Using Spatial Correlation Analysis to Identify Ancillary Data for Soil Fertility Mapping.

The most common soil fertility sampling scheme used by farmers and consultants is the 2.5-acre grid. Even though this sampling is considered a cost-effective approach for crop consultants and farmers, it often fails to capture the spatial variability of soil fertility within a field. The use of ancillary data, such as Sentinel Satellite Imagery and/or soil electrical conductivity, could potentially correlate with the spatial and temporal variability of soil nutrients. Therefore, these ancillary variables could be used to identify zones of variability and guide soil sampling. This study seeks to evaluate various methods of guiding soil sampling schemes by using remotely sensed ancillary data. For this project, data retrieved from Sentinel-2 Satellite, National Agriculture Imagery Program, soil electrical conductivity, and yield monitors are analyzed to assess their relationship with commonly amended soil properties (P, K, and pH) in production crop fields in Alabama. Data from two farmers’ fields from three growing seasons is currently being analyzed. Factorial kriging and spatial multiple linear regression analysis methods are currently under evaluation to identify spatial and temporal correlation and to predict soil macronutrient spatial variability from ancillary data. These methods and ancillary data will then be used to create a prototype decision support system (DSS) to create site-specific recommendations for soil fertility sampling strategies.