Geology topics

Michael T. Slattery

Publications and source records attributed to Michael T. Slattery.

3 recordsLinked to original sources

USGS2016

FishVis, A regional decision support tool for identifying vulnerabilities of riverine habitat and fishes to climate change in the Great Lakes Region

Climate change is expected to alter the distributions and community composition of stream fishes in the Great Lakes region in the 21st century, in part as a result of altered hydrological systems (stream temperature, streamflow, and habitat). Resource managers need information and tools to understand where fish species and stream habitats are expected to change under future conditions. Fish sample collections and environmental variables from multiple sources across the United States Great Lakes Basin were integrated and used to develop empirical models to predict fish species occurrence under present-day climate conditions. Random Forests models were used to predict the probability of occurrence of 13 lotic fish species within each stream reach in the study area. Downscaled climate data from general circulation models were integrated with the fish species occurrence models to project fish species occurrence under future climate conditions. The 13 fish species represented three ecological guilds associated with water temperature (cold, cool, and warm), and the species were distributed in streams across the Great Lakes region. Vulnerability (loss of species) and opportunity (gain of species) scores were calculated for all stream reaches by evaluating changes in fish species occurrence from present-day to future climate conditions. The 13 fish species included 4 cold-water species, 5 cool-water species, and 4 warm-water species. Presently, the 4 cold-water species occupy from 15 percent (55,000 kilometers [km]) to 35 percent (130,000 km) of the total stream length (369,215 km) across the study area; the 5 cool-water species, from 9 percent (33,000 km) to 58 percent (215,000 km); and the 4 warm-water species, from 9 percent (33,000 km) to 38 percent (141,000 km). Fish models linked to projections from 13 downscaled climate models projected that in the mid to late 21st century (2046–65 and 2081–2100, respectively) habitats suitable for all 4 cold-water species and 4 of 5 cool-water species under present-day conditions will decline as much as 86 percent and as little as 33 percent, and habitats suitable for all 4 warm-water species will increase as much as 33 percent and as little as 7 percent. This report documents the approach and data used to predict and project fish species occurrence under present-day and future climate conditions for 13 lotic fish species in the United States Great Lakes Basin. A Web-based decision support mapping application termed “FishVis” was developed to provide a means to integrate, visualize, query, and download the results of these projected climate-driven responses and help inform conservation planning efforts within the region.

Great Lakes Region↗

USGS2014

Development of a spatially universal framework for classifying stream assemblages with application to conservation planning for Great Lakes lotic fish communities

Classifications are typically specific to particular issues or areas, leading to patchworks of subjectively defined spatial units. Stream conservation is hindered by the lack of a universal habitat classification system and would benefit from an independent hydrology-guided spatial framework of units encompassing all aquatic habitats at multiple spatial scales within large regions. We present a system that explicitly separates the spatial framework from any particular classification developed from the framework. The framework was constructed from landscape variables that are hydrologically and biologically relevant, covered all space within the study area, and was nested hierarchically and spatially related at scales ranging from the stream reach to the entire region; classifications may be developed from any subset of the 9 basins, 107 watersheds, 459 subwatersheds, or 10,000s of valley segments or stream reaches. To illustrate the advantages of this approach, we developed a fish-guided classification generated from a framework for the Great Lakes region that produced a mosaic of habitat units which, when aggregated, formed larger patches of more general conditions at progressively broader spatial scales. We identified greater than 1,200 distinct fish habitat types at the valley segment scale, most of which were rare. Comparisons of biodiversity and species assemblages are easily examined at any scale. This system can identify and quantify habitat types, evaluate habitat quality for conservation and/or restoration, and assist managers and policymakers with prioritization of protection and restoration efforts. Similar spatial frameworks and habitat classifications can be developed for any organism in any riverine ecosystem.

Michigan, Minnesota, New York, Ohio, Wisconsin↗

USGS2013

Broad-scale patterns of Brook Trout responses to introduced Brown Trout in New York

Brook Trout Salvelinus fontinalis and Brown Trout Salmo trutta are valuable sport fish that coexist in many parts of the world due to stocking introductions. Causes for the decline of Brook Trout within their native range are not clear but include competition with Brown Trout, habitat alteration, and repetitive stocking practices. New York State contains a large portion of the Brook Trout's native range, where both species are maintained by stocking and other management actions. We used artificial neural network models, regression, principal components analysis, and simulation to evaluate the effects of Brown Trout, environmental conditions, and stocking on the distribution of Brook Trout in the center of their native range. We found evidence for the decline of Brook Trout in the presence of Brown Trout across many watersheds; 22% of sampled reaches where both species were expected to occur contained only Brown Trout. However, a model of the direct relationship between Brook Trout and Brown Trout abundance explained less than 1% of data variation. Ordination showed extensive overlap of Brook Trout and Brown Trout habitat conditions, with only small components of the hypervolume (multidimensional space) being distinctive. Subsequent analysis indicated higher abundances of Brook Trout in highly forested areas, while Brown Trout were more abundant in areas with relatively high proportions of agriculture. Simulation results indicated that direct interactions and habitat conditions were relatively minor factors compared with the effects of repeated stocking of Brown Trout into Brook Trout habitat. Intensive annual stocking of Brown Trout could eliminate resident Brook Trout in less than a decade. Ecological differences, harvest behavior, and other habitat changes can exacerbate Brook Trout losses. Custom stocking scenarios with Brown Trout introductions at relatively low proportions of resident Brook Trout populations may be able to sustain healthy populations of both species within their present range.

New York↗