EGU24-10886, updated on 08 Mar 2024
https://doi.org/10.5194/egusphere-egu24-10886
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Explainable machine learning revealing the mechanism behind drought events in northern Italy: the case of the 2022 drought

Chenli Xue1,2, Aurora Ghirardelli2, Jianping Chen1, and Paolo Tarolli2
Chenli Xue et al.
  • 1China University of Geosciences, School of Earth Sciences and Resources, Beijing, China (xueclucas@163.com)
  • 2University of Padova, Department of Land, Environment, Agriculture and Forestry, Legnaro (Padova), Italy (paolo.tarolli@unipd.it)

Drought is a complex natural hazard involving multiple variables that, depending on the measured parameters, can be categorized into meteorological, hydrological or agricultural drought. Among them, agricultural drought, which refers to soil moisture deficits that fail to meet crop growth, has been attracting more attention for severely threatening food security worldwide. In the context of climate change and the increased occurrence of drought events, it is crucial to monitor drought drivers and progression to plan the subsequent efforts in drought prevention, adaptation, and migration. However, the comprehensive knowledge of agriculture drought still needs to be clarified. Previous works often focused on precipitation or evapotranspiration and failed to capture other potential drivers of drought. This study proposes a novel framework to comprehensively monitor agricultural drought with ensemble machine learning by constructing an integrated agriculture drought index with high temporal-spatial resolution. In addition, the Shapley Additive Explanation (SHAP) explainable model was applied to reveal the driving mechanism behind the drought event that occurred in northern Italy in the summer of 2022. Results indicate that the proposed explainable ensemble machine learning model could effectively reflect the evolution of agricultural drought with spatially continuous maps on a weekly scale. The SHAP analysis demonstrated that the severe agricultural drought in the summer of 2022 was closely related to meteorological indicators, namely precipitation and land surface temperature, crucial in controlling soil moisture. Moreover, the new findings also revealed that soil textures could significantly affect agricultural drought. By combining explainable ensemble machine learning and various earth-observation data involving meteorology, soil, geomorphology, and vegetation conditions, the study constructed an integrated index to monitor and assess agricultural drought in northern Italy. The proposed research framework could effectively contribute to improving the methodology in agricultural drought research, potentially bringing more instructive insights for drought prevention and mitigation.

How to cite: Xue, C., Ghirardelli, A., Chen, J., and Tarolli, P.: Explainable machine learning revealing the mechanism behind drought events in northern Italy: the case of the 2022 drought, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10886, https://doi.org/10.5194/egusphere-egu24-10886, 2024.