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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gong, Junchao, Han, Tao, Chen, Kang, Bai, Lei
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917786434928640
author Gong, Junchao
Han, Tao
Chen, Kang
Bai, Lei
author_facet Gong, Junchao
Han, Tao
Chen, Kang
Bai, Lei
contents Numerical Weather Prediction (NWP) system is an infrastructure that exerts considerable impacts on modern society.Traditional NWP system, however, resolves it by solving complex partial differential equations with a huge computing cluster, resulting in tons of carbon emission. Exploring efficient and eco-friendly solutions for NWP attracts interest from Artificial Intelligence (AI) and earth science communities. To narrow the performance gap between the AI-based methods and physic predictor, this work proposes a new transformer-based NWP framework, termed as WeatherFormer, to model the complex spatio-temporal atmosphere dynamics and empowering the capability of data-driven NWP. WeatherFormer innovatively introduces the space-time factorized transformer blocks to decrease the parameters and memory consumption, in which Position-aware Adaptive Fourier Neural Operator (PAFNO) is proposed for location sensible token mixing. Besides, two data augmentation strategies are utilized to boost the performance and decrease training consumption. Extensive experiments on WeatherBench dataset show WeatherFormer achieves superior performance over existing deep learning methods and further approaches the most advanced physical model.
format Preprint
id arxiv_https___arxiv_org_abs_2409_16321
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle WeatherFormer: Empowering Global Numerical Weather Forecasting with Space-Time Transformer
Gong, Junchao
Han, Tao
Chen, Kang
Bai, Lei
Artificial Intelligence
Machine Learning
Atmospheric and Oceanic Physics
Numerical Weather Prediction (NWP) system is an infrastructure that exerts considerable impacts on modern society.Traditional NWP system, however, resolves it by solving complex partial differential equations with a huge computing cluster, resulting in tons of carbon emission. Exploring efficient and eco-friendly solutions for NWP attracts interest from Artificial Intelligence (AI) and earth science communities. To narrow the performance gap between the AI-based methods and physic predictor, this work proposes a new transformer-based NWP framework, termed as WeatherFormer, to model the complex spatio-temporal atmosphere dynamics and empowering the capability of data-driven NWP. WeatherFormer innovatively introduces the space-time factorized transformer blocks to decrease the parameters and memory consumption, in which Position-aware Adaptive Fourier Neural Operator (PAFNO) is proposed for location sensible token mixing. Besides, two data augmentation strategies are utilized to boost the performance and decrease training consumption. Extensive experiments on WeatherBench dataset show WeatherFormer achieves superior performance over existing deep learning methods and further approaches the most advanced physical model.
title WeatherFormer: Empowering Global Numerical Weather Forecasting with Space-Time Transformer
topic Artificial Intelligence
Machine Learning
Atmospheric and Oceanic Physics
url https://arxiv.org/abs/2409.16321