SINENVAP: An algorithm that employs kriging to identify optimal spatial interpolation models in polygons

Abstract The aim of the SINENVAP algorithm is to facilitate the estimation of spatial interpolations within polygons, by using simple, ordinary, and universal kriging. This algorithm is available as a function of the EcoIndR package, which is available as an RWizard application and an R package on CRAN. The main strengths of this algorithm include: the possibility of using different file formats for polygon variable and coordinate inputs (CSV, EXCEL, RData, shape or ASC), compatibility with UTM or decimal coordinates, estimation of optimal grid cell size, the possibility of selecting only points inside polygons from the entire dataset, the application of a jitter function or to estimate the mean value of the variable for duplicated coordinates, reservation a percentage of data for validation, selection of those grid coordinates nearest the data coordinates reserved for validation, the possibility of fitting 13 different models into the semivariogram, automatic selection of the model that best predicts the data reserved for validation through the use of seven accuracy measures, the possibility of using countries, regions, departments, river basins, or even alpha shape distribution as polygons, and finally, depiction of contour plots with the spatial interpolation of the variable and the error within polygons. The spatial interpolation of the temperature in North America and the distribution of a virtual species are used as examples of this algorithm's potential to perform spatial interpolations on both large and small scales.