Can big data inform invasive dreissenid mussel risk assessments of habitat suitability?

Invasion risk assessments of habitat suitability provide insight on early detection effort allocation; however, sufficient data are rarely available to inform assessments. We explored tradeoffs of leveraging big data from the National Water Quality Portal (WQP), a standardized water quality database in the United States, to inform calcium- and pH-based risk assessments of invasive mussel (Dreissena polymorpha and Dreissena rostriformis bugensis) habitat suitability in the Pacific Northwest's Columbia River Basin. We evaluated risk assessment sensitivity to alternative metrics of summarizing WQP data, tested if the large number of WQP observations resulted in accurate risk prediction of sites lacking WQP data, and characterized the spatial distribution of suitable habitat. Risk assessments were insensitive to how data were summarized at a site. Predictive accuracy was low when interpolating risk to sites lacking data. High-risk sites based on calcium clustered in two water basins, whereas high-risk sites based on pH were at similarly high frequencies. Finally, we found that data gaps still exist in the Columbia River Basin despite the large volumes of WQP data. We conclude that WQP-based risk assessments of habitat suitability could be considered as a starting place for estimating dreissenid invasion risk within an adaptive framework, rather than as a final solution.