 |
|
|||||||
| A multi agency project funded by US EPA's STAR Program | ||||||||
Results: Birds of the Coastal Zone
Ecological Condition (Cobs) Based on Coastal Zone Birds in the Eastern Deciduous Forest Province
Project Summary
Coastal regions of the Laurentian Great Lakes reflect conditions in the surrounding watershed and exert an important influence on the lakes themselves. Indicators of ecological health or integrity in the coastal zone therefore provide valuable information about the effects of human activities on the Great Lakes ecosystem. As part of the GLEI project, researchers from the University of Wisconsin-Green Bay, University of Minnesota Duluth’s Natural Resources Research Institute, and Cornell University conducted an extensive survey of birds in the U.S. portion of the Great Lakes coastal zone, defined as the area within 1 km of the shoreline. Data were collected at 171 randomly selected coastal segments (from 762 total) stratified across a multivariate gradient of environmental stressors (Danz et al. 2005), 76 of which were located in the Eastern Deciduous Forest Ecological Province. For each segment, we conducted transect surveys consisting of standard 10 min bird counts at 15 roadside sampling points, each separated by at least 500 m. Selected sites were sampled twice, yielding a combined sample size of 2544 point counts at 194 bird census routes. The primary objective of this work was to develop multispecies indicators of ecological condition that can be applied at both local and basin-wide scales.
Results
We encountered 187 bird species during the two-year (2002-03) field effort. The numbers of species per transect in the Eastern Deciduous Forest Ecological Province (mean = 43.7 species, s.d. = 8.7) were remarkably similar to those in the Laurentian Mixed Forest Ecological Province (mean = 43.6 species, s.d. = 9. 6), but species composition differed significantly between the two regions. Human impacts associated with the bird survey transects were assessed by GIS/remote sensing analysis of Landsat 5 and Landsat 7 satellite images. Proportions of cover in 5 general land use categories (natural vegetation, residential, commercial/industrial, agricultural, and roads) were evaluated at different distances (100 m, 500 m, 1 km, 3 km, 5 km) from the 15 bird survey points. Multivariate analysis was used to generate a formula for calculating “reference condition” based on land cover, ranging from 0 (maximally impacted by humans) to 10 (minimally impacted by humans). The frequency of each species among the 15 point counts was used as the probability of encountering the species in the coastal segment. Probabilities were plotted against the reference condition to describe the species’ response to environmental stress. A simple, 4 parameter mathematical function was used to describe each relationship (Figure 1). These SSD functions contain information about the response of each species to human impacts as well as the overall probability of finding the species in the study area. Our results demonstrate that these relationships might vary by geographic region. Once derived, the SSD functions can be used to calculate the ecological condition of new sites.
Calculating Ecological Condition
The method for calculating ecological condition (C) follows an iterative or computerized “trial-and-error” process described by Hilborn and Mangel (1997). The first step is to derive parameters of SSD functions for species of interest. This is done through expert opinion or preliminary work like ours. We have derived statistically significant functions for 50 species in the Eastern Deciduous Forest Ecological Province, but robust multivariate indicators can be derived using a smaller number of species. We recommend a list of 25 species showing the strongest SSD functions and representing a range of habitat preferences. Important species on this list include Veery, Ovenbird, Red-eyed Vireo, Black-capped Chickadee, Chipping Sparrow, Red-bellied Woodpecker, American Redstart, and Canada Warbler (positively associated with condition) and Rock Pigeon, House Sparrow, European Starling, Common Grackle, and Ring-billed Gull (negatively associated with condition). Given data on frequencies or presence/absence of the selected species, the next step is to compute (through iteration) the value of C that best “fits” the observed occurrence data. Absence of a species can be as important as presence, so surveys need to be complete (i.e., comparable to the method used for deriving the SSD functions) for the target species. The iterative process for calculating C can be implemented with the help of familiar computer software like Microsoft Excel.
Ecological Condition (Cobs) Based on Coastal Zone Birds in the Laurentian Mixed Forest Province
Project Summary
Coastal regions of the Laurentian Great Lakes reflect conditions in the surrounding watershed and exert an important influence on the lakes themselves. Indicators of ecological health or integrity in the coastal zone therefore provide valuable information about the effects of human activities on the Great Lakes ecosystem. As part of the GLEI project, researchers from the University of Wisconsin-Green Bay, University of Minnesota Duluth’s Natural Resources Research Institute, and Cornell University conducted an extensive survey of birds in the U.S. portion of the Great Lakes coastal zone, defined as the area within 1 km of the shoreline. Data were collected at 171 randomly selected coastal segments (from 762 total) stratified across a multivariate gradient of environmental stressors (Danz et al. 2005), 95 of which were located in the Laurentian Mixed Forest Ecological Province. For each segment, we conducted transect surveys consisting of standard 10 min bird counts at 15 roadside sampling points, each separated by at least 500 m. Selected sites were sampled twice, yielding a combined sample size of 2544 point counts at 194 bird census routes. The primary objective of this work was to develop multispecies indicators of ecological condition that can be applied at both local and basin-wide scales.
Results
We encountered 187 bird species during the two-year (2002-03) field effort. The numbers of species per transect in the Laurentian Mixed Forest Ecological Province (mean = 43.6 species, s.d. = 9. 6) were remarkably similar to those in the Eastern Deciduous Forest Ecological Province (mean = 43.7 species, s.d. = 8.7), but species composition differed significantly between the two regions. Human impacts associated with the bird survey transects were assessed by GIS/remote sensing analysis of Landsat 5 and Landsat 7 satellite images. Proportions of cover in 5 general land use categories (natural vegetation, residential, commercial/industrial, agricultural, and roads) were evaluated at different distances (100 m, 500 m, 1 km, 3 km, 5 km) from the 15 bird survey points. Multivariate analysis was used to generate a formula for calculating “reference condition” based on land cover, ranging from 0 (maximally impacted by humans) to 10 (minimally impacted by humans). The frequency of each species among the 15 point counts was used as the probability of encountering the species in the coastal segment. Probabilities were plotted against the reference condition to describe the species’ response to environmental stress. A simple, 4 parameter mathematical function was used to describe each relationship (Figure 1). These SSD functions contain information about the response of each species to human impacts as well as the overall probability of finding the species in the study area. Our results (e.g., Figure 1) demonstrate that these relationships might vary by geographic region. Once derived, the SSD functions can be used to calculate the ecological condition of new sites.
Calculating Ecological Condition
The method for calculating ecological condition (C) follows an iterative or computerized “trial-and-error” process described by Hilborn and Mangel (1997). The first step is to derive parameters of SSD functions for species of interest. This is done through expert opinion or preliminary work like ours. We have derived statistically significant functions for 72 species in the Laurentian Mixed Forest Ecological Province, but robust multivariate indicators can be derived using a smaller number of species. We recommend a list of 25 species showing the strongest SSD functions and representing a range of habitat preferences. Important species on this list include Ovenbird, Black-throated Green Warbler, Red-eyed Vireo, American Redstart, Hermit Thrush, Winter Wren, White-throated Sparrow, and Nashville Warbler (positively associated with condition) and House Sparrow, European Starling, Common Grackle, Rock Pigeon, Red-winged Blackbird, and House Finch (negatively associated with condition). Given data on frequencies or presence/absence of the selected species, the next step is to compute (through iteration) the value of C that best “fits” the observed occurrence data. Absence of a species can be as important as presence, so surveys need to be complete (i.e., comparable to the method used for deriving the SSD functions) for the target species. The iterative process for calculating C can be implemented with the help of familiar computer software like Microsoft Excel.
|
|