Explainable machine learning for the prediction and assessment of complex drought impacts.

Drought is a common and costly natural disaster with broad social, economic, and environmental impacts. Machine learning (ML) has been widely applied in scientific research because of its outstanding performance on predictive tasks. However, for practical applications like disaster monitoring and assessment, the cost of the models failure, especially false negative predictions, might significantly affect society. Stakeholders are not satisfied with or do not "trust" the predictions from a so-called black box. The explainability of ML models becomes progressively crucial in studying drought and its impacts. In this work, we propose an explainable ML pipeline using the XGBoost model and SHAP model based on a comprehensive database of drought impacts in the U.S. The XGBoost models significantly outperformed the baseline models in predicting the occurrence of multi-dimensional drought impacts derived from the text-based Drought Impact Reporter, attaining an average F2 score of 0.883 at the national level and 0.942 at the state level. The interpretation of the models at the state scale indicates that the Standardized Precipitation Index (SPI) and Standardized Temperature Index (STI) contribute significantly to predicting multi-dimensional drought impacts. The time scalar, importance, and relationships of the SPI and STI vary depending on the types of drought impacts and locations. The patterns between the SPI variables and drought impacts indicated by the SHAP values reveal an expected relationship in which negative SPI values positively contribute to complex drought impacts. The explainability based on the SPI variables improves the trustworthiness of the XGBoost models. Overall, this study reveals promising results in accurately predicting complex drought impacts and rendering the relationships between the impacts and indicators more interpretable. This study also reveals the potential of utilizing explainable ML for the general social good to help stakeholders better understand the multi-dimensional drought impacts at the regional level and motivate appropriate responses.

Developing a remotely sensed drought monitoring indicator for Morocco.

Drought is one of the most serious climatic and natural disasters inflicting serious impacts on the socio-economy of Morocco, which is characterized both by low-average annual rainfall and high irregularity in the spatial distribution and timing of precipitation across the country. This work aims to develop a comprehensive and integrated method for drought monitoring based on remote sensing techniques. The main input parameters are derived monthly from satellite data at the national scale and are then combined to generate a composite drought index presenting different severity classes of drought. The input parameters are: Standardized Precipitation Index calculated from satellite based precipitation data since 1981 (CHIRPS), anomalies in the day-night difference of Land Surface Temperature as a proxy for soil moisture, Normalized Difference Vegetation Index anomalies from MODIS data and Evapotranspiration anomalies from surface energy balance modeling. All of these satellite-based indices are being used to monitor vegetation condition, rainfall and land surface temperature. The weighted combination of these input parameters into one composite indicator takes into account the importance of the rainfall based parameter (SPI). The composite drought index maps were generated during the growing seasons going back to 2003. These maps have been compared to both the historical, in situ precipitation data across Morocco and with the historical yield data across different provinces with information being available since 2000. The maps are disseminated monthly to several main stakeholders groups including the Ministry of Agriculture and Department of Water in Morocco.