Date Awarded

Fall 2019

Document Type


Degree Name

Doctor of Philosophy (Ph.D.)


Virginia Institute of Marine Science


David Johnson

Committee Member

Michael Unger

Committee Member

William Clements

Committee Member

Wayne Landis


A threshold can be defined as the point where small changes in an environmental driver produce an abrupt change within a biological system. These changes can occur at different levels of organization, from organisms to ecosystems. Although thresholds seem to be receiving more attention by ecotoxicologist, not much is known about how contaminants cause or affect thresholds at the landscape level, such as habitat fragmentation thresholds. Habitat fragmentation thresholds can occur due to rapid changes in the landscape structure after a certain amount of habitat is lost, which can cause abrupt effects on the movement of organisms, population abundance and community composition. In this dissertation, threshold effects were investigated using computer simulations, laboratory and field experiments, focusing on the interaction effects of habitat fragmentation and contaminants. In Chapter I, different statistical methods and experimental designs are compared to estimate thresholds of routinely used ecotoxicological tests. In Chapter II, a laboratory experiment was performed to estimate how contamination might interact with fragmentation and affect the movement of organisms. In Chapter III, a manipulative field experiment was conducted to evaluate how contamination can interact with fragmentation in a more natural setting. In Chapter IV, a class of random walkers (Lévy walks) were used to investigate threshold effects of fragmentation on animals with different movement strategies. Chapter I shows that threshold estimation might not be reasonable in all scenarios, such as in datasets without a steep slope. Therefore, threshold models should be used carefully in routine ecotoxicology essays or when there are biological reasons for suspecting the existence of a threshold. In Chapter II and III, both mercury exposure and habitat loss affected the movement behavior of organisms. Even though nonlinear responses were observed in the speed and mean directional changes of organisms, the data did not provide evidence of a threshold response. Both chapters also show that depending on the scale and the fragmentation process, organisms might respond differently to habitat fragmentation. Nonetheless, mercury exposure consistently reduced the movement of the marsh periwinkle in both scenarios. Chapter IV shows that organisms with different movement behaviors might be affected by fragmentation differently, which could result in a strong threshold response. For instance, organisms performing Brownian walk were the most affected by fragmentation but without a strong threshold response. Lévy walkers with μ≈2 presented the strongest threshold response. These evidences combined suggest that contamination may interact with habitat fragmentation and reduce the functional connectivity of landscapes by altering the movement of organisms. This would decrease the search efficiency of organisms and make them more susceptible to the effects of habitat fragmentation. Future studies should investigate the possible interaction effects of theses stressors at higher levels of biological organization.


© The Author