Abstract
A spatial autocorrelation and cross-correlation statistic was used to study spatial relationships of blue grama (Bouteloua gracilis (H.B.K.) Lag. ex Steud) with other site characteristics of a mixed grass community in southern Arizona. Biomass and density of blue grama and other species, as well as terrain and soil variables were measured. A cross-correlation statistic, which Moran's I is a special case, was used to test for presence of spatial cross-correlations between pair-wise combinations of both vegetation and site characteristics. The test statistic showed that biomass of blue grama, other grasses and forbs, soil rockiness and the slope-aspect index were not spatially correlated within the study area, but instead were randomly distributed over the area. Spatial autocorrelations were observed for density of blue grama, soil pH, percent clay and elevation. Soil pH and clay content were spatially correlated with elevation over the study area as indicated by the significant cross-correlation. Results of other studies of blue grama were also used to interpret the spatial distribution of this species with respect to soil and site characteristics. This study revealed that responses of blue grama to site variables may best be understood through spatial autocorrelation and cross-correlation analysis rather than through the unidirectional approach of most gradient analysis studies. That is, simultaneous analysis of blue grama and environmental factors over multi-directions were made possible using spatial cross-correlation techniques.
