The Modern Era of “Big Data” in GIS: Multi-Scale Modeling of Species Distributions, Hydrology, and Gene Flow

Spatial modeling projects in ecology and hydrology often seek predictions or functional relationships without specific a priori knowledge of the best predictors or most appropriate spatial scales.  Unprecedented availability of GIS and remote sensing data allows for seemingly endless combinations of landscape characteristics and spatial scales to be analyzed.  This provides for exciting new research opportunities, but also new challenges in the collection and analysis of spatial data.  To overcome workflow bottlenecks associated with acquiring and extracting GIS data at appropriate spatial scales, a centralized national geodatabase and automated data collector—Geodata Crawler—is being developed to rapidly tabulate customized multi-scale landscape data (i.e. from watersheds, riparian zones, or stream paths).  To accommodate potentially unwieldy datasets, multi-model comparison, machine learning, and ensemble methods will be discussed as analytical solutions.  Three ongoing projects provide examples that have utilized various spatial scales and modeling approaches to address specific research questions: (1) Habitat associations and key spatial scales for conservation of Arkansas’ endemic Sulphur Springs diving beetle; (2) Predicting natural flow regimes and degrees of hydrologic alteration at un-gauged streams; (3) Identifying barriers to gene flow among populations of Bluehead Sucker, a sensitive fish in the Colorado River basin.