Performance of Multi-Source Remote Sensing Soil Moisture Products over Punjab Pakistan During 2022–2023
Saba ul Hassan,Munawar Shah,Rasim Shahzad,Bushra Ghaffar,Bofeng Li,José Francisco de Oliveira‑Júnior,Khristina Maksudovna Vafaeva,Punyawi Jamjareegulgarn
DOI: https://doi.org/10.1007/s00704-024-05082-7
2024-01-01
Theoretical and Applied Climatology
Abstract:The Global Navigation Satellite System Reflectometry (GNSS-R) has emerged as a valuable tool for terrestrial remote sensing applications, particularly in the context of land Surface Soil Moisture (SSM) detection. The high-resolution capability of GNSS-R complements traditional satellite-based active and passive missions but the product reliability and robustness evaluations are still absent due to an efficient retrieval algorithms. In this study, we addressed this lack of reliability and robustness by comprehensively assessing the SSM retrievals from CYclone Global Navigation Satellite System (CYGNSS) data with the satellite-based microwave radiometry products Soil Moisture Active Passive (SMAP) and Modern Era Retrospective-Analysis for Research and Applications (MERRA2) over Punjab in various seasons. ERA5 model-based products for the same period in 2022–2023. Our study reveals a distinct seasonal average SSM variation during autumn (0.20 cm3/cm3), followed by winter values of 0.19 cm3/cm3. Subsequently, the minimum SSM values are observed during summer (0.11 cm3/cm3) and an increase in spring to 0.13 cm3/cm3. Moreover, a strong positive linear relationship (0.74) is evident between SMAP and ERRA 5 in contrast to a low correlation (0.03) between MERRA2 and both the SMAP and ERRA 5. Additionally, SMAP demonstrates moderate and weak correlation of 0.53 and 0.03 with CYGNSS and MERRA2, respectively. The CYGNSS exhibits moderate correlations (0.46) with ERRA 5 and SMAP and a weaker association (0.14) with MERRA2. Our analysis concluded that MERRA2 (Bias = 0.20 cm³/cm³, ubRMSD = 0.25 cm³/cm³, RMSE = 0.12 cm³/cm³, SD = 0.13 cm³/cm³, MAE = 0.04 cm³/cm, R = 0.03) SSM product performs poorly as compared to SMAP (Bias = 0.03 cm³/cm³, ubRMSD = 0.03 cm³/cm³, RMSE = 0.04 cm³/cm³, SD = 0.05 cm³/cm³, MAE = 0.03 cm³/cm³, R = 0.74) and CYGNSS (Bias = -0.01 cm³/cm³, ubRMSD = 0.09 cm³/cm³, RMSE = 0.07 cm³/cm³, SD = 0.06 cm³/cm³, MAE = 0.05 cm³/cm³, R = 0.46) products. This study provides accurate future predictions of SSM with delineating the limitations of GNSS-R in comparison to remote sensing and model values. The findings from this study have also significant implications for the advancement of GNSS-R applications in agriculture and crop management.