A High-Resolution (1km) Future Drought Prediction Dataset via Causal Deep Learning (SPEI, 2023–2030)

This dataset provides high-resolution medium- to long-term drought predictions spanning from 2023 to 2030, generated utilizing advanced causal deep learning methodologies. It features a fine spatial resolution of 1 kilometer and includes the Standardized Precipitation Evapotranspiration Index (SPEI) across multiple timescales: 1-month (SPEI1), 3-month (SPEI3), 6-month (SPEI6), and 12-month (SPEI12). These diverse timescales effectively capture a wide spectrum of drought conditions, from short-term meteorological anomalies to long-term hydrological deficits, making the data essential for climate resilience planning, agricultural risk assessment, and proactive water resource management. By employing a causal deep learning framework rather than relying solely on traditional statistical forecasting, this approach explicitly models the underlying cause-and-effect mechanisms driving climate extremes, ensuring that future projections remain physically robust and reliable under shifting environmental conditions. The development of this dataset was financially supported by the Wuhan Natural Science Foundation Exploration Project (Chenguang Project) under grant number 2024040801020279. Researchers and practitioners utilizing this drought prediction data for academic or applied studies are kindly requested to properly cite the dataset and acknowledge this funding source in their related publications and research outputs.本数据集提供2023年至2030年的高分辨率中长期干旱预测结果，该预测基于先进的因果深度学习方法生成。该数据集空间分辨率精细至1千米，涵盖多时间尺度下的标准化降水蒸散指数（Standardized Precipitation Evapotranspiration Index，SPEI），包括1个月尺度（SPEI1）、3个月尺度（SPEI3）、6个月尺度（SPEI6）及12个月尺度（SPEI12）。这些多样化的时间尺度能够有效捕捉从短期气象异常到长期水文缺水的全范围干旱状况，使得该数据集成为气候韧性规划、农业风险评估及主动水资源管理的核心支撑数据。本研究采用因果深度学习框架而非仅依赖传统统计预测方法，显式建模驱动气候极端事件的内在因果机制，确保未来预测结果在环境变化条件下仍具备物理合理性与可靠稳定性。 本数据集的研发得到了武汉市自然科学基金探索项目（晨光计划）资助，项目编号为2024040801020279。恳请使用该干旱预测数据开展学术或应用研究的科研人员与从业者，在相关出版物及研究成果中正确引用本数据集，并注明本资助来源。