Combined Drought Index Using High-Resolution Hydrological Models and Explainable Artificial Intelligence Techniques in Türkiye

Eyyup Ensar Başakın, Paul C. Stoy*, Mehmet Cüneyd Demirel, Mutlu Ozdogan, Jason A. Otkin

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

We developed a combined drought index to better monitor agricultural drought events. To develop the index, different combinations of the temperature condition index, precipitation condition index, vegetation condition index, soil moisture condition index, gross primary productivity, and normalized difference water index were used to obtain a single drought severity index. To obtain more effective results, a mesoscale hydrologic model was used to obtain soil moisture values. The SHapley Additive exPlanations (SHAP) algorithm was used to calculate the weights for the combined index. To provide input to the SHAP model, crop yield was predicted using a machine learning model, with the training set yielding a correlation coefficient (R) of 0.8, while the test set values were calculated to be 0.68. The representativeness of the new index in drought situations was compared with established indices, including the Standardized Precipitation-Evapotranspiration Index (SPEI) and the Self-Calibrated Palmer Drought Severity Index (scPDSI). The index showed the highest correlation with an R-value of 0.82, followed by the SPEI with 0.7 and scPDSI with 0.48. This study contributes a different perspective for effective detection of agricultural drought events. The integration of an increased volume of data from remote sensing systems with technological advances could facilitate the development of significantly more efficient agricultural drought monitoring systems.

Original languageEnglish
Article number3799
JournalRemote Sensing
Volume16
Issue number20
DOIs
Publication statusPublished - Oct 2024

Bibliographical note

Publisher Copyright:
© 2024 by the authors.

Keywords

  • agricultural drought
  • combined drought indices
  • crop yield
  • SHAP
  • XGBoost

Fingerprint

Dive into the research topics of 'Combined Drought Index Using High-Resolution Hydrological Models and Explainable Artificial Intelligence Techniques in Türkiye'. Together they form a unique fingerprint.

Cite this