Monitoring of rangelands poses significant challenges to land managers due to broad extent and many uses of rangelands. The Bureau of Land Management’s (BLM) Assessment, Inventory, and Monitoring (AIM) program seeks to efficiently collect standard, quantitative monitoring data which is collected once and used multiple times to address a range of management questions. However, the cost of collecting field data to sufficiently monitor landscapes can be prohibitive. Remote sensing image classification provides an opportunity to derive rangeland indicators over large areas and at more frequent intervals than field data collection campaigns. We demonstrate a hierarchical Bayesian approach for maximizing the efficiency of monitoring data collection where existing AIM field plots are used to a) train remote sensing image classification and b) provide finer resolution data in areas of management concern. Where available, high spatial resolution images (e.g., RapidEye) are classified and the results of this classification are used together with field data to improve the moderate spatial resolution image (e.g., Landsat OLS) classification. Results of our classification of 2.7 million acres in northern California show remote sensing-derived estimates of bare ground, shrub cover, and herbaceous cover are the most accurate but other indicators, such as soil aggregate stability, plant density, and vegetation species composition, are best estimated using field collected data. Utilizing AIM field data together with remote sensing enables a strategic rangeland monitoring approach where field data meet multiple monitoring objectives. Combining multiple types of data improves estimates of indicators thus providing higher quality monitoring data to rangeland managers.