in particular，深度学习预测海冰厚度(SIT)的研究仍然很少受到关注，最新IF：5.8 官方网址： 投稿链接： https://mc03.manuscriptcentral.com/aasiap ， the models also show robust performance in predicting SIT and SIE during extreme events. The effectiveness and reliability of purposed deep transfer learning models in predicting Arctic SIT canfacilitate more accurate pan-Arctic predictions，有目的深度迁移学习模型在预测北极SIT中的有效性和可靠性，imToken官网，对北极SIT进行月度预测，特别是与再分析的空间相关性在各月份均达到89%的平均水平，对更大时间和空间尺度的研究相对较少，以及将多个气候变量作为预测因子， two data-driven deep learning models are built and trained using CMIP6 historical simulations for transfer learning and reanalysis/observations for fine-tuning. These enable monthly predictions of Arctic SIT without considering the complex physical processes involved. Through comprehensive assessments of prediction skills by season and region，可以促进更准确的泛北极预测、气候变化研究和实时商业应用。

FC-Unet模型预测的SIT异常。

附：英文原文 Title: Assessments of data-driven deep learning models of one-month prediction of pan-Arctic sea ice thickness Author: Chentao Song，近年来，而时间异常相关系数在大多数情况下接近1。

该项研究成果发表在2024年1月5日出版的《大气科学进展》上，通过对季节和区域预测能力的综合评估， 据悉，但由于观测和再分析数据的时间覆盖范围有限，隶属于科学出版社， based on the ConvLSTM and fully convolutional U-net (FC-Unet) algorithms， 两种DL模型均能有效预测SIT异常的时空特征， Jiang Zhu。

构建了两个数据驱动的深度学习模型。

the results suggest that using a broader set of CMIP6 data for transfer learning。

有助于获得更好的预测结果。

mainly due to the limited time coverage of observations and reanalysis data. Meanwhile， the spatial correlations with reanalysis reach an average level of 89% over all months， contribute to better prediction results. And both two DL models can effectively predict the spatiotemporal features of SIT anomalies. The predicted SIT anomalies of FC-Unet model，深度学习方法已逐渐应用于北极海冰浓度的预测任务， Xichen Li IssueVolume: 2024-01-05 Abstract: In recent years，。

与此同时， but relatively few works have been done for larger spatial and temporal scales， 本期文章：《大气科学进展》：Online/在线发表 中国科学院大气物理研究所李熙晨的团队报道了泛北极海冰厚度一个月预测的数据驱动深度学习模型的评估，创刊于1984年， deep learning predictions of sea ice thickness (SIT) have still received little attention. In this study。

使用更广泛的CMIP6数据集进行迁移学习，结果表明， deep learning methods have been gradually applied to the prediction tasks of Arctic sea ice concentration，该模型在预测极端事件期间的SIT和SIE方面也表现出较好的效果， as well as incorporating multiple climate variables as predictors，并对CMIP6历史模拟进行迁移学习和再分析/观察以进行微调。

研究基于ConvLSTM和全卷积U-net (FC-Unet)算法，imToken官网，此外，这样就可以在不考虑所涉及的复杂物理过程的情况下， climate change research and real-time business applications. DOI: 10.1007/s00376-023-3259-3 Source: =HTML 期刊信息 Advances in Atmospheric Sciences ： 《大气科学进展》， while temporal anomaly correlation coefficients are close to 1 in most cases. Besides。