Effects of sampling unit resolution on the estimation of spatial autocorrelation

Sensitivity of Moran's I spatial autocorrelation coefficient in estimating the magnitude of autocorrelation of species abundance data is examined when different quadrat sizes and shapes are employed. Since the measure of plant abundance is a function of the quadrat size employed, it is expected that the estimation of spatial autocorrelation based on such quantitative data will also be affected by the quadrat size used. Woody plant abundance data are used to illustrate the effects of spatial resolution (quadrat size and shape) on the estimation of spatial autocorrelation. It is found that across quadrat sizes, while the magnitude of spatial autocorrelation varies. the overall shape of the correlograms remains consistent. Specifically, the intensity of spatial autocorrelation increases with quadrat size up to the quadrat sizes of 200-225 m(2), where it reaches a plateau. Then. given that larger quadrat sizes are more Likely to include more environmental variability, the intensity of spatial autocorrelation decreases at the largest quadrat size. With these data, the shape of the quadrat affects the estimation of spatial autocorrelation as well as the degree of anisotropy in the spatial patterns identified.