发表机构
National Institute of Advanced Studies; Indian Institute of Science; School of Development, Azim Premji University（国家高级研究院; 印度科学学院; 阿齐姆·普雷姆吉大学发展学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究以印度四个城市为对象，对比10种方法测试ConvLSTM的一日提前IMDAA降雨场预测性能，发现ConvLSTM未始终更优，架构选择需严格基准测试。
AI 中文摘要
卷积长短期记忆网络（ConvLSTMs）广泛应用于降水预测，但其性能证据大多来自密集、高频的雷达序列。本研究测试卷积循环是否能提升小型每日再分析网格的一日提前降雨场预测效果。分析了1998-2020年6-9月印度季风数据同化与分析（IMDAA）场，涉及班加罗尔、德里、加尔各答和孟买四个城市。比较了10种朴素、统计、树基及神经方法，输入仅大气数据与降雨历史加大气数据两种类型。评估指标涵盖完整场、域平均降雨、空间异常及高降雨日表现。结果显示，ConvLSTM未始终优于更简单的替代方法：FC-LSTM在班加罗尔、加尔各答、孟买的域平均降雨误差数值最低，而持续性方法在德里表现最佳；ConvLSTM仅在孟买的空间异常误差数值最低，孟买的降雨场表现出更强的短期空间连续性，且降雨历史输入对三种神经架构均有提升，不过ConvLSTM与FC-LSTM的差异较小。神经模型在高降雨日低估降雨强度且预测的阈值突破次数过少，而持续性方法在所有城市的检测性能最高。事后分析表明，所选模型对最新输入日最敏感，孟买的近期滞后敏感性更广泛。这些发现说明，仅网格输入不足以支撑选用ConvLSTM，架构选择需在平均、空间及高降雨性能上进行严格基准测试。
英文摘要
Convolutional long short-term memory networks (ConvLSTMs) are widely used for precipitation forecasting, but most evidence for their performance comes from dense, high-frequency radar sequences. This study tests whether convolutional recurrence improves one-day-ahead rainfall-field prediction on small daily reanalysis grids. Indian Monsoon Data Assimilation and Analysis (IMDAA) fields for June-September 1998-2020 were analysed for Bengaluru, Delhi, Kolkata and Mumbai. Ten naive, statistical, tree-based and neural approaches were compared using atmospheric-only and rainfall-history-plus-atmospheric inputs. Performance was assessed for complete fields, domain-mean rainfall, spatial anomalies and high-rainfall days. ConvLSTM did not consistently outperform simpler alternatives. FC-LSTM produced the numerically lowest domain-mean rainfall error in Bengaluru, Kolkata and Mumbai, whereas persistence performed best in Delhi. ConvLSTM produced the numerically lowest spatial-anomaly error only in Mumbai, where rainfall fields showed greater short-term spatial continuity and rainfall-history inputs improved all three neural architectures. The difference between ConvLSTM and FC-LSTM was nevertheless small. Neural models underestimated rainfall magnitude and predicted too few threshold exceedances on high-rainfall days, while persistence achieved the highest detection performance in every city. Post-hoc analyses showed that the selected models were most sensitive to the latest input day, with broader recent-lag sensitivity in Mumbai. These findings show that gridded inputs alone do not justify ConvLSTM and that architecture choice should follow strong benchmarking across average, spatial and high-rainfall performance.
