Ecological niche modeling is quickly becoming a prevalent technique in the attempt to predict species’ ranges. Most standard GIS-based niche models use a limited set of presence-absence data combined with coarse resolution environmental variables to identify and map suitable habitat on a regional spatial scale. We propose a system in which very fine-scale presence-absence data permits the use of more restrictive environmental factors and hence a more powerful predictive model. Using high resolution infra-red aerial photography for the state of Indiana we generated a highly detailed species range map for individual populations of eastern hemlock (Tsuga canadensis). Eastern hemlock is an uncommon species in Indiana, occurring only in small, dense stands that are disjunct from the main range and often widely separated from each other. Our species range map was used as the basis for presence-absence data in a predictive GIS-based model in which we use fine-scale environmental factors which were previously considered too variable to incorporate. This type of model is a potentially powerful resource for ongoing conservation efforts with a species threatened by range-wide extirpation by the hemlock woolly adelgid, a non-native insect pest.

Results/Conclusions

E.B. Little’s range maps, the standard for eastern tree species range data, indicate 8 regions of eastern hemlock habitat within the state of Indiana that total 3,570.38 km² in area. Our technique allowed us to create a map which greatly reduces this estimate of potential habitat by accurately portraying the precise locations of existing eastern hemlock stands. We identified 203 unique stands of eastern hemlock with a total area of 3.13 km² (0.0878% of Little’s area). In addition, 45.6% of the area of our range map is located outside of the boundaries delimited by Little’s map. Presence-absence data at this level of precision has permitted us to use fine-scale environmental factors, such as soil type, slope, aspect, and proximity to water, in the development of an ecological niche model which predicts potential, but currently unoccupied, habitat on a fine spatial scale. Our results show the value of high resolution aerial photography as baseline data for ecological niche modeling and the feasibility of fine-scale predictive habitat modeling.