Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model

TL;DR

This paper introduces a new deep learning model called TrajGRU for precipitation nowcasting, addressing limitations of previous models, and provides a comprehensive benchmark dataset and evaluation protocol to advance research in this field.

Contribution

The paper proposes TrajGRU, a novel recurrent model that learns location-variant motion, and establishes a benchmark with dataset, loss, and evaluation protocol for precipitation nowcasting.

Findings

TrajGRU outperforms ConvLSTM in precipitation nowcasting tasks.

The benchmark dataset enables standardized evaluation of models.

The new training loss improves forecasting accuracy.

Abstract

With the goal of making high-resolution forecasts of regional rainfall, precipitation nowcasting has become an important and fundamental technology underlying various public services ranging from rainstorm warnings to flight safety. Recently, the Convolutional LSTM (ConvLSTM) model has been shown to outperform traditional optical flow based methods for precipitation nowcasting, suggesting that deep learning models have a huge potential for solving the problem. However, the convolutional recurrence structure in ConvLSTM-based models is location-invariant while natural motion and transformation (e.g., rotation) are location-variant in general. Furthermore, since deep-learning-based precipitation nowcasting is a newly emerging area, clear evaluation protocols have not yet been established. To address these problems, we propose both a new model and a benchmark for precipitation nowcasting. Specifically, we go beyond ConvLSTM and propose the Trajectory GRU (TrajGRU) model that can actively learn the location-variant structure for recurrent connections. Besides, we provide a benchmark that includes a real-world large-scale dataset from the Hong Kong Observatory, a new training loss, and a comprehensive evaluation protocol to facilitate future research and gauge the state of the art.