FastMapping: a tool to automate spatial variability mapping

CÓRDOBA, M.; PACCIORETTI, P.; AGUATE, F.; BRUNO, C.; BALZARINI, M.

Victoria

Conferencia; XXVIIIth International Biometric Conference; 2016

Development and use of new information technologies that allow us to capture different types of data associated with spatial localization have been promoted in the last decades. The optimal use of data obtained from those technologies depends on the capacities for efficiently and simply exploring and analyzing (mapping) spatial variability. FastMapping was developed to automate, on a user-friendly platform, large part of the geostatistical analysis focused on spatial variability in continuous domains. The application was created using the libraries Shiny and Shinythemes in R software. The fits of the models for spatial data are obtained using the procedures available in the libraries automap, fields, geoR and raster. FastMapping allows adjustment of several spatial correlation models and automatic selection of the one that best performs in spatial prediction of the studied phenomenon. For this purpose, the software performs cross-validation of the predictive capacity of each adjusted spatial correlation model. Once the model of best prediction RMSE is selected, the adjusted experimental and theoretical semivariograms can be visualized, as well as the spatial variability map obtained by kriging interpolation and the map of prediction variance. FastMapping allows us to make adjustments of models with and without a mean trend, as well as ordinary (without trend) and universal (with trend) kriging interpolation. Predictions can be point or block type, both in a global or local neighborhood (kriging neigborhood). With FastMapping .txt extension files can be opened and results can be exported to a CSV table which includes the coordinates of the points to be predicted, the predictive values and prediction variance. In addition, the spatial variability map can be copied and saved, and exported as a GeoTIFF file. Here we illustrate the use of the software developed to obtain a spatial variability map from an intensively recorded wheat yield database (n=6252) in a precision agriculture plot. The availability of R-based applications, with a user- friendly interface, allows us to improve the methodological and computational bases of spatial data analysis.