An interpretable eXtreme Gradient Boosting–SHAP framework for spatial estimation of hydrological drought across multiple time scales
Abstract
Hydrological drought poses significant challenges to water resource management, hydropower production, and ecosystem sustainability, even in traditionally water-abundant regions such as Sweden. Monthly streamflow data spanning from 1961 to 2025 were used to calculate the streamflow drought index (SDI) at multiple time scales (SDI-1, SDI-3, SDI-6, and SDI-12) at a target station (Gävunda krv) in Sweden using eXtreme Gradient Boosting (XGB) models driven by SDI values from seven neighbouring streamflow stations (Grötsjön, Höljes krv, Långhag, Ljusne strömmar krv, Nybro, Rolfsta, and Skallböle krv) as spatial predictors. The XGB models were trained on approximately 85% of the data and tested on the remaining 15%, achieving strong agreement between observed and estimated SDI and consistently low errors in both training and testing phases. Model performance generally improved with increasing accumulation period during the training phase, indicating that longer SDI time scales exhibit stronger spatial coherence among neighbouring stations. Scatter plot and time series analyses confirmed the models' ability to accurately reproduce the temporal dynamics of major drought events, including the extreme 2018 northern European drought, with no evidence of overfitting. SHapley Additive exPlanations (SHAP) analysis was applied to interpret the trained XGB models and, through beeswarm summary plots, to identify which neighbouring stations most strongly influenced drought estimations at Gävunda krv across different time scales and hydrological conditions. The analysis showed that drought at Långhag station was consistently the dominant predictor of drought at Gävunda krv across all four time scales, whereas the second-ranked contributor shifted from Nybro (SDI-1) to Höljes krv (SDI-3 and SDI-6) and then Grötsjön (SDI-12). Rolfsta and Skallböle krv ranked among the least influential stations, most consistently at the longer accumulation periods (SDI-6 and SDI-12). The results demonstrate that the proposed XGB–SHAP framework offers a powerful, accurate, and interpretable tool for spatial hydrological drought estimation in Nordic hydroclimatic settings, with practical implications for drought early warning systems and water resource management in Sweden.
Document Type: Article
Cited as: Mohammadi, B. An interpretable eXtreme Gradient Boosting–SHAP framework for spatial estimation of hydrological drought across multiple time scales. Sustainable Earth Resources Communications, 2026, 2(2): 80-91. https://doi.org/10.46690/serc.2026.02.03
DOI:
https://doi.org/10.46690/serc.2026.02.03Keywords:
Streamflow drought index, water resources, eXtreme Gradient Boosting, SHapley Additive exPlanations, hydrological drought, SwedenReferences
AghaKouchak, A., Huning, L. S., Sadegh, M., et al. Toward impact-based monitoring of drought and its cascading hazards. Nature Reviews Earth & Environment, 2023, 4(8): 582-595.
Ahmadi, B., Moradkhani, H. Revisiting hydrological drought propagation and recovery considering water quantity and quality. Hydrological Processes, 2019, 33(10): 1492-1505.
Akbarian, M., Saghafian, B., Golian, S. Monthly streamflow forecasting by machine learning methods using dynamic weather prediction model outputs over Iran. Journal of Hydrology, 2023, 620: 129480.
Chen, T., Guestrin, C. Xgboost: A scalable tree boosting system. Presented at Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, San Francisco, California, USA, 2016.
Deulkar, A. M., Dixit, P. R., Londhe, S. N., et al. Comparative assessment of artificial neural networks (ANNs), long short term memory network (LSTM) and hydrologic engineering Centre-Hydrologic modelling system (HEC-HMS) for runoff modelling. Water Resources Management, 2025, 39(5): 2049-2068.
Dikshit, A., Pradhan, B. Interpretable and explainable AI (XAI) model for spatial drought prediction. Science of the Total Environment, 2021, 801, 149797.
Dikshit, A., Pradhan, B., Huete, A. An improved SPEI drought forecasting approach using the long short-term memory neural network. Journal of environmental management, 2021, 283, 111979.
Gebrechorkos, S. H., Peng, J., Dyer, E., et al. Global high-resolution drought indices for 1981–2022. Earth System Science Data, 2023, 15: 5449–5466.
Gebrechorkos, S. H., Sheffield, J., Vicente-Serrano, S. M., et al. Warming accelerates global drought severity. Nature, 2025, 642(8068): 628-635.
Gevaert, A. I., Veldkamp, T. I., Ward, P. J. The effect of climate type on timescales of drought propagation in an ensemble of global hydrological models. Hydrology and Earth System Sciences, 2018, 22(9): 4649-4665.
Gul, E., Staiou, E., Safari, M. J. S., et al. Enhancing meteorological drought modeling accuracy using hybrid boost regression models: A case study from the Aegean region, Türkiye. Sustainability, 2023, 15(15): 11568.
Hagenlocher, M., Meza, I., Anderson, C. C., et al. Drought vulnerability and risk assessments: state of the art, persistent gaps, and research agenda. Environmental Research Letters, 2019, 14(8): 083002.
Kumar, V., Kedam, N., Sharma, K. V., et al. Advanced machine learning techniques to improve hydrological prediction: A comparative analysis of streamflow prediction models. Water, 2023, 15(14): 2572.
Li, M., Yao, Y., Feng, Z., et al. Hydrological drought prediction and its influencing features analysis based on a machine learning model. Natural Hazards and Earth System Sciences, 2025, 25(11): 4299-4316.
Liu, C., Yang, C., Yang, Q., et al. Spatiotemporal drought analysis by the standardized precipitation index (SPI) and standardized precipitation evapotranspiration index (SPEI) in Sichuan Province, China. Scientific reports, 2021, 11(1): 1280.
Liu, Y., Just, A. (2023). SHAPforxgboost: SHAP Plots for 'XGBoost'. R package version 0.2.0. Available online: https://cran.r-project.org/package=SHAPforxgboost (accessed on 25 June 2026).
Lundberg, S. M., Erion, G., Chen, H., et al. From local explanations to global understanding with explainable AI for trees. Nature machine intelligence, 2020, 2(1): 56-67.
Lundberg, S. M., Lee, S. I. A unified approach to interpreting model predictions. Presented at 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA, 2017.
Ma, L., Huang, Q., Huang, S., et al. Propagation dynamics and causes of hydrological drought in response to meteorological drought at seasonal timescales. Hydrology Research, 2022, 53(1): 193-205.
Mayer, M. (2023). shapviz: SHAP visualizations. R package version 0.10.3. Available online: https://CRAN.R-project.org/package=shapviz (accessed on 25 June 2026).
McKee, T. B., Doesken, N. J., Kleist, J. The relationship of drought frequency and duration to time scales. In Proceedings of the 8th Conference on Applied Climatology, Anaheim, California, USA, 17-22 January 1993.
Mishra, A. K., Singh, V. P. A review of drought concepts. Journal of hydrology, 2010, 391(1-2): 202-216.
Mliyeh, M. M., Aqnouy, M., Laaraj, M., et al. Assessing drought dynamics in a semi-arid basin: a multi-index approach using hydrological and remote-sensing indicators. Environmental Sciences Europe, 2025, 37(1): 1-19.
Mohammadi, B., Abdallah, M., Oucheikh, R., et al. Enhancing streamflow drought prediction: integrating wavelet decomposition with deep learning and quantile regression neural network models. Earth Science Informatics, 2025, 18(2): 232.
Mokhtar, A., Jalali, M., He, H., et al. Estimation of SPEI meteorological drought using machine learning algorithms. IEEE Access, 2021, 9: 65503-65523.
Nalbantis, I. Evaluation of a hydrological drought index. European water, 2008, 23(24): 67-77.
Nalbantis, I., Tsakiris, G. Assessment of hydrological drought revisited. Water resources management, 2009, 23(5): 881-897.
Naumann, G., Alfieri, L., Wyser, K., et al. Global changes in drought conditions under different levels of warming. Geophysical Research Letters, 2018, 45(7): 3285-3296.
Ni, L., Wang, D., Wu, J., et al. Streamflow forecasting using extreme gradient boosting model coupled with Gaussian mixture model. Journal of Hydrology, 2020, 586: 124901.
Niazkar, M., Menapace, A., Brentan, B., et al. Applications of XGBoost in water resources engineering: A systematic literature review (Dec 2018–May 2023). Environmental Modelling & Software, 2024, 174: 105971.
Niu, W. J., Feng, Z. K. Evaluating the performances of several artificial intelligence methods in forecasting daily streamflow time series for sustainable water resources management. Sustainable Cities and Society, 2021, 64, 102562.
Osman, A. I. A., Ahmed, A. N., Chow, M. F., et al. Extreme gradient boosting (Xgboost) model to predict the groundwater levels in Selangor Malaysia. Ain Shams Engineering Journal, 2021, 12(2): 1545-1556.
Pokhrel, Y., Felfelani, F., Satoh, Y., et al. Global terrestrial water storage and drought severity under climate change. Nature Climate Change, 2021, 11(3): 226-233.
Slater, L., Blougouras, G., Deng, L., et al. Challenges and opportunities of ML and explainable AI in large-sample hydrology. Philosophical transactions. Series A, Mathematical, physical, and engineering sciences, 2025, 383(2302): 20240287.
Tareke, K. A., Awoke, A. G. Hydrological drought analysis using streamflow drought index (SDI) in Ethiopia. Advances in Meteorology, 2022, 2022(1): 7067951.
Van Loon, A. F. Hydrological drought explained. Wiley Interdisciplinary Reviews: Water, 2015, 2(4): 359-392.
Vicente-Serrano, S. M., Beguería, S., López-Moreno, J. I. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. Journal of climate, 2010, 23(7): 1696-1718.
Vicente-Serrano, S. M., Quiring, S. M., Peña-Gallardo, M., et al. A review of environmental droughts: Increased risk under global warming?. Earth-Science Reviews, 2020, 201, 102953.
Wilhite, D. A. Drought as a natural hazard: concepts and definitions. In Droughts, Routledge, pp. 3-18, 2016.
Xu, T., Liang, F. Machine learning for hydrologic sciences: An introductory overview. WIREs Water, 2021, 8(5): e1533.
Zhang, H., Loaiciga, H. A., Sauter, T. A novel fusion-based methodology for drought forecasting. Remote Sensing, 2024, 16(5): 828.
Zhang, R., Chen, Z. Y., Xu, L. J., et al. Meteorological drought forecasting based on a statistical model with machine learning techniques in Shaanxi province, China. Science of the Total Environment, 2019, 665: 338-346.
Zhang, X., Hao, Z., Singh, V. P., et al. Drought propagation under global warming: Characteristics, approaches, processes, and controlling factors. Science of the Total Environment, 2022, 838: 156021.
Zounemat-Kermani, M., Kheimi, M. Explainable Artificial Intelligence in Hydrology: A Review. Water Resources Management, 2026, 40(3): 106.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Babak Mohammadi
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.