Drought Prediction

Motivation:

The US Drought Monitor provides a snapshot of drought status in the US and classifies drought into 6 categories, from 0 - no drought, to D1 - D4, representing increasing severity of drought. Drought status for different regions of the US is determined by combining quantitative data inputs with qualitative assessments from drought experts after analyzing current environmental conditions. NOAA provides Drought Outlooks for 1 and 3 months into the future that predict whether the drought will persist, diminish, or increase in severity; however, a spatially discretized probabilistic outlook of the drought status does not yet exist. 

Research Goals & Objectives:

With support from the NSF Computational and Data-Enabled Science and Engineering Program, our research team is developing a data-driven statistical model that predicts the drought status. The methodology is computationally efficient and accounts for both spatial and temporal correlations in the high-dimensional, multivariate input dataset. Results from the pilot model are described here, and the input dataset we derived is described here and hosted on the data dryad platform. In addition to developing the model, we are applying it to understand the extent to which it can predict flash droughts, and working with members of the North Carolina Drought Advisory Council to explore how the predictions can support drought response efforts. Finally, we are installing CoCoRaHS rain gauges at local K-12 schools to inspire students to be citizen scientists.

Research Team:

• Dr. Rob Erhardt (PI, WFU Statistical Sciences)

• Dr. Staci Hepler (Co-I, WFU Statistical Sciences)

• Dr. Courtney Di Vittorio (Co-I, WFU Engineering)

• Dr. Lauren Lowman (Co-I, WFU Engineering)