Delineating site‑specifc management zones on pasture soil using a probabilistic and objective model and geostatistical techniques

Abstract

In recent years, diferent algorithms have been utilised to delineate management zones, areas with similar properties, within agricultural felds. However, there are few applications in pasture systems. In this work, the formulation of the Rasch model, as an objective and probabilistic technique to integrate diferent soil properties, provided measures of pasture soil fertility that were used to analyse spatial variability throughout a feld. To illustrate the proposed approach, a case study was conducted in a pasture feld. Ten soil properties (sand, silt, and clay contents, moisture content, pH, organic matter, nitrogen, phosphorus, potas‑ sium, and soil apparent electrical conductivity) were measured at 76 locations in a pasture feld; after their integration in the model, a classifcation of all sampling locations accord‑ ing to pasture soil fertility was determined, and the infuence of each soil property on the soil fertility was highlighted, with the soil moisture, clay, and sand contents and nitrogen being the most infuential properties and the silt content being the least infuential property. Then, an ordinary kriging algorithm was used to estimate pasture soil fertility throughout the feld, and homogeneous zones were delimited from the kriged map. The possibility of using probability maps to determine management zones and provide information for hazard assessments of pasture soil fertility in the feld was also shown. Finally, NDVI data at each sampling location were utilised to verify the diferences between the management zones