Land surface controls on drought termination in Belgium

Droughts are impactful climate extremes with proven dramatic consequences on economy, ecosystems and society. Numerous research has been devoted to exploring land surface controls on meteorological drought onset and evolution. However, the importance of land conditions may be equally important for drought termination, yet the latter remains much less understood. Drought demise is often abrupt, can lead to extreme rainfall and floods, and is generally hard to capture using traditional monthly drought metrics. A better predictability of the end of a drought can not only help better anticipate the duration of droughts, but also significantly improve risk assessment and water resource management during dry extremes.

In this study, we explore the existence of a positive or negative feedback between the decreasing soil moisture and the probability of drought termination. As test cases, multiple droughts in Belgium during the period of 1981–2015 are selected. As a first step, we compose a data set of past droughts based on precipitation and soil moisture from ECMWF reanalysis data and identify the drought termination days. Next, multiple simulations of the drought termination days are executed with the CLASS4GL mixed-layer model framework, in which the influence of changing soil moisture conditions is evaluated. Finally, the sensitivity of drought demise to soil moisture is assessed based on multiple soil moisture–atmosphere coupling metrics and revealed sensitivity relationships. The obtained results highlight the importance of realistic representation of land–atmosphere feedbacks and soil moisture for drought evolution and termination, and could be used to inform drought prediction efforts or pave the way for effective geoengineering solutions designed to mitigate the increasing risk of dry climate extremes in the future.