Spatially correlated counts and proportions appear in a variety of disciplines, such as natural resources
and epidemiology. Various models have been proposed to analyze non-normal spatial data. Binomial
cokriging is one such model, first proposed to predict the underlying spatial rate or probability
distribution based on observed sample proportions of a rare disease [1]. The model has been applied to
the analysis of childhood cancer rates [2] and used to develop a test statistic for cancer “hot spots” [3].
Poisson kriging was proposed in 2006 to model spatially correlated counts observed over
heterogeneous areas [4]. This model used a weighting scheme to adjust the relative importance of
observed counts to favor those taken over a larger observation effort. Binomial cokriging includes no
such weighting effort, which may assign over importance to locations with a relatively small sample size.