Droughts severity score prediction using explainable machine learning methodologies

Kurzfassung

Climate Change and Crisis drives the climate to more extreme weather events. As an example, air, land surface and canopy surface temperatures are increasing, and this is affecting the industry of agriculture in many and different ways. For example, damages and losses on crops and pastures are shown up or wide spreading along a region, as well as shortages in water resources and limitations on water usage from farmers are suggested or imposed. Eastern Mediterranean, Middle Easte and North Africa (EMMENA) region is one of the most affected regions. United States (US) have developed the Drought Monitor (https://droughtmonitor.unl.edu/, visited on 05/03/2024) which is responsible for the monitoring of drought events in the different counties. US Monitor and their categorization in five different categories (from abnormally dry to exceptional drought) based on their drought score. In this work, Artificial Intelligence (AI) methodologies are applied in an open-access Kaggle dataset (https://www.kaggle.com/datasets/cdminix/us-drought-meteorological-data, visited on 28/02/2024) which combines soil, meteorological and drought scores, collected from the Harmonized World Soil Database, NASA Langley Research Centre POWER Project and the US Drought Monitor, respectively, for drought score prediction. The main objective of this work is to apply explainable Machine Learning (ML) techniques, for drought score predictions and raise the awareness for drought events in the wider EMMENA region.