Simulating abundances of stable water isotopologues, that is, molecules differing in their isotopic composition, within climate models allows for comparisons with proxy data and, thus, for testing hypotheses about past climate and validating climate models under varying climatic conditions. However, many models are run without explicitly simulating water isotopologues. We investigate the possibility of replacing the explicit physics-based simulation of oxygen isotopic composition in precipitation using machine learning methods. These methods estimate isotopic composition at each time step for given fields of surface temperature and precipitation amount. We implement convolutional neural networks (CNNs) based on the successful UNet architecture and test whether a spherical network architecture outperforms the naive approach of treating Earth’s latitude-longitude grid as a flat image. Conducting a case study on …
%0 Journal Article
%1 wider2023towards
%A Wider, Jonathan
%A Kruse, Jakob
%A Weitzel, Nils
%A Bühler, Janica C
%A Köthe, Ullrich
%A Rehfeld, Kira
%D 2023
%I Cambridge University Press
%J Environmental Data Science
%K xack
%P e35
%T Towards learned emulation of interannual water isotopologue variations in General Circulation Models
%V 2
%X Simulating abundances of stable water isotopologues, that is, molecules differing in their isotopic composition, within climate models allows for comparisons with proxy data and, thus, for testing hypotheses about past climate and validating climate models under varying climatic conditions. However, many models are run without explicitly simulating water isotopologues. We investigate the possibility of replacing the explicit physics-based simulation of oxygen isotopic composition in precipitation using machine learning methods. These methods estimate isotopic composition at each time step for given fields of surface temperature and precipitation amount. We implement convolutional neural networks (CNNs) based on the successful UNet architecture and test whether a spherical network architecture outperforms the naive approach of treating Earth’s latitude-longitude grid as a flat image. Conducting a case study on …
@article{wider2023towards,
abstract = {Simulating abundances of stable water isotopologues, that is, molecules differing in their isotopic composition, within climate models allows for comparisons with proxy data and, thus, for testing hypotheses about past climate and validating climate models under varying climatic conditions. However, many models are run without explicitly simulating water isotopologues. We investigate the possibility of replacing the explicit physics-based simulation of oxygen isotopic composition in precipitation using machine learning methods. These methods estimate isotopic composition at each time step for given fields of surface temperature and precipitation amount. We implement convolutional neural networks (CNNs) based on the successful UNet architecture and test whether a spherical network architecture outperforms the naive approach of treating Earth’s latitude-longitude grid as a flat image. Conducting a case study on …},
added-at = {2025-02-20T12:14:38.000+0100},
author = {Wider, Jonathan and Kruse, Jakob and Weitzel, Nils and Bühler, Janica C and Köthe, Ullrich and Rehfeld, Kira},
biburl = {https://puma.scadsai.uni-leipzig.de/bibtex/23d2d20bd104c232b8119b9c34854470c/scadsfct},
citation = {Environmental Data Science 2, e35, 2023},
interhash = {56326feb12a29bb73051991586080cd0},
intrahash = {3d2d20bd104c232b8119b9c34854470c},
journal = {Environmental Data Science},
keywords = {xack},
pages = {e35},
publisher = {Cambridge University Press},
timestamp = {2025-07-29T10:29:54.000+0200},
title = {Towards learned emulation of interannual water isotopologue variations in General Circulation Models},
volume = 2,
year = 2023
}
