Representing climate, disturbance, and vegetation interactions in landscape models

• Landscape models must simulate interactions to project climate change responses.
• Suggestions for including interactions into landscape models are given.
• Models should be spatial, mechanistic, multi-scaled, and parsimonious.
• Landscape models need better data, software platforms, and testing methods.

The prospect of rapidly changing climates over the next century calls for methods to predict their effects on myriad, interactive ecosystem processes. Spatially explicit models that simulate ecosystem dynamics at fine (plant, stand) to coarse (regional, global) scales are indispensable tools for meeting this challenge under a variety of possible futures. A special class of these models, called landscape models (LMs), simulates dynamics at intermediate scales where many critical ecosystem processes interact. The complicated dependencies among climate, disturbance, and vegetation present a difficult challenge for LMs, however, because their simulation must reconcile processes and their interactions that occur at different spatial and temporal scales. In the absence of these interactions, key thresholds in ecosystem responses to changes in climate may go undetected or misrepresented. In this paper, we present a general strategy for constructing the next generation of LMs that ensures that interactions are modeled at appropriate scales of time and space, and that, when possible, processes representing these interactions are simulated mechanistically. We identify six key questions to frame this strategy and then provide guidance and possible solutions on the structure and content needed in future LMs to ensure that climate-vegetation-disturbance interactions are incorporated effectively.