In collaboration with DrivenData and the United States Bureau of Reclamation (USBR), I led a team to develop a spatially-aware, explainable machine learning forecast seasonal streamflow in the Western United States.

Project Overview

Motivation	Machine learning models can be used to improve water supply predictions, but institutions are reluctant to use them due to concerns about explainability
Model	An ensemble machine learning model that uses a variety of high quality spatial data to produce explainable seasonal streamflow forecasts
Client	United States Bureau of Reclamation (USBR)
Status	Complete
Outcome	TBD

Accurate seasonal water supply forecasts are crucial for effective water resources management in the Western United States. This region faces dry conditions and high demand for water, and these forecasts are essential for making informed decisions. They guide everything from water supply management and flood control to hydropower generation and environmental objectives.

Yet, hydrological modelling is a complex task that depends on natural processes marked by inherent uncertainties, such as antecedent streamflow, snowpack accumulation, soil moisture dynamics, and rainfall patterns. To maximize the utility of these forecasts, it’s essential to provide not just accurate predictions, but also comprehensive ranges of values that effectively convey the inherent uncertainties in the predictions.