Using a patch dynamics approach to model cottonwood forest populations in river floodplains

Background/Question/Methods River floodplains are dynamic environments where strong physical and abiotic forces interact. In large, alluvial river systems, periodic flooding drives channel meandering and cutoff events, which set the physical template for the establishment of biotic communities. As a result, the riparian ecosystem in these settings is frequently composed of a mosaic of different vegetation types of various ages. Predicting the occurrence and dynamics of these communities is a particular challenge, both because of the stochastic nature of the flood disturbance regime and the local variability inherent within in. Pioneer riparian communities are particularly sensitive to both the favorable resource conditions provided by floods and by their strong effects on plant mortality. Niche models that use spatial factors such as land elevation or time since floodplain creation adequately predict the occurrence of vegetation types that are not dependent on disturbance, but typically overestimate suitable habitat area for pioneer forest, and hence, its actual landscape distribution. This is not because of a poor understanding of their habitat needs, dispersal limitation, or specialized pests. Rather, niche models do not adequately quantify the temporal and spatial variability of resources and mortality agents that drive seedling recruitment, which is the limiting life stage for riparian tree populations. Furthermore, niche models fail to predict vegetation distributions especially in abandoned channels and other complex floodplain habitats in which the physical processes that drive landscape pattern are not included in the models. Because pioneer riparian trees are keystone species in young riparian ecosystems, understanding the specific processes that drive their recruitment is critical for predicting the effects of changes to the disturbance regime from flow regulation, climate change, and landscape modifications. To address these needs, we are taking a mechanistic, patch-dynamic approach to predict the occurrence and population dynamics of Fremont cottonwood (Populus fremontii), a pioneer tree in riparian systems throughout semi-arid regions of North America. The model couples a physical model of channel meandering and cutoff with a stage-structured population model with transition probabilities that dynamically model the climatic, hydrologic, and topographic factors critical for tree recruitment. Results/Conclusions TBD