Technical Abstract: Despite the dedication of significant human, financial and technological resources, dryland degradation continues unabated in both the developed and developing world. Many of the causes have been described and the consequences extensively documented. Here, we argue that the failure of attempts to stop or reverse dryland degradation can be explained by a failure to 1) recognize when ecosystems have crossed ecological, edaphic o geomorphic thresholds, and 2) identify and address the properties and/or processes at relevant scales that confer resistance and resilience. We illustrate how simulation models can be used to address these limitations using examples from one type of model, an individual-based gap dynamics model of grasslands and shrublands (ECOTONE). We used ECOTONE to predict the effects of climatic fluctuations and disturbance frequency on local and regional patterns in species dominance and composition. Results show that patterns in dominance between 2 perennial grasses (Bouteloua gracilis, Bouteloua eriopoda) at the biome transition zone are not predictable based on responses within each biome. An increase in disturbance frequency at the ecotone shifted plant communities to dominance by the short-lived B. eriopoda from communities codominated by B. gracilis or the shrub Larrea tridentata. We propose a strategy that addresses problems in semiarid and arid ecosystems more effectively than previous approaches based on an integrated, hierarchical modeling approach. This approach employs multiple tools, such as geographic information systems, state and transition models, remote sensing, and expert knowledge, in an iterative approach with simulation models. This strategy is applicable to research on processes of dryland degradation and management projects to stop or reverse degradation.