Growth parameter k and location affect body size responses to spatial protection by exploited rockfishes

Introduction

For centuries, many Indigenous cultures have recognized that protecting parts of the ocean from exploitation mitigates human impacts on biodiversity and contributes to fishery sustainability (Johannes, 1978; Jones, Rigg & Lee, 2010; Ban et al., 2018). Consistent with the knowledge and practices of these cultures, there is growing evidence that marine reserves or other forms of spatial protection promote recovery from exploitation (Baskett & Barnett, 2015), and that spillover of adults and larvae from protected areas can enhance adjacent fisheries (Di Lorenzo, Claudet & Guidetti, 2016; Baetscher et al., 2019).

The benefits of spatial protection often are quantified as increases in the density, biomass, and body sizes of exploited species (Baskett & Barnett, 2015). Understanding factors affecting body size distributions is particularly important because fisheries remove the largest individuals, often at great ecological cost (Strong & Frank, 2010; Hixon, Johnson & Sogard, 2014). Larger individuals tend to occupy higher trophic positions (e.g., Trebilco et al., 2016; Olson et al., 2020), and their removal may alter food web structure (Shackell et al., 2010; Zgliczynski & Sandin, 2017). Further, for many fish taxa the relationship between fecundity and body size is hyperallometric (a power function with exponents >  1), such that larger females contribute disproportionately more offspring (per unit of body size) than smaller females (Dick et al., 2017; Barneche et al., 2018). Size differences between protected and fished areas, therefore, can signal the extent to which spatial protection promotes species recoveries and restores food webs.

The benefits of spatial protection, however, can differ across species, and understanding this variation may help predict reserve performance (Jennings, 2000; Claudet et al., 2010; Kaplan et al., 2019). Species with longer lifespans often have slower growth rates, later maturity, and other traits which increase their vulnerability to exploitation and reduce their recovery rates during fishery reprieves (Jennings, 2000). Consequently, the benefits of spatial protection may require more time to manifest in longer-lived species (Starr et al., 2015; Kaplan et al., 2019).

Variation in parameter k of the von Bertalanffy growth function, which depicts the rate at which the asymptotic body size is approached, should influence the time required by different species to restore a larger size structure after the implementation of spatial protection (Jennings, 2000; Kaplan et al., 2019). Additionally, species characterized by smaller movements and greater site fidelity are more likely to remain within a protected area (Hannah & Rankin, 2011) and may benefit more from spatial protection than more mobile species (Kramer & Chapman, 1999; Moffitt et al., 2009). These predictions, however, do not apply to unfished species, for which environmental variation is a primary driver of population characteristics (Claudet et al., 2010; Caselle et al., 2015).

Rockfish (Sebastes spp.), a genus of marine fish with diverse life history traits, allow tests of these ideas. In the northeast Pacific, maximum lifespans range across rockfish species from two decades to two centuries, and interspecific variation in body size and growth parameters are similarly wide (Love, Yoklavich & Thorsteinson, 2002). While most rockfishes associate with rocky reefs, their behavior ranges from sedentary to very mobile and from demersal to benthopelagic. Most rockfishes are exploited yet some small planktivores are unfished (Love, Yoklavich & Thorsteinson, 2002).

In British Columbia, Canada, commercial and recreational sectors overfished rockfishes during the latter part of the 20th century, diminishing their abundance and body sizes (Yamanaka & Logan, 2010; Eckert et al., 2017). During the 2000s federal legislators implemented more conservative management and established a network of spatial fishery closures known as Rockfish Conservation Areas (RCAs) where commercial and recreational fishers cannot use bottom trawls, groundlines, or hook-and-line gear (Yamanaka & Logan, 2010; Appendix S1). For some species, however, declines in size and age structure have continued and their biomass remains depressed (McGreer & Frid, 2017) and references within). RCAs can help reverse these trends, yet prior studies in British Columbia, conducted only 5–7 years after RCA implementation, found no evidence of larger or more abundant rockfishes inside protected sites (Haggarty, Shurin & Yamanaka, 2016; Olson, Trebilco & Salomon, 2019). As time since RCA implementation progresses, size and biomass increases are more likely to be detected (Starr et al., 2015; Keller et al., 2019), particularly for shorter-lived species with greater k values (Kaplan et al., 2019).

Analyses of RCA benefits must account for variation in habitat characteristics which, independently of exploitation, affect the distribution and sizes of rockfish. For some demersal species, relative abundance may increase with topographic structural complexity and larger or older individuals tend to use greater depths than smaller or younger individuals (Love, Yoklavich & Thorsteinson, 2002). Benthopelagic species, however, have a wider range of vertical movements into the water column (Love, Yoklavich & Thorsteinson, 2002; Hannah & Rankin, 2011) and their relationships to depth or topographic structural complexity might be more variable.

Canada is among the countries where Indigenous people are using their traditional knowledge and science to improve marine conservation (Jones, Rigg & Lee, 2010; Ban et al., 2018; Ban, Wilson & Neasloss, 2020). Since 2013, the Central Coast Indigenous Resource Alliance (CCIRA)—comprised of the Wuikinuxv, Nuxalk, Heiltsuk and Kitasoo/Xai’xais First Nations—has been surveying rockfish populations and their habitats (Frid et al., 2018) inside and outside RCAs of British Columbia’s Central Coast (Fig. 1; Appendix S2). Data collected to date represent RCAs that were 8 to 15-years-old and encompass the earliest stage expected for the benefits of spatial protection to manifest for rockfishes (Keller et al., 2019; Kaplan et al., 2019).

Within this temporal context, we used CCIRA’s visual survey data to test the hypothesis that body size responses to spatial protection vary according to species life history traits, movement behavior, and susceptibility to fishing, while controlling for depth and topographic structural complexity. We predicted that the body sizes of exploited species would be larger inside RCAs than in adjacent fished areas, but the strength of this effect would decrease for species which require more time to reach their asymptotic body size (i.e., have lower k values) (Jennings, 2000; Kaplan et al., 2019) or that are more mobile (Kramer & Chapman, 1999; Moffitt et al., 2009). We also predicted that the body sizes of Puget Sound rockfish (S. emphaeus), a small planktivore unlikely to be caught by fishing gear (Love, Yoklavich & Thorsteinson, 2002), would not differ between protected and fished sites (Claudet et al., 2010; Caselle et al., 2015).

In addition to testing a priori predictions, we examined differences in conservation effectiveness between RCA locations. Though exploratory, this aspect of our research was motivated by the need to understand whether some RCAs are performing poorly and the potential ways to improve their management or design (Haggarty, Martell & Shurin, 2016; DFO, 2019a; DFO, 2019b).

Materials & Methods

Results

Of 13,224 observations for exploited species (Appendix S11), 73% were from high mobility-benthopelagic species, and 27% from low mobility-demersal species. The proportion of individuals that had reached or exceeded length at 50% maturity was 47% for low mobility-demersal species and 5% for high mobility-benthopelagic species (Appendix S12).

The effects of spatial protection on the body sizes of exploited species depended on growth parameter k and RCA location, but not on behavioral class (Fig. 2; Appendices S13A, S14A). At two RCAs (West Aristazabal, Goose Island), body sizes were larger at protected than control sites, but these differences were greater for slower- than for faster-growing species (Fig. 3). At three RCAs (Fish Egg Inlet, McMullin Group, Smith Sound), body sizes did not differ between protected and control sites, regardless of k values. At the remaining RCA (Kitasu Bay), species with greater k values were larger at control than at protected sites while species with lower k values had similar sizes in both treatments (Fig. 3). Independently of RCA treatment, the body sizes of high mobility-benthopelagic species increased with depth and decreased with greater topographic structural complexity. In contrast, the body sizes of low mobility-demersal species increased with both depth and topographic structural complexity (Fig. 4).

For the unfished Puget sound rockfish, body sizes increased with depth and topographic structural complexity but did not differ between RCA treatments or locations (Fig. 2B, Appendices S13B, S14B, S15).

Discussion

We predicted that the body sizes of exploited species would be larger inside RCAs than in adjacent fished areas, but the strength of this effect would decrease for species which require more time to reach their asymptotic body size (i.e., have lower values for growth parameter k) or that are more mobile. Our results were inconsistent with these predictions. At two RCAs (Goose Island, West Aristazabal) the body sizes of exploited rockfishes were larger at protected than at adjacent fished sites, but these differences increased with lower k values. One potential explanation for this result is that body size differences between protected and exploited sites could be influenced by an interaction between parameter k and the intensity of exploitation (see Jennings, 2000; Kaplan et al., 2019); exploitation rates at control sites may have not been high enough to generate strong RCA effects for species with high k values. Also, most rockfishes shift ontogenetically to deeper depths (Love, Yoklavich & Thorsteinson, 2002), and the depths that we sampled (≤35 m) encompassed primarily immature fish. Data collected at deeper depths via remotely operated vehicles (e.g., Haggarty, Shurin & Yamanaka, 2016) are required to more fully understand how variation in growth parameters affect responses to spatial protection.

Also contrary to our predictions, interspecific differences in movement behavior did not affect responses to spatial protection by exploited species. One potential explanation is that the proportion of immature fish was greater for high mobility-benthopelagic species (95%) than for low mobility-demersal species (53%). Most mature individuals from high mobility-benthopelagic species may have already undergone ontogenetic depth shifts below the reach of our dive surveys (Love, Yoklavich & Thorsteinson, 2002), reinforcing the need to collect data at deeper depths to strengthen inferences. Additionally, high mobility-benthopelagic species tended to have higher k values (Table 1), which might have confounded mobility effects.

The relative body sizes of high mobility-benthopelagic species increased with depth and decreased with greater topographic structural complexity, while the body sizes of low mobility-demersal species had the opposite relationship to these habitat variables. Perhaps the different responses reflect the distribution of resources being tracked (high mobility-benthopelagic species feeding more on pelagic prey near the surface: Love, Yoklavich & Thorsteinson, 2002) and different antipredator strategies (low mobility-demersal species relying more on crypsis and refuges in complex reefs).

The larger body sizes for exploited species within the Goose Island and West Aristazabal RCAs likely reflect reduced fishery mortality rather than environmental variation, as body sizes for the unfished Puget Sound rockfish did not differ between protected and fished sites. Given that fecundity increases disproportionately with body size (Dick et al., 2017) and that larvae exports can occur from protected to fished areas (Baetscher et al., 2019), these RCAs may already be contributing to fishery sustainability; such benefits should increase over time (Starr et al., 2015; Kaplan et al., 2019).

At four of six RCAs that we examined, however, rockfish body sizes at protected sites were similar to or smaller than those at adjacent fished sites. There are at least three potential, non-mutually exclusive explanations that might explain this result. First, the effectiveness of a protected area generally increases with its size and age (Molloy, McLean & Côté, 2009). In our sample, RCA age varied little, but RCA size ranged widely. The two RCAs that included larger fish within their boundaries (relative to adjacent areas) also were the two largest RCAs that we examined (Appendix S2). While more RCAs of varying sizes need to be examined, it is plausible that RCA size influenced our results.

Second, despite analyses controlling for depth and topographic structural complexity, some of the variation in RCA performance might reflect oceanographic differences between control and protected sites (see Caselle et al., 2015). Within a rockfish species, growth rates are faster in oceanic areas—which have lower temperatures and higher salinity—than in inland coastal waters (West, Helser & O’neill, 2014). Oceanographic context, therefore, may partly explain why we found no protection benefits at the Fish Egg Inlet RCA; this RCA encompasses inland coastal waters but local geography constrained control sites to more oceanic conditions (Fig. 2), where growth rates likely are faster (West, Helser & O’neill, 2014). Perhaps also reflecting oceanographic factors, at the Kitasu Bay RCA a control site just outside its boundary encompassed a reef with outstanding rockfish productivity, which might explain why individuals from faster-growing species were larger outside this RCA. Similarly, the Goose Island and McMullin Group RCAs share a boundary, yet only the former had larger fish in protected than in control sites; the control sites for the McMullin Group include several locations where conditions are more oceanic than at the control sites for Goose Island, which may have influenced this result (Fig. 2).

Third, while analyses are lacking for the extent to which legal and illegal fisheries occur within RCAs of the Central Coast, the intensity of these activities might have varied between locations. Studies in southern British Columbia documented a high incidence of illegal recreational fisheries within RCAs (Lancaster, Dearden & Ban, 2015; Haggarty, Martell & Shurin, 2016). Fisher compliance tended to decline with fishing effort adjacent to the RCA and with the RCA’s proximity to fishing lodges. Further, lower compliance was associated with fewer resources for enforcement and smaller RCA sizes (Haggarty, Martell & Shurin, 2016). Importantly, RCAs were created to protect rockfish from the highest-risk commercial and recreational fisheries, yet trap fisheries for invertebrates and mid-water trawls for groundfish remain legal within them (Appendix S1), likely causing rockfish mortalities (DFO, 2019b). The extent to which illegal and legal fisheries may have contributed to the lack of larger fish sizes at two thirds of the RCAs that we examined requires further investigation. Further, responses to spatial protection might be more likely to be detected if areas adjacent to reserves are fished more intensely (see Jennings, 2000; Kaplan et al., 2019), and future analyses should consider these effects.

Indigenous fisheries for traditional foods are allowed in RCAs, yet these fisheries are regulated by traditional laws derived from the long-term knowledge and stewardship principles that have allowed sustainable harvests over centuries (Jones, Rigg & Lee, 2010; Ban et al., 2018; Ban, Wilson & Neasloss, 2020). Hereditary chiefs remain the traditional stewards of their territories who, with technical staff, develop food fishing policies consistent with traditional laws. Such policies are operationalized via Coastal Guardian Watchmen who communicate with food fishers and patrol the waters for compliance, mitigating impacts from traditional fisheries in protected areas. Additionally, Central Coast First Nations run their own catch-monitoring programs, which they use to inform the governance and management of their traditional fisheries (Ban et al., 2018; Ban, Wilson & Neasloss, 2020). Thus, it is unlikely that Indigenous traditional fisheries contributed to the lower performance of some RCAs. Consistent with this notion, other studies in British Columbia found that the relative abundance of Dungeness crab (Cancer magister) increased over time after spatial fishery closures for commercial and recreational fishers where implemented, despite traditional Indigenous fisheries continuing within these closures (Frid, McGreer & Stevenson, 2016; Burns et al., 2020).

Conclusions

At two spatial fishery closures that were only 13 to 15 years-old when most data were collected, we found that the body sizes of 12 species of exploited rockfishes were larger at protected than at fished sites. Differences between protected and fished sites were greater for species with lower values for growth parameter k, which was the opposite of what we expected. Also contrary to expectation, interspecific differences in movement behavior did not affect body size responses to spatial protection. Nonetheless, life history and behavioral frameworks guided the prediction and interpretation of species differences in responses to spatial protection, which is essential for the adaptive management of protected areas (Jennings, 2000; Claudet et al., 2010; Kaplan et al., 2019).

Canada’s federal fishery management agency is reviewing the effectiveness of RCAs (DFO, 2019a; DFO, 2019b) and our results can inform this process. For four of six RCAs we found no evidence that fish were larger in protected than in adjacent fished sites. Perhaps these deficiencies in conservation effectiveness could be mitigated by increasing resources for compliance monitoring and enforcement, and potentially by modifying some RCA boundaries to increase the proportion of more productive, oceanic habitats within such boundaries.

Critically, the trophic position of rockfishes increases with their individual body sizes (Trebilco et al., 2016; Olson et al., 2020). Larger size classes of yelloweye rockfish (S. ruberrimus)—a long-lived demersal species inherent to Indigenous diets and undergoing size declines (Eckert et al., 2017; McGreer & Frid, 2017)—occupy a particularly high trophic position in rocky reefs (Olson et al., 2020). Ensuring that RCAs effectively restore large size structures for exploited species, therefore, is essential for ecosystem-based fishery management.

Supplemental Information