Location: Hydrology and Remote Sensing Laboratory
Title: An intercomparison of drought indicators based on thermal remote sensing and NLDAS-2 simulations with U.S. drought monitor classifications Authors
|Hain, C -|
|Otkin, J -|
|Zhan, X -|
|Mo, K -|
|Svoboda, M -|
|Wardlow, B -|
|Pimstein, A -|
Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: March 2, 2012
Publication Date: N/A
Interpretive Summary: Agricultural drought can appear quickly, driven by prolonged heatwaves and high winds in combination with normal or lower than average rainfall. Such "flash drought" events are not well captured by standard drought indicators based solely on precipitation deficits. The Evaporative Stress Index (ESI), derived from remote sensing estimates of evapotranspiration (ET), measures anomalies in water use, and therefore is well suited to identify flash drought occurrences. This paper compares spatial and temporal patterns in the ESI with drought classifications in the US Drought Monitor for 2000-2011, and with soil moisture anomalies predicted by a standard set of land-surface models run by the National Oceanic and Atmospheric Administration in support of operational drought monitoring. The set of indicators agreed well, and in combination provide independent evidence of emerging drought signals. The ESI in several cases provided advanced detection of incipient drought, suggesting it may be a valuable component of a Drought Early Warning System.
Technical Abstract: Comparison of multiple hydrologic indicators, derived from independent data sources and modeling approaches, may improve confidence in signals of emerging drought – particularly during periods of rapid onset. This paper compares the Evaporative Stress Index (ESI) - a diagnostic fast-response indicator describing evapotranspiration (ET) deficits derived within a thermal remote sensing energy balance framework - with prognostic estimates of soil moisture (SM), ET and runoff anomalies generated with the North American Land Data Assimilation System (NLDAS). Widely used empirical indices based on thermal remote sensing (Vegetation Health Index; VHI) and precipitation percentiles (Standardized Precipitation Index; SPI) were also included to assess relative performance. Spatial and temporal correlations computed between indices over the contiguous U.S. were compared with historical drought classifications recorded in the U.S. Drought Monitor (USDM). Based on correlation results, improved forms for the ESI were identified, incorporating a Penman-Monteith reference ET scaling flux and implementing a temporal smoothing algorithm at the pixel level. Of all indices evaluated, anomalies in the NLDAS ensemble-average SM provided the highest correlations with USDM drought classes, while the ESI yielded the best performance of the remote sensing indices. The VHI provided reasonable correlations, except under conditions of energy-limited vegetation growth during the cold season and at high latitudes. Change indices computed from ESI and SM time series agree well, and in combination offer a good indicator of change in drought severity class in the USDM – often preceding USDM class deterioration by several weeks. Results suggest that a merged ESI-SM change indicator may provide valuable early warning of rapidly evolving “flash drought” conditions.