首页|期刊导航|内蒙古农业大学学报(自然科学版)|基于OCO-3与ERA5数据的双分支大气CO2浓度反演网络

基于OCO-3与ERA5数据的双分支大气CO2浓度反演网络OA

Dual-branch Feature-enhanced Network for Atmospheric CO2 Concentration Retrieval Based on OCO-3 and ERA5 Data

中文摘要英文摘要

针对卫星遥感数据易受轨道、云层和气溶胶等因素干扰,导致在复杂气溶胶场景下XCO2反演精度低等问题,提出了一种用于大气二氧化碳浓度反演的双分支特征增强网络CCFM-DBNet(CO₂ Concentration Feature-Merging Dual-Branch Net-work).首先,在光谱分支中将全局分组坐标注意力模块(GGCA)嵌入2D-CNN主干,强化对关键CO2吸收波段的响应,抑制气溶胶和水汽干扰;其次,在气象分支中设计多尺度残差注意力模块(MSRA),结合多尺度卷积、残差连接与通道注意力机制,实现气象垂直剖面特征的高效建模.最后,将两个分支的输出特征向量在特征维度进行拼接,实现物理信息的互补与XCO₂浓度的端到端预测.实验结果表明,本方法的决定系数R2较于基线提高了0.65%,RMSE与MAE分别下降了7.83%、9.02%,验证了该方法在提高XCO2反演精度上的有效性.

Atmospheric CO₂ retrieval from satellite observations is often hindered by orbital variability,clouds,and aerosols,result-ing in degraded accuracy under complex aerosol conditions.To address these challenges,this study introduced a CO₂ Concentration Feature-Merging Dual-Branch Network(CCFM-DBNet),a dual-branch feature-enhanced architecture designed for precise atmo-spheric CO₂ inversion.The spectral branch integrated a Global Grouped Coordinate Attention(GGCA)module into a 2D convolu-tional backbone to strengthen responses to key CO₂ absorption bands while suppressing aerosol and water-vapor interference.The me-teorological branch employed a Multi-Scale Residual Attention(MSRA)module that combines multi-scale convolutions,residual connections,and channel attention to effectively capture vertical meteorological profiles.Feature vectors from both branches were fused along the feature dimension,enabling complementary physical information and achieving end-to-end XCO₂ prediction.Experi-mental results demonstrated that the CCFM-DBNet increased the coefficient of determination(R²)by 0.65%and reduced RMSE and MAE by 7.83%and 9.02%,respectively,compared with the baseline model.These improvements confirmed the method's effec-tiveness in enhancing XCO₂ retrieval accuracy under complex atmospheric conditions.

邓天民;赵西崇;彭扬洋

重庆交通大学交通运输学院,重庆 400074重庆交通大学交通运输学院,重庆 400074重庆交通大学信息科学与工程学院,重庆 400074

资源环境

XCO2反演双分支特征增强网络OCO-3ERA5全局分组注意力

XCO₂ retrievalDual-branch feature-enhanced networkOCO-3ERA5Global grouped coordinate attention

《内蒙古农业大学学报(自然科学版)》 2026 (5)

89-100,12

重庆市技术创新与应用发展专项重点项目(CSTB2022TIAD-KPX0118)国家重点研发计划项目(2022YFC3800502)重庆市技术创新与应用发展专项重点项目(CSTB2022TIAD-KPX0113)

10.16853/j.cnki.1009-3575.2026.05.010

评论