This book is a welcome addition to the growing literature on spatial statistics, with the distinctive feature that it is aimed at “the intersection of at least three traditionally separate academic disciplines: statistics, epidemiology and geography” (p. 4). Spatial statistics is a subject with a fascinating history. From a theoretical perspective, the core of the subject is the formulation and analysis of stochastic processes indexed by (typically two-dimensional) space rather than by one-dimensional time, in which context the increase in dimensionality is less important than the loss of a natural ordering to the index set. My favorite teaching example of this point, which has its origins in an article by Brook (1964), is the following. Consider the first-order temporal autoregressive process Yt , conventionally defined as Yt = αYt−1 + Zt , (1)
[1]
D. Krige.
A statistical approach to some basic mine valuation problems on the Witwatersrand, by D.G. Krige, published in the Journal, December 1951 : introduction by the author
,
1951
.
[2]
D. Brook.
On the distinction between the conditional probability and the joint probability approaches in the specification of nearest-neighbour systems
,
1964
.
[3]
Geoffrey S. Watson.
Trend Surface Analysis and Spatial Correlation
,
1969
.
[4]
J. Besag.
Spatial Interaction and the Statistical Analysis of Lattice Systems
,
1974
.
[5]
N. Cressie.
Fitting variogram models by weighted least squares
,
1985
.
[6]
J. Besag,et al.
Bayesian analysis of agricultural field experiments
,
1999
.