Submitted to: Applied Statistics In Agriculture Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: April 30, 2003
Publication Date: April 30, 2003
Citation: Meek, D.W., Sauer, T.J. 2003. Suggestions for presenting Kriging data. In: Proceedings of the 15th Applied Statistics In Agriculture Conference, April 27-30, 2003, Manhattan, Kansas. p. 258-276. Technical Abstract: Kriging maps are often part of the reported analyses in many environmental research studies including those our agency is working on in the area of precision/sustainable farming. All to often important details on the underlying variography and/or kriging procedures are omitted. Likewise the content and form of presenting kriging results vary greatly. Often features of the underlying variability are not reaidly seen. Instead of reviewing poor practice in current literature, we offer guidelines for reporting the methodology and presenting the results with the use of soil test phosphorus (STP) measures from a real world pasture study. Relevantly, the stationarity assumption for the varigram is argued; computational aspects for both the model and empirical variogram development are reported; and similarly, computational aspects for the kriging surface are reported. In short, enough detail is reported to understand and reproduce the analyses. Standard practice for presenting kriging results should include both the kriging estimates and the associated standard error map. Various planar and three dimensional plots are shown and discussed. Enphasis is on developing quality gray-scale planar maps for conventional publications. Ideally, for both recommended plots, patterns and unique features of the surfaces' variability are revealed.