Correcting estimates of net primary production: are we overestimating plant production in rangelands?

Abstract

This paper addresses the issue of the effect of random errors in field estimates of net primary production (NPP). This is a critical subject in range management because field estimates of plant production are regularly used to determine stocking rates, range condition, and animal consumption. What we show in this paper is that random errors associated with field estimates of NPP can result in a positive bias and thus an overestimation of NPP. Depending on the case, this overestimation has been reported as high as 700%. We present examples with overestimations in the 200% to 400% range. The overestimation in NPP increases with increases in biomass variances, frequency of sampling, and number of taxonomic (species) and tissue (live, dead, etc) components sampled. We (1) outline in nonmathematical terms the reasons behind overestimation in NPP and the analytical solutions designed to correct them; and (2) present applications of the analytical solution for adjustments to concrete cases. The adjustments for overestimation outlined in this paper do not guarantee an accurate estimate of NPP but eliminate an unneeded source of error. A computer program (for IBM(TM) compatible) designed to implement the necessary adjustments is available from the authors free of charge (send a blank diskette).