The iconic ecosystems of western Washington, including the Olympic Peninsula, Puget Sound, and North Cascades, support diverse and abundant ecosystems, species and habitats, and provide a range of natural resources and services for human communities. However, the new and variable conditions that are emerging due to rapid climate change are expected to significantly alter the natural systems that wildlife and human communities both depend on. In the face of mounting climate and current land-use stressors, it is critical that we develop climate change adaptation strategies that help sustain Washington’s unique ecosystems, species and habitats.

To help address this challenge, components of the Yale Framework were used to structure an investigation of how and where important aquatic freshwater and coastal marine ecosystem habitats from the coastal and inland mountains to the Puget Sound (Western Washington) are likely to be affected by climate change. Changes in coastal habitats due to sea level rise and the potential impact to existing biodiversity hotspots, as well as the influence of land use – namely shoreline armoring and projected development – on more resilient coastal habitats were also considered. We are currently working with state agency and non-profit organization partners in the region to use these maps to prioritize conservation opportunities and actions for each ecosystem, as well as inform adaptation planning across ecosystem types.

Objectives

1. Identify watersheds likely to be more or less impacted by changing climate conditions
2. Identify watersheds likely to continue to harbor focal fish species
3. Assess how projected changes in wildfire regimes could potentially exacerbate hydrologic changes in watersheds

Geographic Location

Puget Sound lowlands in Western WA

Principal Investigator

Jessi Kershner

Ecosystem Type

Freshwater and coastal marine ecosystems

Framework focus

• Protect current patterns of biodiversity
• Protect future patterns of biodiversity
• Maintain ecological functions and processes
• Protect climate refugia

EcoAdapt focused their assessment of the Yale Framework on freshwater and coastal marine ecosystems of the Puget Sound region. This project relied on existing spatial datasets for the region – both climate and ecological – to create innovative, comparative maps. Geographic Information System (GIS) tools were used to relate multiple spatial and temporal data features (historic, current, and modeled) in order to identify areas of conservation priority. Spatial climate data for freshwater ecosystems were used to map current and future patterns in climate to identify freshwater watersheds likely to remain relatively stable in terms of summer thermal stress and winter flooding stress. To locate the watersheds most likely to continue to harbor species in the future, focal species distributions were comparatively mapped with future climate patterns. Using an MC1 dynamic vegetation model developed by the Pacific Northwest Research Station, projected fire patterns were also compared with future stable watersheds to identify stable areas likely to be impacted. Sea Level Affecting Marshes Model (SLAMM) maps for nine locations around Puget Sound were spatially compared with existing near shore biodiversity assessments and shoreline armoring to identify biodiversity hotspots at risk from sea level rise impacts as well as locations where areas of biodiversity may be able persist under changing conditions by shifting inland. Additional components of the Western Washington project not discussed in this comparison can be found in the final Western Washington Project Report.

For freshwater ecosystems, projected changes in hydrologic variables were modeled using the Variable Infiltration Capacity (VIC) hydrologic model and included the timing of flows, summer low flows (7Q10), winter high flows, precipitation regime (i.e., whether a watershed is snow, rain, or transitional). A series of maps were generated for each of these variables including historic (1915-2006) and projected (2040s, 2080s) conditions, percent change compared to historic (2040s, 2080s), and mean percent change (2040s, 2080s) summarized across a watershed (HUC 4 level). Hydrologic variables and stream temperature projections (Hamlet, 2011) were summarized across watersheds to create maps of magnitude of change for 2040 and 2080. These maps categorize watersheds into those likely to experience low, moderate, and high magnitudes of change compared to historic hydrologic conditions. The current spatial distributions of seven focal fish species (Chinook, coho, pink, sockeye, and chum salmon; bull trout; and steelhead) were overlaid on the magnitude of change maps to identify those watersheds likely to continue to harbor these species in the future. Using the MC1 vegetation distribution model developed by the MAPSS team at the Pacific Northwest Research Station, projected fire patterns were also compared with the magnitude of change maps to identify watersheds likely to lose substantial amounts of vegetation, potentially exacerbating hydrologic climate impacts.

The coastal marine assessment involved the use of the SLAMM tool, which was run using a projected 1-meter change in sea level by 2100 for nine locations in Puget Sound. Changes in habitat types were mapped for 2050 and 2075 and compared to initial condition maps (1977). Coastal biodiversity data from The Nature Conservancy’s Ecoregional Assessments for the Puget Trough area were added to the habitat change maps to identify vulnerable biodiversity hotspots. Unfortunately, the spatial resolution of the ecoregional assessment data did not match the resolution of the SLAMM data, so we are currently investigating additional biological data layers for comparison. Shoreline armoring locations and projected population growth and development (2010) were also added to the habitat change maps to identify where nearby disturbances may undermine predicted stable habitats.

The technical use of the tool(s)

Freshwater

To identify watersheds likely to be more or less impacted by changing hydrologic conditions (precipitation regime, low flows, high flows, change in timing of flows), we used coastal PNW stream flow metric data from the USDA-FS Rocky Mountain Research Station (2011) and snow water equivalent data from the University of Washington Climate Impacts Group (2009). Snow water equivalent (SWE) data, which uses the ratio of peak snow water equivalent to cumulative winter precipitation (Oct-Mar), were summarized across watersheds (HUC 4). Change maps for SWE were created by subtracting baseline from future models and then reclassified into low, medium, and high risk based on the watershed type (i.e., whether snow, rain, or transitional). For stream flow metrics, the tabular data were joined on ‘ComID’ to the National Hydrology Dataset Plus (nhdflowline layer). Data used included: (1) Channelflow (Q1.5) – the 1.5-year flow, sometimes considered the channel-forming flow; (2) CtrFlowMass (CFM) – timing of the center of the mass of flow, or the day of the water year at which 50% of the year’s flow has passed; and (3) Flow7q10 (7Q10) – the 7-day low flow with a 10-year return interval.

• Center of Mass of Flow – a shift in timing field was added based on the change in central flow mass field and reclassified to 0 to -15 days, -15 to -25 days, and -25 to -50 days. Polyline data were converted to raster layers and zonal statistics were calculated to summarize across watersheds (HUC 4).
• Low Flow (7Q10) – a percent change in low flow field was added based on the change in low flow from historic to projected conditions and reclassified. Polyline data were converted to raster layers and zonal statistics were calculated to summarize across watersheds (HUC 4).
• Channel Flow – a percent change in channel flow field was added based on the change in flow from historic to projected conditions and reclassified. Polyline data were converted to raster layers and zonal statistics were calculated to summarize across watersheds (HUC 4).

For a more detailed methodology on climate projections, how uncertainty was dealt with, and the VIC hydrologic modeling results, please refer to the Western US Stream Flow Metric Dataset (2011), The Washington Climate Change Impacts Assessment (2009), Littell et al. (2010), Elsner et al. (2010), and Wenger et al. (2010). We are currently in the process of revising our methods for stream temperature projections and will provide an update soon.

To assess the potential impacts of wildfire in watersheds, we used the MC1 dynamic vegetation model developed to simulate the potential biosphere impacts and biosphere-atmosphere feedbacks from climatic change (MAPSS). Using the A2 emissions scenario, fire projections were calculated from three GCMs for baseline (1961-1990), mid-century (2035-2045), and late-century (2075-2085). Data represent the probability of a proportion of each grid cell (~8km) burned by fire.

StreamNet is a cooperative information management and data dissemination project focused on fisheries and aquatic related data and services in the Columbia River basin and the Pacific Northwest. We used their data on seven focal fish species to create a single map depicting overlap in species distribution by converting polylines from the database to rasters and summing. This resulted in a dataset where values range from 0 (no focal species) to 7 (all focal species overlap in a particular location), and was compared to the magnitude of change map to identify potentially stable watersheds with high focal fish species biodiversity.

Coastal

Coastal habitat change maps were created using the SLAMM 1-meter change by 2100 data created by the National Wildlife Federation and Warren Pinnacle consulting (2007). The SLAMM classification scheme was reclassified into more general groups (from 26 habitat types to 10) using custom python scripts, which were also used to convert the ascii data to rasters and defined the projection as NAD 83 State Plane Washington North in meters. The initial condition (1977 in most cases) was subtracted from the future conditions (2050 and 2075). Change maps were then reclassified into binary data; 0 representing no change and 1 representing change from any habitat type to another. Statistics for gain/loss in a particular habitat type were calculated by importing raster data pixel counts into Excel.

To identify potentially vulnerable biodiversity hotspots, priority nearshore conservation areas, as identified by The Nature Conservancy’s Ecoregional Assessment for the Georgia Basin and Puget Trough, were added to the habitat change maps (Floberg, 2004). To assess the impacts of nearby land use on potentially stable coastal habitats, we added armoring locations identified by the Puget Sound Nearshore Ecosystem Restoration Project (PSNERP) to the change maps. We are in the process of adding projected population development information from the EPA Integrated Climate and Land Use Scenarios (ICLUS) project to these maps.

Final products for this work included a report detailing summary maps and a discussion of adaptation planning across ecosystem types, with associated guidance and recommendations on how the results could be used to inform conservation and management opportunities for Puget Sound stakeholders. Detailed methodologies with discussion of data and model uncertainties will be included. The final mapping products for this work include:

• Maps for each of the five hydrologic variables
• Maps of combined hydrologic variables including: (1) low flows and temperature; (2) high flows, precipitation regime, and change in timing of flows; and (3) all hydrologic variables
• Maps of combined hydrologic variables and biodiversity or focal species data
• Map of combined hydrologic variables and projected fire patterns
• Maps of coastal habitat changes for nine Puget Sound locations
• Maps of coastal habitat changes and biodiversity data
• Maps of coastal habitat changes and shoreline armoring or projected human population growth and development
• Combined maps of hydrologic and coastal habitat changes to facilitate planning across ecosystem types

Because this project represents a broad range of habitats, it allowed us to better understand how differences in systems (terrestrial, freshwater, and coastal marine), data types and availability, and climate may influence the efficacy of the Framework and a planner’s ability to use it. Similarly, comparing results from assessments that apply different approaches (e.g., projecting sea level rise, identifying areas that will continue to harbor species in the future, mapping potential future patterns of fire) allowed us to evaluate the strengths and weaknesses of each approach and its utility in developing particular adaptation strategies. Finally, developing adaptation strategies across ecosystem types (mountains to the sea) is rare and possibly non-existent, thus our work not only provides a unique opportunity to showcase the Framework’s utility, but may also provide an important case study for holistic adaptation planning.

Key Stakeholders and Participants

• Local Resource Managers (e.g., Washington Department of Fish and Wildlife, Washington Department of Natural Resources)
• Non-Governmental Organizations (e.g., Sierra Club, National Wildlife Federation)
• Researchers (e.g., University of Washington)
• Expert Panel Members from the University of Washington Climate Impacts Group, WA Dept of Fish and Wildlife, Western Washington University, University of Washington School of Forest Resources, North Pacific Landscape Conservation Cooperative, National Wildlife Federation, Sierra Club – WA State Chapter, and Seattle City Light Environmental Affairs Division