A reliable climate model based on Gaussian process

Yu Pan; Xin Zhou; Fupeng Sun; Yin Ni; Xingming Gao; Huafei Sun

doi:10.1117/12.2692833

23 August 2023 A reliable climate model based on Gaussian process

Yu Pan, Xin Zhou, Fupeng Sun, Yin Ni, Xingming Gao, Huafei Sun

Proceedings Volume 12784, Second International Conference on Applied Statistics, Computational Mathematics, and Software Engineering (ASCMSE 2023); 127840E (2023) https://doi.org/10.1117/12.2692833
Event: 2023 2nd International Conference on Applied Statistics, Computational Mathematics and Software Engineering (ASCMSE 2023), 2023, Kaifeng, China

Abstract

Climate data is of great essence in various fields of researches. As direct observations are limited by sensing resolution and accuracy, a considerable amount of climate data is generated by theoretical computation. However, current mainstream climate models suffer from either limited accuracy or discrete outputs. In this paper, we propose a method of generating climate data by utilizing the properties of the Gaussian Process and carefully designing its kernel structure. Different from previous work, our model considers both temporal and spatial dimensions as inputs, hence being able to generate continuous predictions on certain areas or time intervals and achieve high accuracy. We conduct experiments on the famous ERA5 dataset and compare our method with classic model. Our model outperforms the classic one in terms of error scores in both spatial and temporal tasks, and comparison of uncertainty growth shows our advantage.

Citation Download Citation

Yu Pan, Xin Zhou, Fupeng Sun, Yin Ni, Xingming Gao, and Huafei Sun "A reliable climate model based on Gaussian process", Proc. SPIE 12784, Second International Conference on Applied Statistics, Computational Mathematics, and Software Engineering (ASCMSE 2023), 127840E (23 August 2023); https://doi.org/10.1117/12.2692833

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available