Get reliable rangeland science

An analysis is made of the selection and application of reference areas in the derivation of indices of range condition. In present
usage there are deficiencies arising from bias in reference area selection, the absence of correction for natural variation, the designation
of 'climax' as 'best' and the choice of grazing pressure as the only gradient of vegetation change.
A system of nlultiple reference areas is proposed, based on a sampling of land units and the analysis of the principle vectors contributing
to vegetation change. These areas would be chosen ;IS a statisticai sample of the land unit they represent and the calculation
of indices of difference or change would be corrected for errors arising from natural variatio~i and measurement error. Reference areas
would be chosen on the basis of their potential productivity in terms of the principal land use.