Climate Science Program

Contact us

Got a question or comment? Contact us at (515) 294-9871or e-mail

Mark S. Kaiser




Office:1456 Wilson Hall and 642 Science II

Phone: 515-294-8871






Department: Statistics

Research Interests: Spatial statistics (especially Markov random field models), hierarchical models, applications in limnology, ecology, fisheries science, and meteorology

Brief description of current research:

Dr. Kaiser's research focuses on the development of Markov random field models for applications in spatial problems and problems with other complex dependence structures. Such models are formulated by specifying a univariate conditional distribution for each random variable on a lattice (e.g., grid). While such models for Gaussian (normal) conditional distributions are well developed, models for other distributions present a number of challenges, not the least of which is ensuring that a joint distribution exists that corresponds with the specified conditionals. Models for binary conditionals, such as would
be used to model the probabilities of events (e.g., rainfall), Poisson conditionals, such as would be used to model counts, and beta conditionals, such as would be used to model continuous proportions (e.g., percent cover) have now been formulated. Currently, applications of the statistical methodology developed are underway that include the spatial/temporal spread of invasive plant species, spatial/temporal modeling of relative humidity in a region including Iowa, and spatial/temporal modeling of the occurrence of bean pod mottle virus in soybeans.

Recent publications:
Kaiser, M.S. and Cressie, N. (2000), The construction of multivariate
distributions from Markov random fields.
Journal of Multivariate Analysis, 73, 199-220.

Kaiser, M.S., Daniels, M.J., Furukawa, K. and Dixon, P. (2002), Analysis
of particulate matter air pollution using
Markov random field models of spatial dependence. Environmetrics 12,

Kaiser, M.S., Cressie, N. and Lee, J. (2002), Spatial mixture models
based on exponential family conditional
distributions. Statistica Sinica 12, 449-474.