Monitoring global land surface drought based on a hybrid evapotranspiration model

Abstract

The latent heat of evapotranspiration (ET) plays an important role in the assessment of drought severity as one sensitive indicator of land drought status. A simple and accurate method of estimating global ET for the monitoring of global land surface droughts from remote sensing data is essential. The objective of this research is to develop a hybrid ET model by introducing empirical coefficients based on a simple linear two-source land ET model, and to then use this model to calculate the Evaporative Drought Index (EDI) based on the actual estimated ET and the potential ET in order to characterize global surface drought conditions. This is done using the Global Energy and Water Cycle Experiment (GEWEX) Surface Radiation Budget (SRB) products, AVHRR-NDVI products from the Global Inventory Modeling and Mapping Studies (GIMMS) group, and National Centers for Environmental Prediction Reanalysis-2 (NCEP-2) datasets. We randomly divided 22 flux towers into two groups and performed a series of cross-validations using ground measurements collected from the corresponding flux towers. The validation results from the second group of flux towers using the data from the first group for calibration show that the daily bias varies from −6.72W/m2 to 12.95W/m2 and the average monthly bias is −1.73W/m2. Similarly, the validation results of the first group of flux towers using data from second group for calibration show that the daily bias varies from −12.91W/m2 to 10.26W/m2 and the average monthly bias is −3.59W/m2. To evaluate the reliability of the hybrid ET model on a global scale, we compared the estimated ET from the GEWEX, AVHRR-GIMMS-NDVI, and NECP-2 datasets with the latent heat flux from the Global Soil Wetness Project-2 (GSWP-2) datasets. We found both of them to be in good agreement, which further supports the validity of our model’s global ET estimation. Significantly, the patterns of monthly EDI anomalies have a good spatial and temporal correlation with the Palmer Drought Severity Index (PDSI) anomalies from January 1984 to December 2002, which indicates that the method can be used to accurately monitor long-term global land surface drought.