Citation:

Figary, S., N. Detenbeck, AND C. O'Donnell. Guiding riparian management in a transboundary watershed through high resolution spatial statistical network models. JOURNAL OF ENVIRONMENTAL MANAGEMENT. Elsevier Science Ltd, New York, NY, 278(2):111585, (2021). https://doi.org/10.1016/j.jenvman.2020.111585

Impact/Purpose:

The Atlantic salmon population has been nearly decimated in the Saint John River, a larger international watershed straddling the border of Maine and New Brunswick, Canada. In May 2017, Federal agencies including US EPA and Canadian Federal departments signed a statement of cooperation with the Maliseet Tribal/First Nation leaders to affirm their mutual objective to cooperate in the restoration of the portion of the Walastook/Saint John River that lies in the state of Maine and in the province of New Brunswick. It is important to maintain and restore coldwater habitats in streams to support coldwater fish such as the Atlantic salmon. In 2017 EPA published a statistical model to describe the factors affecting the temperature regime of streams and rivers in New England. Due to data constraints, that model did not include portions of northern Maine watersheds that crossed the border with Canada. More recently, data have become available allowing the extension of the New England Stream Temperature model to international watersheds such as the Meduxnekeag tributary to the Saint John River. This effort allowed refinement of the model to a finer spatial scale to inform the potential effects of reforestation of riparian zones along the river. We were able to demonstrate and evaluate some new refinements to SSN temperature modeling in this project. We started by identifying gaps in available monitoring data and demonstrated an efficient method for identifying new sampling locations for the Maliseet tribe to fill in those monitoring gaps. We were able to demonstrate the high quality of models that can be developed using the new high resolution NHDPlus stream line geospatial data set (r2 = 0.88 – 0.96, root mean square prediction error < 1 degree C) and at a spatial resolution that can meaningfully inform riparian buffer management. Using the shade.xls spreadsheet model, we were able to estimate different indicators of shaded solar radiation based on differing buffer widths (30 meters vs 90 meters) and lengths of influence upstream. The monthly median temperature models performed best with shaded solar radiation aggregated over a 3 hour travel time upstream, while the growing season daily maximum prediction models performed best with shaded solar radiation aggregated over a 1 km upstream buffer. In evaluating restoration scenarios, we were able to demonstrate that most of the shading benefit would accrue from restoration of a 30 meter buffer width, with some (but much less) additional benefit of restoring up to 90m buffer width. We were also able to illustrate important differences in thermal regime distribution between a relatively wet (2011) and dry (2010) year, showing the importance of maintaining base flow for protection of thermal regimes, e.g., through use of stormwater green infrastructure practices. We were also able to illustrate the critical significance of restoring headwater reaches to provide refuges for coldwater species during times of thermal stress (see 30m restoration scenario as compared to max growing season temperatures). Results will be used by the Maliseet tribe in their watershed restoration activities to support existing brook trout populations and for restoration of Atlantic salmon.

Description:

The United States Environmental Protection Agency and the Houlton Band of Maliseet Indians (HBMI) built a stream temperature spatial statistical network (SSN) model for the Meduxnekeag Watershed. The headwaters of the Meduxnekeag Watershed are in Maine, United States of America and the outlet is in New Brunswick, Canada, creating an additional challenge because many datasets are constrained to political boundaries. The release of the High-Resolution National Hydrology Dataset Plus included transboundary watersheds and enabled creation of fine resolution (1:24,000) SSN temperature models consistent with management scales for riparian buffers. SSN models were developed for July, August, and September median stream temperatures and the growing season maximum (GSM). Fitted SSN models had relatively high R2 values (0.88–0.96) and all final models included significant parameters for shade-attenuated solar radiation, reference flow, air temperature, and bankfull depth or width. Fitted models predicted stream temperatures during a dry (2010) and wet (2011) year. Monthly models predicted the fewest cold water (99% cold water reaches, and August results were intermediate between July and September. GSM predictions found 81% of stream reaches could not support salmonid survival (>27.0 °C) in the dry year and 59% of the reaches were warmwater (22.5–27.0 °C) in the wet year. The model was used to predict stream temperatures following restoration scenarios of a forested 30-m or 90-m buffer of stream segments bordered by agricultural or developed land. The restoration scenarios expanded cold water habitat based on monthly median temperatures and decreased the habitat area with GSM above survival thresholds, with little difference in effectiveness of the two buffer widths. These results will guide riparian restoration projects by the HBMI to expand habitat for cold water fishes.

URLs/Downloads:

Record Details: