Filling of Missing Data in Atmospheric Series with Linear Krigeage

Abstract

Numerous investigations indicate an important relationship between atmospheric variables and various indicators of environmental contamination that affect public health [1]. Most of the series of meteorological variables used, as independent variables, have defects, Being the most common the lack of specific data or missing data lagoon [2]. There are numerous methods to solve this problem, Most of them require an assumption of normality in the behavior of the data and do not consider their spatial dependence. Through geostatistics, There is an alternative solution to this problem, Making use of regionalized variables, of random functions, of the variogram and Kriging, which is defined as the best linear unbiased estimator [5]. Finalize analysis with method validation, through the cross-validation method “leave one out” [6]. It is concluded that in annual series of atmospheric variables, with time-zone temporal resolution, that the Krigeage presented errors under 5%.