Abstract:Numerical Weather Prediction (NWP) models are extensively utilized worldwide and have played a pivotal role in weather forecasting. Precipitation is subject to various intricate factors, rendering it one of the most challenging factors to predict. Additional post-processing steps are required to reduce biases and achieve reliable predictions for precipitation-related decision-making. In this work, we propose a transformer-enhanced spatiotemporal neural network called TransLSTMUNet for short- and medium-range precipitation post-processing. Firstly, TransLSTMUNet employs convolutional operators to extract localized meteorological features. Secondly, it capitalizes on transformer architecture to enrich these extracted features with a broader, global perspective of spatial information. Thirdly, TransLSTMUNet leverages ConvLSTM to further enhance the features with temporal information. Furthermore, to address challenges posed by imbalanced distribution of precipitation intensity, we design a novel loss function called quantile weighted mean squared error (QWMSE). QWMSE simultaneously considers both normal and intense precipitation during the model' s training phase. In the experiments, the THORPEX Interactive Grand Global Ensemble (TIGGE) dataset provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) is used as input for post-processing the precipitation forecasts. The experiments show that the precipitation forecasts post-processed by TransLSTMUNet exhibit the best overall performance compare with the eight post-processing baselines. It significantly improves the forecast performance of TIGGE forecasts. Specifically, TransLSTMUNet enhances the accuracy (ACC) metric by 12.14 % and increases the threat scores (TS) for 24-hour accumulated precipitation of 0.1 mm, 10.0 mm, 25.0 mm, and 50.0 mm by 8.30 %, 9.77 %, 31.60 %, and 51.25 % respectively. By effectively integrating the strengths of convolutional and transformer methodologies, the proposed TransLSTMUNet model offers a novel approach for post-processing precipitation forecasting. This model design has the potential to inspire various other research avenues within the hydrological domain and beyond.

Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data

Transformer Neural Processes -- Kernel Regression

Spatio-Temporal Transformer Network for Weather Forecasting

Translation Equivariant Transformer Neural Processes

NAST: Non-Autoregressive Spatial-Temporal Transformer for Time Series Forecasting

A Hierarchical Spatial Transformer for Massive Point Samples in Continuous Space

NTDformer: A Multi-Scale Forecasting Model for Non-Stationary Multilevel Time Series

Transformer-enhanced spatiotemporal neural network for post-processing of precipitation forecasts

WeatherFormer: Empowering Global Numerical Weather Forecasting with Space-Time Transformer

A Novel Multiscale Transformer Network Framework for Natural Gas Consumption Forecasting

Leveraging 2D Information for Long-term Time Series Forecasting with Vanilla Transformers

Enhancing Time Series Forecasting: A Hierarchical Transformer with Probabilistic Decomposition Representation

GRAformer: A Gated Residual Attention Transformer for Multivariate Time Series Forecasting

STC-ViT: Spatio Temporal Continuous Vision Transformer for Weather Forecasting

Multi-Task Spatial-Temporal Transformer for Multi-Variable Meteorological Forecasting

Efficient Large-Scale Traffic Forecasting with Transformers: A Spatial Data Management Perspective

Interpretable Transformer Neural Network Prediction of Diverse Environmental Time Series Using Weather Forecasts

TENT: Tensorized Encoder Transformer for Temperature Forecasting

In-Context In-Context Learning with Transformer Neural Processes

GTHP: a novel graph transformer Hawkes process for spatiotemporal event prediction

Analyzing and Exploring Training Recipes for Large-Scale Transformer-Based Weather Prediction