Mathematical aspects of the kriging applied on landslide in Halenkovice (Czech Republic)

Abstract

Kriging is one of the geostatistical techniques for spatial data analysis that is usually used for a modelling of natural phenomena or a creation of digital elevation models. In this paper, we introduce kriging methods in the context of a landslide modelling in time. The proposed procedure, as well as most of the statistical methods, is designed for complete data sets, i.e. the observations at the beginning and the end of the study are available. In order to use all the information from the data and to avoid the loss of information after omitting observations with missing values, the algorithm for imputation of missing data values based on kriging techniques is proposed. The methodology was verified by the landslide modelling in Halenkovice, Czech Republic, during the period from June 2008 to March 2010. The obtained results showed a potential of kriging methods for the landslide modelling.

Keywords

• universal kriging
• variogram
• landslide
• missing values
• data imputation