Mountain systems are characterized by strong environmental gradients, rugged topography and extreme spatial heterogeneity in ecosystem structure and composition. Consequently, most mountainous areas have relatively high rates of endemism and biodiversity, and function as species refugia in many areas of the world. Mountains have long been recognized as critical entities in regional climatic and hydrological dynamics but their importance as terrestrial carbon stores has only been recently underscored (Schimel et al. 2002; this volume). Mountain ecosystems, therefore, are globally important as well as unusually complex. These ecosystems challenge our ability to understand their dynamics and predict their response to climatic variability and global-scale environmental change.

To meet this challenge, mountain scientists increasingly are modeling the vast array of relationships that comprise ecosystem dynamics. Dynamic modeling can examine the interactions between land management strategies and climatic change to develop appropriate responses to future human demands on mountain systems. Modeling provides spatially and temporally explicit, quantified results that can be validated in the field, thus providing feedback to our understanding of ecosystem dynamics. Modeling results, particularly maps and other visual tools, also give a concrete dimension to our understanding of the scale and magnitude of potential future changes. Modeling alerts scientists and land managers to apparently counter-intuitive outcomes of ecosystem responses to climate change or management decisions. For instance, in an early modeling exercise for northwest Montana, USA, Running and Nemani (1991) found that streamflow in a warmer future climate decreased by 30% in the Swan Range even when precipitation was increased by 10% in a particular climate change scenario. This unexpected response was due to enhanced forest growth, and increased evapotranspiration, resulting from the earlier snowmelt and extended growing season. There is a rich legacy of models that address climate and weather, hydrology, forest growth (e.g. gap dynamics and succession), forest fires (e.g. fuel loading) and land cover change (cf. Bugmann et al., this volume). Much less common, however, are attempts to fully integrate models from various disciplines to create a robust system that adequately addresses the entire range of ecosystem dynamics. In addition, fine-resolution modeling of entire mountain ranges (i.e. regional ecosystem scale) is not as common as global or continental scale modeling or watershed/catchment scale modeling. However, this is the scale that is germane to policy decisions such as in the western US and Canada, i.e. in those areas that contain most of the mountainous terrain of North America. This paper describes our efforts to implement an integrated regional modeling approach while characterizing potential future responses of a mountain ecosystem to climate change. Our study area was Glacier National Park in northwestern Montana, USA. Glacier Park is a 4082 km” mountain wilderness that straddles the continental divide and contains over 150 summits of up to 3150 m elevation in the Lewis and Livingston mountain ranges.