Get reliable science

Grasslands support a wide variety of species and provide environmental services such as water capture, atmospheric carbon sequestration, soil maintenance, among others. The effects of the global environmental change on grassland regions may jeopardize the benefits provided by these ecosystems. The weather monitoring throughout the grassland regions of the Mexican state of Chihuahua is recorded by a group of weather stations; however, their non-homogeneous distribution represents a limitation for the climate analysis. The aim of this study was to assess and compare the precision of three interpolation methods on the estimation of the spatial behavior of the mean annual temperature for the state of Chihuahua. Data from 57 weather stations located in the state were used and analyzed in a Geographic Information System. Data from 90% of the stations were used for the estimation of the interpolated temperature while the rest of them were utilized for validation. The interpolation techniques used consisted of two deterministic (IDW and Spline- Radial Base) and one geostatistic (Kriging) techniques. Results from the root mean square error (RMSE), the estimated effectiveness of prediction (E), and the coefficient of determination (R2) showed that the highest precision was reached by the IDW method with values of 0.885, 1.49 and 85.0, respectively. The Kriging method reached values of 0.79, 2.2 and 67.36 and the Spline method reached values of 0.86, 1.93 and 74.95 for R2, RMSE and E, respectively. The study showed the advantages of the IDW method as a tool in the analysis of the spatial behavior of the mean annual temperature in the state of Chihuahua, especially in large areas of grassland in the state where the coverage of weather stations is poor. It was also observed that the precision of the Kriging method gets compromised when the density of weather stations is low.