Estimation of mustard and wheat phenology using multi-date Shannon entropy and Radar Vegetation Index from polarimetric Sentinel- 1

Abstract

Dual-pol Shannon Entropy captures the dynamic crop growth parameters. It was used to evaluate biophysical parameters of important crops using Sentinel-1 data. Improved phenology information is vital input to crop-growth models. Mustard experienced increase as it advanced in phenology but for wheat, decreased due to absorption. A significant relationship at initial crop stages (height < 150 cm and biomass < 5 kg m⁻²) R² =0.65 between Shannon Entropy (SE) and crop parameters, though less strong (R²=0.33) for entire mustard cycle and wheat (plant height < 80 cm and biomass < 6 kg m⁻²) was observed. The SE appeared sensitive to low-medium biomass and useful in monitoring crop phenology in low biomass and initial phenophases. The dynamic crop profile is manifested conjunctively as Radar Vegetation Index (RVI). Random Forest Regression (RFR) and Support Vector Regression (SVR) were evaluated to predict the phenophases. Phenology responses till peak/beyond were found to perform close to ground observations.

Keywords:

• Sentinel 1 observables
• Shannon entropy
• SAR polarimetry
• machine learning
• crop parameters
• Rabi crops

Disclosure statement

No potential conflict of interest was reported by the author(s) .

Additional information

Funding

Authors would like thank the Airborne- SAR AO for the funds and Mr. Om Pal, HARSAC for his support in ground data collection.