Estimating Maize Planting Date By Combining Climate, Satellite-Sensed NDVI and Census-Based Agricultural Data.

Planting date (PD) is an essential management information, and it is commonly required input data by crop model-derived estimates of yield production and crop water demand. The PD varies annually depending on climate conditions, crop variety, and the scheduling of agricultural activities. Crop phenophases and weather variability are strongly interconnected, therefore, the planting timing has a considerable effect on yield. Multiple attempts have been made to estimate PD date at a regional scale, with varying levels of success, including climate-rule-based, census-based, and Earth observation-based approaches. In this study, we argue that we can improve the PD estimative reliability by combining the different approaches. We developed a method to estimate maize PD in Nebraska, USA by coupling gridded layers of soil temperature and moisture, and a time series of the Normalized Difference Vegetation Index (NDVI) from TERRA and Aqua satellites, with U.S. National Agricultural Statistics Service Crop Progress and Condition Reports (NASS-CPR). Meteorological data were obtained from the Parameter-elevation Relationships on Independent Slopes Model (PRISM), soil data from Soil Survey Geographic (SSURGO), and crop mapping data from the Cropland Data Layer (CDL) database. PD data from 150 agricultural areas in Nebraska during the period from 2007 to 2012 were used to validate the method. The proposed method improves spatial and temporal PD estimates with more accurate results through the application of crop meteorological forcing data, while maintaining the reliability using census-based and time series of satellite observation.  The combination of the observed and model-based data has the potential to improve planting date estimates, not only for maize but other crops as well.