Skip to main content
ARS Home » Southeast Area » Florence, South Carolina » Coastal Plain Soil, Water and Plant Conservation Research » Research » Publications at this Location » Publication #72028
Title: ACCOUNTING FOR NON-STATIONARITY IN MEAN AND VARIANCE OF SPATIAL YIELD DATA IN THE SE COASTAL PLAIN

Author
item Sadler, Edward
item Busscher, Warren

Submitted to: Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 7/15/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Widespread interest in precision agriculture, or site-specific farming, has brought pressure on researchers to provide producers and other interests with appropriate data for analysis of yield maps and management recommendations. Although rigorous analytical techniques exist for several types of spatial data, restricting assumptions eliminate many simpler procedures from consideration for use on many data of interest. One particularly troublesome assumption in the standard analysis of spatial yield data is the assumption of stationarity of the mean and variance, which requires that no consistent trend in either exists in the space examined. Previously, analyses were conducted with acknowledgment of the possible limitation, but no clearly objective test for stationarity existed. The GSLIB library (Deutsch and Journel, 1992) provides methods for analysis of data regardless of whether the assumptions are met. Yield data from 14 crop-years on a single field are re-analyzed with the advanced tools. Comparisons with the previous analyses and judgments about the suitability of the simpler tools will be presented.