1 Introduction

The northern coastal temperate rainforest of southeastern Alaska and British Columbia is the largest contiguous coastal temperate rainforest in the world, and yet there is limited synthesized information on transboundary climate change implications (Fig. 1). Steep, sloping watersheds where average winter temperatures are within a few degrees of freezing (Simpson et al. 2005), drain the Coast Range into estuaries of the eastern Pacific Ocean. This sharp environmental gradient results in a high degree of climatic sensitivity and changes in ecological communities over relatively short distances, our understanding of which would benefit from increased interdisciplinary collaboration. This paper represents the synthesis efforts of 15 regional resource experts to identify important social and ecological implications in the context of a long-term climate modeling exercise.

Fig. 1
figure 1

A map of the study area with the eight biogeoclimatic provinces, three vegetation zones, and current habitat totals (ha) for the coastal temperature rainforest of southeast Alaska, USA, and northern coastal British Columbia, Canada

Full size image

The governing watershed dynamics in this ecoregion can be grouped into three general categories: glacial melt, snowmelt, or rainfall systems (Edwards et al. 2013; Fig. 2). In lowland forest streams, the annual pattern of watershed discharge tracks the annual pattern of precipitation. These watersheds are predominantly found on island provinces (sections 2, 4, 5, and 6 in Fig. 1). Snowmelt watersheds are characterized by a spring snowmelt peak, a summer discharge minimum and an autumnal flow peak. These watersheds can be found on central parts of mountainous islands and in mainland provinces (1, 3, 7, and 8). Finally, high-elevation glacial watersheds export little water during winter and early spring, and have their main peak discharge during late spring to early fall. Glacial watersheds are found almost exclusively in mainland provinces (1, 3, and 8). Within this framework, a primary control on the timing and magnitude of stream flow is the proportion of the watershed that lies above the rain-snow transition zone, making the hydrology of this region particularly sensitive to changes in snowline elevation. For instance, water temperatures in rainfall dominated watersheds are 5–12 °C warmer during summer than snowmelt and glacial watersheds (Fellman et al. 2014).

Fig. 2
figure 2

Characteristic hydrographs for the three watershed types (glacial, snowmelt-dominated, and rain-dominated) in the coastal temperate rainforest of southeast Alaska and northern coastal British Columbia. Hydrographs are long term averages (30–60 years) from three individual representative streams in southeast Alaska, and are compared with long-term average precipitation in Juneau, Alaska, USA

Full size image

Although species diversity is low relative to other rainforests globally, ecosystem productivity is extraordinarily high (Alaback 1996). For example, the total carbon stock in the forest and soils of the Tongass National Forest in southeast Alaska is 8 % of that in the conterminous United States (Leighty et al. 2006). Southeast Alaska (Halupka et al. 2003) and northern British Columbia (Temple 2005) contain a combined estimate of 7700 anadromous salmon streams with six Pacific salmonids (Oncorhynchus spp.). In southeast Alaska, salmon support an annual commercial and sport fisheries harvest of 30–70 million fish per year with a value estimated at $600 million to $1 billion (Waters 2010). Species distributions are consistent with island biogeography theory with a wide range of genetically isolated species or sub-species (Cook and MacDonald 2013). Rural human communities (approximately 50 % of the region’s population) are isolated by waterways and steep terrain, and rely on ferry, barge, and plane services for transportation. These communities rely on ecosystem services related to fish, wildlife, forests, water and scenery (Beier et al. 2008b; Crone and Mehrkens 2013; Patterson and Coelho 2009).

We synthesized an expert review of climate change implications for this region based on an analysis of projected temperature, precipitation, and snowfall. Each author was provided with a series of maps for each climate variable including the raw data from which to draw inferences. Implications for two general biophysical systems were considered—hydroecological (freshwater hydrology and fishes) and terrestrial ecological (vegetation and wildlife), and concluded with their associated ecosystem services. Five IPCC CMIP5 global climate models (GCMs) that performed well for the North Pacific were analyzed (Walsh et al. 2008; Radic and Clarke 2011): CCSM4, GFDL-CM3, GISS-E2-H, IPSL-CM5B-LR, and MRI-CGCM3. These were run for two representative concentration pathways (RCPs): RCP4.5 assumes medium-low global anthropogenic emissions and RCP8.5 assumes high emissions. The spatial climate modeling software ClimateWNA 4.85 (Wang et al. 2012b) and GIS were used to conduct the analysis at a 1-km resolution for the 2080s (2071–2100; 30-year normal period). Precipitation values were rounded to the nearest 10 mm. This climate tool is based on the PRISM climate model (Daly et al. 2002) and provides a historical mean (1961–1990) to which predicted changes can be compared. Biogeoclimatic zones with similar energy flow, vegetation and soils developed by the BC Forest Service (Meidinger and Pojar 1991) and the US Forest Service (USFS 2008) were used to stratify our analysis by eight geographic provinces and three vegetation zones: alpine (Coastal Alpine), subalpine (Mountain Hemlock), and lowland forest (Northern Coastal Hemlock/Pacific Silver Fir).

2 Climate model results

Our analyses projected an overall regional increase in mean annual temperature (MAT) and mean annual precipitation (MAP), and a decrease in annual precipitation as snow (PAS) (Figs. 3 and 4). Based on projections for five IPCC CMIP5 global climate models (GCMs) and two representative concentration pathways (RCPs), predicted mean annual temperature will increase from a current average (1961–1990) of 3.2 °C to 4.9–6.9 °C (5 GCM range; RCP4.5 scenario) or 6.4–8.7 °C (RCP8.5), mean annual precipitation is predicted to increase from 3130 mm to 3210–3400 mm (3–9 % increase; RCP4.5) or 3320–3690 mm (6–18 % increase; RCP8.5), and total precipitation as snow is anticipated to decrease from 1200 mm to 940–720 mm (22–40 % decrease; RCP4.5) or 720–500 mm (40–58 % decrease; RCP8.5) by the 2080s (2071–2100; 30-year normal period).

Fig. 3
figure 3

Histograms of mean annual temperature (MAT), mean annual precipiation (MAP), and precipitation as snow as water equivalent (PAS). These were stratified by eight biogeoclimatic provinces and three vegetation zones: alpine, subalpine and lowland forest. The historical normals (1961–1990) are presented next to a five global climate model ensemble average with range bars (CCSM4, GFDL-CM3, GISS-E2-H, IPSL-CM5B-LR, and MRI-CGCM3) from the IPCC CMIP5 scenarios RCP4.5 and RCP8.5 for the 2080s (2071–2100; 30-year normal period)

Full size image
Fig. 4
figure 4

A map series of potential climate change showing the current mean annual temp (MAT), mean annual precipitation (MAP), and precipitation as snow as water equivalent (PAS) compared to corresponding projections for the 2080s (2071–2100; 30-year normal period) using a five global climate model ensemble average (CCSM4, GFDL-CM3, GISS-E2-H, IPSL-CM5B-LR, and MRI-CGCM3) from the IPCC CMIP5 scenarios RCP 4.5 and RCP 8.5

Full size image

Projected changes in climate variables generally followed a high to low pattern along two gradients, from north to south and from mainland to the outer coast; with northern mainland areas showing the greatest change in temperature and precipitation and southern island provinces exhibiting the least (Figs. 3 and 4). For example, averages changes for MAT by the 2080s across vegetation zones showed a north-south and mainland to outer coast trend. Yakutat showed the largest average increase of 2.9 °C (0.9 °C SD over 5 GCMs) in RCP4.5 and Mainland increased 4.7 °C (1.0 °C SD) in RCP8.5; Haida Gwaii exhibited the smallest increase with 2.2 °C (0.7 °C SD) to 3.6 (0.8 °C SD) in projected change. The pattern of change for MAP primarily followed a mainland mountain to low-elevation outer coast trend. Yakutat received the largest average increase in MAP with 290 mm in RCP4.5 (160 mm SD) and 490 mm (220 mm SD) in RCP8.5; Nass Kitimat changed the least with increases of 130 mm in RCP4.5 (100 mm SD) to 270 mm in RCP8.5 (190 mm SD). Changes in PAS showed a mainland to outer coast trend, with Yakutat showing the largest decrease of 540 mm (120 mm SD) in RCP4.5 and 830 mm (130 mm SD) in RCP8.5; Haida Gwaii showed the smallest decline of 280 mm (60 mm SD) in RCP4.5 and 380 mm (30 mm SD) in RCP8.5.

Projected increases in MAT showed a uniform pattern of change on average across vegetation zones with 2.5 °C (0.8 SD; RCP4.5) and 4.2 °C (1.0 SD; RCP8.5). Alpine showed the largest increases in MAP and largest decreases in PAS. Projections for MAP in alpine showed an increase of 240 mm (120 mm SD) in RCP4.5 and 460 mm (200 mm SD) in RCP8.5. Alpine PAS decreased by 540 mm (140 mm SD) in RCP4.5 and 830 mm (120 mm SD) in RCP8.5.

3 Implications for freshwater hydrology and fishes

Currently, the average freshwater discharge in our study area is approximately 780 km3/year, or roughly twice the magnitude of the annual discharge of the Mississippi River Basin (Morrison et al. 2011; Neal et al. 2010). Projected declines in PAS are anticipated to moderate the storage of water as snow, thereby changing the magnitude and timing of seasonal snowmelt runoff (Shanley and Albert 2014). As the rain-snow transition zone increases in elevation in response to increased MAT, less PAS will be stored in seasonal snowfields or ice (Fig. 4). Therefore, runoff patterns are expected to transition toward lower elevation watershed types, shifting from glacial to snow melt and from snow melt to rainfall-dominated (Fig. 2). In southeast Alaska, snow melt-dominated watersheds have already shifted towards higher winter stream flows and lower summer time stream flows during the warm phase of the Pacific Decadal Oscillation (PDO; Neal et al. 2002), and based on modeling results, we anticipate this trend will be more prevalent in the future.

Whereas the greatest projected decreases in PAS are for alpine habitats in the mainland provinces (1, 3, 7 and 8), southern island provinces (5 and 6) may show the first signs of transitioning from snow to rain-dominated systems because their mean winter temperature is closer to the freezing threshold. Mainland provinces (1, 3, 7 and 8), however, may be more prone to rain-on-snow and flooding events with the increasing spring and summer temperatures melting subalpine and alpine snowpacks (Fig. 4).

Glacial watersheds currently contribute approximately 30 % of total freshwater discharge in southeast Alaska (Neal et al. 2010), and 95 % of glaciers in this region are losing volume, some at among the highest rates on earth (Berthier et al. 2010; Larsen et al. 2007; Pike et al. 2008). As increased MAT causes glacier mass balance to become more negative, stream flow in glacier-dominated watersheds will initially increase and then eventually decrease as glaciers retreat to occupy only a small fraction (<10 %) of the watershed (Casassa et al. 2009; Moore et al. 2009). This pattern initially may be most pronounced in the southern portions of mainland provinces (3 and 8) with smaller glaciers. In watersheds with a high percentage of ice cover (1 and 3), the initial runoff increase could be substantial (>50 %; Adalgeirsdottir et al. 2006) and will likely exceed climate-driven runoff changes expected from other components of the water budget (Jansson et al. 2003). However, in the long term, loss of ice is anticipated to contribute to lower water yields. Increases in climate-driven discharge are already evident in the few glacial watersheds with temporally extensive stream flow records (Neal et al. 2002).

Runoff from glaciers also has pronounced impacts on water temperature and clarity (Fellman et al. 2014; Hood and Berner 2009; Moore et al. 2009) and riverine concentrations of bioavailable organic matter (Bhatia et al. 2010; Hood et al. 2009), phosphorus (Hodson et al. 2004; Hood and Scott 2008), and micronutrient iron (Schroth et al. 2011). The unique suite of solutes associated with glacial runoff can influence biogeochemistry and productivity in downstream freshwater and near-shore marine ecosystems (e.g., Fellman et al. 2010). Stream flow and stream temperature affect production and timing of benthic fauna that forms much of the food base for freshwater fishes in lowland forest streams (Wipfli 1997; Wipfli and Baxter 2010). Changes in flow, temperature, and nutrient dynamics in freshwater ecosystems influence fish abundance across life-history stages (Bryant 2009). For example, high-flows during the winter incubation period increase salmonid embryo mortality, causing steep declines in adult returns (e.g., pink salmon [O. gorbuscha] and coho salmon [O. kisutch] in Glacier Bay; Milner et al. 2013).

Salmonids are a cold-water species with high critical thermal maxima (e.g., ~29 °C in juvenile coho salmon; Konecki et al. 1995), although most salmonid functions (incubation, feeding, growth) are compromised when temperatures exceed ~12–15 °C (Richter and Kolmes 2005). Dramatic declines in gamete viability have been observed in pink salmon in southeast Alaska when stream temperatures exceed 15 °C during the spawning run (A. Gharrett, pers. communication, 2013). As ectotherms, fish are strongly governed by temperature. Demographic studies have shown that salmonid run timing has shifted and is more temporally compressed (Kovach et al. 2012, 2013, 2015). For example, coho salmon currently return to Auke Creek in southeast Alaska to spawn 17 days earlier than they did in the 1970s (Kovach et al. 2013). However, increased summer temperatures in Auke Lake increase the biomass of juvenile coho and sockeye salmon (O. nerka) smolts the following fall (Kovach et al. 2014), implying that moderately higher temperatures will positively affect some salmon life history stages while negatively impacting others. The diversity of salmon species and life histories will likely be an important source of resilience to climate change throughout the northern coastal temperature rainforest in the future (Kovach et al. 2015).

Winter high flows may be particularly important to monitor for scour of salmon eggs from the gravel (Montgomery et al. 1999; Shanley and Albert 2014). The nests of larger females should be more resistant to scour; however increasing temperatures during the marine phase can lead to smaller adult size in salmonids (Jonsson et al. 2012), thus compounding the effects of warming. According to the climate projections, scour may initially be most pronounced in northern mainland provinces (1 and 3) with increased MAT and MAP. Acute temperature-driven physiological and from snow melt to rainfall-dominated likely to occur during summer months in provinces with a higher likelihood of low flows and diminished snowpack (5, 6 and 7), although increased MAP could mediate lower flows in some watersheds. Temperature-related phenological shifts are expected throughout a fish’s life cycle, but phenological changes are likely to be most pronounced during spawning migrations (mid-summer to fall). Marine-derived nutrients and energy are anticipated to be less available to those species whose own phenologies do not track shifts in these migrations. However, because marine subsidies propagate through freshwater and riparian ecosystems (Wipfli and Baxter 2010), effects might occur in other seasons as well.

Warmer temperatures are expected to slowly influence the lowland forest landscape towards a greater proportion of deciduous trees, especially Alnus spp. (Implications for terrestrial ecosystems—vegetation and wildlife section, this paper), leading to allochthonous litter inputs that are more concentrated in autumn and higher food quality for riparian and aquatic food webs (Wipfli 1997; Wipfli and Baxter 2010). Alnus, an N fixer, rapidly colonizes deglaciating watersheds and will therefore potentially increase nitrate abundance in adjacent streams through increases in soil nitrogen pools (Hood and Scott 2008).

Among glacial watersheds in mainland provinces (1, 3, and 8), near-term increases in summer glacial runoff is anticipated to buffer increases in water temperatures, potentially reducing the likelihood that marginal glacial streams become optimal for spawning and development, but in the long term, diminishing glacial influence may also lead to warmer, more thermally optimal streams supporting increased production of food for stream-rearing salmonids as well as conditions for salmon development and spawning (Fellman et al. 2014; Milner et al. 2009). For example, the warm phase of the PDO has also been correlated with high commercial harvest of salmon in Alaska (Mote et al. 2003). The effect of climate change on local marine subsidies will depend on the amount of suitable spawning habitat, species’ phenologies, and fluctuations in anadromous fish populations.

Sea-level is projected to rise under current climate change scenarios, due to thermal expansion of seawater, melting of land ice, and changes in liquid storage on land (Church et al. 2013). IPCC projections for 2081–2100 range from 0.4 to 0.63 m above 1986–2005 levels (ibid), and vary considerably geographically due to variable changes in ocean temperature and salinity related to changes in ocean circulation. In northern southeast Alaska (1, 2, and northern 3), however, sea level is decreasing by as much as ~3 cm/year as a result of isostatic rebound from glacier mass loss (Larsen et al. 2005). Sea-level changes have important implications for coastal tidal wetlands and salmonid rearing habitats (e.g., Bryant 2009). Some models predict that 20–60 % of the world’s tidal wetlands will be lost this century due to sea level rise (e.g., Craft et al. 2009).

4 Implications for terrestrial ecosystems—vegetation and wildlife

A key feature of current vegetation patterns is the disturbance ecology of frequent wind, snow and ice storms from fall through early spring (Alaback et al. 2013). These winter storms not only constrain growing seasons for vegetation, but also have important effects on associated soil ecosystems (D’Amore and Hennon 2006), nutrient cycling (D’Amore et al. 2010) and forest regeneration rates (Alaback 1982). For example, Alaska yellow-cedar (Callitropsis nootkatensis) decline appears to be closely related to changes in transient snow—reduced snowpacks affect soil insulation, causing fine root damage and mortality in areas with poor drainage (Hennon et al. 2012).

Given projected increases in MAT, we expect distribution shifts in individual species and ecological communities (Hamann and Wang 2006; Wolken et al. 2011). Some wide spread genetically diverse species, such as western hemlock (Tsuga heterophylla) and Sitka spruce (Picea sitchensis), are expected to colonize new habitats rapidly (Farr and Harris 1979; Gavin and Hu 2006). Other species, such as the cedars and many understory plants are expected to change more slowly, due to both regeneration requirements and low genetic diversity (Gilliam 2007; Bunnell and Kremsater 2012) that may lead to new plant communities (Oakes et al. 2014).

Projected increases in MAP suggest that lowland coastal forests will remain characteristically wet places (Fig. 3). Hydrologic changes (previous section, this paper) in the occurrence and intensity of flooding events could lead to changes in the nature and extent of riparian forest disturbance dynamics (Naiman et al. 1998). Assuming that seasonality of precipitation does not change significantly, fire will remain unimportant as a disturbance agent, but higher MAT is anticipated to increase the incidence and severity of insects and disease in lowland forests (Haughian et al. 2012; Hennon et al. 2012). Exceptions to this general rule are anticipated where periodic droughts and fires do occasionally occur i.e., in the southern provinces (6, 7 and 8) (Haughian et al. 2012). This impact is projected to increase in severity, as MAT is also projected to increase (Figs. 3 and 4). Decomposition rates are also anticipated to increase, and some wetlands (bogs, fens etc.) may become productive forests, as has occurred in the past (e.g., Neiland 1971).

Subalpine and alpine habitats are anticipated to undergo more rapid changes than lowland forests. For example, Wang et al. (2012a) suggest that British Columbia’s coastal alpine habitat will shift upward in elevation more than 200 m by the 2080s. Alpine habitat MATs are projected to exceed current subalpine habitat MAT by the 2080s (temperatures that trees currently grow) in all the island provinces (2, 4, 5, and 6) in RCP4.5, and most of the mainland provinces (3, 7, and 8) with the exception of Yakutat in RCP8.5. At the same time, accumulated PAS in alpine habitats will on average decline by 15–54 % across provinces for RCP4.5 and 27–74 % in RCP8.5. The combination of warmer alpine MAT and lower snowpacks are anticipated to raise the elevation of treeline (Wang et al. 2012a), moving treeline to mountaintops, and resulting in a regional loss of high elevation tundra ecosystems. Subalpine vegetation can occur at relatively low elevations due to the persistence of snow in the late spring and early summer, and the high frequencies of snow avalanches (due to steep topography and wet heavy snow). With projected reductions in snowpack, however, the lower boundary of the subalpine habitats is anticipated to increase in elevation.

Climate, vegetation, and weather interact to affect wildlife in the region through food availability and habitat quality, which affect species fitness and demography. Populations of animals and resource selection in northern climates are often determined by the weather severity (Brinkman et al. 2011; Person and Russell 2010). For example, the availability of salmon greatly influences habitat quality for brown bear (Ursus arctos) (Hilderbrand et al. 1999). When salmon populations move or decline as a result of weather and climate-driven changes in hydrology, brown bear populations will also move or decline. Persistent deep alpine and subalpine snow forces mountain goats (Oreamnos americanus) and Sitka black-tailed deer (Odocoileus hemionus sitkensis) to move to lowland forest habitat (Schoen and Kirchhoff 1990; White et al. 2011); condensing winter range and reducing habitat carrying capacity. Deep snow also increases energetic costs by hindering locomotion and restricting movement among forage patches (Parker et al. 1984). Cold, rainy weather in the spring potentially increases thermoregulatory demands and parental investment in wildlife species.

Future climate predictions, however, may cumulatively reduce the severity of ecological and niche limitations on wildlife. Projected increases in MAT may expand lowland forest habitat available to ungulates during the winter and reduce energetic costs (Parker et al. 1984). Future projections of PAS suggest that deep snow may no longer be a limiting factor for deer in the southern and island provinces (5, 6 and 7). Where PAS is still anticipated in the future (1, 2, 3 and 4), snow accumulation and persistence may decline and alleviate winter stress for wildlife. Further, a longer growing season may increase food availability for wildlife in the spring when energy demand of winter-stressed individuals is high. Any negative effects of potentially increasing rain-on-snow and other icing events may be short lived because winter temperatures in the region typically fluctuate close to freezing.

Wildlife habitats associated with coastal and estuarine wetlands are anticipated to undergo changes due to sea-level rise and an increase in the frequency and severity of storm surges (Michener et al. 1997). The direct ecological effects of sea-level rise include coastal inundation, flood damage, increased erosion, and saline water intrusion (Nicholls and Cazenave 2010); with associated changes in wetland community structure (Craft et al. 2009). Wildlife species dependent on these habitats, such as brown bears (Schoen et al. 1994), Vancouver Canada geese (Branta canadensis fulva; Fox 2008), and shorebirds (Iverson et al. 1996) are anticipated to see declines in essential seasonal habitats in southern provinces not experiencing isostatic rebound (6, 7, 8 and southern 3, 4, and 5; Larsen et al. 2005).

Endemic species may be particularly vulnerable to the effects of climate change due to habitat change coupled with limited migration potential in the naturally fragmented and glacier-bounded island archipelago. As MAP and MAT increase in the island provinces (2, 4, 5, 6 and portions of 7), which tend to be the hotspots of endemism in this area, critical aspects of ecological niches are anticipated to also change, and these conditions may trigger extirpations (Cook and MacDonald 2013).

Temperature-related shifts of wildlife ranges northward (Root et al. 2003) may increase interspecific competition, especially for endemic species. However, the isolated geography may slow immigration to the region compared to other parts of North America. Little is known how interactions in wildlife communities may change as the climate becomes more tolerable for southern species currently at the northern extent of their range.

5 Implications for ecosystem service benefits

Increased MAT and MAP coupled with decreased PAS are anticipated to primarily affect ecosystem services that regulate water quality, water quantity, greenhouse gas fluxes and nutrient fluxes (or subsidies). Changing hydrologic regimes could alter the distribution and productivity of fisheries, dependability on hydropower, and desirable tourism activities (e.g., wildlife viewing)—three key pillars of local and regional economies in southeast Alaska and northern coastal British Columbia (Crone and Mehrkens 2013).

The legacy of industrial-scale forest management in southeast Alaska and northern coastal British Columbia has continued influence on both the supply and demand of forest products in the region. Harvesting of wood fiber and fuel is seen as less sensitive to projected climatic changes than the existing constraints on logging in the region (Albert and Schoen 2013; Beier 2011). However, the climate-driven decline of Alaska yellow-cedar across the region has resulted in localized but geographically widespread areas (2, 4, and 5) of high-value salvage logs that may be processed locally into high value wood products (Beier et al. 2008a). Accelerated forest growth rates, if driven by MAT warming, may in turn increase the quantity but decrease the tight grain quality (especially for spruce and cedar) of forest products from second-growth forests, relative to those harvested from primary forests during the 20th century. Greater attention to greenhouse emissions and renewable energy alternatives may promote more widespread use of pre-commercial second-growth forests for biomass energy.

Climate change impacts on ecosystem services related to fish and wildlife populations will undoubtedly influence their many beneficiaries in the region, which include subsistence, commercial and recreational users (Beier et al. 2008b). Wild food sources remain vital to cultural activities for residents, especially in remote First Nation and Alaskan Native communities. Commercial fisheries remain essential in many communities in addition to a rapidly growing guide and outfitter sector (Colt et al. 2007). For fisheries, climate changes that disrupt the regular timing and distribution of salmon runs (Kovach et al. 2013) may reduce fishing success. Conflicts among fisheries may emerge if potentially negative impacts reduce the allowable catch. Related changes that disrupt estuarine systems and near-shore food webs (Schoch et al. 2013) could also impact the availability of shellfish that support locally important subsistence and personal use fisheries. On land, Sitka black-tailed deer are the most harvested game species in southeast Alaska (Brinkman et al. 2011) and because their population movements are sensitive to changing snow patterns, hunters may experience increased difficulty locating deer on traditional hunting grounds or in accessible areas (Brinkman et al. 2009). Overall, if traditional hunting and fishing grounds become less productive or predictable due to climate change, it remains to be seen whether residents, business owners and visitors will continue to either seek or ‘capture’ these ecosystem service benefits.

6 Synthesis and conclusion

Our collaborative synthesis of potential climate change impacts across biophysical systems highlights the coupling of social and ecological dynamics that characterize the northern coastal temperate rainforest of North America. The sharp environmental gradient from icefields to coastal estuaries creates a unique climate sensitivity to warmer and seasonally wetter climate projections with elevated snowlines. Globally, projected increases in days above freezing are only comparable to the southern Andes in South America as well as central and northern parts of Europe and Asia (Meehl et al. 2004; Sillmann et al. 2013). The southern Andes and northern Europe contain temperate rainforest ecosystems (Dellasala et al. 2011) that may provide opportunities for shared learning internationally. Finer scale monitoring and systematic placement of additional long-term hydro-meteorological stations will improve capacity to evaluate assumptions and uncertainty associated with global climate models that do not account for the significance of local variability (e.g., Wright and Stevens 2012). Greater research integration (e.g., interagency collaborations) will provide better insights into the complexity and coupling within and between ecosystems and climate for policy and planning direction. Most information needs cited as a part of this review process (Table 1) relate to improved modeling of changes in hydrology, and the social and ecological consequences of those changes. Immediate recognition and prioritization of our findings is critical because of the rapid rate of change anticipated, and the relatively high dependence on the cultural and provisional ecosystem services in the northern coastal temperate rainforest of southeast Alaska and British Columbia.

Table 1 Information needs cited by disciplinary experts involved in this review paper for increasing understanding of climate change implications for the northern coastal temperate rainforest region of North America
Full size table