Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study evaluates the effectiveness of convolutional long short-term memory networks (ConvLSTMs) for one-day-ahead rainfall-field prediction across four major Indian cities, comparing them against naive, statistical, tree-based, and other neural methods. The analysis, based on Indian Monsoon Data Assimilation and Analysis (IMDAA) fields from 1998 to 2020, reveals that ConvLSTM does not consistently outperform simpler models, with FC-LSTM achieving the lowest domain-mean rainfall error in three cities and persistence performing best in Delhi. These results indicate that the choice of architecture should be guided by comprehensive benchmarking rather than assumptions about the superiority of ConvLSTM based on input types alone.
ConvLSTM's performance in rainfall prediction is underwhelming, often overshadowed by simpler models like FC-LSTM and persistence across major Indian cities.
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.