Shirk P.L., Linden D.W., Patrick D.A., Howell K.M., Harper E.B., Vonesh J.R.
Department of Biology, Virginia Commonwealth University, Richmond, VA, United States; Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI, United States; Division of Forestry, Natural Resources and Recreation, Paul Smith's College, Paul Smith's, NY, United States; Department of Zoology and Wildlife Conservation, University of Dar es Salaam, Dar es Salaam, Tanzania; Department of Biology, University of Florida, Gainesville, FL, United States; Department of Wildlife Ecology and Maine Cooperative Fish, Wildlife Research Unit, University of Maine, Orono, ME, United States
Shirk, P.L., Department of Biology, Virginia Commonwealth University, Richmond, VA, United States, Department of Biology, University of Florida, Gainesville, FL, United States; Linden, D.W., Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI, United States, Department of Wildlife Ecology and Maine Cooperative Fish, Wildlife Research Unit, University of Maine, Orono, ME, United States; Patrick, D.A., Division of Forestry, Natural Resources and Recreation, Paul Smith's College, Paul Smith's, NY, United States; Howell, K.M., Department of Zoology and Wildlife Conservation, University of Dar es Salaam, Dar es Salaam, Tanzania; Harper, E.B., Division of Forestry, Natural Resources and Recreation, Paul Smith's College, Paul Smith's, NY, United States; Vonesh, J.R., Department of Biology, Virginia Commonwealth University, Richmond, VA, United States
Aim: We map estimated historical population declines resulting from species-specific models of sensitivity to habitat fragmentation for three forest-dependent chameleons. Location: East Usambara Mountains, Eastern Arc Mountains, Tanzania. Methods: We surveyed three chameleon species (Rhampholeon spinosus, Rhampholeon temporalis and Trioceros deremensis) along 32.2 km of transects and used a hierarchical, distance-sampling model to estimate densities. The model included habitat characteristics at the landscape (patch) and local (transect) scales while accounting for detectability. By analysing the model in a Bayesian framework, we were able to propagate error through the entire analysis and obtain exact solutions despite small sample sizes. We then used our estimated relationships between habitat and density to project chameleon population sizes across current and historical land cover maps of the study area (230 km2), giving an estimate of the impact of anthropogenic habitat alteration on these species. Results: Species' densities increased in larger patches and further from patch edges and varied seasonally. Local vegetation characteristics had significant relationships with expected chameleon densities, though effect sizes were small. Estimates of total current population sizes varied by two orders of magnitude among species, but each was 49-79% higher than detection-naïve estimates. All three declined from estimated historical levels by approximately 60% in the study area, approximately one-third more than would be expected from forest loss alone. Remaining populations of the study species are predominantly located in protected nature reserves, so the future of these species will likely be determined by the degree of protection offered by the nature reserves. Main conclusions: Habitat loss and fragmentation have greatly reduced forest-dependent chameleon population sizes in the East Usambara Mountains. Populations of these species in other areas are experiencing higher rates of habitat loss. Efforts aimed at ensuring the efficacy of protected forests may be a key to conserving remaining populations. © 2014 John Wiley & Sons Ltd.