Predictive model of phosphate availability in soil by using QGIS software
Carlos Roberto Martínez Martínez, Carlos Eduardo Somoza Vargas
Site-specific fertilizer plans are effective to reduce waste and costs, considering that farm soils are not spatially uniform due to a considerable the amount of biochemical interactions that rule nutrients availability. In Salvadorian soils, phosphorus is a limiting macro-nutrient and farmers often use diammonium phosphate (DAP) as a palliative amendment, however, they disregard spatial variability and therefore causing fertilizer misuse. In this research, 55 soil samples were taken from a 30 hectare field and submitted to a lab to be treated with a dose of DAP in order to measure how much phosphate remained available through time in a root-free scenario. It was demonstrated that responses were not strictly proportional to the values of pH and Effective Cation-Exchange Capacity (ECEC), meaning that this variables were not sufficient to sustain traditional assumptions about phosphate availability across a field; the higher value of Pearson correlations for pH and ECEC data versus P availability measures was 0.4, and the maximum R2 was 0.21, indicating that a statistical predictive model was not trustable. To contribute for a better agronomic analysis, a predictive model of phosphate availability is proposed in this paper. Data obtained from treated samples was processed to produce semivariogram graphs which fitted at least one of the standard models of linear, circular or spherical patterns. This also indicated that Kriging interpolation was convenient for representing data and establishing predictive models. Then, GIS software was used to create maps that modeled the stage of P availability trough time. The contribution of this work is the proposal of geostatistical software tools to improve soil nutrients requirement analysis based on soil samples, allowing the modeling of an entire farmland for a more precise fertilizer plans development.
Carlos Roberto Martínez Martínez, Carlos Eduardo Somoza Vargas. Predictive model of phosphate availability in soil by using QGIS software. International Journal of Academic Research and Development, Volume 5, Issue 6, 2020, Pages 30-35