发表机构
KNMI; Centrum Wiskunde & Informatica; Delft University of Technology; Utrecht University（荷兰皇家气象研究所; 数学与计算机科学中心; 代尔夫特理工大学; 乌特勒支大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
EC-EarthFlow利用流匹配模型仿真EC-Earth3气候模拟，以低成本自回归预测日温度场，再现变率、空间格局及长期趋势，并保持长期稳定性。
AI 中文摘要
我们介绍了EC-EarthFlow，一种生成式流匹配模型，用于仿真物理气候模型EC-Earth3的模拟结果。该模型在EC-Earth3（1950-2166年，SSP2-4.5情景）的瞬态模拟数据上训练，根据前几天的温度以及年平均温度来预测第二天的温度场。预测采用自回归方式进行，滚动周期从一个月到延长季节不等。仅利用这一感兴趣的变量，我们就能以远低于物理模型的计算成本，再现EC-Earth3的每日变率、空间格局、年循环和长期趋势。我们证明，EC-EarthFlow在长时间推理期间保持稳定，并且能够学习EC-Earth3中模拟的物理关系。
英文摘要
We introduce EC-EarthFlow, a generative flow matching model that emulates simulations from the physical climate model EC-Earth3. The model is trained on transient simulations from EC-Earth3 (1950-2166, SSP2-4.5) to predict the day ahead temperature field from the previous days temperature as well as annual mean temperature. Predictions are made auto-regressively with rollout periods of between a month and an extended season. Using only this variable of interest, we are able to reproduce the daily variability, spatial patterns, annual cycle and long-term trend from EC-Earth3 at a substantially lower computational cost than the physical model. We demonstrate that EC-EarthFlow is stable for long inference periods, and that it can learn the physical relationships as simulated in EC-Earth3.
