- 1Aalto University, School of Engineering, Espoo, Finland (pihla.seppala@aalto.fi)
- 2Finnish Environment Institute (Syke), Helsinki, Finland
- 3Gulbali Institute, Charles Sturt University, New South Wales, Australia
- 4Pellervo economic research PTT, Helsinki, Finland
Droughts are among the most devastating natural hazards, driving conflict, migration, and socioeconomic changes worldwide. Compound droughts – where meteorological, hydrological, and soil moisture (agricultural) drought co-occur – have greater ecological and socio-economic impacts than individual drought types. However, existing knowledge about global-scale compound droughts is limited, as research is mostly focusing on smaller areas and the propagation of meteorological drought to other types, typically considering just two different drought types.
Here, we use an ensemble of 9 model outputs from the ISIMIP3a experiment (H08, WaterGAP 2.2e, Miroc-Integ-Land, forced with 20CRv3-ERA5, 20CRv3-W5E5, GWSP3-W5E5 reanalysis datasets) with daily outputs of precipitation, soil moisture and discharge to compute empirical drought indices. Focusing on severe drought events with index value (intensity) below -1.5, we analyse event characteristics as well as probability and duration of compounding for 1961–2020.
We found significant variability in duration and probability across different hydrological regions (hydrobelts) and drought indices, with results sensitive to the drought type used as basis for the comparison. The largest differences in duration and probability between hydrobelts occurred with soil moisture drought as the basis of analysis, while meteorological drought as the base showed the smallest differences. Compound drought durations were longer in the Southern Hemisphere, particularly near the equator. Soil moisture and hydrological droughts had longer median durations than meteorological droughts and therefore higher probabilities of compounding. The high probabilities were concentrated in northern latitudes and Asia for soil moisture drought and were globally more evenly distributed for hydrological. Analysing the influence of ENSO revealed longer durations and higher probabilities globally during El Niño compared to La Niña months. The uncertainty in the probability of compounding shows large spatial variation and was found to depend on the model, climate forcing, drainage basin size and the hydrobelt.
Our results may help prepare regional or national drought management plans by providing insights into the spatial characteristics and probability of compound droughts. However, until the uncertainty in global modelling is addressed, and new methods or simulations are provided, the benefit is limited.
How to cite: Seppälä, P., Kallio, M., Ahopelto, L., Fallon, A., Kinnunen, P., Heino, M., and Kummu, M.: Global models show strong spatial variation in compound drought occurrence, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18679, https://doi.org/10.5194/egusphere-egu25-18679, 2025.
Comments on the supplementary material
AC: Author Comment | CC: Community Comment | Report abuse