Tuesday, August 3, 2010: 10:10 AM
412, David L Lawrence Convention Center
Background/Question/Methods
Understanding the possible responses and vulnerability of vegetation to climate change is a critical issue facing public and private land management agencies and land conservation organizations. A common strategy for estimating the potential future geographic distribution of vegetation is to use either mechanistic or empirical models. Mechanistic models are based on physiologically-based algorithms and rules that relate to the tolerance of a particular species or vegetation type (DGVMs) to changing environmental conditions. Correlative or statistical species distribution models (SDMs) estimate the environmental conditions that are suitable by associating known species presence with a suite of environmental variables that are expected to drive species presence and abundance and coincide with physiological tolerances. Models assumptions and structure contribute to the uncertainty of the projections. While there have been numerous comparisons among statistical modeling approaches in the literature, comparisons that include mechanistic models are few. We illustrate the possible responses of vegetation to climate change using two statistically based species distribution models (Maxent and BRT) and a dynamic global vegetation model (MC1) under four GCM/emission scenario combinations.
Results/Conclusions
We compare and contrast model output across two landscapes, one in eastern Oregon and one in northern Arizona. While certain projections among the models are strikingly consistent, important differences are noted. We specifically discuss the issues of uncertainty and the usability of our combined results by managers in terms of conservation and to develop effective adaptation strategies.
Understanding the possible responses and vulnerability of vegetation to climate change is a critical issue facing public and private land management agencies and land conservation organizations. A common strategy for estimating the potential future geographic distribution of vegetation is to use either mechanistic or empirical models. Mechanistic models are based on physiologically-based algorithms and rules that relate to the tolerance of a particular species or vegetation type (DGVMs) to changing environmental conditions. Correlative or statistical species distribution models (SDMs) estimate the environmental conditions that are suitable by associating known species presence with a suite of environmental variables that are expected to drive species presence and abundance and coincide with physiological tolerances. Models assumptions and structure contribute to the uncertainty of the projections. While there have been numerous comparisons among statistical modeling approaches in the literature, comparisons that include mechanistic models are few. We illustrate the possible responses of vegetation to climate change using two statistically based species distribution models (Maxent and BRT) and a dynamic global vegetation model (MC1) under four GCM/emission scenario combinations.
Results/Conclusions
We compare and contrast model output across two landscapes, one in eastern Oregon and one in northern Arizona. While certain projections among the models are strikingly consistent, important differences are noted. We specifically discuss the issues of uncertainty and the usability of our combined results by managers in terms of conservation and to develop effective adaptation strategies.