Identifying early drought signals and propagation time based on land vegetation atmosphere indicators over the Mississippi and California watersheds
Drought typically initiates with precipitation deficiency over an extended time span and gradually manifests its adverse impact on the hydrological processes, agricultural activities, and socioeconomic aspects. The direct and indirect impact of drought on agriculture can be substantial causing economic loss, environmental collapse, and threatening food security and geopolitical stability of a region. Lack of rainfall (meteorological drought) does not immediately transfer into such damage, and the shifting time from meteorological to agricultural drought provides a unique opportunity to anticipate and mitigate the negative impacts. Nevertheless, the complexity of the land-vegetation-atmospheric continuum and its processes poses challenges and uncertainties to capture drought signals on agricultural sections based on meteorological drought indicators. Many efforts on identifying the propagation time have ignored such complexity and evaluated the agricultural drought based on a single source of drought information. In this study, we considered three integrated drought indices to represent the land-vegetation-atmospheric continuum and explored their coherence and lag time with respect to the commonly used precipitation indicator (standardized precipitation index) using wavelet analysis.
The indices included the Vegetation Drought Response Index (VegDRI), Surface Soil Moisture Percentile (SSMP), and Evaporative Demand Drought Index (EDDI). The VegDRI represents drought condition by integrating satellite-based observations of vegetation condition with climatic data, SSMP uses satellite-based observations of soil moisture and shows the probability of wet or dry conditions based on a long historical average, and EDDI represents the atmospheric conditions by combining the land surface temperature, humidity, wind speed, and solar radiation. We evaluated ten years of these indicators over two watersheds in the U.S. with contrasting climate and agronomic practices, including the upper Mississippi region and the California region.
​
We are currently working on this topic and Initial assessments over the Mississippi watershed showed the advantage of using EDDI as the precursor of drought signals in the agricultural section given its high coherency with the bi-weekly SPI and its faster manifestation of drought (21 days on average). We found a 3-day lag time from the appearance of atmospheric stress to soil moisture stress condition. The vegetation was impacted 16 days after the initial signs of drought on the atmospheric condition.