Adebayo et al. (2026) Sub-seasonal forecasting of cropland productivity anomalies using satellite soil moisture in water-limited environments

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-09-05
• Authors: Adebowale Daniel Adebayo, Catherine Nakalembe
• DOI: 10.1016/j.rse.2026.115645

Research Groups

• Department of Geographical Sciences, University of Maryland

Short Summary

This study explores the potential of satellite root-zone soil moisture to provide skillful sub-seasonal forecasts of crop productivity anomalies in drought-prone croplands across Eastern and Southern Africa. The research demonstrates that incorporating SMAP RZSM into a ConvLSTM model can improve forecast accuracy, with the added value of RZSM concentrated in water-limited environments.

Objective

• Investigate the relationship between satellite root-zone soil moisture (RZSM) and near-infrared reflectance of vegetation (NIR𝑣) anomalies to predict crop productivity in African croplands

Study Configuration

• Spatial Scale: 9 km resolution, covering drought-prone croplands across Eastern and Southern Africa
• Temporal Scale: 8-day composites, with a reference input window of 48 days (6 lead steps)

Methodology and Data

• Models used: ConvLSTM encoder-decoder architecture
• Data sources:
  • SMAP Level 4 soil moisture product
  • VIIRS Nadir BRDF-Adjusted Reflectance product for NIR𝑣 computation
  • Cropland mask from Khan et al. (2026)
  • Aridity index from Zomer et al. (2022)

Main Results

• The study demonstrates that incorporating SMAP RZSM into a ConvLSTM model can improve forecast accuracy, with the added value of RZSM concentrated in water-limited environments.
• The optimal lag between RZSM and NIR𝑣 anomalies is found to be around 20-25 days.
• The study shows that the ConvLSTM model outperforms simpler baselines, including persistence and tree-based models.

Contributions

• This study provides new insights into the relationship between satellite root-zone soil moisture and near-infrared reflectance of vegetation anomalies in African croplands.
• The results demonstrate the potential of using SMAP RZSM to improve crop productivity forecasts in water-limited environments.

Funding

• This research was funded by the NASA Terrestrial Hydrology Program (Grant Number: NNX16AQ71G) and the University of Maryland's Department of Geographical Sciences.

Citation

@article{Adebayo2026Subseasonal,
  author = {Adebayo, Adebowale Daniel and Nakalembe, Catherine},
  title = {Sub-seasonal forecasting of cropland productivity anomalies using satellite soil moisture in water-limited environments},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115645},
  url = {https://doi.org/10.1016/j.rse.2026.115645}
}