Addressing Crop Dormancy for Phenology-Based Field-Level Detection of Winter Wheat Planting Dates in the United States

Satellite-derived phenology has been proven valuable in determining crop planting dates (PDs). Potential crop dormancy of winter wheat adds challenges in the phenology-based PD detection. Here, we aim to detect field-level winter wheat PDs using satellite-derived phenological information with consideration of potential crop dormancy. Specifically, we categorized winter wheat growth into two scenarios: with and without dormancy. We performed phenological modeling for the two scenarios across eight major winter wheat states in the U.S. from 2017 to 2022 using ~3-day and 30-m harmonized Landsat and Sentinel-2 (HLS) time series. We unified the time lags between phenological stages and PDs for the two scenarios according to the accumulated heat units. Our predicted PD maps were validated against field-level data from the USDA reports, demonstrating high accuracy with R² of 0.77, a mean absolute error (MAE) of 6.30 days, and a root mean square error (RMSE) of 9.73 days. Incorporating detailed information on winter wheat fields, such as intended use, irrigation practice, and crop varieties, can further improve the accuracy. Winter wheat PDs showed no significant trends during this period and were significantly correlated to the harvest dates of previous cash crops ( P -value <0.05). Increased precipitation and temperature in early fall were associated with delayed planting of winter wheat, and farmers tended to plant earlier in fields with lower clay content. This framework has high flexibility for applying to other winter crops to derive field-level PDs from satellite time series and supports sustainable agricultural management practices related to crop planting.