Date of Award:
Master of Science (MS)
Understanding how natural variation in flow regimes influences stream ecosystem structure and function is critical to the development of effective stream management policies and actions. Spatial variation in flow regimes is well understood for stream reaches in mesic regions, but a more robust characterization of flow regimes in arid regions is needed, especially to support biological monitoring and assessment programs. Methods are specifically needed that can accurately predict the flow regime expected at ungauged reaches. We used long-term (41 y) records of mean daily flow from 287 stream reaches in the arid western USA to develop and compare several alternative classifications. We also evaluated how accurately we could predict flow regime classes from topographic, soil, and climatic data. Over the 41-y record examined (1972 – 2013), the 287 stream reaches varied continuously from being always wet (perennial) to being dry most days. We explored 5 hierarchical levels of classification and interpreted the 5-group classification to include ephemeral, nonperennial snowmelt-driven, perennial snowmelt-driven, nonperennial rain-driven, and mixed perennial/nonperennial, rain-driven flow regime classes. We created a second set of 4 classifications based on the percentages of days and years with zero flows. We then built random forest classification models to predict class membership, in addition to 2 random forest regression models to directly predict the mean percent of days in a year with zero flow and the number of years with zero flow. Ephemeral and perennial stream reaches were predicted with less error than stream reaches with intermediate nonperennial days or years. The regression models explained ~ 50% of the variation in both percent of zero flow days in a year and percent of zero flow years. These models would predict flow regimes at ungauged reaches in Arizona, identifying ephemeral flow regimes. Maps based on these predictions were generally consistent with qualitative expectations of how flow regimes varied spatially across the state, but larger Arizona stream reaches were predicted with more error than smaller stream reaches. These results represent a promising step toward more effective stream assessment and management in arid regions.
Merritt, Angela M., "Classification and Prediction Models for Natural Streamflow Regimes in the Arid Southwestern USA" (2020). All Graduate Theses and Dissertations. 7974.
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 .