Oral Presentation AMOS Annual Meeting and International Conference on Tropical Meteorology and Oceanography

Flash drought in CMIP5 models (#135)

David Hoffmann 1 2 , Ailie Gallant 1 2 , Julie Arblaster 1 2 3
  1. Monash University, Clayton, VIC, Australia
  2. ARC Centre of Excellence for Climate Extremes, Melbourne
  3. National Center for Atmospheric Research, Boulder

Drought is a stochastic natural hazard and an inherent part of climate variability. The evolution of drought occurs over various time scales from sub-seasonal to multi-decadal and is subdivided into different drought types. The newly emerging term ‘Flash drought’ describes the rapid onset of drought on sub-seasonal time scale. It is of particular interest for the agriculture as it can rapidly deplete soil moisture for crop growth within several weeks and poses great potential for economic damage.

To better understand the driving processes of flash drought we evaluate the importance of evapotranspiration (ET) by comparing two different drought indices estimating this hazard using meteorological and hydrological parameters from CMIP5. The Standardised Precipitation Index (SPI) has become popular drought indices due to its simplicity. Criticism arises in terms of loose connection to ground conditions but has been found to be sufficient for identifying soil moisture drought as well as drought indices including ET in a previous study. The Evaporative Demand Drought Index (EDDI) is ET-based and calculated from daily aggregated evaporative demand. Its disadvantage is the multitude of sparse input parameters, which makes it challenging to apply it on a global domain. Soil moisture is a key proxy for flash drought as it has to be depleted quickly in order to be classified as a flash drought. However, it is also difficult to measure it over a wide area and existing sparsely stationary measurements are not representative for bigger domains and statistical purposes.

Consequently, we have obtained outputs from CMIP5 models to conduct this study. Here, soil moisture estimates and indices calculated from model output are correlated over the available time period to date. Results will show the capabilities of both drought indices to identify flash drought in CMIP5 models and weather a simply precipitation only drought index is sufficient.