Interval-valued kriging for geostatistical mapping with imprecise inputs (Q2069049)

Kriging is one of the most applied tool in geostatistics. Due to various limitations the used data are often imprecise and should be modelled as interval-valued data or fuzzy-valued data. This paper proposes a modification to Diamond's interval-valued kriging [\textit{P. Diamond}, Math. Geol. 20, No. 3, 145--165 (1988; Zbl 0970.86500)]. In contrast to Diamond it plays a key role that the authors use not a interval-valued variance and covariance but a real-valued one which goes back to \textit{R. Körner} [Fuzzy Sets Syst. 92, No. 1, 83--93 (1997; Zbl 0936.60017)]. The resulting interval-valued kriging overcomes many of the mathematical and computational difficulties of previous attempts. The numerical implementations are presented in R packages. The proposed method is applied to kriging of design ground snow loads in Ohio.