Severe reduction in spawning days and larval abundance of walleye pollock under future warming in the western East/Japan Sea

The collapse of walleye pollock catch in the Korean fishing region during the late 1980s remains unresolved despite enormous efforts to recover its stock. To investigate the future fate of walleye pollock in the western East/Japan Sea (EJS) in the late 21st century, we implemented a dynamical downscaling approach by developing high-resolution regional ocean climate models (1/20°) with the selected Coupled Model Intercomparison Project 6 (CMIP6) global climate models. We analyzed the changes in the spawning days and larval distribution of walleye pollock in the western EJS. Under the SSP5–8.5 global warming scenario, the suitable spawning period ratio of walleye pollock was drastically reduced by 76%. The severe reduction in spawning days in the western EJS was mainly attributed to ocean warming caused by a decrease in atmospheric surface cooling. Moreover, our particle tracking experiment showed substantial loss of eggsand larvae south of 38° N. Our study projected a drastic collapse of walleye pollock in the western EJS and proposed a variety of collapse patterns based on local circulation under future warming conditions. These findings can help the local fishing industry adapt and assist neighboring governments in planning future fisheries management strategies.


Introduction
Walleye pollock (Gadus chalcogramma), one of the cold-water species inhabits the North Pacific Ocean's coastal regions, extending from the East/Japan Sea (EJS) to the central coast of California [1-5] 6 .The annual capture of walleye pollock was 3.4 million tons in 2018, ranked second among fish species caught worldwide [6].Predictions of fish abundance provide valuable information for local communities, helping them plan various livelihood activities and adapt to the climate change [7].Therefore, forecasting walleye pollock, which is in high demand, can provide significant advantages.
The East Korea Bay (EKB) off the coast of Korea is known as a major spawning ground for walleye pollock in the EJS (figure 1) [8][9][10], and the species living in the Korean fishing area spawn in the bay [10,11].In contrast to the stable catches in Japan and Russia, the catch in Korea has significantly declined since the 2000s and has not yet fully recovered [4,9].Due to the importance of walleye pollock in commercial fisheries, the decline in pollock stocks has posed a critical challenge in managing Korean fisheries.Therefore, the Korean government is making efforts to restore walleye pollock resources in Korean fishing area.Physical environmental changes in habitats may be linked to the reduction in walleye pollock catches [12][13][14].Specifically, after the late 1980s, a warmed regime influenced the western EJS ecosystem [14][15][16].Both global warming and periodic climate variability could contribute to driving ecosystem changes.
Intergovernmental Panel on Climate Change reported that the global mean Sea Surface Temperature (SST) increased by 0.60 [0.44 to 0.74] °C from 1980 to the period of 2011-2020 [17].A trend for SST over the entire EJS shows a more significant rise of 0.26 °C per decade compared to the global trend (0.11 °C per decade) [18].The SST was projected to increase between the period of 1995-2014 and 2081-2100 by 2.89 [approximate range from 2.01 to 4.07] °C on global average under the shared socioeconomic pathway 5-8.5 (SSP5-8.5)scenario [17].Global warming has already exerted a substantial influence on marine ecosystems, and this impact is anticipated to intensify in the future [19,20].
In the early life stage of walleye pollock, eggs and larvae drift passively towards nursery grounds carried by ocean currents.During this transportation process, they can be susceptible to changes in their physical environment [21][22][23][24].Previous research has indicated that the survival of walleye pollock in the early life stages can influence recruitment success in various regions [25][26][27][28].To simulate the transport process for the eggs and larvae, Particle tracking experiments based on ocean circulation models can be useful tools [29][30][31][32][33]. Transport processes of walleye pollock have been explored in several regions using particle tracking models [14,[34][35][36][37].Despite several studies being undertaken, a high-resolution future simulation resolving the physical environment of western EJS region are currently unavailable.
To investigate the future changes in walleye pollock spawning days andlarval transport in the western EJS, we evaluated the performance of the Coupled Model Intercomparison Project 6 (CMIP6) global models and selected four models for dynamical downscaling.Subsequently, we developed high-resolution regional ocean climate models (RCMs) using the selected global models as boundary forcings under the SSP5-8.5 scenario.We then simulated eggand larval transport patterns using the Larval TRANSport Lagrangian model (LTRANS) based on the results from each RCM ensemble members in January and February, the primary spawning period for walleye pollock in the western EJS.

Climate ocean models
To conduct dynamical downscaling, we acquired global climate models (GCMs) driven by observed greenhouse gas concentrations until 2014 and subsequently by high-end emission (SSP5-8.5)scenarios to 2100 from CMIP6 database [17,38,39].We used 21 models provided all the oceanic and atmospheric variables required for deriving the boundary values for the RCM.The evaluation of CMIP6 GCMs was conducted following a previous study by Kim, Kim, Jeong, Lee, Byun and Cho [40] in order to select appropriate models to apply dynamical downscaling.The GCM evaluation method is described in detail in Supplementary section 1.
The Regional Ocean Modeling System (ROMS) was used to dynamically downscale the physical environment surrounding the spawning grounds of walleye pollock [46].The domain of RCM (15°-52°N, 115°-164°E) covers the Northwestern Pacific Ocean and its marginal seas with 1/20°grid (Supplementary figure 2).We initialized the RCMs with GCMs in January 1985 with 10 years spin-up.The RCMs were continuously simulated until 2100 under historical and SSP5-8.5 radiative forcing.The downscaled simulations from the GCMs are referred to as RCM-CNE (CNRM-ESM2-1), RCM-ECV (EC-Earth3_Veg), RCM-ACC (ACCESS-CM2), and RCM-CNH (CNRM-CM6-1-HR).The ensemble mean of the four RCMs is referred to as the RCM-ENS.We used the optimum interpolation sea surface temperature (OISST) based on the Advanced Very High-Resolution Radiometer satellite SST [47] and the altimetry-based geostrophic current from the Copernicus Marine Environment Monitoring Service (CMEMS; https://doi.org/10.48670/moi-00148) to validate the RCMs.The detailed configuration of the RCMs is described in Supplementary section 2.

Particle tracking experiment
We conducted a particle tracking experiment using LTRANS with RCM results to estimate changes in spawning days and larval transport [48].During January and February, walleye pollock spawns the majority of eggs in the western EJS.The eggs are floated to the surface due to their lower density than surrounding seawater [8,49].In the experiment, the eggs were placed on surface grids, taking their density into account.The spawning grids were determined by their suitable spawning temperature range (2 °C-5 °C) and the appropriate habitat depths (50-500 m) for walleye pollock [10,21,[50][51][52].We considered the potential spawning ground located in the EKB between 36°N and 41°N (figure 1(b)).The schematic of the tracking experiment is represented in Supplementary figure 3. Walleye pollock eggs and hatched larvaes were tracked for 30 days initialized by the first day of January to the end of February during the historical and future periods.We released eggs 2,369 times at each suitable spawning grid in the RCMs.
Suitable spawning grids for the historical (1995-2014) and future (2081-2100) periods are presented in section 3. We estimated the number of suitable spawning days in grids corresponding to spawning temperature and depth conditions in the EKB.The suitable spawning period ratio (SSPR) in each grid cell was defined as the ratio of the suitable spawning days to the total releasing days (2,369 days) in January and February in the historical and future period.SSPR = 1 indicated that walleye pollock spawned every day, whereas an SSPR = 0 indicated that the walleye pollock could not spawn.In this study, it was assumed that the time of hatching was two weeks after spawning [53,54] and that the swimming ability of larvae developed after 5 weeks [55].To focus on the passive transport of eggs and exclude the swimming ability during the earliest stages of development, we analyzed the total number of eggs and hatched larvae in each grid 15 and 30 days after spawning.

Estimation of cumulative heat uptake, surface net heat flux, and lateral heat transport
Estimations of changes in heat content, surface net heat flux, and oceanic heat transport can shed light on the primary drivers of future warming in the EKB region.The cumulative heat uptake (ΔHC, unit: J) to a depth of 500 m in each grid was calculated as follows: Here, m represents the total number of monthly time steps and n represents the number of vertical grid points.The model output was interpolated at 10-m intervals for calculation.T denotes the model temperature (°C).The horizontal grid intervals are represented by dx and dy, and dz is fixed at 10 m.The constant , a equal to 4.1868, is used to convert from calories (cal) to joules (J).The cumulative heat uptake represents the monthly change in heat content during a specific period, and a positive (negative) sign indicates an increase (decrease) in heat content up to 500 m.The net surface heat flux in the ROMS model is the sum of the shortwave radiation, longwave radiation, latent heat, and sensible heat fluxes.The cumulative surface net heat flux was estimated by integrating the monthly surface net heat flux over a specific period.A positive sign indicated a net downward flux (heating), whereas a negative sign indicated a net upward flux (cooling) during a specific period.In the spawning grounds over the EKB, the heat uptake up to 500 m was determined by the surface net heat flux and lateral heat transport by oceanic currents, because the spawning grounds were located near the western coast of the EJS, where the depth was shallower than 500 m.Thus, in this study, the cumulative lateral heat transport was evaluated by subtracting the cumulative surface net heat flux from the cumulative heat uptake.Positive/negative signs represent the lateral heat gain/loss.

Physical environment changes
The simulation of SST was essential to the experiment as it determines the suitable condition for walleye pollock spawning.Figure 2(a) depicts the average OISST for the period 1995-2014.In the EKB region, warm water supplied by the EKWC (>10 °C) is present around 38-39°N, wherease cold water (<5 °C) suitable for the Walleye pollock spawning is found in the northern coastal region.The RCM-ENS was comparable with the observations with a pattern correlation coefficient of 0.94, and the time series of the spatial mean SST over the EKB from the RCM-ENS closely matched the OISST (Supplementary figure 4).The RCM-ENS shows colder SST in the coastal EKB, where the observational error of the OISST is large (0.4) near the coastal region (figure 2(c)).A cold coastal SST pattern was also shown in a previous study using reanalysis data [14].
According to the SSP5-8.5 scenario, the RCM-ENS projects warmer conditions over the EKB compared to the historical SST at the end of the 21st century (figure 2(e)).In all RCMs, the majority of the locations with temperatures below 5 °C vanishes, and the 10 °C isotherm shifted further north.Figure 2(g) depicts a temperature increase of roughly 4 °C in the southern offshore near EKB.Northern EKB experiences an even smaller temperature increase; it is greater than 3 °C.
Horizontal advection plays an important role in ecosystem variability, particularly in areas with strong boundary currents that convey temperature anomalies and change nutrient availability [56].Figure 2(b) shows the satellite altimetry-based surface geostrophic current obtained by CMEMS.The strong northward EKWC, one of the primary currents along the Korean coast, transports heat from the southern region to the EKB (figure 1(a)).The EKWC separates from the coast at approximately 38°N and flows eastward [57][58][59].Although low-resolution GCMs have had difficulty in adequately resolving the EKWC, RCMs can replicate its paths [60,61].RCM-ENS demonstrates a strong northward EKWC along the coastline south of 39°N, which is similar with observations showing pattern correlation coefficients of 0.53 and 0.58 for zonal and meridional velocity, respectively (figure 2(d)).
The future mean surface currents near the EKB over the RCM-ENS are shown in figure 2(f).The current speed along 130°E becomes faster in the future than in the present time.Thus, the anticyclonic circulation linked to the northward EKWC strengthens in the southern offshore region of the RCM-ENS, whereas the cyclonic circulation was enhanced in the EKB (figure 2(h)).
Although the results from the RCM-ENS indicated general changes in ocean states in the future, ocean dynamics should be interpreted using a single model system.The increase in SST over the EKB was particularly enhanced when the EKWC extends towards the northern region.However, this increase was relatively moderate when the separation of the EKWC was restricted around 39°N.This suggests that future warming can evolve under two scenarios based on different ocean dynamics.The detailed surface temperature and current of each RCM and the temporal differences between the historical and future periods are described in Supplementary section 4.

Changes in spawning days and larval abundance of walleye pollock
The SSPRs and abundances of eggs and larvae from the RCM-ENS are displayed in figure 3.Because of the warm water supplied by the EKWC, the temporal mean SSPR is only 0.14 in the southern EKB (<39°N) and the SSPR in the northern EKB (>39°N) is 0.44 during the historical period (figure 3(a)).
Surface currents cause the eggs and larvae to become widely dispersed over time.Walleye pollock larvae are the most abundant in the middle region of the EKB along the eastern coast of North Korea after 15 days from spawning in RCM-ENS (figure 3(b)).Owing to the longer transportation time of the surface current, the hatched larvae spread farther from the spawning region after 30 days.More walleye pollock larvae were concentrated in the northern EKB than in the southern EKB after both 15 and 30 days of spawning (figures 3(b), (c)).
At the end of the 21st century, SSPR declined substantially in the RCM-ENS (figure 3(d)).The mean SSPR in the northern EKB was 0.13 during the future period, which is a 70% decrease from the historical period.The SSPR was only 0.04 in the southern EKB, suggesting that walleye pollock may not spawn on 96% of the days in the southern EKB by the end of the 21st century under the high-emission global warming scenario.The majority of larvae moved to the center of the EKB and were then transported offshore owing to the separation of the EKWC (figures 3(e), (f)).The larval abundance after 15 days still demonstrated an asymmetric distribution, with more larvae in the northern EKB, owing to the higher spawning rate in the north.Our model projects that poor spawning conditions in the EKB in the future will result in a decline in the overall population.Larvae 30 days after spawning followed similar patterns as 15 days after spawning; however, they were transported further offshore.
To examine the decreasing pattern under the warming scenario, we present walleye pollock egg and larval differences between historical and future periods in figure 4. In the EKB, the spatially averaged SSPR decreases to 0.27 (figure 4(a)), which was much larger in the northern EKB (0.31) than in the southern EKB (0.17).Compared with the results during the historical period, the entire EKB decreased by 75.5% (figure 4(b)).The southern EKB   had an SSPR decline of 85.3%, which was higher than that of the northern EKB (71.0%).This asymmetric decrease can be attributed to the meridionally distinct EKB temperature pattern separated by the subpolar front.Compared to the warm southern EKB, the cold northern EKB maintained a higher spawning rate.However, the warm southern region can quickly reach temperatures that are too high for walleye pollock to spawn under global warming.Therefore, spawning may be more problematic in the southern EKB under warming scenarios.
The abundance of walleye pollock larvae decreased (figures 4(c)-(f)).The majority of larvae vanished in the central EKB around the spawning grounds.However, a small region off the coast of the EKB in the vincinity of the separation latitude of the EKWC (around 39°N) remained, where the abundance increased or weakly decreased marginally.This suggests that enhancing the EKWC promotes offshore transport of eggs and larvae, further hindering their survival.
Regarding changes in the spawning days caused by future warming in each RCM, two different SSPR patterns were projected depending on local circulation patterns.In moderate warming due to the restricted separation of the EKWC around 39°N, the spawned eggs were present, although the number of the eggs substantially decreased.In contrast, most of the spawned eggs disappeared because of the excessive warming caused by EKWC overshooting.The SSPR and particle tracking results of each RCM and the temporal differences between historical and future periods are detailed in Supplementary section 5.

Discussion
In this study, we analyzed the effects of future global warming on the spawning days and larval transport of walleye pollock in the western EJS.We implemented dynamic downscaling to simulate local environmental changes, which are limited to low-resolution global climate models.The RCM-ENS projected a SST increase of 4.8 °C over the EKB between the historical and future periods (figure 2(g)).This is considerably higher than the global mean SST increase in selected GCMs (3.2 °C) and CMIP6 multi-model ensembles (2.9 °C) from IPCC Assessment Report 6 [17].The RCMs successfully simulated that the transported eggs and larvae of the walleye pollock were concentrated in the northern EKB during the historical period.Under the SSP5-8.5 scenario, the results of the RCMs showed substantial reductions in walleye pollock spawning and larval dispersion by the end of the 21st century.
Our findings indicate that the abundance of eggs and larvae in the western EJS walleye pollock will drastically reduce in the future.They used to spawn during the coldest months (January and February) in the western EJS, so an increase in future minimum temperature in the western EJS could lead to a decrease in SSPR, and the spawning day could even vanish in the southern EKB.The reduced number of young walleye pollock could cause recruitment failure in the western EJS.Recruitment failure is also associated with decrease in spawning stock biomass which can also lead to the decline of the total walleye pollock stock.Furthermore, in a warmer future, the spawning of walleye pollock in the western EJS could be adjusted by the cyclonic circulation, which supplies cold water.In this case, since most of the eggs should be transported by the cyclonic circulation, most of the eggs spawned in the southern EKB should be transported offshore by the EKWC, although other eggs spawned in the northern EKB would be transported to the nursery in the western EJS.Thus, this limited eggs and larval transport pattern may also contribute to the acceleration of the recruitment failure in the southern EKB.Evolutionary adaptation to warming is generally thought to be a slower process than migration [62].As a survival strategy, walleye pollock may permanently shift their spawning area northward in search of a colder environment.Previous studies on other marine species have demonstrated a projected poleward movement at both the leading and lagging edges into waters that align with their thermal tolerance [63][64][65][66].As a result of permnantly shift of walleye pollock habitats and increas in temperature, other warm-temperature favoring species, such as common squid and jack mackerel, may proliferate in the south of EKB.
The ocean warming pattern of the CMIP6 multi-model ensemble represents spatially nonuniform trends [67].Heat uptake at the ocean surface and interior heat transport governed the spatial warming trends.By 2100, the added heat at the air-sea interface will globally dominate the overall future warming patterns, and ocean dynamics, such as current, eddy mixing, and small-scale diffusion, will play an important role in heat redistribution.A previous study focusing on the EKB also showed that atmospheric forcing and oceanic changes from the lateral boundary condition primarily determine the temperature and circulation patterns [68].Figure 5 shows the cumulative heat uptake, surface net heat flux, and lateral heat transport up to 500 m in the RCM-ENS during historical and future periods and the differences between both periods.Positive cumulative heat uptake indicated that the heat content increased during both the historical and future periods (figures 5(a), (d)); however, the increase was greater near the spawning ground during the future period (figure 5(g)).To estimate the effects of the surface net heat flux and lateral heat transport on the changes in heat uptake, their cumulative values were calculated for each period.During both periods, heat was primarily delivered to the EKB through lateral heat transport (figures 5(c), (f)), whereas heat was released to the atmosphere via the surface net heat flux (figures 5(b), (e)).This phenomenon was attributed to the influence of the EKWC, which transports warm water from lower latitudes, resulting in apparent heat loss from the warmer water to the atmosphere [69].At the end of the 21st century, there was a significant increase in heat uptake near the spawning grounds compared to the historical period (figure 5(g)).Differences in the cumulative surface net heat flux and lateral heat transport suggest the contributions of surface heat flux and lateral heat transport to the enhancement of heat uptake during the future period compared to that during the historical period (figures 5(h), (i)).A positive sign in in figure 5(h) over the EKB indicates a distinct reduction in the surface net heat flux during the future period, indicating that the weakening of the cooling owing to the decrease in the surface net upward flux in the future plays a major role in the increase in heat uptake over the EKB.
Based on the heat budget analysis conducted using the RCM-ENS, it is evident that the net surface heat flux plays a significant role in the future warming of the EKB.However, it should be noted that not all members of the RCM exhibit the same circulation pattern over the bay.In other words, the RCMs projected two different circulation patterns over the EKB, suggesting two scenarios for surface warming (figure 6).If cyclonic circulation predominates in the EKB, the cold water from the northern region mitigates the warming effect because of the surface heat flux.This results in a suitable spawning temperature, allowing walleye pollock to spawn considerably because of moderate warming (figure 6(b)).Nonetheless, when the EKWC overshoots, walleye pollock cannot spawn due to excessive warming by oceanic heat transport and eventually become extinct in the EKB owing to the combined effect of warm water conveyed by the EKWC and global warming (figure 6(c)).These two scenarios suggest that the reduction in spawning grounds in the EKB at the end of the 21st century is notable because of a decrease in heat transfer from the ocean to the atmosphere, and changes in local circulation patterns could modulate the severity of the reduction.The local circulation and warming patterns in each RCM are described in Supplementary section 4.
The projection results under global warming would have several implications for fishermen and fisheries management.Firstly, they may assist fishermen in scheduling their livelihood activities.For instance, while pollock is vulnerable to climate change, other species like jack mackerel are less susceptible to global warming [70].Our results can help inform whether fishermen should continue to focus on catching other species or make efforts to catch walleye pollock in the future.Additionally, they can aid in making long-term fisheries management policy decisions in the Republic of Korea.The Korean government has attempted to release a large number of eggs and larvae produced by pollock farming into the coastal waters to restore walleye pollock stocks in western EJS.The study results could aid the government in determining whether to continue similar efforts in the future and, if so, in which areas.Given that both of our moderate warming and excessive warming cases resulted in a decrease in walleye pollocks in the study area, the Korean government may consider it risky to release eggs in the future.However, they can expect relatively promising results if there would be moderate warming pattern in the future.The strategy can be chosen based on whether the near-term prediction follows an extreme warming pattern or a moderate warming pattern.

Conclusion
According to our future model simulation, 76% of the mean Suitable Spawning Period Ratio (SSPR) in the East Korea Bay (EKB) will significantly decrease owing to the reduction in the heat transfer from the ocean to the atmosphere compared to the value in the historical period for the RCM-ENS.In addition, two different walleye pollock collapse patterns are possible even under the same global warming scenario according to the local circulation characteristics around spawning grounds.
While predictions for the entire global lower trophic ecosystem using climate models are actively researched, studies on the future changes in higher trophic level species are relatively limited.This limitation may stem from inaccuracies caused by the coarse grid of GCMs and the complex parameterization of higher trophic models.Our interdisciplinary approach employing dynamical downscaling, ensemble techniques, and particle tracking is expected to contribute to predicting the distribution of higher trophic level species.Although this study is confined to the coastal areas of the western part of the EJS, our approach can be applied to other regions and vulnerable species affected by climate change.For instance, the significant fluctuation in squid catches in the EJS requires further investigation into how the physical environmental changes in the ocean due to global warming may impact the squid's habitat ground and early life stages within the EJS [71].
In this study, we addressed winter temperature, circulation, and heat flux changes using small ensemble members.To decrease uncertainty, large ensembles of RCMs should be used in the future to investigate the primary drivers of changing patterns in the physical environment.In addition, we only considered the effects of temperature and current on the transport of walleye pollock eggs and larvae.The combination of the dynamic downscaled hydrodynamic model and the individual-based model with biotic factors can address the biological mechanism changes in an ecosystem and expand the research field beyond passive particle tracking experiments.

Figure 1 .
Figure 1.(a) Major currents in the East/Japan Sea (EJS).The TC, EKWC, and NKCC represent the Tsushima, East Korea Warm, and North Korea Cold Currents, respectively.(b) Study area indicating the spawning ground (red shading) that the walleye pollock in the East Korea Bay (EKB) and Korean fishing area (blue shading) adapted from [14].

Figure 2 .
Figure 2. Temporal mean Sea Surface Temperature (SST; units °C) and surface velocities around the EKB.The first column (a), (c), (e), (g) and second column (b), (d), (f), (h) represent SST and velocities, respectively.The first (a), (b) and second row (c), (d) represent the satellite observation data and ensemble mean of RCMs (RCM-ENS) for the historical period (1995-2014), respectively.The Optimum Interpolation Sea Surface Temperature (OISST) and altimetry reprocessed geostrophic current by Copernicus Marine environment Monitoring Service (CMEMS) were used as observation data (a) and (b), respectively.The third row (e), (f) represents the RCM-ENS horizontal maps for the SSP5-8.5 scenario (2081-2100).(g) and (h) represent the difference map between the future and historical periods.

Figure 3 .
Figure 3. Suitable Spawning Period Ratio (SSPR) of walleye pollock (a), (d) and transported larvae after 15 days (b), (e) and 30 days (c), (f) from the release.The first row (a), (b), (c) and second row (d), (e), (f) are the results of RCM-ENS for the historical and future periods, respectively.SSPR is defined as the ratio between the successful spawning days divided by the total releasing days.The mean number of transported larvae over each period are represented as colour on the grids (b), (c), (e), (f).

Figure 4 .
Figure 4. Changes in the SSPR (a) and the larval abundance (c, e) between the historical and future periods.(c) and (e) represent abundances 15 and 30 days after spawning, respectively.(b) indicates a decreased ratio (%) in the future compared to the historical SSPR.(d) and (f) represent a decreased ratio compared to the historical abundances 15 and 30 days after spawning.To exclude the negligible areas where abundances are close to zero in the historical period, we displayed the region where abundances exceeded the median value for the entire region during the historical period.

Figure 5 .
Figure 5. Cumulative heat uptake up to 500m in the RCM-ENS during (a) the historical and (d) future periods.Cumulative (b), (e) surface net heat flux and (c), (f) lateral heat transport during (b), (c) the historical and (e), (f) future periods.Difference of (g) cumulative heat uptake, (h) surface net heat flux and (i) lateral heat transport between the future and historical periods.Positive/ negative sign means more heat gain/loss from surface net heat flux or lateral heat transport for the future period.

Figure 6 .
Figure 6.Schematics of walleye pollock collapse patterns based on local circulations in the future.(a) indicates an environment favorable to walleye pollock spawning in the western EJS.(b) represents a pattern of moderate warming, where EKWC separates at low latitude and northern cold water from cyclonic circulation block excessive warming, resulting in less collapse of walleye pollock despite the same global warming.(c) illustrates the pattern of excessive warming that the EKWC predominates in the EKB.Walleye pollock disappear from the western EJS because of the combination of global warming and warm water transported by the EKWC.