A machine learning approach to identify barriers in stream networks demonstrates high prevalence of unmapped riverine dams.

Restoring stream ecosystem integrity by removing unused or derelict dams has become a priority for watershed conservation globally. However, efforts to restore connectivity are constrained by the availability of accurate dam inventories which often overlook smaller unmapped riverine dams. Here we develop and test a machine learning approach to identify unmapped dams using a combination of publicly available topographic and geospatial habitat data. Specifically, we trained a random forest classification algorithm to identify unmapped dams using digitally engineered predictor variables and known dam sites for validation. We applied our algorithm to two subbasins in the Hudson River watershed, USA, and quantified connectivity impacts, as well as evaluated a range of predictor sets to examine tradeoffs between classification accuracy and model parameterization effort. The random forest classifier achieved high accuracy in predicting dam sites (true positive rate = 89%, false positive rate = 1.2%) using a subset of variables related to stream slope and presence of upstream lentic habitats. Unmapped dams were prevalent throughout the two test watersheds. In fact, existing dam inventories underestimated the true number of dams by ∼80-94%. Accounting for previously unmapped dams resulted in a 62-90% decrease in dendritic connectivity indices for migratory fishes. Unmapped dams may be pervasive and can dramatically bias stream connectivity information. However, we find that machine learning approaches can provide an accurate and scalable means of identifying unmapped dams that can guide efforts to develop accurate dam inventories, thereby informing and empowering efforts to better manage them.

Environmental flows in the context of unconventional natural gas development in the Marcellus Shale.

Quantitative flow-ecology relationships are needed to evaluate how water withdrawals for unconventional natural gas development may impact aquatic ecosystems. Addressing this need, we studied current patterns of hydrologic alteration in the Marcellus Shale region and related the estimated flow alteration to fish community measures. We then used these empirical flow-ecology relationships to evaluate alternative surface water withdrawals and environmental flow rules. Reduced high-flow magnitude, dampened rates of change, and increased low-flow magnitudes were apparent regionally, but changes in many of the flow metrics likely to be sensitive to withdrawals also showed substantial regional variation. Fish community measures were significantly related to flow alteration, including declines in species richness with diminished annual runoff, winter low-flow, and summer median-flow. In addition, the relative abundance of intolerant taxa decreased with reduced winter high-flow and increased flow constancy, while fluvial specialist species decreased with reduced winter and annual flows. Stream size strongly mediated both the impact of withdrawal scenarios and the protection afforded by environmental flow standards. Under the most intense withdrawal scenario, 75% of reference headwaters and creeks (drainage areas <99 km2 ) experienced at least 78% reduction in summer flow, whereas little change was predicted for larger rivers. Moreover, the least intense withdrawal scenario still reduced summer flows by at least 21% for 50% of headwaters and creeks. The observed 90th quantile flow-ecology relationships indicate that such alteration could reduce species richness by 23% or more. Seasonally varying environmental flow standards and high fixed minimum flows protected the most streams from hydrologic alteration, but common minimum flow standards left numerous locations vulnerable to substantial flow alteration. This study clarifies how additional water demands in the region may adversely affect freshwater biological integrity. The results make clear that policies to limit or prevent water withdrawals from smaller streams can reduce the risk of ecosystem impairment.