biomass inside territories with randomly chosen plots, effects of ovenbird predation on invertebrate biomass and size distributions, and the alignment of ovenbird territories with small-scale (microtopographic) gradients in soil moisture.

Forest Structure From Synthetic Aperture Radar Imagery. We developed a geometric model of forest vertical structure-based on field measurements-that is suitable for several objectives: (1) analyzing the distribution of forest interior birds and forest invertebrates; (2) predicting the spatial extent of forests of varying vertical complexity; and (3) integrating a single metrics or small number of metrics with remote sensing imagery to map forest vertical structure. The model employs coarse measurements of the three-dimensional structure of a forest to estimate for any location a series of metrics that provide an overall synthetic estimation of the complexity of the forest. The model uses analytical geometry to determine the volume of canopy space filled by various structural levels in the canopy (e.g., canopy, subcanopy, shrub layer). This information is then used to estimate the distribution of canopy elements (e.g., leaf area) within a structural level and then determine the degree of complexity of the canopy. Structural complexity is measured using Gini coefficient (G), a measure of inequality among canopy layers, as well as the Shannon Weiner index (H'), a measure of the evenness and diversity of a canopy. Also reported is VLAI, which is a measure of the dispersion of leaf area by volume (i.e., canopy volume divided by leaf area). Different types of forests exhibit markedly different structural attributes. The model very accurately recreates the appearance of the forest in the field and provides a limited number of new metrics to describe that distribution.

The measures of forest vertical structure have been used to explain bird distributions (see Brian Sturtevant's work), and are currently being used to understand the distribution of forest invertebrates. The forest structural attributes are strongly related to remote sensing measures from Landsat and SAR data, and we anticipate the completion of the development of remote sensing-based spatial models of the distribution of forest vertical structure during the final year of this research project.

Land Use Classification for the Mid-Atlantic. One of our goals for this research project is to examine the relationships among watershed nutrient export, the extent of forest cover, and the spatial distribution of forest cover (especially relative to topography). Our original intent was to use the Interagency Multi-Resolution Land Cover Characterization (MRLC) land use data, but after examining some of the inconsistencies of this database for our topographically diverse field study areas, we embarked on a new classification of land use for the Mid-Atlantic Highlands that will be used for this analysis. During the previous reporting period, we developed a method to rapidly correct multiple images for topographic shading. This method greatly reduced the effects of shading on the classification of Landsat imagery, and also permitted comparison of vegetation on slopes with differing orientations but otherwise similar vegetation. Consequently, we expect an increase in accuracy of the land cover classification. During the current reporting period, all Landsat images for the western two-thirds of the Chesapeake Bay watershed were classified using land use classes that directly correspond to MRLC. Tedious efforts were made to adjust for cloud cover and snow cover. The resulting classified Landsat scenes, which are just being "pieced together" as a complete coverage, will be used to: (1) compare with MRLC to assess the degree of change associated with our new classification; (2) calculate statistics of forest cover, forest fragmentation, and forest distribution relative to topographic position for HUC 8-digit watersheds contained within our classification coverage; and (3) analyze the statistical relationships between nitrogen export (database supplied by the U.S. Environmental Protection Agency) and these forest statistics.

Future Activities:

During the final year of this research project, we will focus on: (1) publication of results from the avian studies; (2) combining our avian habitat quality results with the remote sensing of forest structure to refine our habitat quality models; (3) finishing analyses of surface water quality as a function of forest distribution and topography; and (4) developing indicators to simultaneously reflect both water quality and avian habitat quality. Publication of the results from our avian habitat quality studies is already well advanced and should consist largely of final editing and submission of manuscripts. As noted in this report, our measures of forest structure also are complete. Coordinately, these structural measures with topographic variables to predict avian habitat quality will depend on finding statistically valid relationships between avian diversity or reproductive success and forest structure, followed by regional extrapolation via our extensive SAR data. As noted, the analysis of surface water quality as a function of forest distribution hinges on our reclassification of land use for the Mid-Atlantic Highlands. This process is near completion, and thus statistical analyses represent the major effort for this part of the research project. The results of this analysis will influence our ability to develop indicators that simultaneously reflect avian habitat quality and water quality. Essentially, all of our databases for attempting this synthesis are complete. Thus, after some initial data analyses, we look forward to a year of synthesis that results in testing of our original hypotheses concerning ecological indicators for multiple resources.

Journal Articles on this Report : 2 Displayed | Download in RIS Format

Other project views: All 38 publications 7 publications in selected types All 7 journal articles
Type Citation Project Document Sources
Journal Article Sturtevant BR, Seagle SW. Food availability for forest interior songbirds in the central Appalachians: the role of topography. Landscape Ecology. R826598 (2001)
R826598 (2002)
not available
Journal Article Sturtevant BR, Seagle SW, Townsend PA, Chasstain RC. Forest songbird diversity, density, and fledging success as a function of topography and forest productivity in western Maryland. Ecological Applications. R826598 (2001)
not available

Supplemental Keywords:

water, watersheds, land, animal, ecosystem, indicators, scaling, terrestrial, aquatic, habitat, biology, ecology, hydrology, Environmental Monitoring and Assessment Program, EMAP, surveys, Landsat, remote sensing, Chesapeake Bay, EPA Region 3, Maryland, MD, Mid-Atlantic Highlands, forest interior, fragmentation, topography, digital elevation models, synthetic aperture radar, invertebrate., RFA, Scientific Discipline, Geographic Area, Water, Ecosystem Protection/Environmental Exposure & Risk, Hydrology, Ecology, Water & Watershed, Ecosystem/Assessment/Indicators, Ecosystem Protection, Forestry, Ecological Effects - Environmental Exposure & Risk, Mid-Atlantic, Ecological Risk Assessment, Biology, Geology, EPA Region, Watersheds, Ecological Indicators, risk assessment, remote sensing, landscape indicator, multi-level indicators, stream ecosystems, Region 3, bird habitat, ecosystem indicators, estuarine ecosystems, gypsy moth, Mid-Atlantic Highlands, terrestrial, aquatic ecosystems, water quality, stress responses, defoliation, land use

Progress and Final Reports:

Original Abstract
  • 1999
  • 2000 Progress Report
  • 2002 Progress Report
  • Final Report