Estimating and mapping pasture soil fertility in a portuguese montado based on a objective model and geostatistical techniques

Abstract

Pasture soils can exhibit a high spatial variability which should be characterised to properly manage the yield potential of different within-field areas. Thus, with the aim of proposing an objective methodology to estimate the pasture soil fertility and, later, analyse its spatial pattern, the formulation of the probabilistic Rasch model constitutes a new approach in pasture fields. In this research, a case study was performed to illustrate the proposed method. Consequently, after taking some soil samples (34) and measuring different soil properties (sand, silt, and clay content, organic matter, phosphorus, potassium, moisture content, soil apparent electrical conductivity, elevation, and slope), the use of the Rasch model provides a integrated measure of pasture soil fertility at each sampling location, which can be computed using geostatistical algorithms to map its spatial distribution throughout the field. After verifying that data fit the model reasonably, the main outputs of the Rasch model were a ranking of all sampling locations according to the pasture soil fertility and another ranking of the soil properties according to their influence on the soil fertility, being the topographical properties (slope and elevation) the most influential. Later, the ordinary kriging algorithm was utilised to estimate soil fertility throughout the pasture field and the probability kriging algorithm was used to provide information for hazard assessment of pasture soil fertility, being both kriged maps the basis to delineate homogeneous zones. Finally, vegetation indices and pasture yield data at sampling points were employed to check that two zones previously determined were different. The analysis of zonal differences in pasture systems can lead to an optimal application of inputs and a more cost-effective management, with the associated environmental, economic, and energetic benefits.