Evidence that spillover from Marine Protected Areas benefits the spiny lobster (Panulirus interruptus) fishery in southern California

Marine Protected Areas (MPAs) are designed to enhance biodiversity and ecosystem services. Some MPAs are also established to benefit fisheries through increased egg and larval production, or the spillover of mobile juveniles and adults. Whether spillover influences fishery landings depend on the population status and movement patterns of target species both inside and outside of MPAs, as well as the status of the fishery and behavior of the fleet. We tested whether an increase in the lobster population inside two newly established MPAs influenced local catch, fishing effort, and catch-per-unit-effort (CPUE) within the sustainable California spiny lobster fishery. We found greater build-up of lobsters within MPAs relative to unprotected areas, and greater increases in fishing effort and total lobster catch, but not CPUE, in fishing zones containing MPAs vs. those without MPAs. Our results show that a 35% reduction in fishing area resulting from MPA designation was compensated for by a 225% increase in total catch after 6-years, thus indicating at a local scale that the trade-off of fishing ground for no-fishing zones benefitted the fishery.

www.nature.com/scientificreports/ documented a 245% increase in lobster caught within reserves that was linked to an 87% increase in nearby fished areas. The net benefit to the fishery was attributed in both cases to the spillover of relatively large lobsters from MPAs. Such studies are critical for demonstrating the ability of MPAs to meet their fishery objectives because without them the implementation of MPAs risks being perceived as having unfulfilled expectations and a lack of credibility in their purported value as an effective tool for fisheries management 32 . The long history of coastal MPAs in California (CA) provides a unique setting for studying their effects. Small coastal MPAs have existed in the state since the 1930s. Passage of the Marine Life Protection Act in 1999 led to the establishment of an additional 124 MPAs, encompassing 2197 km 2 of protected marine habitat, distributed along 1900 km of coastline 33 . Studies conducted within CA's MPAs have focused primarily on understanding the biological and ecological consequences of protection. For example, researchers studying a network of MPAs at the Channel Islands in southern CA documented significant positive effects of protection on a wide variety of fish and invertebrate species within and immediately outside of reserve borders 34,35 . One of the largest responses was observed in the CA spiny lobster (Panulirus interruptus), which increased dramatically in both biomass and size within multiple MPAs after only five years (2003)(2004)(2005)(2006)(2007)(2008) of protection 21 , and exhibited spillover into adjacent fishing zones 36 . The rapid response of lobster to MPA protection was rather surprising because the lobster fishery in CA is considered relatively sustainable 37 . Whether the gain in lobster abundance in fishing zones adjacent to MPAs is sufficient to compensate for the reduction in fishable area created by the designation of MPAs has yet to be determined.
The establishment of a network of no-take zones in 2012 along the mainland coast of southern CA, where CA spiny lobster is heavily fished, provided us with a unique opportunity to study the influence of MPAs and spillover on the lobster fishery and to determine its potential benefit to the commercial lobster fishery. Prior observations of spillover of lobster from MPAs within five years of establishment at the offshore Channel Islands 33 suggested that a similar pattern would emerge along the adjacent mainland coast, especially because the mainland fishery is more heavily fished than that at the Channel Islands 27,38 . Therefore, we tested the hypothesis that the establishment of MPAs along the mainland coast of the Santa Barbara Channel would lead to increased lobster biomass and density inside MPAs, which in turn would lead to increased landings in the region despite the reduction in fishable area caused by the establishment of the MPAs. Our approach involved analyzing: (1) fishery independent data collected by divers within and outside of MPAs, immediately before and six years after MPA establishment, to determine the effects of the MPAs on lobster density, size and biomass, and (2) fishery dependent landings data of lobster catch and fishing effort, also collected six years before and after MPA establishment, to assess the effects of the MPAs on the fishery. Our results are important to fishers, who were assured spillover would eventually enhance lobster catches, and fishery managers, who worked in collaboration with fishers to established reserves with fishery benefits.

Results
Fishery independent estimates of lobster abundance and size. The density of lobster increased significantly over time in both fished and unfished plots, however the rate that density increased in unfished MPA plots was three times higher than that in fished plots ( Fig. 1a; F 1.248 = 5.53, P = 0.020 for protection status × time interaction). By contrast, the mean carapace length of lobster did not change significantly over time in either fished or unfished plots ( Fig. 1b; F 1.182 = 2.12, P = 0.147 for protection status × time interaction), although it's worth noting the non-significant trends in fished and unfished plots showed opposite patterns with size increasing over time in the unfished plots and decreasing in the fished plots. Similar patterns of non-significance and opposing directional trends were observed for changes in the median carapace length of lobster (F 1.182 = 2.16, P = 0.143, data not shown), which is more sensitive to skewed distributions.
Observed changes in lobster biomass following MPA establishment closely resembled those of lobster density ( Fig. 1a vs. c). Biomass increased significantly over time following MPA establishment in both fished and unfished plots (F 1.131 = 4.65, P = 0.033 and F 1.118 = 62.45, P < 0.0001 for fished and unfished plots, respectively), with threefold higher increases in unfished MPA plots compared to fished plots ( Fig. 1c; F 1.248 = 12.14, P < 0.001 for protection status × time interaction).

Fishery dependent estimates of lobster catch and fishing effort. Analyses of commercial fishing
block landing data showed that the effects of fishing block on the total annual catch of lobster differed significantly between the six-year periods before and after MPA establishment ( Fig. 2a; F 3.40 = 4.86, P = 0.006 for block × time period interaction). Average annual lobster landings more than doubled in the 6 years after MPA establishment in the block with the two MPAs (#654) despite a 35% reduction in the fishable area (Table 1). In contrast, lobster landings in the three blocks without MPAs remained relatively unchanged during the six years following MPA establishment.
Annual fishing effort estimated by the number of traps pulled increased throughout the study region following MPA establishment with the largest increase occurring in block 654 ( Fig. 2b; F 3.40 = 4.31, P = 0.010 for block × period interaction). CPUE differed significantly among the fishing blocks ( Fig. 2c; F 3.40 = 4.54, P = 0.008), but not between the periods before and after MPA establishment (F 1.40 = 0.84, P = 0.850). Block 654 with the two MPAs consistently had the highest CPUE irrespective of the timing of MPA establishment. The nearly identical values of CPUE in block 654 before and after MPA establishment suggests the doubling of landings in this block in the six years following the establishment of the two MPAs resulted from a doubling of the fishing effort rather than an increase in lobster stock.
The average annual catch in the study region summed across all four fishing blocks increased by 57% in the six years following MPA establishment while the fishing effort increased by 73%. By contrast, the average annual CPUE of the region declined slightly from 0.68 (± 0.051 SE) to 0.61 (± 0.058 SE) pounds per pull.

Discussion
Spiny lobster density and biomass increased in MPAs and fished areas in the 6 years following the establishment of the two no fishing zones. The increase in lobster abundance across the region observed in our fishery independent data may be related to a suite of factors that we did not test, including enhanced lobster recruitment related to El Nino events in 2006-2007, and 2014-2016 39 . However, the increase within the MPAs was four times greater for density, and almost twice as much for biomass than in the fished areas, indicating lobster abundance increased dramatically in the absence of fishing. Similar increases in lobster populations have been observed within established no-fishing zones at the Channel Islands 40 and elsewhere 12,41 , but over a substantially longer period of time (15-23 years). That the change in mean size of lobsters did not vary markedly inside vs. outside of MPAs is interesting but not unexpected. Prior work comparing lobster sizes inside vs. outside MPAs in the Channel Islands has been equivocal concerning effects on mean carapace length: Iaccheci et al. 41 found little difference for relatively long-established MPAs while Kay et al. 21 found large positive effects of MPAs only 5 years after they were established.
Cessation of fishing appears to have been the primary mechanism driving the relatively large difference in lobster density and biomass inside vs. outside of the MPAs. Nevertheless, other ecological mechanisms may have also played a part, including spatial variation in lobster habitat and recruitment, and differences in food availability. Prior work in the Santa Barbara Channel found that the intensity of lobster spillover from MPAs across multiple reserves was driven mainly by cessation of fishing, and to a lesser degree by the spatial arrangement of the caves, cracks, and crevices where lobsters gather to shelter from their predators 36 . While we did not quantify the amount and quality of lobster habitat in this study, we have no reason to believe that those habitat metrics changed over the course of the study. Of course, habitat availability is a key factor in spillover because as lobster density increases inside reserves preferred shelters become crowded leading to the movement of lobsters to We have no direct evidence that lobster populations increased inside the MPAs due to greater food availability. An indirect ecological effect of ceasing to fish lobsters on temperate latitude rocky reefs is an increase in the abundance of macroalgae-specifically giant kelp in the SB Channel-due to the decline of sea urchins, which graze on kelp and are preyed upon by lobsters 42 . In turn, increased kelp abundance can lead to greater abundances of many species of invertebrates, some of which are lobster prey 43 . We do not have the data to test this idea but other work in our region indicates that lobster prey species tend to decline instead of increase inside MPAs, apparently due to increased lobster forging 44 . In summary, indirect ecological effects related to   www.nature.com/scientificreports/ the cessation of fishing may have contributed to the large increase in lobster abundance in the MPAs, but do not adequately explain the dramatic differences we report between fished and unfished areas. Fishery-dependent catch data showed that lobster landings in the fishing block 654 with MPAs increased by ~ 225% after the reserves were implemented compared with an average increase of 19% in the three blocks without MPAs. Importantly, the large increase in catch in fishing block 654 occurred despite a 35% decrease in fishable area, and resulted from a 250% increase in effort, as catch efficiency (i.e., CPUE) remained relatively constant. The pattern of increased effort and constant CPUE implies there was a substantial increase in the abundance of legal sized lobsters in the fishable area of fishing block 654 after the MPAs were established, which is consistent with the results of the diver surveys. By contrast, the CPUE in the three fishing blocks without MPAs declined by an average of 13% in the six years after the MPAs were established. We predict that this pattern was driven mainly by the most efficient fishers who shifted their effort to fish close to the MPA borders. To our knowledge, this is the largest increase in lobster catch associated with MPA spillover ever recorded.
The most plausible mechanism leading to the increase in lobster abundance and catch in the fishing block with MPAs was the movement of legal-sized lobsters from within the reserve boundaries to areas outside where they were caught. Such movement from the Naples SMCA and Campus Point SMCA may be facilitated by extensive rocky reefs, the preferred habitat of spiny lobster, that extend uninterrupted from within the MPA's boundaries to areas outside where fishing occurs. Such habitat corridors are well-known to facilitate lobster movement 45 . Prior work indicates a lack of such habitat corridors reduces, or even impedes, lobster movement and spillover 21,38 . Prior work in the northern Channel Islands also suggests that substantial increases in population abundance within reserves takes place within 5 years of reserve establishment, and that increases in abundance are positively associated with greater movement of legal-sized lobsters 21 . Considered together, this is strong circumstantial evidence that substantial increases in lobster abundance within multiple MPAs over a six-year period increased the local commercial catch through spillover.
Interestingly, fishing effort also increased in nearby fishing blocks without MPAs (albeit at a much smaller rate) between 2012 and 2017, indicating that there was an overall increase in fishing effort in the region following the establishment of MPAs. This is consistent with results from visual surveys of deployed lobster traps (which are marked by surface floats), which implied that the increase in fishing pressure was concentrated at the borders of the two MPAs in fishing block 654. For example, the average number of traps counted along the eastern border of the Campus Point SMCA during the second week of the lobster fishing season nearly doubled from 46 ± 16 traps per hectare in 2009-2012 to 89 ± 14 traps per hectare in 2014-2019 (Authors, unpublished data). The higher fishing effort and catch coupled with a relatively constant CPUE in block 654 suggests fishers were attracted to the MPAs because of enhanced catch from spillover. This is born out in interviews with fishers who acknowledged that they purposefully fish the borders of the Naples and Campus Point SMCAs to maximize their catch (H. Lenihan, personal communications).
Theory suggests that MPAs benefit fisheries the most when they are poorly managed and overfished prior to reserve implementation 31,46,47 . Empirical evidence so far supports this prediction 19,30,48,49 . For example, Halpern et al. 50 used statistical analyses of 17 MPA-fishery systems to suggest that spillover can replenish a highly depleted fishery that targets a mobile species when: (1) the gradient of decay in population abundance across the MPA's border is steep, (2) catch is greater than the population growth rate outside the reserve, and (3) there are multiple reserves subsidizing the fishery, such as in a reserve network. A synthesis of field studies by Goñi et al. 19 also reported patterns consistent with MPAs replenishing heavily exploited fisheries in which the abundance of target species in fished areas was far below that in no-fishing zones 51 . In addition to heavy fishing pressure, other important elements stimulating high catch outside of MPAs appear to be habitat characteristics that promote the movement of animals across MPAs borders 26,36 , good fishing habitat in close vicinity (e.g., < 800 m) to MPA borders 36,50 , and a behavioral response by the fishery to fish near the borders.
Our results provide one of the few examples that link increased abundance of a target species within reserves to increased catch within the fishery, albeit at a local scale (see review by Goñi et al. 18 ). In his review Hilborn 52 reiterated the growing consensus, gleaned from models and empirical work 19,31,50,[53][54][55] , that MPAs increase the catch outside of no-fishing zones when fishing pressure is very high and stocks are seriously overexploited. By contrast, we have shown in this study that MPAs can increase catch in a relatively sustainable fishery. Certainly, the CA spiny lobster fishery is heavily exploited but remains sustainable, due in large part to management efforts designed to control fishing effort, and oceanic connectivity that delivers larvae to southern CA from less-heavily exploited fishing grounds in Baja California, Mexico 37,56 . An important general lesson re-emerges from our results: close collaboration between fishers, scientists, policy makers, and fishery managers enhances marine ecosystem management 10,57,58 . California's MPA-fishery system is an integrated, well-enforced network of no-fishing zones designed to protect productive subtidal rocky reefs that are essential habitat to populations of many target species, including CA spiny lobster. The lobster fishers helped to design the CA MPA network, and the State dedicates substantial funding to effective enforcement. Our results indicate that by responding to the no-fishing zones by fishing near MPA borders, the fishery is experiencing a net benefit of no-fishing zones at a local scale. The next step is to test whether southern California's MPA network is enhancing total catch across the entire CA spiny lobster fishery.  (Fig. 3). At each site divers recorded the number and visually estimated the size (i.e., carapace length) of lobster in 1200 m 2 plots (n = 2 to 8 plots per site depending on the available reef habitat). Using handheld lights divers thoroughly searched the benthos, including areas of dense vegetation, crevices, ledges, and other structures used by lobster to shelter, to obtain a census of lobster in each plot (n = 36 plots total). Surveys were conducted annually from 2012 to 2018 during daylight hours in late August to mid-September prior to the start of the lobster fishing season to control for possible confounding effects of habitat type, storms, season, and time of day on lobster abundance (sensu Iacchei et al. 41 ). Details of the sampling methodology and data can be found in Reed 60 . We classified plots as fished (n = 17) or unfished (n = 19) depending on whether or not they were located in an MPA. The effects of protection status (fish vs. unfished) and time since MPA establishment on the absolute change in lobster density, mean carapace size, and biomass since MPA established were determined in separate ANCOVAs in which protection status was considered a fixed factor and time since MPA establishment a covariate. The mass of each lobster (wet g) was calculated as 0.001 × L 2.914 where L is the carapace length in mm 61 . Lobster biomass was calculated as the summed mass of all lobster observed in a plot. Lobster survey data were not collected prior to 2012 when MPAs were established.

Methods
Fishery dependent landings data. CA fisheries are managed by the CA Department of Fish and Wildlife (CDFW), who has divided the entire coastline into rectangular fishing blocks (~ 140 km 2 ), from which commercial lobster fishers, and other fisheries, are required to record and log all their catch (https ://nrm.dfg.ca.gov/FileH andle r.ashx?Docum entID =67449 &inlin e). Annual fishing block data of commercial landings (wet kg caught) and fishing effort (number of traps pulled) of spiny lobster in the study region were obtained from CDFW for the six fishing seasons prior to MPA establishment (2006-2011) and the six seasons after MPA establishment www.nature.com/scientificreports/ (2012-2017). We defined a fishing season by the year in which it started (e.g., the 2012 season extended from October 2012 through March 2013). Fishing block data are based on landings weighed at the dock by the processor, who records the data on a "fish ticket" that is submitted to the CDFW. Fishermen are required to assign their landings to a specific fishing block and report the fishing effort (i.e., number of traps pulled) allocated to their catch. Catch-per-unit-effort (CPUE) is defined by CDFW as the number of legal lobsters per trap pull. Herein, we define CPUE as the weight of lobsters pulled per traps pulled. The study region included four commercial fishing blocks that extend along the mainland coast of the Santa Barbara Channel (Fig. 3). The two MPAs (Naples SMCA and Campus Point SMCA) are located in a single block (#654) and their designation in 2012 constituted a 35% reduction in the fishable area in this block ( Table 1). The CA Marine Life Protection Act had no effect on the amount of fishable area in the other three fish blocks in the study region (# 652, 653, 655).
The effects of fishing block and period (i.e., before vs. after MPA establishment) on annual catch, annual effort, and annual catch-per-unit-effort (CPUE) were evaluated in separate fixed factor ANOVAs. Landings and effort data were log transformed to meet the assumptions of normality and homoscedasticity. All statistical analyses were conducted in SPSS-SYSTAT software.
Ethical approval. All methods were carried out in accordance with University of California and National