LAUSR

Abstract Estimates of plant biomass are helpful for many applications in invasive plant science and management, but measuring biomass can be time-consuming, costly, or impractical if destructive sampling is inappropriate. The objective of this study was to assess feasibility of developing regression equations using a fast, nondestructive measure (cover) to estimate aboveground biomass for red brome (Bromus rubens L.), a widespread nonnative annual grass in the Mojave Desert, USA. At three study sites, including one measured for three consecutive years, B. rubens cover spanned 0.1% to 85% and aboveground biomass 1 to 321 g m–2. In log10-transformed linear regressions, B. rubens cover accounted for 68% to 96% of the variance in B. rubens biomass among sites, with all coefficients of determination significant at P < 0.05. For every doubling of percent cover, biomass was predicted to increase by 78%, 83%, and 144% among the three sites. At the site measured for three consecutive years, which ranged in rainfall from 65% to 159% of the long-term average, regression slopes each year differed from other years. Regression results among sites were insensitive to using cover classes (10 classes encompassing 0% to 100% cover) compared with simulated random distribution of integer cover within classes. Biomass of B. rubens was amenable to estimation in the field using cover, and such estimates may have applications for modeling invasive annual plant fuel loads and ecosystem carbon storage.