Glacier retreat creating new Pacific salmon habitat in western North America

Glacier retreat poses risks and benefits for species of cultural and economic importance. One example is Pacific salmon (Oncorhynchus spp.), supporting subsistence harvests, and commercial and recreational fisheries worth billions of dollars annually. Although decreases in summer streamflow and warming freshwater is reducing salmon habitat quality in parts of their range, glacier retreat is creating new streams and lakes that salmon can colonize. However, potential gains in future salmon habitat associated with glacier loss have yet to be quantified across the range of Pacific salmon. Here we project future gains in Pacific salmon freshwater habitat by linking a model of glacier mass change for 315 glaciers, forced by five different Global Climate Models, with a simple model of salmon stream habitat potential throughout the Pacific Mountain ranges of western North America. We project that by the year 2100 glacier retreat will create 6,146 (±1,619) km of new streams accessible for colonization by Pacific salmon, of which 1,930 (±569) km have the potential to be used for spawning and juvenile rearing, representing 0 to 27% gains within the 18 sub-regions we studied. These findings can inform proactive management and conservation of Pacific salmon in this era of rapid climate change.


nature research | reporting summary
April 2020 Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
Our study projects the extent of future gains in Pacific salmon freshwater habitat throughout the Pacific mountain ranges of western North America. We linked a model of glacier mass change for almost 600 glaciers, forced by five different Global Climate Models, with a simple model of salmon stream habitat potential across a 623,000 km2 study region. We then quantified where throughout our study region would experience gains in salmon habitat for the years 2050 and 2100. The design structure is factorial.
DEMs used in this analysis can be downloaded from NASA (https://asterweb.jpl.nasa.gov/gdem.asp) or Shuttle Rada Topography Mission (SRTM) with a spatial resolution of~30m and vertical resolution of~¬¬+ 5 m. Watershed boundary data for the USA were sourced from USGS (https://www.usgs.gov) and from the Freshwater atlas of BC for British Columbia (https://www2.gov.bc.ca/gov). Glacier outline data was obtained from the Glacier Inventory v6.0 (https://www.glims.org/RGI/). Pacific salmon presence data were available from the Anadromous Waters Catalogue (AWC; www.adfg.alaska.gov/sf/SARR/AWC). Present-day rivers were represented by building a synthetic stream network using ArcGIS 10.6 and previously mentioned DEMs. Ice thickness distribution was calculated at a grid resolution of 25 to 200 m (depending on glacier area) using a simple dynamic model that considers glacier mass turnover and ice flow mechanics, and by inverting the glaciers' surface topography. For Glacier retreat data, we used the Global Glacier Evolution Model (GloGEM), which computes glacier mass balance and associated geometry changes for each individual glacier in the study region. GloGEM is forced with temperturea and precipitation time series from an ensemble of five Global Climate Models (GCMs). The five GCM included: CanESM2, CSIRO-Mk3-6-0, GFDL-CM3, MIROC-ESM, and MPI-ESM-LR, and we presented projections from Representative Concentration Pathways (RCPs): RCP4.5 and RCP8.5, which correspond to plausible scenarios for the rate of change in the concentration of atmospheres CO2 and other greenhouse gas emissions.
We used spatial analysis to determine our findings. Therefore, did not use any statistical methods to predetermine sample size. Our sample size was determined by the total number of glaciers within the study region (n =~300), and the number of subregions (n= 18) within our study region. We considered glacier retreat and ice thickness for~300 glaciers within the Pacific mountain ranges of North American. We summarized these findings by 18 subregions, which were selected based on their watershed delineations, and where there were known salmon present in these watersheds. All watersheds that contained less than 1.5% glacier cover were included in our analysis. We ran 5 GCM model projections to determine timing of glacier retreat. These 5 GCM were chosen based on previous literatures assessments of most suitable projections for western North America.
There are no start and stop dates associated with the data collection as the data were sourced from Open sources. However, the glacier outline data refer roughly to the years 2009 + 2 for Alaska, and 2004 + 5 for Western Canada. All data obtained from open sources range from southern British Columbia, Canada to Alaska, USA.
There were no data excluded from the analysis.
We did not conduct any experimental designs, therefore do not have reproducibility to verify. We did run sensitivity analyses on stream segment lengths and total number of glaciers selected in our analyses. This was conducted to determine the error associated with our findings. In our analysis, we determined the total number of glaciers accessible as well as future salmon-accessible stream kilometers for salmon by using stream segment lengths of~500m, but also ran a sensitivity analysis using segment lengths of 250m, 400m, 600m and 750m for glacier selection and 250m and 750m segment length for future stream kilometers.
We did not need to consider randomization given that we did not conduct a statistical analysis but used spatial analysis instead. We did assess future stream kilometer gains for salmon across 18 sub regions, which were determined using the delineation of watershed boundaries.
We did not need to consider blinding, as we did not conduct our analysis using an experimental design.