标题: A process model-guided transfer learning framework for mapping global gross primary production
作者: Guan, XB (Guan, Xiaobin); Li, YY (Li, Yuyu); Chen, JM (Chen, Jing M.); Ma, YM (Ma, Yongming); Shen, HF (Shen, Huanfeng)
来源出版物: AGRICULTURAL AND FOREST METEOROLOGY 卷: 372 文献号: 110678 DOI: 10.1016/j.agrformet.2025.110678 Published Date: 2025 SEP 15
摘要: Accurate estimation of gross primary production (GPP) is critical for global carbon cycle research. The current mainstream approaches, i.e., process-based models, are limited by insufficient parameter representation and high computing costs, and the burgeoning machine learning methods still suffer from inadequate training samples and poor transferability when applied to GPP estimation. Therefore, in this paper, we propose a process modelguided transfer learning approach for global GPP estimation, taking the low-resolution (0.5 degrees) estimates from the Biosphere-atmosphere Exchange Process Simulator (BEPS) model as the source domain and eddy covariance (EC) data as the target domain. After joint constraint from the two domains, relatively high-accuracy GPP estimation at a resolution of 0.05 degrees can be achieved after downscaled pre-training and fine-tuning based on EC tower data. The results indicate that the proposed framework can significantly improve the accuracy of GPP estimation, compared to a direct machine learning method based on only EC tower data (Delta R2 = 0.05, Delta RMSE = -1.02 g C m-2month-1) and the original BEPS estimates (Delta R2 = 0.05, Delta RMSE = -14.14 g C m-2month-1). The results of the temporal validation and regional cross-validation also show consistent results, indicating the superior spatio-temporal expandability of the proposed method. Furthermore, when compared with other global GPP products, the new global GPP product built in this study can effectively correct the underestimation/ overestimation in high/low GPP regions in the existing machine learning-based GPP products (e.g., FLUXCOM GPP), especially in the area near the equator, and shows higher consistency with solar-induced chlorophyll fluorescence (SIF)-based and model-based GPP products. In addition to the new global GPP product, the results of this study also prove the reliability of combining a process-based model and a machine learning model in GPP estimation.
作者关键词: GPP; BEPS model; Machine learning; Downscaling; Transfer learning
KeyWords Plus: NET PRIMARY PRODUCTIVITY; BOREAL ECOSYSTEM; AMAZON FOREST; CARBON; MODIS; FLUXES; GPP; PHOTOSYNTHESIS; VEGETATION; ATMOSPHERE
地址: [Guan, Xiaobin; Li, Yuyu; Ma, Yongming; Shen, Huanfeng] Wuhan Univ, Sch Resource & Environm Sci, Wuhan 430079, Peoples R China.
[Chen, Jing M.] Univ Toronto, Dept Geog & Planning, Toronto, ON M5S 3G3, Canada.
[Chen, Jing M.] Fujian Normal Univ, Sch Geog Sci, Fuzhou 350117, Peoples R China.
[Shen, Huanfeng] Minist Educ, Key Lab Geog Informat Syst, Wuhan 430079, Peoples R China.
[Shen, Huanfeng] Minist Nat Resources, Key Lab Digital Mapping & Land Informat Applicat, Wuhan 430079, Peoples R China.
通讯作者地址: Shen, HF (通讯作者)，Wuhan Univ, Sch Resource & Environm Sci, Wuhan 430079, Peoples R China.
Shen, HF (通讯作者)，Minist Educ, Key Lab Geog Informat Syst, Wuhan 430079, Peoples R China.
Shen, HF (通讯作者)，Minist Nat Resources, Key Lab Digital Mapping & Land Informat Applicat, Wuhan 430079, Peoples R China.
电子邮件地址: [email protected]
影响因子：5.7
