Linking spatial data and predictive models to forecast alternative options and futures in managed landscapes

Knowing where species occur is elemental to any management plan developed to protect these species. Habitat prediction models, developed from statistical relationships between species observations and bio-physical characteristics of a landscape, allow managers to target landscape features rather than rely on sparse observation records in prioritizing areas for protection. Many bio-physical influences, such as aspect, slope, and elevation, are static, but others such as nearby land use, upstream water quality/quantity, and proximity to other protected areas can be altered by human behavior. This implies that a proper conservation area design must account not only for where species are now, but where they are expected to be given expected landscape changes. We combine iterative simulation techniques and graph theory to develop a spatially explicit model that forecasts the habitat impacts of various landscape reconfigurations in the Great Basin. We use this model to show the impacts of water drawdown, of wildfire, and of human development on bird and butterfly distributions and habitat connectivity.