Chen et al. (2025) A global long-term (2002–2022) C-band vegetation optical depth record retrieved after merging AMSR-E, AMSR2 and WindSat

Identification

• Journal: International Journal of Applied Earth Observation and Geoinformation
• Year: 2025
• Date: 2025-11-20
• Authors: Dongbo Chen, Lei Fan, Jian Peng, Gabriëlle De Lannoy, Jean‐Pierre Wigneron, Frédéric Frappart, Shengli Tao, Mengjia Wang, Xiaojun Li, Xiangzhuo Liu, Huan Wang, Qiangqiang Yuan, Xiuzhi Chen, Yao Xiao, Philippe Ciais
• DOI: 10.1016/j.jag.2025.104961

Research Groups

• Chongqing Jinfo Mountain Karst Ecosystem National Observation and Research Station, School of Geographical Sciences, Southwest University, China
• State Key Laboratory of Remote Sensing Science, Beijing Normal University, China
• Chongqing Engineering Research Center for Remote Sensing Big Data Application, School of Geographical Sciences, Southwest University, China
• Department of Remote Sensing, Helmholtz Centre for Environmental Research-UFZ, Germany
• Remote Sensing Centre for Earth System Research, Leipzig University, Germany
• Department of Earth and Environmental Sciences, KU Leuven, Belgium
• ISPA, UMR 1391, INRAE Nouvelle-Aquitaine, Universit´e de Bordeaux, France
• Institute of Ecology, College of Urban and Environmental Sciences, and Key Laboratory for Earth Surface Processes of the Ministry of Education, Peking University, China
• School of Geo-Science and Technology, Zhengzhou University, China
• School of Geodesy and Geomatics, Wuhan University, China
• School of Atmospheric Sciences, School of Geography and Planning, Sun Yat-sen University, China
• Laboratoire des Sciences du Climat et de l’Environnement, LSCE/IPSL, CEA-CNRS-UVSQ, Universit´e Paris Saclay, France

Short Summary

This study developed a global, long-term (2002–2022) C-band Vegetation Optical Depth (C-VOD) dataset by merging observations from AMSR-E, AMSR2, and WindSat sensors using a combined inter-calibration method. The resulting merged C-VOD exhibited substantially improved temporal consistency across sensors, reducing global discrepancies between AMSR-E and AMSR2 from 6.20 % to 0.34 %.

Objective

• To design a method to eliminate systematic bias among the AMSR-E, AMSR2, and WindSat C-band brightness temperature (TB) observations.
• To assess the consistency of the merged C-band TB over the three sensor periods.
• To retrieve a merged C-VOD dataset from the merged TB using the C-MEB model and assess its temporal consistency by examining its changes and relationships with aboveground biomass (AGB), canopy height, normalized difference vegetation index (NDVI), and enhanced vegetation index (EVI) across the three sensor periods.

Study Configuration

• Spatial Scale: Global, with data aggregated to a 0.25° spatial resolution for analysis.
• Temporal Scale: Long-term, spanning 21 years from 2002 to 2022.

Methodology and Data

• Models used: C-band Microwave Emission of the Biosphere (C-MEB) model (based on τ-ω radiative transfer theory) for VOD retrieval. For comparison, the Land Parameter Retrieval Model (LPRM) (Version 6) was used for the VODCA product.
• Data sources:
  • Satellite Brightness Temperature (TB): AMSR-E (Aqua, 2002–2011), AMSR2 (GCOM-W1, 2012–present), and WindSat (Coriolis, 2003–2020) C-band (6.8 GHz to 6.925 GHz) descending-orbit Level 3 (AMSR-E/AMSR2) and Level 1C (WindSat) data.
  • Reanalysis Data: ERA5 (ECMWF) skin temperature and soil temperature (0–7 cm and 28–100 cm) at 0.25° spatial resolution.
  • Land Cover Map: MODIS MCD12Q1 (IGBP classification) at 500 m spatial resolution, resampled to 0.25°.
  • Forest Disturbance Data: Global Forest Change (GFC) (Landsat, 30 m) and Tropical Moist Forest (TMF) (Landsat, 30 m) for identifying undisturbed dense forests.
  • Evaluation Datasets:
    • VOD Product: VODCA C-VOD (2002–2018).
    • Biomass: CCI AGB (Version 4, 100 m), GlobBiomass AGB (100 m), Saatchi AGB (1 km).
    • Vegetation Structure: Canopy height (Sentinel-2, GEDI LiDAR, 10 m).
    • Optical Vegetation Indices: MODIS NDVI and EVI (MOD13A2 V6.1, 1 km, 16-day).
    • Productivity: FLUXNET2015 Gross Primary Production (GPP) (site-level).
    • Leaf Area Index: MODIS LAI (MCD15A3H V6.1, 500 m, 4-day).

Main Results

• The merged C-band TB (2002–2022) demonstrated high consistency, with strong annual temporal correlations with skin temperature in undisturbed dense forests (H polarization: R = 0.90; V polarization: R = 0.86).
• The combined inter-calibration method (linear regression for sparse vegetation, linear rescaling for dense vegetation) effectively reduced systematic TB biases between sensors; for example, the H-polarization bias between AMSR-E and AMSR2 TB in evergreen broadleaf forests (EBF) was reduced from 1.50 K to 0.35 K.
• The merged C-VOD exhibited significantly improved long-term temporal consistency across the three sensors, reducing global discrepancies between the AMSR-E (2003–2010) and AMSR2 (2013–2020) periods from 6.20 % to 0.34 %.
• Paired t-tests confirmed reliable cross-sensor continuity for the merged C-VOD, showing no significant differences (P-value > 0.01) at global and most vegetation-type scales.
• Compared to the VODCA product, the merged C-VOD showed higher temporal correlations with NDVI (in 54.44 % of pixels) and EVI across more vegetated areas globally, and improved site-level GPP correlations (outperforming VODCA at 51.02 % of sites).
• The merged C-VOD demonstrated more stable and slightly higher spatial correlations with various vegetation variables (AGB, canopy height, NDVI, EVI) across different sensor periods, exhibiting less saturation and higher sensitivity under dense vegetation conditions than VODCA.

Contributions

• Developed a novel combined inter-calibration method that adaptively uses linear regression for sparse vegetation and linear rescaling for dense vegetation to merge C-band TB from AMSR-E, AMSR2, and WindSat.
• Generated a consistent, global, and long-term (2002–2022) C-band VOD dataset that effectively addresses systematic biases and observational gaps between different microwave sensors.
• Demonstrated significant improvements in both temporal and spatial consistency of the new C-VOD product compared to existing merged products (e.g., VODCA), particularly in reducing inter-sensor discrepancies and enhancing correlations with key vegetation indicators.
• Provided a more reliable and stable long-term record of C-VOD, which is crucial for monitoring global vegetation dynamics, canopy water content, and biomass variations in the context of climate change research.

Funding

• National Natural Science Foundation of China (grant no. 42301401)
• Graduate Scientific Research and Innovation Foundation of Chongqing (grant no. CYS240144)
• Open Fund of State Key Laboratory of Remote Sensing Science (grant no. OFSLRSS202411)
• China Postdoctoral Science Foundation (grant no. 2024T170823)

Citation

@article{Chen2025global,
  author = {Chen, Dongbo and Fan, Lei and Peng, Jian and Lannoy, Gabriëlle De and Wigneron, Jean‐Pierre and Frappart, Frédéric and Tao, Shengli and Wang, Mengjia and Li, Xiaojun and Liu, Xiangzhuo and Wang, Huan and Yuan, Qiangqiang and Chen, Xiuzhi and Xiao, Yao and Ciais, Philippe},
  title = {A global long-term (2002–2022) C-band vegetation optical depth record retrieved after merging AMSR-E, AMSR2 and WindSat},
  journal = {International Journal of Applied Earth Observation and Geoinformation},
  year = {2025},
  doi = {10.1016/j.jag.2025.104961},
  url = {https://doi.org/10.1016/j.jag.2025.104961}
}