Transfer Learning in Junction With a Light Use Efficiency Model for Estimating Grassland Gross Primary Production

Transfer Learning in Junction With a Light Use Efficiency Model for Estimating Grassland Gross Primary Production

It is significant to simulate grassland gross primary production (GPP) to understand the terrestrial carbon budget over Inner Mongolia (IMG), China. Nevertheless, there is not sufficient <italic>in situ</italic> GPP data over this region. In this study, we proposed a novel model-based tr...

Full description

Saved in:
Bibliographic Details
Main Authors: Ruiyang Yu, Yunjun Yao, Qingxin Tang, Xueyi Zhang, Changliang Shao, Joshua B. Fisher, Jiquan Chen, Xiaotong Zhang, Yufu Li, Jia Xu, Lu Liu, Zijing Xie, Jing Ning, Jiahui Fan, Luna Zhang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Grassland gross primary production (GPP)
light use efficiency (LUE)
remote sensing
transfer learning (TL)
Online Access:https://ieeexplore.ieee.org/document/10923714/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is significant to simulate grassland gross primary production (GPP) to understand the terrestrial carbon budget over Inner Mongolia (IMG), China. Nevertheless, there is not sufficient <italic>in situ</italic> GPP data over this region. In this study, we proposed a novel model-based transfer learning (MTL) approach with generative adversarial networks-long short-term memory (GAN-LSTM) and light use efficiency (LUE) models to derive grassland GPP over IMG, China. We first used 25 grassland eddy covariance sites over the conterminous United States to establish the GAN-LSTM model and then fine-tuned it with six sites over IMG to estimate water constraints that were embedded into the LUE model to predict GPP. We then compared it with instance-based transfer learning and nontransfer learning approaches. Against the six IMG EC sites, the GPP estimates of MTL-LUE outperformed the other approaches with a lower root-mean-square error median (1.35 g C m<sup>&#x2212;2</sup> d<sup>&#x2212;1</sup>) and a higher Kling-Gupta efficiency of 0.54. An innovation of this approach is that MTL-LUE mitigates the effect of limited training samples on the machine learning-based LUE hybrid model for GPP estimates over IMG.
ISSN:1939-1404
2151-1535

Similar Items