Leveraging meteorological legacy effects and machine learning to improve daily gap-free normalized difference vegetation index reconstruction: a critical basis for fine-scale carbon sink capacity monitoring
4 affiliations
- 1Shaanxi Key Laboratory of Qinling Ecological Intelligent Monitoring and Protection, School of Life Science and Technology, Northwestern Polytechnical University, Xi’an, 710012 China
- 2Xi’an Institute for Innovative Earth Environment Research, Xi’an, 710061 Shaanxi Province China
- 3National Observation and Research Station of Regional Ecological Environment Change and Comprehensive Management in the Guanzhong Plain, Xi’an, 710061 China
- 4Earth Systems Research Center, Institute for the Study of Earth, Oceans, and Space, University of New Hampshire, Durham, 03824 NH USA
Abstract
Background: The daily Normalized Difference Vegetation Index (NDVI) is a critical indicator of terrestrial carbon sequestration capacity, essential for accurately quantifying vegetation carbon sink functions and their dynamic responses to climate change. We previously developed a long-term daily NDVI dataset across China, but the over smoothed polynomial fitting method constrained reconstruction accuracy in complex scenarios, hindering precise fine-scale carbon sink monitoring. To address this limitation, we improved the reconstruction framework by integrating meteorological legacy effects and the machine learning algorithm, with a focus on validating its application value for net primary productivity (NPP) estimation.
Results: The reconstructed daily gap-free NDVI (1982–2023) shows strong consistency with original valid NDVI, achieving a national average R2 of 0.9, percentage bias (PB) of − 0.09%, and root mean square error (RMSE) of 0.03. Over 97% of China’s reconstructed NDVI reach high-quality (R2 > 0.8, |PB|<1%, RMSE < 0.08), representing a 13.9% R2 improvement over our previous dataset. The national daily average NDVI exhibits a continuous increasing trend during the study period, primarily driven by Southwest, Central, Southeast, South, and Northeast China. Attribution analysis indicates that meteorological legacy effects serve as the primary drivers for daily NDVI reconstruction across China. Temperature-related factors including average temperatures over the past 10, 30 and 60 days dominate 68.94% of China, while 60-day cumulative precipitation contributes 12.16%. NPP estimation validation indicates that our daily NDVI-derived NPP shows high agreement with in-situ measurements (R2 = 0.82, PB = 5.18%, RMSE = 0.19 kg C m–2 yr–1), significantly outperforming coarse-temporal-resolution NDVI-based products (R2 < 0.28).
Conclusions: This study provides a robust long-term daily gap-free NDVI dataset for China, serving as a critical basis for fine-scale terrestrial carbon sink capacity monitoring. Integrating meteorological legacy effects into NDVI reconstruction enhances the accuracy of vegetation dynamic simulation and improves the quantification of terrestrial carbon cycle processes, particularly for fine-scale NPP estimation. Our findings highlight the superiority of high-temporal-resolution NDVI in capturing fine-scale vegetation phenological dynamics, addressing the limitations of conventional coarse-resolution products and supporting more reliable ecological and climate change research.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s13021-026-00487-5.
Related
- Spatially evaluated statistical reconstruction of Quaternary vegetation with PaleoVeg
- Combined active-passive remote sensing inversion of grassland aboveground biomass on the Qinghai-Tibet plateaus and exploration of its driving mechanisms with optical-radar-meteorological systems
- Soil Enzyme Activities and Microbial Nutrient Limitation Between Restored Vegetation and Primary Forest on the Loess Plateau
- A multimodal deep learning for rice crop health monitoring using vegetation index feature selection
- Fixed-ground multisensor monitoring resolves longitudinal vegetation condition and fuel-state structure in Southern California chaparral and oak scenes
BibTeX
@article{li2026leveraging,
title = {{Leveraging meteorological legacy effects and machine learning to improve daily gap-free normalized difference vegetation index reconstruction: a critical basis for fine-scale carbon sink capacity monitoring}},
author = {Huiwen Li and Yue Cao and Jingfeng Xiao and Fei Yu and Rui Wang and Xinyi Hao and Zuoqiang Yuan},
journal = {Carbon Balance and Management},
year = {2026},
month = jul,
doi = {10.1186/s13021-026-00487-5},
url = {https://doi.org/10.1186/s13021-026-00487-5},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Leveraging meteorological legacy effects and machine learning to improve daily gap-free normalized difference vegetation index reconstruction: a critical basis for fine-scale carbon sink capacity monitoring
AU - Huiwen Li
AU - Yue Cao
AU - Jingfeng Xiao
AU - Fei Yu
AU - Rui Wang
AU - Xinyi Hao
AU - Zuoqiang Yuan
PY - 2026
DA - 2026/07/26
JO - Carbon Balance and Management
DO - 10.1186/s13021-026-00487-5
UR - https://doi.org/10.1186/s13021-026-00487-5
AB - Background: The daily Normalized Difference Vegetation Index (NDVI) is a critical indicator of terrestrial carbon sequestration capacity, essential for accurately quantifying vegetation carbon sink functions and their dynamic responses to climate change. We previously developed a long-term daily NDVI dataset across China, but the over smoothed polynomial fitting method constrained reconstruction accuracy in complex scenarios, hindering precise fine-scale carbon sink monitoring. To address this limitation, we improved the reconstruction framework by integrating meteorological legacy effects and the machine learning algorithm, with a focus on validating its application value for net primary productivity (NPP) estimation. Results: The reconstructed daily gap-free NDVI (1982–2023) shows strong consistency with original valid NDVI, achieving a national average R2 of 0.9, percentage bias (PB) of − 0.09%, and root mean square error (RMSE) of 0.03. Over 97% of China’s reconstructed NDVI reach high-quality (R2 > 0.8, |PB|<1%, RMSE < 0.08), representing a 13.9% R2 improvement over our previous dataset. The national daily average NDVI exhibits a continuous increasing trend during the study period, primarily driven by Southwest, Central, Southeast, South, and Northeast China. Attribution analysis indicates that meteorological legacy effects serve as the primary drivers for daily NDVI reconstruction across China. Temperature-related factors including average temperatures over the past 10, 30 and 60 days dominate 68.94% of China, while 60-day cumulative precipitation contributes 12.16%. NPP estimation validation indicates that our daily NDVI-derived NPP shows high agreement with in-situ measurements (R2 = 0.82, PB = 5.18%, RMSE = 0.19 kg C m–2 yr–1), significantly outperforming coarse-temporal-resolution NDVI-based products (R2 < 0.28). Conclusions: This study provides a robust long-term daily gap-free NDVI dataset for China, serving as a critical basis for fine-scale terrestrial carbon sink capacity monitoring. Integrating meteorological legacy effects into NDVI reconstruction enhances the accuracy of vegetation dynamic simulation and improves the quantification of terrestrial carbon cycle processes, particularly for fine-scale NPP estimation. Our findings highlight the superiority of high-temporal-resolution NDVI in capturing fine-scale vegetation phenological dynamics, addressing the limitations of conventional coarse-resolution products and supporting more reliable ecological and climate change research. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s13021-026-00487-5.
SN - 1750-0680
ER -