Characterizing the spatial structure of vegetation communities in the Mojave Desert using geostatistical techniques

Geostatistical techniques are used to evaluate spatial characteristics of v egetation communities mapped within a portion of the Mojave Desert, Califor nia. Spatial structures of different vegetation communities were characteri zed using nugget, range and sill parameters of spherical model variograms. Model variograms reveal different vegetation communities have distinctive s patial properties that are quantified effectively by the parameters of thes e models. Model variograms were fit to experimental variograms calculated f rom plant width values of ground based vegetation data collected for 96 100 0 x 2 m transects. The maximum model variogram range for all communities wa s 39 m, suggesting transects must be at least this long to classify them un ambiguously into the communities currently recognized. Experimental variogr ams were calculated for all plants in each of eight mapped vegetation commu nities. The Bush Seepweed Series has a distinctively long range, for exampl e, and the Creosote Bush Series has a distinctively high nugget and sill. T hese findings support use of high-resolution remote sensors and geostatisti cs for determining vegetation community structure. Spatial pattern informat ion produced by these methods could improve broad scale vegetation classifi cations produced by low-resolution remote sensing systems. (C) 2000 Elsevie r Science Ltd. All rights reserved.