ESTIMATION OF VEGETATIVE COVER IN AN ARID RANGELAND BASED ON SOIL-MOISTURE USING COKRIGING

Publisher Website
Google Scholar
Abstract
Soil moisture and vegetative cover (ephemeral, perennial forb and perennial grass) were measured at 30-m intervals along a 2730-m transect. The spatial correlation of vegetative cover and soil moisture, and their cross-correlations, were examined. A total of 610 estimates of vegetative cover were generated at 3-m intervals along the transect, using both ordinary kriging and cokriging methods. Jackknifed estimates of mean-reduced error and reduced variance were used to validate the cokriging models. The average kriging and cokriging variances and the mean sum of squares (SSQ) were used to compare the two methods. Based on SSQ comparisons, cokriging is the best method to estimate spring ephemeral cover (for observations from all stations) and perennial spring grass cover (for observations from alternate stations). Cokriging estimations for spring perennial grass cover and for spring ephemeral cover produced 7.14 and 2.13%, respectively, improvement over kriging. Cokriging gave better estimations when the number of observations for perennial spring grass cover and fall perennial forb and perennial grass covers were decreased. © Williams & Wilkins 1992. All Rights Reserved.

This publication has 0 references indexed in Scilit: