Abstract

Some of the most visible effects of Anthropogenic climate change are shifts in resource (water) availability, and disturbance regimes across the world. Ecosystems across the world are now being forced to adapt to amplified disturbances, or in some cases, completely new disturbance types. These shifts have major implications for ecosystem stability and functioning, and more broadly for nutrient cycling across the world. In the western USA, wildfire is playing a larger and more dramatic role than before, altering vegetation communities, landscape stability, and soil and stream chemistry. While fire has always played an important role in semi-arid ecosystems, the scale and intensity at which wildfires are happening now raises serious questions about how human (and other) lives will be affected by this dramatic alteration of historical disturbance regimes. To better predict how these changes will impact not only water availability and quality, but also ecosystem functioning and stability, better tools are needed to reveal how these complex processes and systems interact. Machine learning provides a unique opportunity for modeling complex system behavior, long-term trends, and spatial dynamics. Here, we showcase how machine learning can be applied to water quality modeling and prediction (chapter 1), and the recovery of microbial communities and stream chemistry after an intense “megafire” in the western USA (chapter 2). Both chapters provide case studies on how machine learning can be applied to predicting and interpreting the impacts of increased disturbance on ecosystems we depend on.

Degree

College and Department

Life Sciences

Rights

https://lib.byu.edu/about/copyright/