A frame-based spatially explicit model of subarctic vegetation response toclimatic change: comparison with a point model

An important challenge in global-change research is to simulate short-term transient changes in climate, disturbance regime, and recruitment that driv e long-term vegetation distributions. Spatial features (e.g., topographic b arriers) and processes, including disturbance propagation and seed dispersa l, largely control these short-term transient changes. Here we present a fr ame-based spatially explicit model (ALFRESCO) that simulates landscape-leve l response of vegetation to transient changes in climate and explicitly rep resents the spatial processes of disturbance propagation and seed dispersal . The spatial model and the point model from which it was developed showed similar results in some cases, but diverged in situations where interaction s among neighboring cells (fire spread and seed dispersal) were crucial. To pographic barriers had little influence on fire size in low-flammability ve getation types, but reduced the average fire size and increased the number of fires in highly flammable vegetation (dry grassland). Large fires were m ore common in landscapes with large contiguous patches of two vegetation ty pes while a more heterogeneous vegetation distribution increased fires in t he less flammable vegetation type. When climate was held constant for thous ands of years on a hypothetical landscape with the same initial vegetation, the spatial and point models produced identical results for some climates (cold, warm, and hot mesic), but produced markedly different results at cur rent climate and when much drier conditions were imposed under a hot climat e. Spruce migration into upland tundra was slowed or prevented by topograph ic barriers, depending on the size of the corridor. We suggest that frame-b ased, spatially explicit models of vegetation response to climate change ar e a useful tool to investigate both short- and long-term transients in vege tation at the regional scale. We also suggest that it is difficult to antic ipate when non-spatial models will be reliable and when spatially explicit models are essential. ALFRESCO provides an important link between models of landscape-level vegetation dynamics and larger spatio-temporal models of g lobal climate change.