Phenologically Corrected Crop Condition Mapping and Assessment with Vegetation Index Time Series

Vegetation index (VI) time series from remotely sensed data have been used to assess the crop condition by comparing the current year's conditions with the averaged VI time series over multiple years. However, due to the yearly differences in planting dates and climate conditions, averaging VI on the same calendar day may not reflect the general crop condition at the same growth stages. In this study, phenologically corrected Enhanced Vegetation Index (EVI2) was generated to assess crop condition across years at the same growth stages. Crop emergence dates were first generated using the within-season emergence (WISE) algorithm based on routine harmonized Landsat and Sentinel-2 data covering central Iowa. Different years of EVI2 time series from 20182022, generated from the Harmonized Landsat-Sentinel (HLS) dataset were aligned and then averaged based on the crop emergence dates as a baseline of normal crop condition. Then, the crop conditions from other years are assessed are compared with respect to this baseline. Comparisons based on growth degree days (GDD) instead of calendar days were also carried out to reflect the variance in crop development rate. The validation of the assessed crop conditions is performed on the National Agricultural Statistics Service (NASS) Crop Progress & Condition (CPC) gridded layers. The assessment results indicated that they are consistent with the CPC layers. Further evaluation of yield anomaly estimates with and without phenological correction indicates that the method could facilitate crop yield prediction as well as providing within-season crop condition monitoring information at high spatial and temporal resolution.