Retrieving crop phenology at field-scale in the Nile Delta based on Sen2Like and PlanetScope data

This is the authors' abstract. We don't add key points for this paper.

• Crop phenology is retrieved with Sen2Like in double-cropping farms. • Estimated metrics are evaluated with a reference crop calendar. • Sen2Like shows strong agreement with metrics derived from PlanetScope. Monitoring crop phenology is crucial for improving agricultural management and supporting food security, especially in low- and middle-income (LMIC) countries like Egypt, where smallholder farms account for most food production. This study evaluates the potential of fused Sentinel-2 and Landsat 8/9 data, generated with the Sen2Like processor, for retrieving key phenological metrics at field-scale in a multi-cropping agricultural system of the Nile Delta. Forty-two fields were analysed over six years, focusing on wheat, maize, rice and clover in El Gharbia Governorate. Phenological metrics were derived using the Decomposition and Analysis of Time-Series Software (DATimeS) and were evaluated based on a reference crop calendar. Results showed that 91% of crop seasons were detected using the Sen2Like fused product compared to 88% using Sentinel-2 alone. For both datasets, the Start of Season (SOS) for summer crops presented a root mean squared error (RMSE) of about 35 days from the calendar date, while the SOS for winter crops had an RMSE of about 14 days. Due to its daily temporal coverage, PlanetScope data was used for additional evaluation. Metrics extracted with Sen2Like and PlanetScope presented a good agreement, with an RMSE of 1 day for metrics occurring during spring months and an RMSE of 6 days for winter metrics. Moreover, mapping key phenological stages provided insights into the crops’ spatial and temporal distribution across the study area. This study demonstrates the value of Sen2Like in capturing field-scale phenology with implications for improving agricultural practices and farm sustainability.