Overview

We approached the question of what climate change portends for groundfish fleets on the U.S. West Coast using a coupled social-ecological approach. We define coupled social-ecological risk due to climate change as the combination of exposure to projected environmental or ecological change and the social vulnerability of the affected community. Social vulnerability is defined in terms of sensitivity and adaptive capacity. We assessed fleet-specific risk in two ways (Fig 2). First, we evaluated risk if fleets change target species while continuing to fish in current fishing grounds (the adapt in-place assessment). Second, we assessed risk if fleets shift fishing grounds while targeting current species (the adapt on-the-move assessment). This evaluation builds on the general framework of the Intergovernmental Panel on Climate Change (IPCC) [3], and more recent reviews and developments introduced by [9,26,46–48]. We define each groundfish fleet as the collection of vessels landing groundfish caught using bottom trawl gear and delivered to buyers in the same port group (S1 Table). We note that this definition of a groundfish fleet is inclusive of vessels with federal permits for the fishery and vessels participating in state-managed bottom trawl fisheries that capture federally-managed groundfish incidentally.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Conceptual framework to consider coupled social-ecological risk due to climate change.

(a) Assuming fleets change target species while remaining in current fishing grounds (adapt in-place); (b) assuming fleets shift fishing grounds while targeting current species (adapt on-the-move). We define coupled social-ecological risk due to climate change as the combination of exposure to projected environmental or ecological change and the sensitivity and adaptive capacity (i.e., social vulnerability) of the affected community, or more formally, Risk = (Exposure²+Vulnerability²)^1/2 (Eq 7) where Vulnerability = (Sensitivity²+(Lack of Adaptive Capacity) ²)^1/2 (Eq 6). This approach is adapted from frameworks in [3,20]. In both panels, redder colors indicate higher exposure due to warming.

https://doi.org/10.1371/journal.pclm.0000285.g002

For the adapt in-place assessment, we estimated exposure as the amount of thermal change expected between the periods 1990–2020 and 2065–2095 within the present-day fishing grounds used by each fleet. We estimated the flexibility dimension of adaptive capacity based on an index of diversification, defined as realized opportunities to participate in multiple fisheries in each port group from 2011–2019, and encompassing a recent period of consistent management regulations [37].

For the adapt on-the-move assessment, we estimated exposure as the projected extent of horizontal (change in latitude and/or longitude) and vertical (change in depth) displacement of near-bottom isotherms representative of present-day fishing grounds for each fleet between the periods 1990–2020 and 2065–2095 (S1 and S2 Figs; [49]). We estimated the flexibility dimension of adaptive capacity based on an index of mobility, defined based on documented distances of fishing grounds from landing ports during 2011–2019.

For both the adapt in-place and adapt on-the-move assessments, we defined sensitivity as the economic dependence of each groundfish fleet on bottom trawl groundfish relative to total commercial fishing revenue, including pink shrimp, Dungeness crab, and Pacific whiting, generated by those fleets within the U.S. Exclusive Economic Zone and state waters during the period of 2011–2019. This approach assumes that more economically-dependent fleets are more susceptible to harm if climate change negatively affects bottom trawl groundfish. To estimate overall risk due to climate change for groundfish fleets, we calculated a social vulnerability index based on the sensitivity and adaptive capacity estimates, and combined it with estimates of exposure. We describe these calculations in detail below.

Defining fishing footprints

The foundation of this risk assessment is the location of fishing grounds for each groundfish fleet. We defined the spatial footprints of each of 14 fleets based on fishery-dependent catch data available from logbooks from 2011–2019 in Washington, Oregon, and California. We retrieved these data from the Pacific Fisheries Information Network (PacFIN; http://pacfin.psmfc.org). To connect these data with specific fishing communities, we associated footprints with port groups of landing for each bottom trawl tow in the database (following [50,51]; S1 Table). There are nearly 300 ports where groundfish are landed and the distinction between ports can often be as small as two different sides of a small bay. The port groupings were developed by the PFMC for biennial groundfish harvest specifications. In addition, aggregating individual ports into port groups is necessary to provide a feasible set of geographic areas for a coastwide climate risk analysis. Finally, analysis at the individual port-level would violate confidentiality requirements, because there are often fewer than three buyers in any one port.

We pre-processed the logbook data to remove problematic hauls prior to development of footprints (https://zenodo.org/record/7916821). Specifically, we included hauls lasting at least 0.2 hours but not more than 24 hours, and removed hauls with coordinates outside of the U.S. EEZ, and those on land or outside of a customary catch depth (>2,000 m) or area (defined based on locations of bottom trawl tows during the period 2010–2015). We evaluated the depth reported for each haul using the Imap R package (https://github.com/John-R-Wallace-NOAA/Imap), which overlays hauls with the National Geophysical Data Center (NGDC) bathymetry (at a resolution of 3 arc-seconds, or ~90m at the Equator) [52–54]. We retained hauls reporting a depth within 250 m of the NGDC depth, assuming accurate reported haul locations. However, we assumed that if reported depths were inaccurate by >250 m, the haul locations were likely to be similarly erroneous. Finally, we assumed that failure to report depth was not indicative of positional error, but a simple misstep on the skipper’s part, so we acquired the missing depth from NGDC based on the geocoordinates of the set (start) point for each haul. Combined, these filters reduced the size of the logbook dataset by ~4% across all years (S2 Table).

For each fleet, we extracted all tows from the period 2011–2019 from the logbook data, excluding fleets with fewer than 3 vessels reporting logbook data during that time period. We used the summed weight of landed catch of all groundfish species actively managed or listed as ecosystem component species in the groundfish fishery management plan used by the PFMC (Tables 3–1, 3–2 in https://www.pcouncil.org/documents/2016/08/pacific-coast-groundfish-fishery-management-plan.pdf/), along with the geocoordinates of trawl set points, to create a kernel density surface [33]. We calculated kernel density with a 10 km bandwidth, using the density.ppp function in the sp package in R [55]. The kernel density allowed us to define the footprint of each fleet, using a percent volume contour that represents the boundary of the area that contains 75% of the volume of the kernel density distribution. The percent volume contour was determined using the getvolumeUD function in the adehabitat package in R [56]. The position of each fleet’s footprint on the coast was relatively unchanged by the choice of the 50, 75, 90, or 95 percent volume contour (S4 Fig), and would not influence the rank order exposure of fleets, or the relationships between exposure and latitude, described below given the large-scale patterns of projected bottom temperature change, horizontal displacement, and vertical displacement (S1 and S2 Figs).

Exposure

Poor ocean bottom conditions are the most relevant hazard for the life stages of groundfish species caught with bottom trawl gear, and temperature is an established predictor of groundfish species’ range shifts [57]. We obtained projected bottom temperatures–the basis for a regional assessment of hazard–from an ensemble of regional downscaled ocean projections [11] produced using the Regional Ocean Modeling System (ROMS; S1 Fig). The ROMS domain spans the California Current ecosystem from 30°-48°N latitude and from the coast to 134°W longitude at 0.1° degree (~7–11 km) horizontal resolution with 42 terrain-following vertical layers. The regional projections were forced with output from three Earth System Models (ESMs) contributing to phase 5 of the Coupled Model Intercomparison Project (CMIP5): Geophysical Fluid Dynamics Laboratory (GFDL) ESM2M, Hadley Center HadGEM2-ES (HADL), and Institut Pierre Simon Laplace (IPSL) CM5A-MR. While we only used the high-emissions Representative Concentration Pathway (RCP) 8.5 scenario, which is the highest-emission scenario and one which appears to be increasingly unlikely [58], the ESMs were chosen to bracket the spread of potential future change. Specifically, GFDL and HADL represent low and high ends of the spectrum, respectively, for the projected magnitude of warming in the CMIP5 ensemble [11,59]. The relatively weak warming in GFDL under RCP8.5 is comparable to the CMIP5 ensemble mean warming under RCP4.5. We focused on 30-year historic (1990–2020) and future (2065–2095) periods to best capture interdecadal variability [59] in ocean conditions characteristic of the California Current ecosystem.

We estimated exposure based on analysis of projected bottom temperatures within each fleet’s fishing footprint. For the adapt in-place assessment, we calculated exposure e_{adaptin−place,p} for each fleet operating out of port group p as the thermal state change normalized by historic thermal variability within each fishing footprint, addressing the question: if the footprint of fishing effort for a fleet remains stationary, how much will the environment change within it relative to the scale of variability it normally experiences?

To obtain estimates of e_{adaptin−place, ESM,p} for each ESM we spatially joined bottom temperature projections to the fleet footprints (using the sf library in R; [60]), and calculated the mean and standard deviation in bottom temperature during the historic period, t_{historic,ESM,p,c} and σ_{historic,ESM,p,c}, respectively, and the mean bottom temperature during the future period, t_{future,ESM,p,c}, for each ROMS cell c within each footprint. We estimated exposure as the difference in the average future and historic temperatures across all cells within each footprint, and , divided by the average standard deviation in historic bottom temperature across all cells within each footprint, , or (1)

Therefore, the units for this exposure metric are essentially standard deviations of temperature change relative to the historic baseline.

For the adapt on-the-move assessment, we calculated exposure for each fleet based on horizontal (change in latitude and/or longitude) and vertical (change in depth) displacement of isotherms representative of present-day fishing grounds (S2 Fig). Displacement is a metric that characterizes environmental change in terms of the minimum distance that must be traveled to track constant temperature contours [49], addressing the question: if the footprint of fishing effort for a fleet moves to find a future environment that matches the historical one, how far will it have to go? In the case of bottom temperature, we calculated both horizontal and vertical displacement for each ROMS cell. We excluded ROMS cells in which >10% of their area was inaccessible to the trawl fishery due to presence of untrawlable habitat or the most recent spatial fishery regulations (2020-present; S2 Text, S3 Fig). Sensitivity analysis revealed that the choice of the 10% threshold for inaccessible habitat did not qualitatively change conclusions. To capture movement on finer spatial scales than the 0.1° degree resolution of the ROMS output, displacements were interpolated to capture the minimum distance required (i.e., it is not necessary to move a full 0.1° degree to the next grid cell if a partial movement would account for the temperature change). As with e_{adaptin−place,ESM,p}, we joined the summaries of displacement to the fleet footprints, and calculated the average value of horizontal and vertical displacement for each fleet and ESM, or e_{adapton−the−move,ESM,Hd,p} and e_{adapton−the−move,ESM,VD,p}, respectively. The units for the horizontal and vertical displacement metrics are in kilometers that would have to be shifted to maintain an isotherm.

Sensitivity

We calculated sensitivity in the same way for both the adapt in-place and adapt on-the-move assessments, focusing on the economic dependence of fleets on bottom trawl groundfish. To obtain information on fisheries landings by port group, on 3 October 2022 we downloaded data for all bottom trawl groundfish vessels for the period 2011–2019 from PacFIN’s comprehensive fish tickets table. We calculated sensitivity s of vessel v in year y to changes in revenue r (adjusted for inflation to 2021 USD) from the bottom trawl-caught groundfish gbt in port group p in relation to all fisheries f and port groups in which it participates as (2)

We calculated annual sensitivity of each fleet S_{f = gbt,p,y} based on the median value of s_{f = gbt,p,y} across vessels for each year and port group as (3)

Adaptive capacity

Adaptive capacity is a complex and multifaceted concept, defined by the Intergovernmental Panel on Climate Change as “[t]he ability of a system to adjust to climate change (including climate variability and extremes), to moderate potential damages, to take advantage of opportunities, or to cope with the consequences” ([61], p. 9). Evaluating adaptive capacity comprehensively requires assessment of multiple domains, including assets, flexibility, organization, learning, and agency [29–31]. Here we focused on the flexibility domain as it pertains to coping capacity, the “ability to react to and reduce the adverse effects of experienced hazards” ([62], p. 72). Specifically, we quantified diversification and mobility within the groundfish fleets, equating reduced diversification and mobility with reduced capacity to cope and adapt.

Assessment of risk due to climate change

We integrated measures of exposure, sensitivity, and adaptive capacity of the groundfish fleets on the U.S. West Coast to evaluate coupled social-ecological risk to climate change. Our definitions follow those of the IPCC [62], such that high exposure to climate change, given the hazard of projected warming bottom temperatures [11], and high vulnerability, together imply high risk. Vulnerability is defined broadly as “the propensity or predisposition to be adversely affected” ([3], p. 5), and here we calculate it by integrating our measure of sensitivity (economic dependence) with our measures of adaptive capacity (diversification or mobility).

Specifically, we calculated median exposure values based on thermal change relative to historic variability, horizontal displacement, and vertical displacement across the 3 ESMs for each fleet, and rescaled the median exposure values to index values of E^*_{p, thermal change}, E^*_{p, horizontal displacement}, and E^*_{p, vertical displacement} such that their minimum values were 0 and their maxima were 1 (the maximum thermal change relative to historic variability, horizontal displacement, and vertical displacement expected across all fleets). We calculated the average value of across 2011–2019 and rescaled it to create a sensitivity index S^*_p with a minimum value of 0 and a maximum value of 1, with 1 reflecting the maximum observed across all fleets. For each of the measures of adaptive capacity, we calculated their average annual values across 2011–2019, and rescaled the resultant quantities such that their minimum values were 0 and their maxima were 1, with 1 reflecting the minimum diversification or mobility observed across all fleets. This reversal of scale converted these indices into measures of a relative lack of capacity to cope and adapt, due to a relative lack of diversification D^*_p and relative lack of mobility M^*_p.

We calculated vulnerability of each fleet under the adapt in-place assessment V_{p, adapt n-place} and under the adapt on-the-move assessment V_{p, adapt on-the-move}, as the Euclidean distance to the origin of the location represented by sensitivity S^*_p and either D^*_p or M^*_p values, such that (6A) and (6B)

With this calculation, we assume vulnerability to be equally affected by sensitivity and adaptive capacity. Following [46] (their Fig 2, right), we represented this vulnerability to climate change visually, and used it to distinguish between fleets of greater or lesser concern and those that are potential adapters or have high latent risk.

Our ultimate interest was in the combined risk due to climate change of each fleet under the adapt in-place assessment R_{p, adapt in-place} and under the adapt on-the-move assessment R_{p, adapt on-the-move}. Specifically, we defined this integrated measure of exposure and vulnerability as the Euclidean distance to the origin of the location associated with each value of E*_p,i and vulnerability V_p,j, (7A) (7B)

With these calculations, we assume risk to be equally affected by exposure and vulnerability, and interpret fleet risk relative to other fleets in this analysis, rather than capturing an absolute measure of risk.

Geographical patterns

To evaluate whether there were geographical patterns in the exposure, sensitivity, adaptive capacity, and risk metrics, we conducted regressions of these variables against latitude. Specifically, we used the glmmTMB package to evaluate (i) the fixed effects of latitude on thermal change relative to historic variability, horizontal displacement, or vertical displacement for each ESM separately; (ii) the fixed effect of latitude and the random effect of year on sensitivity, diversification, and mobility; and, (iii) the fixed effect of latitude on each of the risk metrics. In all of the models, we weighted the regressions by the number of vessels composing each fleet. For the sensitivity and diversification models, we used a logit link and the ordered beta family because the data represent proportions. For the mobility model, we used a log link and the Gaussian family to adequately capture the long tail in the distribution of landings-weighted distance from port across fleets, and included splines (number of knots = 3). All other models used an identity link and the Gaussian family. While the convention when plotting regressions is to have the explanatory variable on the x-axis, we decided to plot latitude on the y-axis because it provides a more intuitive representation of poleward and equatorward shifts in fishing fleets operating off the U.S. West Coast. To evaluate the leverage of individual fleets in these analyses, we re-ran the regressions described above using leave one out cross validation (LOOCV; see S3 Text for details).

Building climate resilience for fishing fleets

Parsing risk into its constituents (exposure, sensitivity, and adaptive capacity, under two contrasting adaptation strategies) suggests different types of interventions that can be implemented to reduce risk. Communities may have similar risk scores, but contrasting sources of risk, and therefore may respond favorably to customized interventions. Mitigating risk may require more proactive efforts to improve adaptive capacity, such as fisheries portfolio diversification or enhancing fleet mobility, or to reduce sensitivity through expansion of revenue streams, among other solutions [29,46,68]. For example, in California, there are existing precedents for enhancing adaptive capacity for fleets with latent risk (low sensitivity and low adaptive capacity). For instance, following the implementation of individual fishing quotas in 2011, members of the Fort Bragg, Morro Bay, Monterey, and Santa Barbara fleets organized quota risk pools with the support of local government and non-government organizations to navigate bycatch constraints, thereby enhancing resilience within the new regulatory environment [69].

In contrast, the suite of interventions for fleets that are potential adapters (because they have higher adaptive capacity and sensitivity, e.g., Fort Bragg or Astoria) are more likely to focus on a reduction in sensitivity. Livelihood diversification (e.g., through mariculture or tourism activities) can dampen sensitivity while also improving adaptive capacity in-place, whereas improving access to fish for other target species and in new (or previously closed) fishing grounds are more exclusively directed at reducing sensitivity [46,68]. Finally, there are interventions that could rescale the risk landscape across all fleets, such as recent efforts to create increased market share for groundfish [70]. Increased consumer demand for a diversity of groundfish could increase profit margins, augment financial safety nets for fishers, and provide an opportunity to take advantage of currently underutilized and abundant stocks. However, creation of market demand in specific areas requires resolution of mismatches between locations of fishery landings, seafood processing, and seafood markets (e.g., through accurate mapping of seafood supply chains and rescuing of stranded capital; [71]). In addition, market demand interventions may exacerbate ecological risk if they incentivize localized depletion of stocks to meet growing local demand [68,72].

Historical contingencies in management, market, and ecological forces provide important context for evaluating the most useful interventions, regardless of whether risk due to climate change is higher or lower for these fleets. These forces create a geography of pre-existing vulnerability, akin to that documented in other regions where shrinkage and disappearance of fishing communities has occurred [73] or where implementation of new management measures has set the stage for responses to subsequent shocks [25,74]. For the bottom trawl groundfish fishery on the U.S. West Coast, revenue has become more concentrated within fewer fleets over the last several decades, a trend that continued throughout the 2011–2019 period we focused on in this study. Furthermore, the narrower continental shelf available to California fleets has led to smaller fishing footprints (areal extent) and a lower projected exposure to expected ocean warming for equatorward groundfish fleets (Fig 4), which also tend to be composed of smaller, less mobile vessels (Fig 5 and S7 Fig, [73]). These trends are a result of the biogeographic context in which each fleet operates, a changed regulatory environment, historical impacts to more equatorward groundfish stocks [75], and various other factors (e.g., geographic locations of buyers, processors, and associated infrastructure; [37,45]). As in other fisheries (e.g., Dungeness crab; [76]), practices that level the playing field for the many smaller vessels composing equatorward groundfish fleets may help to reduce their climate risk. In contrast, for more poleward groundfish fleets that have high sensitivity, it may be more effective to employ approaches that bolster other dimensions of adaptive capacity such as organization, e.g., via social capital building to create cooperatives [46]. Each fleet’s history complicates the many possible paths forward, but potential futures are made less opaque with the information we have provided here on climate risk.

Future directions for assessing climate risk in fisheries

Our approach to understanding spatial heterogeneity in climate risk for fishing fleets in general, and on the U.S. West Coast in particular, highlights opportunities for future research. The data and methods we used to estimate exposure, sensitivity, and adaptive capacity, and to combine them into a risk index, deserve further examination. For instance, we found that estimates of exposure based on horizontal displacement of bottom isotherms are highly uncertain (S8 Fig). This result underscores the challenge of generating expectations about future ocean conditions and use, and brings into question how other environmental factors that affect species distributions, such as dissolved oxygen [77,78] may change and interact with the behavior of fishing fleets [79–82]. Another avenue of future research is integrating expectations for other fisheries in the participation networks (S5 Fig, [38,63]) that are likely to experience climate effects, which will add complexity to estimates of adaptive capacity. For example, Dungeness crab fisheries at higher latitudes may be negatively impacted by ocean acidification effects by the late 21st Century [51], and numerous Pacific salmon (Oncorhynchus spp.) populations along the U.S. West Coast are highly vulnerable to climate impacts at multiple life history stages [83]. An extension of this work could connect species distributions projected using dynamically downscaled ESM outputs (e.g., [84–86]) to fishing footprints directly, using expected changes in the resources themselves within customary use areas to derive estimates of exposure. Such an approach could capture the potential for more equatorward species moving into footprints while others move out ([87–89]; but see [90]), and would also need to address the potential for fleets to capitalize on these changes under existing regulations. There is also the question of how best to identify fishing areas, or footprints, for estimating exposure. Here we identified the primary fishing grounds where the majority of harvested biomass is extracted based on vessel landings by port. Alternative approaches could use metrics such as revenue [91], fisher days [33], or could define fishing areas specific to vessel home ports [23].

There are also alternative approaches for describing sensitivity and adaptive capacity. For example, rather than focus solely on economic dependence on a target species relative to all other commercial fisheries, it would be informative to quantify the economic dependence of fleets on target species relative to all other income streams including those outside of commercial fisheries. Such data are not necessarily widely available, though household survey research in small-scale fisheries provides a template for pursuing this line of inquiry [92–94]. Additionally, the sensitivity and adaptive capacity of crew on fishing vessels may be quite different than for captains or owners. Strong social identity related to participation in particular fisheries could affect fishers’ willingness or ability to adapt by shifting to new fisheries or livelihood activities [95,96]. Ideally, future work to understand risk of fishing communities will embrace a participatory approach in which notions of community, vulnerability, and adaptive capacity are co-developed [97] and considered alongside perceptions of other risks beyond climate change [98]. Approaches such as fisheries learning exchanges may have the added benefit of building trust amongst stakeholders to allow for increases in flexibility in response to climate change, without jeopardizing ecological sustainability [99].

While we chose to analyze fleets defined by common fishing grounds and ports of landing as one type of community, there are other units of community analysis that are equally or more compelling (e.g., communities-of-place defined shoreside, [100,101]; and fisher networks emergent as communities-of-practice [102,103]). Different rubrics for describing communities may lead to greater or lesser emphasis on mobility and diversification as primary metrics to index adaptive capacity. Being able to fish a larger portfolio of species can buffer fishers’ revenues against change and high variability [65]–but doing so often requires owning multiple permits, which may be cost prohibitive for many participants or difficult to manage given current jurisdictional boundaries [104]. This insight could lead to deeper exploration of geographic gradients in the assets dimension of adaptive capacity.

We do not know whether current levels of diversification and mobility are at an upper bound or if there is room for further adjustment given current costs (fuel consumption, insurance, etc.; [105]). Fishing new species may be constrained by fisheries regulations that are slow to adapt to shifting species distributions [21]. Specifically, for the bottom trawl groundfish fishery, some quota categories are restricted to certain geographic regions, which would be problematic if stocks move out of the designated areas [104]. Similarly, mobility may be limited for smaller-vessel fleets and larger-vessel fleets with more diversified catch, as has been demonstrated on the U.S. East Coast [73]. Diversification and mobility aspects of flexibility are underpinned by enabling conditions that intersect with other domains of adaptive capacity such as assets (e.g., financial resources), learning (e.g., access to knowledge, adaptable skill sets), and organization (e.g., community cohesion), all of which may vary across different community typologies [29,30,46]. Future work to explore these issues, for example through retrospective evaluation of community changes associated with adaptive capacity measures existing prior to a disruptive event [25,74], would be illuminating.

Assessments of risk due to climate change can be used to communicate potential impacts to people, regions, or sectors at local scales [5], and in so doing can provide rationale for medium- to long-term policy decisions intended to improve resilience. This case study provides a practical implementation of the widely-used IPCC risk assessment framework at a geographic scale that is relevant to fishers, communities, and U.S. federal fisheries managers. It achieves this appropriately-scaled outcome by integrating climatic, ecological, and socio-economic data from a regionally large-volume, relatively profitable, lynchpin fishery. These kinds of data are commonly available from many of the largest-volume, greatest-value fisheries globally. However, given that these data were also available for the relatively small fleets we assessed here, this framework may be viable for smaller-scale fisheries as well, especially with creative approaches to generating information streams (e.g., improving understanding of fishing grounds, economic dependence on target species, and mobility via structured surveys and participatory workshops; [97]). Similar analyses for fleets in other regions, coupled with scenario planning efforts [106,107], can provide more comprehensive insight into the risks of climate change for fisheries. This insight can be used to identify regions with the greatest potential to improve resilience to climate change through government-based regional action plans, self-determined actions, and via new legislation for fishery disaster responses (e.g., in the U.S. via the Fishery Resource Disasters Improvement Act) [26,29].

The contrasts observed here among U.S. West Coast groundfish fleets have explanations ranging from physics to market forces, and contingencies fueled by historical and present-day regulations. They add to evidence from the U.S. that more poleward fishing fleets may be at greater risk due to climate change [51,86], in contrast to expectations for greater equatorward risk in other parts of the world, such as Europe [35,36,66]. While the potential for the adapt on-the-move strategy to mitigate greater poleward risk exceeded that for the adapt in-place strategy, our results imply that neither of these within-fisheries flexibility measures are sufficient to disrupt fundamental geographic patterning of risk. Rather, alternative adaptation approaches that build out other attributes of flexibility, including those external to commercial fisheries, and alternative dimensions of adaptive capacity not addressed here, may prove most fruitful for ameliorating latitudinal patterns of climate risk. For example, increased agency for fishers to access new target species entering their fishing grounds, introduction of greater flexibility to shift fishing permits quickly, and organizational support to develop new markets are all aspects of adaptive capacity that can reduce climate risk. Evaluations of climate risk and adaptation approaches that capture these other types of issues need not be more complex, but instead can strive for transparency, replicability, and comparability with this one. While the insights presented here are specific to the U.S. West Coast, they suggest that coupled social-ecological risk assessments like this one offer a promising path forward for evaluating climate adaptation options in other regions around the world.