Meteorological Drought Prediction in the Loess Plateau Based on Three Machine Learning Models

Yingqiang Li; Yongfan Guo; Qingchi Yi; Yanli Pei; Qunqun Li; Wei Wei

doi:10.6911/WSRJ.202604_12(4).0001

Authors

Yingqiang Li
Yongfan Guo
Qingchi Yi
Yanli Pei
Qunqun Li
Wei Wei

DOI:

https://doi.org/10.6911/WSRJ.202604_12(4).0001

Keywords:

Loess Plateau; drought prediction; SVM; RF; CNN-LSTM; SPI; SPEI.

Abstract

Drought, a severe natural hazard, exerts profound impacts on both ecological systems and agricultural productivity. Consequently, advancing drought prediction methodologies is paramount for optimizing regional water resource management and mitigating associated risks. This research centers on the Loess Plateau, leveraging monthly precipitation (P) and potential evapotranspiration (PET) datasets spanning 1960 to 2022 to compute the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) at temporal resolutions of 1-, 3-, 6-, and 12-months. The predictive efficacy of three computational models—Convolutional Long Short-Term Memory (CNN-LSTM), Random Forest (RF), and Support Vector Machine (SVM)—was systematically evaluated across semi-humid, semi-arid, and arid climatic zones. Empirical results elucidate three key findings: (1) CNN-LSTM outperformed its counterparts across all regions and time scales; RF demonstrated superior accuracy relative to SVM at 3- and 6-month lags, whereas SVM yielded stronger predictive power at the 12-month scale. Overall, the performance hierarchy prioritizes CNN-LSTM > RF > SVM for short-to-medium term forecasts, while CNN-LSTM > SVM > RF holds for long-term projections. (2) Model predictive accuracy achieved higher magnitudes when employing SPEI as the predictor compared to SPI within semi-arid and semi-humid domains. Conversely, in arid regions at extended time scales, SPI-12 outperformed SPEI-12, indicating that precipitation-dominated trends exhibit greater relevance for long-term drought forecasting in water-scarce areas. (3) Incorporating SPEI into the predictive framework conferred significantly enhanced performance in semi-arid and arid zones relative to semi-humid regions, further underscoring SPEI’s utility in water-limited landscapes. Collectively, this study elucidates the interrelationships among model architectures, drought indices, and regional climatic characteristics, thereby establishing a scientific basis for developing regional drought early-warning systems and decision-support tools. These outcomes carry practical implications for optimizing agricultural management strategies and regulating water resources.

Downloads

Download data is not yet available.

References

[1] Acharki, S., Singh, S. K., do Couto, E. V., Arjdal, Y., & Elbeltagi, A. (2023). Spatio-temporal distribution and prediction of agricultural and meteorological drought in a Mediterranean coastal watershed via GIS and machine learning. Physics and Chemistry of the Earth, Parts A/B/C, 131, 103425. https://doi.org/10.1016/j.pce.2023.103425.

[2] AhmadEbrahimpour, E., Aminnejad, B., & Khalili, K. (2018). Application of global precipitation dataset for drought monitoring and forecasting over the Lake Urmia basin with the GA-SVR model. International Journal of Water, 12(3), 262. https://doi.org/10.1504/ijw.2018.093674.

[3] Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). Crop evapotranspiration-Guidelines for computing crop water requirements-FAO Irrigation and drainage paper 56. Fao, Rome, 300(9), D05109.

[4] Almikaeel, W., Čubanová, L., & Šoltész, A. (2022). Hydrological Drought Forecasting Using Machine Learning—Gidra River Case Study. Water, 14(3), 387. https://doi.org/10.3390/w14030387.

[5] Alsubih, M., Mallick, J., Talukdar, S., Salam, R., AlQadhi, S., Fattah, Md. A., & Thanh, N. V. (2021). An investigation of the short-term meteorological drought variability over Asir Region of Saudi Arabia. Theoretical and Applied Climatology, 145(1–2), 597–617. https://doi.org/10.1007/s00704-021-03647-4.

[6] Belayneh, A., Adamowski, J., Khalil, B., & Ozga-Zielinski, B. (2014). Long-term SPI drought forecasting in the Awash River Basin in Ethiopia using wavelet neural network and wavelet support vector regression models. Journal of Hydrology, 508, 418–429. https://doi.org/10.1016/j.jhydrol.2013.10.052.

[7] Buttafuoco, G., Caloiero, T., & Coscarelli, R. (2014). Analyses of Drought Events in Calabria (Southern Italy) Using Standardized Precipitation Index. Water Resources Management, 29(2), 557–573. https://doi.org/10.1007/s11269-014-0842-5.

[8] Chiang, F., Mazdiyasni, O., & AghaKouchak, A. (2021). Evidence of anthropogenic impacts on global drought frequency, duration, and intensity. Nature Communications, 12(1). https://doi.org/10.1038/s41467-021-22314-w.

[9] Cortes, Corinna, and Vladimir Vapnik. "Support-vector networks." Machine learning 20.3 (1995): 273-297.

[10] Dixit, S., & Jayakumar, K. V. (2021). A study on copula-based bivariate and trivariate drought assessment in Godavari River basin and the teleconnection of drought with large-scale climate indices. Theoretical and Applied Climatology, 146(3–4), 1335–1353. https://doi.org/10.1007/s00704-021-03792-w.

[11] Dong, T. (2016). Study on Remote Sensing-based Drought Monitoring Method derived from Multi-Source Data. (Doctoral dissertation, Wuhan University).

[12] Feng, P., Wang, B., Liu, D. L., Waters, C., & Yu, Q. (2019). Incorporating machine learning with biophysical model can improve the evaluation of climate extremes impacts on wheat yield in south-eastern Australia. Agricultural and Forest Meteorology, 275, 100–113. https://doi.org/10.1016/j.agrformet.2019.05.018.

[13] Hargreaves, G. H., & Samani, Z. A. (1982). Estimating Potential Evapotranspiration. Journal of the Irrigation and Drainage Division, 108(3), 225–230. https://doi.org/10.1061/jrcea4.0001390.

[14] George H. Hargreaves, & Zohrab A. Samani. (1985). Reference Crop Evapotranspiration from Temperature. Applied Engineering in Agriculture, 1(2), 96–99. https://doi.org/10.13031/2013.26773.

[15] Hao, Z., Hao, F., Singh, V. P., Xia, Y., Ouyang, W., & Shen, X. (2016). A theoretical drought classification method for the multivariate drought index based on distribution properties of standardized drought indices. Advances in Water Resources, 92, 240–247. https://doi.org/10.1016/j.advwatres.2016.04.010.

[16] He, L., Wang, J., Ciais, P., Ballantyne, A., Yu, K., Zhang, W., Xiao, J., Ritter, F., Liu, Z., Wang, X., Li, X., Peng, S., Ma, C., Zhou, C., Li, Z.-L., Xie, Y., & Ye, J.-S. (2023). Non-symmetric responses of leaf onset date to natural warming and cooling in northern ecosystems. PNAS Nexus, 2(9). https://doi.org/10.1093/pnasnexus/pgad308.

[17] He, L., Wang, J., Peñuelas, J., Zohner, C. M., Crowther, T. W., Fu, Y., Zhang, W., Xiao, J., Liu, Z., Wang, X., Li, J.-H., Li, X., Peng, S., Xie, Y., Ye, J.-S., Zhou, C., & Li, Z.-L. (2024). Asymmetric temperature effect on leaf senescence and its control on ecosystem productivity. PNAS Nexus, 3(11). https://doi.org/10.1093/pnasnexus/pgae477.

[18] He, L., Wang, J., Peltier, D. M. P., Ritter, F., Ciais, P., Peñuelas, J., Xiao, J., Crowther, T. W., Li, X., Ye, J.-S., Sasaki, T., Zhou, C., & Li, Z.-L. (2025). Lagged precipitation effects on plant production across terrestrial biomes. Nature Ecology & Evolution, 9(10), 1800–1811. https://doi.org/10.1038/s41559-025-02806-4.

[19] Hou, zhaojun, Wang, B., Zhang, Y., Zhang, J., & Song, J. (2024). Drought Prediction in Jilin Province Based on Deep Learning and Spatio-Temporal Sequence Modeling. https://doi.org/10.2139/ssrn.4737903.

[20] Husak, G. J., Michaelsen, J., & Funk, C. (2006). Use of the gamma distribution to represent monthly rainfall in Africa for drought monitoring applications. International Journal of Climatology, 27(7), 935–944. https://doi.org/10.1002/joc.1441.

[21] Jiao, W., Wang, L., Smith, W. K., Chang, Q., Wang, H., & D’Odorico, P. (2021). Observed increasing water constraint on vegetation growth over the last three decades. Nature Communications, 12(1). https://doi.org/10.1038/s41467-021-24016-9.

[22] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature,521(7553), 436-444..

[23] Li, J., Wang, Z., Wu, X., Xu, C., Guo, S., Chen, X., & Zhang, Z. (2021). Robust Meteorological Drought Prediction Using Antecedent SST Fluctuations and Machine Learning. Water Resources Research, 57(8). Portico. https://doi.org/10.1029/2020wr029413.

[24] Li, M., Zhou, Q., Han, X., & Lv, P. (2024). Prediction of reference crop evapotranspiration based on improved convolutional neural network (CNN) and long short-term memory network (LSTM) models in Northeast China. Journal of Hydrology, 645, 132223. https://doi.org/10.1016/j.jhydrol.2024.132223.

[25] Li, Z., Zheng, F.-L., & Liu, W.-Z. (2012). Spatiotemporal characteristics of reference evapotranspiration during 1961–2009 and its projected changes during 2011–2099 on the Loess Plateau of China. Agricultural and Forest Meteorology, 154–155, 147–155. https://doi.org/10.1016/j.agrformet.2011.10.019.

[26] Liu, W., & Sang, T. (2013). Potential productivity of the Miscanthus energy crop in the Loess Plateau of China under climate change. Environmental Research Letters, 8(4), 044003. https://doi.org/10.1088/1748-9326/8/4/044003.

[27] Liu, Z., Wang, Y., Shao, M., Jia, X., & Li, X. (2016). Spatiotemporal analysis of multiscalar drought characteristics across the Loess Plateau of China. Journal of Hydrology, 534, 281–299. https://doi.org/10.1016/j.jhydrol.2016.01.003.

[28] Manzano, A., Clemente, M. A., Morata, A., Luna, M. Y., Beguería, S., Vicente-Serrano, S. M., & Martín, M. L. (2019). Analysis of the atmospheric circulation pattern effects over SPEI drought index in Spain. Atmospheric Research, 230, 104630. https://doi.org/10.1016/j.atmosres.2019.104630.

[29] McKee, T. B., Doesken, N. J., & Kleist, J. (1993). The relationship of drought frequency and duration to time scales. In Proceedings of the 8th Conference on Applied Climatology (Vol. 17, No. 22, pp. 179-183).

[30] Mishra, A. K., & Singh, V. P. (2010). A review of drought concepts. Journal of Hydrology, 391(1–2), 202–216. https://doi.org/10.1016/j.jhydrol.2010.07.012.

[31] Potop, V., Boroneanţ, C., Možný, M., Štěpánek, P., & Skalák, P. (2013). Observed spatiotemporal characteristics of drought on various time scales over the Czech Republic. Theoretical and Applied Climatology, 115(3–4), 563–581. https://doi.org/10.1007/s00704-013-0908-y.

[32] Schreiner-McGraw, A. P., & Ajami, H. (2021). Delayed response of groundwater to multi-year meteorological droughts in the absence of anthropogenic management. Journal of Hydrology, 603, 126917. https://doi.org/10.1016/j.jhydrol.2021.126917.

[33] Şener, İ. F., & Tuğal, İ. (2025). Optimized CNN-LSTM with hybrid metaheuristic approaches for solar radiation forecasting. Case Studies in Thermal Engineering, 72, 106356. https://doi.org/10.1016/j.csite.2025.106356.

[34] Shahnazi, S., Roushangar, K., & Hashemi, H. (2025). A novel implementation of pre-processing approaches and hybrid kernel-based model for short- and long-term groundwater drought forecasting. Journal of Hydrology, 652, 132667. https://doi.org/10.1016/j.jhydrol.2025.132667.

[35] Sun, L. (2009). Blended drought indices for agricultural drought: risk assessment on the Canadian prairies. https://doi.org/10.22215/etd/2009-08724.

[36] Vicente-Serrano, S. M., Beguería, S., & López-Moreno, J. I. (2010). A Multiscalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index. Journal of Climate, 23(7), 1696–1718. https://doi.org/10.1175/2009jcli2909.1.

[37] Wu, Z., Yin, H., He, H., & Li, Y. (2022). Dynamic-LSTM hybrid models to improve seasonal drought predictions over China. Journal of Hydrology, 615, 128706. https://doi.org/10.1016/j.jhydrol.2022.128706.

[38] Xu, D., Zhang, Q., Ding, Y., & Zhang, D. (2021a). Application of a hybrid ARIMA-LSTM model based on the SPEI for drought forecasting. Environmental Science and Pollution Research, 29(3), 4128–4144. https://doi.org/10.1007/s11356-021-15325-z.

[39] Xu, D., Zhang, Q., Ding, Y., & Huang, H. (2020). Application of a Hybrid ARIMA–SVR Model Based on the SPI for the Forecast of Drought—A Case Study in Henan Province, China. Journal of Applied Meteorology and Climatology, 59(7), 1239–1259. https://doi.org/10.1175/jamc-d-19-0270.1.

[40] Xu, D., Zhang, Q., Ding, Y., & Zhang, D. (2021b). Spatiotemporal Pattern Mining of Drought in the Last 40 Years in China Based on the SPEI and Space–Time Cube. Journal of Applied Meteorology and Climatology, 60(9), 1219–1230. https://doi.org/10.1175/jamc-d-21-0049.1.

[41] Xu, H., Wang, X., Zhao, C., Shan, S., & Guo, J. (2021c). Seasonal and aridity influences on the relationships between drought indices and hydrological variables over China. Weather and Climate Extremes, 34, 100393. https://doi.org/10.1016/j.wace.2021.100393.

[42] Xu, Q., Jia, Q., Li, Y., Zhang, H., & Ren, P. (2025). A self-attention convolutional long and short-term memory network for correcting sea surface wind field forecasts to facilitate sea ice drift prediction. Computers & Geosciences, 205, 105997. https://doi.org/10.1016/j.cageo.2025.105997.

[43] Xu, Y., Zhang, X., Hao, Z., Hao, F., & Li, C. (2021a). Projections of future meteorological droughts in China under CMIP6 from a three‐dimensional perspective. Agricultural Water Management, 252, 106849. https://doi.org/10.1016/j.agwat.2021.106849.

[44] Yuan, X., Zhang, M., Wang, L., & Zhou, T. (2017). Understanding and seasonal forecasting of hydrological drought in the Anthropocene. Hydrology and Earth System Sciences, 21(11), 5477–5492. https://doi.org/10.5194/hess-21-5477-2017.

[45] Zhang, B., Wu, P., Zhao, X., Wang, Y., Wang, J., & Shi, Y. (2012). Drought variation trends in different subregions of the Chinese Loess Plateau over the past four decades. Agricultural Water Management, 115, 167–177. https://doi.org/10.1016/j.agwat.2012.09.004.

[46] Zhang, Q., Miao, C., Gou, J., & Zheng, H. (2023). Spatiotemporal characteristics and forecasting of short-term meteorological drought in China. Journal of Hydrology, 624, 129924. https://doi.org/10.1016/j.jhydrol.2023.129924.

[47] Zhang, X., Duan, Y., Duan, J., Chen, L., Jian, D., Lv, M., Yang, Q., & Ma, Z. (2022). A daily drought index-based regional drought forecasting using the Global Forecast System model outputs over China. Atmospheric Research, 273, 106166. https://doi.org/10.1016/j.atmosres.2022.106166.

[48] Zhang, Y., Ma, Y., Fang, H., & Wang, H. (2025). Investigation on forecast of offshore wind power generation hybrid attention mechanism and bi-directional long short-term memory based on deep learning. Ocean & Coastal Management, 270, 107884. https://doi.org/10.1016/j.ocecoaman.2025.107884.

[49] Zhao, G., Tu, X. J., Wang, T. (2021). Drought prediction based on artificial neural network and support vector machine. Pearl River, 42(4), 1–9. https://doi.org/10.3969/j.issn.1001-9235.2021.04.001.