Data-Driven Modeling of Snowmelt, Streamflow and Ion Export Dynamics

Abstract:

Headwater watersheds in the Rocky Mountains play a critical role in the water resources of the Western United States. Predicting snow and streamflow dynamics – such as snowmelt timing, peak flow timing, annual discharge and annual nitrogen export – is critical for management decisions as well as for watershed science. In this project, we first apply data mining approaches to quantify the effect of air temperature on snowmelt timing. We then calculate the impact of air temperature and snowmelt timing on streamflow discharge and ion export in the East River Watershed in Colorado using the historical record over the last 30 years. We explore the spatiotemporal variability of these variables (snowmelt timing, discharge, nitrogen and element exports) along with the spatial variability of plant species, ecological dynamics and geology using different machine learning methods, such as linear regressions, random forest, and clustering. These models inform research into different functional relationships within the watershed system, offer practical implications for guiding field campaigns and support better watershed management.