Multivariate Spatial Models

  • J. Andrew Royle
Part of the Lecture Notes in Statistics book series (LNS, volume 144)


In studies involving spatial data, it is seldom the case that data for only a single process are collected. Typically, there is great expense associated with establishing spatial monitoring networks or other mechanisms of spatial data collection (e.g., satellites) and so measurements are usually made on two or more variables. Even when networks are established for purposes of monitoring a single physical process, there often exist alternative and disjoint networks that were established for monitoring related processes. Finally, it is seldom the case that a single physical process is uninfluenced by other processes, and so it would be inefficient to neglect information from these other processes even when one’s objective centers on modeling or prediction of the single process. Thus, statistical techniques for multivariate spatial data are critical for effective modeling of spatial processes. The primary objectives of this chapter are to review existing multivariate spatial modeling strategies, discuss their strengths and weaknesses, and to discuss several atmospheric science problems that require multivariate spatial models. For a more thorough exposure to traditional geostatistical approaches, the reader is referred to Wackernagel [Wac95] and the review by Gotway and Hartford [Got96].


Covariance Function Conditional Model Best Linear Unbiased Prediction Prediction Variance Universal Kriging 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag New York, Inc. 2000

Authors and Affiliations

  • J. Andrew Royle
    • 1
  1. 1.U.S. Fish and Wildlife ServiceLaurelUSA

Personalised recommendations