中国海海气通量格点数据产品

本数据集由中科院海洋大科学研究中心共建单位中科院南海海洋研究所王鑫研究员和张荣望助理研究员牵头研制，主要基于卫星遥感和再分析数据，以中国近海浮标、海上平台和通量塔等观测数据为训练数据，采用卷积神经网络深度学习和改进后的COARE 3.5算法等技术研发，有效克服了中等大气比湿条件下的潜热通量低估和海表冷皮效应造成的潜热通量误差等突出问题。与国际主流海气通量产品对比，该产品在中国近海具有显著的区域优势和性能指标优势。

This dataset was led in development by Researcher Wang Xin and Assistant Researcher Zhang Rongwang from the South China Sea Institute of Oceanology, Chinese Academy of Sciences (CAS SCSIO), a co-construction unit of the CAS Ocean Science Research Center. It was primarily constructed using satellite remote sensing and reanalysis data, with in-situ observational datasets including buoys, offshore platforms, and flux towers in the China Seas as training data. Technologies including convolutional neural network (CNN)-based deep learning and the improved COARE 3.5 algorithm were employed during its development. This work effectively resolves key issues such as the underestimation of latent heat flux under moderate atmospheric specific humidity conditions and latent heat flux errors induced by the sea surface skin effect. Compared with internationally mainstream air-sea flux products, this product exhibits notable regional and performance advantages in the China Seas.

搜集汇总

背景与挑战

背景概述

该数据集是中国近海的海气通量格点数据产品，覆盖1980年至2020年，具有0.25°×0.25°的高空间分辨率和月平均时间分辨率。它基于卫星遥感和再分析数据，采用卷积神经网络和改进算法研制，有效解决了潜热通量误差问题，并在中国近海区域表现出优于国际主流产品的性能优势。