Highways and hubs: connectivity for freshwater conservation under climate change

Ian M. McCullough, Katelyn B. S. King and Patrick J. Hanly, Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI, Tyler Wagner, Pennsylvania Cooperative Fish and Wildlife Research Unit, U.S. Geological Survey, Pennsylvania State University, University Park, PA

Background/Question/Methods
Enhancing landscape connectivity has long been a key conservation recommendation for promoting gene flow, migrations, population viability and species range shifts under climate change. Most connectivity studies, however, focus on terrestrial biodiversity and ecosystems, despite the fact that freshwater biodiversity is more threatened worldwide. Furthermore, it is unknown to what extent protected areas support freshwater movement corridors. We analyzed LAGOS-US-NETWORKS, a novel dataset of all 898 lake networks within the conterminous US (representing 18% of US lakes ≥ 1 ha), to ask: 1) How can we describe structural connectivity of freshwater networks within the conterminous US?, 2) How sensitive are freshwater networks to potential disruptions? and 3) How well protected are freshwater networks and highly connected lakes within freshwater networks? We performed a national-scale analysis on a series of graph-based network metrics to identify major types of freshwater networks. We also identified “hub” lakes that are disproportionately important for whole-network connectivity, based on upper quantiles of metrics of network strength and within-network centrality. Finally, we analyzed current protection for networks and hub lakes using the US Protected Areas Database according to strict (i.e., managed for biodiversity) and multi-use (i.e., managed for biodiversity and natural resource extraction) protection.
Results/Conclusions
Freshwater networks contained from 2 to 32811 lakes (median of 3 lakes) and spanned 0 to 1330 km (median of 6 km) north-south, indicating that most lake networks are relatively small. Across all networks, only approximately 1-2.5 network disruptions were enough to undermine maximum north-south connectivity, suggesting that relatively small breaks can have large impacts on large-scale connectivity. We identified 1662 highly connected hub lakes, representing less than 0.01% of US lakes ≥ 1 ha. Approximately 88% and 77% of hub lakes have less than 10% of their watersheds under strict and multi-use protection, respectively. These results indicate that the relatively few lakes disproportionately important for large-scale connectivity in the conterminous US are currently not well protected. Further, just 17-53% of freshwater networks are greater than 10% protected, depending on protection level. Given the high prevalence of small lake networks, many such networks may eventually represent climatic dead ends and may require active management to maintain local- to landscape-scale biodiversity. Conversely, conservation planning efforts might consider focusing on maintaining the relatively few large networks intact to support metapopulations for wide-ranging species and future species range shifts.