Search

Drought events have multiple adverse impacts on environment, society, and economy. It is thus crucial to monitor and characterise such events. Here, we compare the ability of selected state-of-the-art long-term reanalysis and remote-sensing products to represent major seasonal and multi-year drought events in the 2000–2020 period globally. We focus on soil moisture (or agroecological) drought and place the results in the context of trends in dry-season soil moisture. We consider surface and root-zone soil moisture from ERA5, the related ERA5-Land, and MERRA-2 reanalysis products, the ESA CCI remote-sensing surface soil moisture products (encompassing an ACTIVE, a PASSIVE and a COMBINED product), as well as its near real-time counterpart produced within C3S. In addition, we use a new root-zone soil moisture dataset derived from the ESA CCI COMBINED product. Except for ESA CCI surface and root-zone soil moisture, the considered products offer opportunities for drought monitoring since they are available in near real-time. We analyse 18 documented drought events within predefined spatial and temporal bounds derived from scientific literature. Based on standardised daily anomalies of surface and root-zone soil moisture, the drought events are characterised by their severity (the time accumulated standardised anomalies), magnitude (the minimum of the standardised anomalies over time), duration, and spatial extent. Product deviations in drought severity and magnitude are then placed in the context of trends in dry-season soil moisture, and potential reasons for diverging global soil moisture trends in the products are further investigated. All investigated products capture the considered drought events. Overall, responses of surface soil moisture tend to be weakest for the ACTIVE remote-sensing products in all metrics, but most pronounced in the drought magnitudes. Also, MERRA‑2 shows lower magnitudes than the other products. Except for the COMBINED products, the remote-sensing products tend to underestimate the spatial extents of larger droughts. Product differences in drought severity and magnitude for single events are consistent with the differences in dry-season soil moisture trends. These trends are globally diverse and partly contradictory between products. ERA5, ERA5-Land and the COMBINED products show larger fractions of drying trends, MERRA-2 and the C3S ACTIVE and PASSIVE products more widespread wetting trends. MERRA-2 surface air temperature shows regionally negative biases in trends compared to a ground observational product, which suggests that this reanalysis product underestimates drought trends. Also, the comparison with trends in selected land-surface characteristics and bioclimatic indicators shows that dry-season soil moisture trends may be affected by retrieval or modelling artifacts in some cases. In the root zone (based on the reanalysis products and the ESA CCI root-zone soil moisture dataset), the droughts are dampened in magnitude and smaller in spatial extent but show a tendency to prolonged durations. Based on the overall observational evidence and the consideration of the respective limitations of the included products, the present analyses suggest a consistent tendency towards drying during the last two decades in some regions, namely in parts of central Europe, in a region north of the Black Sea/Caspian Sea, in southern Africa, and in parts of Australia, Siberia and South America. [ABSTRACT FROM AUTHOR]