Objective monitoring and prediction system for drought classification over the continental United States

This proposal responds to the 2014 solicitation for CPO’s Modeling, Analysis, Prediction and Projection (MAPP) program. The proposal specifically responds to area 1) of the solicitation, Research to Advance Understanding, Monitoring, and Prediction of Drought, and within that priority, Focus area 2), advancing the development of a national drought monitoring and prediction system. This is a joint proposal from NOAA’s Climate Prediction Center (CPC) and the University of Washington; with parallel submissions from both entities.

The U.S Drought Monitor (USDM) and the seasonal and monthly Drought Outlooks (USDO) are used by water resources managers, government and state agencies in their planning efforts which are intended to reduce the severity of the impacts of drought to U.S. society. The USDM classifies drought into categories D0-D4 (moderate to severe). The USDO predicts the development of drought in terms of changes in the same categories.
One major gap in the drought information system that underlies the USDM and USDO is that the suite of NOAA climate prediction models does not explicitly use the D0-D4 categories. Instead, NOAA’s drought monitoring and prediction capabilities are based on the North American Land Data Assimilation Systems (NLDAS), which use model-predicted soil moisture and runoff (typically expressed as percentiles relative to historical runs of four component land surface models) in lieu of observations (which are not available over domains as large as the continental U.S.). While monitoring systems based on sources like NLDAS are able to detect droughts, they are challenged by classification of drought into the D0 to D4 categories in part due to uncertainties among multiple drought indicators, models and assimilation systems. While the USDM authors use both subjective and objective information in the USDM (the former to incorporate “on the ground” observations into their drought identifications), they have difficulty in the use of NLDAS-based objective information because it is formulated in a fundamentally different manner than the USDM classifications. For the same reason, there is at present no well formulated method of incorporating objective forecasts of drought categories into the USDO.

We propose to explore an objective scheme for drawing boundaries between the D0-D4 classes used by the USDM. Our approach will be based on multiple drought indices that will be derived from NLDAS outputs, from which we will form an ensemble mean index. We will then remap the mean index to a uniform distribution by using the climatology of the ensemble (percentiles) averages. To assess uncertainties in the classifications, we plan to use a concurrence measure among indices. The classification scheme we propose to develop will provide information about drought severity, as well as the representativeness of the ensemble mean index. Forecasts of the indices will be derived using the National Multi-Model Ensemble (NMME) system to force the four land surface models (LSMs) that operate within NLDAS (NMME_LSM). The initial conditions for each LSM will be taken from NLDAS, which drives the LSMs with observed forcings. We expect that the objective drought classification nowcasts and forecasts that we propose to develop based on the NMME_LSM will fill a major gap in the drought information system widely used within the U.S., and will provide drought forecasters with a mechanism to issue reproducible drought forecasts.