Modeling of nitrate ion concentration in Costa Rican surface waters

Abstract

Inappropriate or excessive use of fertilizers mobilizes excess nitrate ions into surface and groundwater through runoff and percolation processes. The impact is a deterioration of ecosystems and water quality. In this research, the concentration of nitrate ions in the surface waters of a territory was modeled under the same land use model, Costa Rica, with 88.2% accuracy. Free access physicochemical data and geographic information were used in the modeling, processed with the Random Forest learning algorithm of the R program. Climatic, soil and terrain topography data that affect mobilization were included. Finally, a Geographic Information System (GIS) was developed with the vector layers obtained and the modeled points were classified into concentration categories related to anthropogenic activity. The results show alert values in the wetlands of the North Caribbean and South Pacific. The relationship between the factors that mobilize the nitrate ion, and the modeled nitrate ion concentration is consistent.