We employed decision-tree mapping models in two formats to establish a time series (2001 - 2015) of sagebrush condition class in the western United States. The formats were predictive and descriptive, and each model produced distinct spatially explicit datasets. The predictive model mapped the probability of sagebrush recovery, tipping point (environmental degradation), or stable classes. The descriptive model mapped rules that were defined by environmental thresholds. The thresholds were defined by the interaction between the independent variables and the dependent variable. Mapping areas of stability and areas of change using machine-learning algorithms allows both the identification of dominant abiotic variables that drive ecosystem dynamics and the variables’ important thresholds. The sagebrush recovery class, on average, covers areas of mid elevations (1602 m) for sagebrush ecosystems, warm 30-yr July maximum temperatures, and 30-yr March precipitation averages equal to 26.26 mm. The tipping point class covers areas with a mean elevation value about 100 m lower than the recovery class. The stable class occurs at the highest elevations of all classes, averaging 1939 m. Both the tipping point and the stable classes were more mesic in March and cooler in July than the recovery class. These defined variable averages can be used to understand current dynamics of sagebrush condition and to predict where future transitions may occur under changing conditions.