Saved in:
Bibliographic Details
Main Authors: Ephrati, Sagy, Woodfield, James
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.21664
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917031454965760
author Ephrati, Sagy
Woodfield, James
author_facet Ephrati, Sagy
Woodfield, James
contents This paper demonstrates the feasibility of trajectory learning for ensemble forecasts by employing the continuous ranked probability score (CRPS) as a loss function. Using the two-scale Lorenz '96 system as a case study, we develop and train both additive and multiplicative stochastic parametrizations to generate ensemble predictions. Results indicate that CRPS-based trajectory learning produces parametrizations that are both accurate and sharp. The resulting parametrizations are straightforward to calibrate and outperform derivative-fitting-based parametrizations in short-term forecasts. This approach is particularly promising for data assimilation applications due to its accuracy over short lead times.
format Preprint
id arxiv_https___arxiv_org_abs_2508_21664
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Trajectory learning for ensemble forecasts via the continuous ranked probability score: a Lorenz '96 case study
Ephrati, Sagy
Woodfield, James
Numerical Analysis
Machine Learning
65M75, 37M05
This paper demonstrates the feasibility of trajectory learning for ensemble forecasts by employing the continuous ranked probability score (CRPS) as a loss function. Using the two-scale Lorenz '96 system as a case study, we develop and train both additive and multiplicative stochastic parametrizations to generate ensemble predictions. Results indicate that CRPS-based trajectory learning produces parametrizations that are both accurate and sharp. The resulting parametrizations are straightforward to calibrate and outperform derivative-fitting-based parametrizations in short-term forecasts. This approach is particularly promising for data assimilation applications due to its accuracy over short lead times.
title Trajectory learning for ensemble forecasts via the continuous ranked probability score: a Lorenz '96 case study
topic Numerical Analysis
Machine Learning
65M75, 37M05
url https://arxiv.org/abs/2508.21664