Fuel consumption predictions are necessary to accurately estimate or model fire effects, including pollutant emissions during wildland fires. Fuel and environmental measurements on a series of operational prescribed fires were used to develop empirical models for predicting fuel consumption in big sagebrush (Artemisia tridentata Nutt.) ecosystems. Models are proposed for predicting fuel consumption during prescribed fires in the fall and the spring. Total prefire fuel loading ranged from 5.3-23.6 Mg*ha-1; between 32% and 92% of the total loading was composed of live and dead big sagebrush. Fuel consumption ranged from 0.8-22.3 Mg*ha-1, which equates to 11-99% of prefire loading (mean=59%). Model predictors include prefire shrub loading, proportion of area burned, and season of burn for shrub fuels (R^2=0.91). Models for predicting proportion of area burned for spring and fall fires were also developed (R^2=0.64 and 0.77 for spring and fall fire models, respectively). Proportion of area burned, an indicator of the patchiness of the fire, was best predicted from the coverage of the herbaceous vegetation layer, wind speed, and slope; for spring fires, day-of-burn 10-h woody fuel moisture content was also an important predictor variable. Models predicted independent shrub consumption measurements within 8.1% (fall) and 12.6% (spring) for sagebrush fires.