On-Farm Characterization of Soil Spatial Variability for Model-Based Site-Specific Management

Managing irrigation water and other resources, such as nitrogen, in a field requires distinct knowledge about the spatial soil variability that is related to the local yield potential. Applying the right amount of irrigation water at the right intensity and time depends on the soil clay content and the slope of the land surface. Hence, knowing the spatial distribution of soil clay content is an important basis for site-specific management decisions. Moreover, soil clay content is a key-input variable in agro-ecosystem models used for predicting crop yield and for weather and climate scenarios. Measuring the soil clay content at a spatial resolution that would be necessary for site-specific management is impossible because the required sampling density would be a 150 by 150 ft grid or finer, which implies an unrealistic effort. Electrical conductivity (EC) in a field can be mapped at a fine spatial resolution of 10 by 50 feet or finer. We hypothesize that EC, when combined with low-density measurements of soil clay content, will effectively support the mapping of soil clay content at a resolution finer than possible for clay measurements. In a previous research project, we collected an intensive soil clay content dataset in a farmer’s field (70 acres) near Princeton, KY, at a resolution of 150 ft (96 points in total) and EC data (7,500 data points in the same field). In the following analysis, EC data were combined with clay content measurements at different spatial densities. Even when the clay content density was reduced from 2 samples per acre to 1 sample every 5 acres, the quality of the geostatistically estimated clay content map did not substantially decrease. The remaining question is: Is the simple combination of 1 clay measurement per 5 acres and EC data valid for other and larger areas within the same agricultural region? The advantage is that clay content and EC only need to be sampled once, their relationship remains valid over time, and the sampling effort is realistic for on-farm conditions. The objective of this project is to validate the sampling strategy and the statistical analysis for another, much larger area. Soil samples in this large area (430 ac.) will be taken at a resolution of 1 sample per 5 acres (80 locations*3 depths). Based on the EC measurements taken at a much finer resolution, a clay content map will be estimated at a resolution finer than the original data set. Thirty additional locations (3 depths) will be sampled for clay. Measurements will be compared with the estimated clay content results to validate the approach. The resulting map can be used as a decision basis for variable-rate irrigation and nitrogen application. Results are also expected to support soil and crop process model parameterization for site-specific management in farmers’ fields, weather and climate scenarios.