Approaches for simulating alternative futures of complex forested landscapes: A review

Certain aspects of the computational methods that can be employed to simulate the development and change of managed and natural landscapes, where the disparate interests of multiple landowner groups should be recognized, are challenging for modelers. Four bibliographic databases were queried using several key phrases related to this topic. Reasonable modeling approaches exist that recognize and emulate landowner behavior through transition probabilities informed through sampling or statistical models, or through knowledge gained by communicating with landowner stakeholders. Assumptions regarding both spatial extent and spatial resolution relate directly to data storage requirements and the capacity of a model to accommodate the desired simulations. The agility of a landscape model to produce information suitable for comparing alternative scenarios depends on the flexibility of search parameters and the capability of the data to adequately represent alternative future states. Verification processes and statistical tests are used to support the credibility of simulated outcomes, as errors and associated uncertainty (random and process-related) can arise based on the data employed and how models are developed. Realistic modeling of landscape sustainability may require integration of natural processes and socio-economic concerns, although often this scope of analysis is lacking or limited. Although there are many options for modeling landscape change, there is no perfect model for addressing all potential future scenarios, and compromises will be made to address the accuracy of data and uncertainty inherent in projected outcomes.