Guardado en:
Detalles Bibliográficos
Autores principales: Prasad, Ayush, Harder, Paula, Yang, Qidong, Sattegeri, Prasanna, Szwarcman, Daniela, Watson, Campbell, Rolnick, David
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2407.12517
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913434437681152
author Prasad, Ayush
Harder, Paula
Yang, Qidong
Sattegeri, Prasanna
Szwarcman, Daniela
Watson, Campbell
Rolnick, David
author_facet Prasad, Ayush
Harder, Paula
Yang, Qidong
Sattegeri, Prasanna
Szwarcman, Daniela
Watson, Campbell
Rolnick, David
contents Climate downscaling, the process of generating high-resolution climate data from low-resolution simulations, is essential for understanding and adapting to climate change at regional and local scales. Deep learning approaches have proven useful in tackling this problem. However, existing studies usually focus on training models for one specific task, location and variable, which are therefore limited in their generalizability and transferability. In this paper, we evaluate the efficacy of training deep learning downscaling models on multiple diverse climate datasets to learn more robust and transferable representations. We evaluate the effectiveness of architectures zero-shot transferability using CNNs, Fourier Neural Operators (FNOs), and vision Transformers (ViTs). We assess the spatial, variable, and product transferability of downscaling models experimentally, to understand the generalizability of these different architecture types.
format Preprint
id arxiv_https___arxiv_org_abs_2407_12517
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Evaluating the transferability potential of deep learning models for climate downscaling
Prasad, Ayush
Harder, Paula
Yang, Qidong
Sattegeri, Prasanna
Szwarcman, Daniela
Watson, Campbell
Rolnick, David
Machine Learning
Climate downscaling, the process of generating high-resolution climate data from low-resolution simulations, is essential for understanding and adapting to climate change at regional and local scales. Deep learning approaches have proven useful in tackling this problem. However, existing studies usually focus on training models for one specific task, location and variable, which are therefore limited in their generalizability and transferability. In this paper, we evaluate the efficacy of training deep learning downscaling models on multiple diverse climate datasets to learn more robust and transferable representations. We evaluate the effectiveness of architectures zero-shot transferability using CNNs, Fourier Neural Operators (FNOs), and vision Transformers (ViTs). We assess the spatial, variable, and product transferability of downscaling models experimentally, to understand the generalizability of these different architecture types.
title Evaluating the transferability potential of deep learning models for climate downscaling
topic Machine Learning
url https://arxiv.org/abs/2407.12517