Date of Award:

5-2022

Document Type:

Dissertation

Degree Name:

Doctor of Philosophy (PhD)

Department:

Watershed Sciences

Committee Chair(s)

Charles Hawkins

Committee

Charles Hawkins

Committee

D. Richard Cutler

Committee

Scott Miller

Committee

Edd Hammill

Committee

Brett Roper

Abstract

The temperature of streams and rivers is changing rapidly in response to a variety of human activities. This rapid change is concerning because the abundances and distributions of many aquatic species in streams and rivers are strongly associated with temperature. Linking observations of temperature effects on species distributions with observations of temperature effects on fitness is important for improving confidence that temperature (and not some other variable) is causing the distributions we observe. Furthermore, producing accurate models of temperature effects on species distributions may allow us to develop tools to diagnose whether or not thermal pollution has impaired aquatic life. Such a diagnostic tool could help us better target management efforts on the specific stressors impairing aquatic life. In chapter two, I describe several laboratory experiments designed to examine the link between the effects of temperature observed in the field with effects of temperature observed in the laboratory. I found that the effects of temperature on survival were correlated with the thermal limits inferred from species distributions, which supports the hypothesis that temperature influences distributions by affecting the survival of species. In chapters three and four, I assessed two techniques that could potentially improve our ability to model relationships between temperature and distributions. In chapter three, I show that methods for dealing with imbalanced data broadly improved our ability to model the relationship between predictor variables (temperature and other variables) and species distributions. In chapter four, I evaluated a recently developed technique (deep artificial neural networks) for modeling large complex datasets. I found that deep artificial neural networks did not improve predictions over that of standard artificial neural networks and random forest models. In chapter five, I developed and evaluated a diagnostic biotic index for diagnosing the likelihood that temperature has affected macroinvertebrate species in streams and rivers. This index showed that 2.6% of streams across the continental United States had species with thermal tolerances higher than expected compared with thermally undisturbed conditions.

Checksum

49d083dcd5f72977027c5d7b0991b529

Download

Share

COinS

Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .