Dissertation/Thesis Abstract

Evaluating violations of modeling assumptions as potential contributors to inconsistent landscape effects within imperiled ecological systems
by Tavernia, Brian G., Ph.D., Tufts University, 2010, 206; 3422363
Abstract (Summary)

Inconsistencies in species-distribution-landscape models call into question the ecological inferences drawn about the influence of landscape features and lead to uncertainty when trying to use these relationships in an applied context. I evaluated violations of modeling assumptions as potential contributors to inconsistent landscape effects, especially urbanization effects, on wetland and forest birds. (1) First, I evaluated the influence of landscape extent and habitat context, factors that commonly differ between studies, on correlations between urban metrics. I found that multivariate associations and bivariate correlations among urbanization measures varied across landscape extents and habitat contexts. These results indicate that urbanization measures are not interchangeable and that inconsistent landscape effects could be due to the use of different metrics. (2) Next, I examined the influence of several spatial, temporal, and life history assumptions on the outcomes of distribution-landscape models relating wetland bird communities to independent landscape gradients, including urbanization. My analyses indicated that wetland bird communities were responding to urbanization in a negative, threshold fashion; specifically, wetland obligates tended to be absent from urban settings. The overall results from my analyses provided evidence that violations of common assumptions (e.g., selection of appropriate extent, lack of time-lagged effects) can impact the outcome of distribution-landscape models, which could lead to inter-study inconsistencies. (3) I also investigated the possibility that the observed threshold effect was due to an increase in the prevalence of the exotic, invasive plant purple loosestrife (Lythurm salicaria). I found that L. salicaria coverage was lower within wetlands but varied less across wetlands in urban settings. Negative responses by some obligate wetland birds suggested L. salicaria may contribute to the observed threshold effect. (4) Finally, I found that resident forest bird species were no more likely than migratory species to respond to landscape features quantified in landscape extents truncated at local movement barriers. This contradicts the hypothesis that migratory birds evaluate habitat features in a top-down, hierarchical manner while resident birds employ a bottom-up strategy, making a series of local-scale movements to evaluate landscapes. My results suggest that migratory status cannot be used as a proxy for habitat evaluation strategy in birds.

Indexing (document details)
Advisor: Reed, J. Michael
Commitee: Bolger, Douglas, Chew, Frances, Orians, Colin, Starks, Philip
School: Tufts University
Department: Biology
School Location: United States -- Massachusetts
Source: DAI-B 71/10, Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Ecology
Keywords: Birds, Exotic species, Habitat fragmentation, Habitat selection, Invasive species, Landscape effects, Urbanization
Publication Number: 3422363
ISBN: 9781124212494
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest