For mobile, landscape view is recommended.
This project aimed to develop and evaluate practical methods to estimate species richness and occupancy of diverse taxonomic groups across space and time and in diverse ecosystems. As reflected in the statement of need to which this project responded, this work is relevant to the needs of the Department of Defense (DoD) to assess and monitor native species and to evaluate the potential effects of land use and management actions on native species. Species richness, or the number of native species, is a common surrogate measure of ecosystem status. Occupancy, the probability that a given location is occupied by a species, can serve as a surrogate measure of the species’ abundance, which in turn is related to probability of persistence.
This project addressed five major objectives consistent with the primary aim. First, researchers assessed relations between environmental variables and species richness at nested spatial extents. Second, they developed guidelines for periodicity of sampling. Third, they tested methods for estimating species richness of multiple taxonomic groups on the basis of spatial variation in occurrence or heterogeneity of one group. Fourth, they examined the extent to which occupancy could provide a foundation for measuring species richness. Fifth, they investigated potential responses of species richness and occupancy to manageable environmental changes, and meaningful scales of sampling for detection of biological effects of environmental change.
As land use changes in the multi-jurisdictional lands surrounding DoD installations in the US, it becomes more challenging to sustain military readiness and conserve biological diversity. Researchers sampled anurans, birds, and butterflies in the Chesapeake Bay Lowlands (including Joint Base Langley–Eustis), and birds and butterflies in the central and western Great Basin (including Hawthorne Army Depot and the Marine Corps Mountain Warfare Training Center), from 2012–2015. Working in ecosystems in which land-cover configurations and drivers of those configurations differ, and on disparate faunal groups, allowed the researchers to evaluate the geographic and taxonomic transferability of their methods and inferences. They also capitalized on existing data and local experience that served the project aims.
The sampling methods allowed the researchers to estimate detection probability—the probability of detecting a species if it is present. If detection probability is not quantified, inferred relations between environmental covariates and species occurrence or demography may be erroneous. They conducted hierarchical analyses of the species richness of birds and butterflies in the central and western Great Basin. They modeled species richness at two nested spatial extents as functions of environmental variables measured at either extent. They used the same data to develop a method to identify indicator species (small sets of species with occurrence patterns that are related to species richness of larger sets) and environmental variables that explain considerable variation in species richness. They conducted novel, rigorous external evaluations of the spatial and temporal transferability of such indicator species. Furthermore, the researchers evaluated the extent to which different ecological processes might drive spatial and temporal variation in species identities.
Researchers examined the responses of estimated detection probability and occupancy of birds in the Chesapeake Bay Lowlands and central and western Great Basin to the duration of sampling. Additionally, they examined whether detection probability and occupancy of birds and anurans in the Chesapeake Bay Lowlands, and butterflies in the Great Basin, differed if sites were sampled repeatedly on a single day versus once each on multiple days.
Researchers explored the extent to which occupancy of butterflies in the Chesapeake Bay Lowlands and central and western Great Basin was associated with vegetation structure and composition, topography, and other environmental attributes and whether assumptions of closure were met with assemblage-level sampling designs. Furthermore, they identified attributes of wetlands and the surrounding terrestrial environment that were associated with anuran populations of high relative abundance versus populations of any abundance class, and evaluated how these attributes relate to different life history stages.
Researchers examined how songbirds in the Chesapeake Bay Lowlands, including species of concern, may respond to changes in vegetation fragmentation and structure that result from urbanization and different levels of use by white-tailed deer. They also explored turnover in species composition of birds and butterflies in the central and western Great Basin to gain insight into meaningful scales of sampling for detection of biological effects of environmental change.
In the central and western Great Basin, species richness of birds and butterflies responded to environmental variables at different spatial extents. Species’ identities were more consistent in space in butterfly than in bird assemblages, whereas spatial nestedness (the extent to which species-poor faunas are statistical subsets of species-rich faunas) often was higher in butterfly than in bird assemblages. In some cases, researchers were able to explain more than 80% and 70% of the variation in species richness of birds and butterflies, respectively, as functions of vegetation and topography. The predictive accuracy and spatial and temporal transferability of models of species richness of birds and butterflies that are based on indicator species can exceed that of models based on environmental variables.
In all of the ecosystems, modest increases in the duration of point counts for birds did not affect inferences based on occupancy models, but might affect estimates of species richness that were not detection-weighted. Given the results, researchers believe that estimation of use may be equally or more informative than estimation of occupancy for exploring species-environment relations.
In many cases, because detection probabilities for many species of birds and butterflies were low, occupancy could not be estimated for more than 50% of the species detected in each ecosystem and year. Although imperfect detection may lead to biased estimates of species occurrence, demographic parameters, and species-environment relations, achieving certain scientific objectives still may require use of occurrence data that are not detection weighted. The results suggest that if anurans are extirpated from wetlands in the Chesapeake Bay Lowlands, the wetlands are unlikely to be recolonized.
The researchers regularly shared data and inferences with environmental managers at the installations on which they worked, who indicated that the results would contribute to watershed management programs and Integrated Natural Resources Management Plans. They also shared data and inferences with DoD partners, including Landscape Conservation Cooperatives. In the Chesapeake Bay Lowlands, the results are relevant to management of white-tailed deer, which have substantial effects on other native species and, via transmission of disease vectors, public health. In the Great Basin, the results are applicable to vegetation treatments that are intended to increase the probability of persistence of Greater Sage-Grouse and benefit hundreds of other species. The results also suggest directions for future research on the extent to which simple measures of species occurrence may provide information on probabilities of species persistence.