Vegetative Treatment Area Geospatial Nutrient Distribution After Eight Years of Operation

Abstract

Cattle feeding operators are interested in alternative runoff control and treatment systems that eliminate the need for long-term liquid storage; however, the long-term sustainability of these systems is yet to be determined. The objective of this study was to evaluate kriging and cokriging of soil data and electromagnetic induction apparent soil conductivity (EMI ECa) data to predict spatial nutrient distribution of a vegetative treatment area used to control feedlot runoff. Once the nutrient spatial distribution is known, a total nutrient mass can be calculated and monitored year to year to evaluate sustainability. Thirty-three soil samples were collected from a 4.5 hectare vegetative treatment area (VTA) to provide for geospatial nutrient predictions. Samples were analyzed for chloride, total nitrogen, and total phosphorus. In addition, transect EMI ECa data were collected from the VTA on 6 m intervals with data continuously collected at one point per second along the transects. This resulted in a data point approximately every 2.5 m. Soils data from the 0 – 15 cm depth were kriged to generate a root mean squared standard error (RMSSE) value as an indicator of the model’s ability to predict spatial nutrient distribution. Soils data were then cokriged with EMI ECa data to evaluate any improvement in prediction. Improvement was based on a reduction in the RMSSE term. Little or no improvement was detected when the entire VTA was evaluated as one field; however, predictions were improved when the VTA was separated into north and south segments to account for chronic nutrient loading in the south segment. A strategy of focusing most or all of the soil samples in the segment that receives the most nutrient loading may improve spatial prediction without increasing sampling cost, but still may not provide adequate spatial prediction. Additional spatial prediction strategies such as multiple linear regression using high density, relatively low-cost EMI ECa may provide more detailed predictions without the need for as many high-cost soil samples as cokriging.