Abstract:Weather forecasting plays a vital role in today's society, from agriculture and logistics to predicting the output of renewable energies, and preparing for extreme weather events. Deep learning weather forecasting models trained with the next state prediction objective on ERA5 have shown great success compared to numerical global circulation models. However, for a wide range of applications, being able to provide representative samples from the distribution of possible future weather states is critical. In this paper, we propose a methodology to leverage deterministic weather models in the design of probabilistic weather models, leading to improved performance and reduced computing costs. We first introduce \textbf{ArchesWeather}, a transformer-based deterministic model that improves upon Pangu-Weather by removing overrestrictive inductive priors. We then design a probabilistic weather model called \textbf{ArchesWeatherGen} based on flow matching, a modern variant of diffusion models, that is trained to project ArchesWeather's predictions to the distribution of ERA5 weather states. ArchesWeatherGen is a true stochastic emulator of ERA5 and surpasses IFS ENS and NeuralGCM on all WeatherBench headline variables (except for NeuralGCM's geopotential). Our work also aims to democratize the use of deterministic and generative machine learning models in weather forecasting research, with academic computing resources. All models are trained at 1.5° resolution, with a training budget of $\sim$9 V100 days for ArchesWeather and $\sim$45 V100 days for ArchesWeatherGen. For inference, ArchesWeatherGen generates 15-day weather trajectories at a rate of 1 minute per ensemble member on a A100 GPU card. To make our work fully reproducible, our code and models are open source, including the complete pipeline for data preparation, training, and evaluation, at <a class="link-external link-https" href="https://github.com/INRIA/geoarches" rel="external noopener nofollow">this https URL</a> .

A comparative study of stochastic and deep generative models for multisite precipitation synthesis

Evaluation of a Multi-Site Weather Generator in Simulating Precipitation in the Qiantang River Basin, East China

Stochastic Generation of Daily Rainfall Events: A Single-Site Rainfall Model with Copula-based Joint Simulation of Rainfall Characteristics and Classification and Simulation of Rainfall Patterns

A New Framework for Multi-Site Weather Generator: a Two-Stage Model Combining a Parametric Method with a Distribution-Free Shuffle Procedure

Hydrological Response to Climate Change: McGAN for Multi-Site Scenario Weather Series Generation and LSTM for Streamflow Modeling

A Generative Deep Learning Approach to Stochastic Downscaling of Precipitation Forecasts

A Multi-Site Stochastic Weather Generator for High-Frequency Precipitation Using Censored Skew-Symmetric Distribution

A Two-Stage Multisite and Multivariate Weather Generator

Deep learning models for generation of precipitation maps based on numerical weather prediction

Generative Modeling of High-resolution Global Precipitation Forecasts

Assessment of the Effectiveness of a Multi-Site Stochastic Weather Generator on Hydrological Modelling in the Red Deer River Watershed, Canada

ArchesWeather & ArchesWeatherGen: a deterministic and generative model for efficient ML weather forecasting

A Multi-State Weather Generator for Daily Precipitation for the Torne River Basin, Northern Sweden/western Finland

Forecasting Spatio-Temporal Renewable Scenarios: a Deep Generative Approach

Generative Data Assimilation of Sparse Weather Station Observations at Kilometer Scales

Precipitation nowcasting with generative diffusion models

A generative adversarial network approach to (ensemble) weather prediction

Links Between the Spatial Structure of Weather Generator and Hydrological Modeling

Spatiotemporally Coherent Probabilistic Generation of Weather from Climate

Experimental Study on Generative Adversarial Network for Precipitation Nowcasting

High-resolution meteorology with climate change impacts from global climate model data using generative machine learning