INTRODUCTION
The recreational fishing industry is growing, with numerous commercial fisheries under increasing recreational exploitation (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996; Cooke & Cowx, Reference Cooke and Cowx2004, Reference Cooke and Cowx2006). In 2000–2001 in Australia, an estimated 3.3 million people participated in recreational fishing. This figure equates to ~20% of the population, 23.2 million individual fishing events and over 20 million fisher days (Henry & Lyle, Reference Henry and Lyle2003). While time-series of catch and effort data are often available for commercial fisheries, this information is rarely available in the recreational sector (Henry & Lyle, Reference Henry and Lyle2003; Bridge & Conron, Reference Bridge and Conron2010).
Surf clam populations are prone to overfishing, with commercial and recreational sectors commonly competing for a declining stock (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996). Significant declines in surf clam abundance have resulted from the cumulative effect of both sectors (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996, Zeichen et al., Reference Zeichen, Agnesi, Mariani, Maccaroni and Ardizzone2002). While examples of recreational surf clam assessment do exist (Schoeman, Reference Schoeman1996; Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000; Hartill et al., Reference Hartill, Cryer and Morrison2005), the extent to which recreational harvest impacts on surf clam populations is still poorly understood (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996).
In Australia, the surf clam, Donax deltoides (known locally as the ‘pipi’), is widely distributed along the south-eastern coastline where they are commonly harvested by recreational fishers (Ansell, Reference Ansell, McLachlan and Erasmus1983; Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000). Genetic analyses have suggested historical population connectivity in south-eastern Australia fisheries (Versace et al., Reference Versace, Matthews, Miller, Bowie, Ierodiaconou, Mills and Lewis2011), with high levels of gene flow implying potentially high larval dispersal (Murray-Jones & Ayre, Reference Murray-Jones and Ayre1997). Pipis form dense aggregations in the intertidal sediments of high energy sandy beaches, constituting up to 98% of the macrobenthic infaunal biomass and playing a crucial role in beach ecology (Ansell, Reference Ansell, McLachlan and Erasmus1983). Population abundance is highly variable based on fluctuating levels of recruitment and mortality (King, Reference King1976; Ansell, Reference Ansell, McLachlan and Erasmus1983).
Stock declines have been observed in commercial fisheries in New South Wales (NSW) and South Australia (SA) (Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000; Murray-Jones & Johnson, Reference Murray-Jones and Johnson2003; Ferguson & Mayfield, Reference Ferguson and Mayfield2006). In NSW, catch and effort statistics have been compared between recreational and commercial sectors to assess the effects of harvesting (Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000). Recreational catch was shown to contribute up to 20% of total catch in NSW and over 90% of total harvest effort, highlighting the importance of understanding the impacts of recreational harvesting (Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000).
In Victoria, south-eastern Australia (Figure 1), pipis are targeted by recreational harvesters in Venus Bay, a large embayment, with no commercial fishery present. Harvest is managed through a combination of input (licensing, equipment restrictions–manual harvest only) and output controls daily catch limits (DCL). In response to anecdotal evidence of a decline in abundance of pipis in Venus Bay, managers have reduced DCLs (from 5 L to 2 L) (Fisheries Victoria, Reference Fisheries2010) to reduce potential effects on stock levels from recreational harvest. To date, no baseline data are available for the effects of this management to be assessed.
Location of the Venus Bay township (dark grey shading) and Coastal Park (light grey shading) along the Victorian coastline. The black circles denote sampling sites 1, 2, 3 and 4. Inset: Location of Venus Bay within Victoria, Australia (−38.678274, 145.790806).
Catch and effort (catch per unit effort) statistics have been used as indicators of stock abundance in commercial surf clam fisheries. In recreational examples, the large number of individual fishers, small scale of resources and a lack of economic value has reduced the effective application of these indicators; although examples do exist (Jaramillo et al., Reference Jaramillo, Pino, Filun and Gonzalez1994; Defeo & De Alva, Reference Defeo and De Alva1995; Schoeman, Reference Schoeman1996; Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000; Hartill et al., Reference Hartill, Cryer and Morrison2005). Fishery independent estimates of stock status present a means to inform sustainable management of surf clam stocks where catch statistics are unknown (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996; Hartill et al., Reference Hartill, Cryer and Morrison2005). While a suite of indicators that can be used in monitoring stock levels exist (including relative abundance, relative biomass, size–frequency and mature/immature ratios), little is known about the relevance of these indicators for use in recreational stock assessment.
The aim of this study is to apply indicators for monitoring a recreational surf clam fishery. In doing so, we investigate spatial and temporal trends in: (i) relative abundance; (ii) relative biomass and size structure; and (iii) ratio of sexual maturity for the Venus Bay surf clam fishery.
MATERIALS AND METHODS
Study site
Venus Bay is located on the south-east coast of the state of Victoria, approximately 175 km from the capital Melbourne. It comprises a wide southward facing bay characterized by a high-energy surf zone with a wide intertidal sand margin (Figure 1). This area provides ideal habitat for pipis (King, Reference King1976; Ansell, Reference Ansell, McLachlan and Erasmus1983) and has an established recreational fishery. Recreational access to the fishery is by foot only (no vehicle access), limited to five public beaches which are distributed in the north-west corner of a large protected Coastal Park within Venus Bay (Figure 1).
Experimental design
Sampling surveys were conducted during the peak recreational harvest season, from November to February (austral summer). Four sites situated along approximately 10 km of the beach were sampled each month. Three of the sites were situated at the most popular beaches in the Venus Bay township, public beach 5 (site 1), beach 3, (site 2) and beach 1 (site 3) (Figure 1). These three sampling sites shared common traits including ease of beach access and close proximity to well-developed tourism facilities (tourist signage, public bathrooms and car parking). Site 4 was located approximately 5 km south-east of site 3 and is publically accessible only by foot along the beach reducing its exposure to recreational harvest (Figure 1).
Within each site, three transects which were perpendicular to the shoreline were sampled. Transects were located 500 m apart and orientated with the middle transect centred on the beach access point for sites 1, 2 and 3. At site 4 where there was no beach access track through the dunes, the centre transect was located 5 km from the centre transect of site 3. Transects extended from the mean high water mark (MHW) to the start of the swash zone (generally around 60 m in distance). When the tide was at its lowest (>60 m distance from MHW), sampling occurred until 60 m from the MHW to standardize sampling effort between sites and across time.
Within each transect, 13 quadrats were sampled. Quadrats were located at 5 m intervals along transects starting at the MHW. A quadrat volume of 0.027 m3 (0.3 m × 0.3 m × 0.3 m) was considered sufficient to account for small-scale pipi variability from patchy distribution (Schlacher et al., Reference Schlacher, Schoeman, Dugan, Lastra, Jones, Scapini and McLachlan2008). Each quadrat was divided into three sand depth-classes: 0–0.1 m, 0.1–0.2 m and 0.2–0.3 m. Sand from each depth-class was excavated by hand and sieved to isolate pipi from the excavated sediments. A sieve mesh of 2 mm was considered appropriate to effectively isolate pipi from coarse sediments (Schlacher et al., Reference Schlacher, Schoeman, Dugan, Lastra, Jones, Scapini and McLachlan2008).
The number of pipis was recorded for each sample. Relative abundance used in stock analysis was calculated as the total number of pipis per transect (no.ts−1).
In order to estimate maturity, shell length was measured to the nearest 0.1 mm using digital Vernier callipers. Shell length was used as a proxy to classify pipi as: immature (<37 mm); and mature (≥37 mm). Due to an absence of local maturity data, length related maturity-classes were based on those described in NSW and SA fisheries (the study site is intermediate to these locations), which demonstrated consistent results across two geographically separate Australian populations (Murray-Jones, Reference Murray-Jones1999; Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000; Ferguson & Mayfield, Reference Ferguson and Mayfield2006). The proportion of maturity was calculated as the percentage of mature pipis present per transect.
To estimate relative biomass, each pipi was weighed in shell (wet weight) and measured to the nearest 0.1 g using a set of digital scales. Weight per unit area was calculated per quadrat. The mean weight per unit area in each transect was used as an estimate of relative biomass (kg.m−2), proportionalized to an area of 1 m2.
Statistical analyses
Each measurement was analysed using a series of restricted maximum likelihood (REML) mixed model analyses that examined fixed effect combinations for month and site after accounting for a random source variation for transect within site.
The measurements sampled at each transect were taken as an experimental unit in all REML analyses. Prior to REML analyses, relative abundance was logarithmically transformed and relative biomass and proportion of maturity were square root transformed to reduce skewness of residuals. For each measurement: relative abundance (no.ts−1), relative biomass (kg.m−2) and proportion of maturity (number of mature pipis/total pipis), a parsimonious model was selected by fitting a saturated model first and then dropping individual terms sequentially based on a Wald F test. The selected parsimonious model was used to achieve the predicted mean values with appropriate back transformation. It is worthwhile to note here that it was not possible to test the third-order interaction among month, site and transect due to the fact that transect was nested in site.
RESULTS
Changes in relative abundance
Abundance across a range of sand depths was not uniform. Pipis consistently occurred in the surface sediment (0–0.1 m), but were rarely observed in deeper sand. Pipis were observed in less than 2.5% of samples at 0.1–0.2 m sand depth and only 1.3% in 0.2–0.3 m sand depth (N = 603).
No significant interaction between month and site was detected while examining the proportion of maturity (Table 1). Relative abundance varied significantly among sampling sites (P = 0.0090). No significant difference was detected among months (Table 1). Mean relative abundance peaked at 50.08 ± 33.72 no.ts−1 in November before declining to 29.66 ± 10.78 no.ts−1 in December and 25.41 ± 9.44 no.ts−1 in January. There was a marginal increase to 35.27 ± 9.23 no.ts−1 in February.
P values of Wald F tests for choosing parsimonious models for each parameter: relative abundance (log10(y + 1) transformed); proportion of maturity (square root transformed); and relative biomass (square root transformed).
Relative abundance varied among months at each site (Figure 2). Sites 1 and 2 demonstrated a consistent trend in relative abundance peaking in November (156.66 ± 123.58 no.ts−1, 44.33 ± 21.45 no.ts−1), decreasing through December (16.00 ± 3.21 no.ts−1, 35.33 ± 32.33 no.ts−1) and January (10.00 ± 4.51 no.ts−1, 5.33 ± 3.84 no.ts−1) before increasing in February (17.66 ± 6.17 no.ts−1, 33.66 ± 24.67 no.ts−1) (Figure 2). Relative abundance peaked in December at sites 3 and 4 (54.00 ± 29.74 no.ts−1, 70.00 ± 24.70 no.ts−1) before declining in January (10.33 ± 4.81 no.ts−1, 13.33 ± 8.84 no.ts−1). Relative abundance further declined at site 3 in February (4.66 ± 2.01 no.ts−1) but increased at site 4 (35.33 ± 5.24 no.ts−1) (Figure 2).
Mean ± standard error relative abundance (no.ts−1) at sampling sites across month in Venus Bay. November is black, December is light grey, January is white and February is dark grey.
Changes in maturity and size structure
A significant difference in the proportion of maturity was observed among sites (P = 0.00026). There was no significant difference in the proportion of maturity among months or month and site interaction (Table 1). The mean proportion of adults present at site 4 (0.698 ± 0.066) was up to six times that observed at the other sites. The mean proportion of adults at sites 1–3 were 0.113 ± 0.040, 0.160 ± 0.061 and 0.097 ± 0.043, respectively.
Differences in size structure between sites reflected the trend in maturity (Figure 3). Size structures at sites 1, 2 and 3, were unimodal with a modal cohort of 10–15 mm present at all three sites (Figure 3). Up to 88.80% of pipis sampled at sites 1, 2 and 3 were <30 mm shell length, with few mature pipis (>37 mm). In contrast, large numbers of mature pipi (>37 mm) were present at site 4 with a modal cohort of 50–55 mm.
Length (shell)–frequency distribution at sampling sites in Venus Bay: (A) site 1 (n = 601); (B) site 2 (N = 353); (C) site 3 (N = 305); (D) site 4 (N = 383).
Changes in relative biomass
Relative biomass varied significantly (P = 0.0043) among months by sites. The trends in relative biomass during each month were differing at each site.
Relative biomass at sites 1 and 2 peaked at 0.11 ± 0.04 kg.m−2 and 0.10 ± 0.05 kg.m−2 respectively in November and declined across December and January before increasing slightly in February (Figure 4). At site 3, relative biomass peaked at 0.085 ± 0.01 kg.m−2 in December before continual decline across January and February. Relative biomass fluctuated monthly at site 4, peaking at 0.64 ± 0.09 kg.m−2 in December and again at 0.55 ± 0.11 kg.m−2 in February. The peak in relative biomass at site 4 was at least 6 times higher than at any other sites, with mean relative biomass at site 4 about 8 times higher than biomass at other sites.
Mean ± standard error of relative biomass (kg.m−2) at sampling sites across month in Venus Bay. November is black, December is light grey, January is white and February is dark grey.
DISCUSSION
The aim of this study was to provide a preliminary assessment of pipi stock and apply a suite of stock indicators in the Venus Bay (Australia) recreational surf clam fishery. Prior to this study, no evaluation of changes in pipi abundance had been undertaken, therefore limited information was available on the abundance or size composition in this fishery. Although a preliminary assessment, the results will provide managers with baseline information and offer a framework for ongoing monitoring of stocks and prioritizing future research in the region. Our findings are discussed individually below.
Availability to harvest
Our results demonstrated that sand habitat was limited to surface sediments (<0.1 m); consistent with results from previous studies (Ansell, Reference Ansell, McLachlan and Erasmus1983; James & Fairweather, Reference James and Fairweather1995; James, Reference James1999). Such shallow depth distribution, while allowing efficient tidal migration, provides an easy target for recreational harvesters, limiting the potential for refuge in deeper sands inaccessible to harvest (McLachlan et al., Reference McLachlan, Dugan, Defeo, Ansell, Hubbard, Jaramillo and Penchaszadeh1996). Tidal movement may provide refuge opportunities for stock in lower-tidal and sub-tidal areas (>60 m from the MHW) which are largely inaccessible to harvest, owing to a heavy surf zone.
Trends in relative abundance
Results demonstrated temporal variation in relative surf clam abundance. Relative abundance declined across the peak harvest months of December and January which may indicate harvest had some effect on longshore abundance. The absence of a consistent decline in relative abundance at site 4 during this period, when compared to sites 1–3, may support this notion given site 4's relatively minimal exposure to harvest. On the basis of our data, there is no evidence that recreational fishing has had an impact on pipi abundance.
Several beach morphological factors, not measured in this study, have been shown to impact on surf clam abundance including: beach slope (Hacking, Reference Hacking1998); beach formation (James, Reference James1999); sediment particle size (de la Huz et al., Reference de la Huz, Lastra and López2002); and wave exposure (Delgado & Defeo, Reference Delgado and Defeo2007). Variation in these factors may confound temporal trends in abundance. Further, longshore abundance of beach macro-infauna has been shown to peak towards the middle of the beach, with lower densities towards the ends of the beach (James & Fairweather, Reference James and Fairweather1996). As the level of harvest effort is likely to decline towards the middle of the beach, as distance to beach access increases (access at north-west beach end), it is difficult to separate the effect of harvest from environmental influences.
Trends in size structure, biomass and ratio of maturity
A difference in relative biomass was observed between sites by months combinations, whereas a difference in sexual maturity was observed only between sites. The sites that were easily accessible to recreational fishers (sites 1, 2 and 3) and subsequently exposed to relatively high harvest had reduced relative biomass compared with site 4. The reduction in relative biomass at sites 1, 2 and 3 was also reflected in the low proportion of mature pipi present at these sites.
The changes in relative biomass and size structure between sites 1–3 (exposed to recreational harvest) and site 4 (minimal recreational harvest), imply the primary effects of recreational harvest were on stock size structure and relative biomass, not abundance.
It is possible that the decline in biomass in areas accessible to harvest may be attributed to fisher preference, with fishers targeting and removing larger pipi (Hartill et al., Reference Hartill, Cryer and Morrison2005). The consistent removal of large individuals (>37 mm) could reduce reproductive potential in the population (Hartill et al., Reference Hartill, Cryer and Morrison2005), which may inhibit future capacity for population recovery. Harvester selectivity has been shown to change with stock availability in surf clam fisheries, with harvesters targeting smaller stock in the absence of larger individuals (Hartill et al., Reference Hartill, Cryer and Morrison2005). Reduced abundance in size-classes >30 mm, at sites 1–3 may reflect changing selectivity as large pipis (>37 mm) decline. This notion is supported by the unimodal size distribution observed at these sites, with a significant 10–15 mm cohort. The reduction in larger individuals suggests there may be potential for recruitment over-fishing, on a local scale.
This result may reflect seasonality of growth and timing of recruitment, despite the small age/size of maturity common to Donax species (King, Reference King1976; Murray-Jones, Reference Murray-Jones1999). The exposure of fishery recruits to at least one season of harvest before achieving maturity may effectively harvest out maturing individuals. Indeed, there is evidence to support this; in the NSW commercial pipi fishery, no mature cohort developed over a period of five years, despite the presence of small pipis (<12 mm) in all years (Murray-Jones & Steffe, Reference Murray-Jones and Steffe2000).
Management implications and future research
The results of this study indicate recreational harvest may affect biomass, the proportion of maturity and size structure of surf clam populations, already managed through established equipment restrictions (no tools: use of hands and feet only), recreational licensing and DCL. However, this study elucidated the difficulties in differentiating the extent to which recreational harvest influences stock abundance from changes induced by environmental variation.
The results from this paper provide baseline information on the abundance, biomass, size, and maturity structure of a surf clam stock that is fished exclusively by the recreational sector, for further use in evaluating the effectiveness of currently reduced DCL.
This paper illustrates that a suite of indictors including relative biomass, size–frequency and proportion of maturity may provide a more holistic assessment of stock in recreationally fished surf clam populations, than relative abundance alone. Further, this study highlights the need for the development of applicable proxy methods to estimate relative biomass in surf clam populations that are not exploited commercially and have no consistent catch and effort data available.
Multiple indicators such as those utilized in this study may be useful for evaluating the effectiveness of current management arrangements. Regular monitoring programmes can provide short-term estimates of available biomass with which it is possible to estimate the relative status of stocks levels across time. Unlike commercial fisheries however, it is difficult to restrict total recreational take due to the potential for a large, unconstrained number of harvesters. While DCL provides a useful means by which to regulate equity of resource access across the recreational sector, it is likely to be insufficient as a tool for preventing stock decline, particularly if the numbers of recreational harvesters within the fishery grow rapidly. Like many recreational surf clam fisheries, the small spatial scale and short summer harvest season in which this fishery operates precludes the effective application of some available management tools. In particular, seasonal closures and rotation of harvest areas may operate on such small scales as to become an ineffective means to reduce the effects of recreational harvest.
Minimum size limits may effectively regulate recreational harvest by protecting part of a population's reproductive capacity when set at the size at which 50% of pipi are sexually mature. This would allow pipi to spawn at least once before being available to harvest and reduce the likelihood of recruit over fishing, increasing reproductive potential. Similarly, maximum size limits may be used to protect large individuals which represent ‘fit’ and adaptable individuals that produce a disproportionate number of gametes (Fenberg & Roy, Reference Fenberg and Roy2008; Johnson & Smee, Reference Johnson and Smee2012) to help secure the reproductive capacity of the population.
To be effective, size limits rely on accurate length at maturity data which is currently lacking in this fishery. Further research should include growth and maturity studies alongside abundance and distribution monitoring. Research that is explicitly designed to assess harvest traits (such as Creel surveys, Angler Diaries) may be useful and provide catch and effort information within recreational fisheries without compulsory data acquisition. Such studies are required to develop long-term recreational harvest strategies.
ACKNOWLEDGEMENTS
The authors acknowledge the funding and support provided by the Department of Primary Industries (Victoria, Australia), which made this project possible. In addition, the authors express their gratitude to Monique Leanne, Dick Brumley, Trudy Shmidt, Adrienne O'Neil and Ty Matthews for their assistance and review. The authors declare that they have no conflicts of interest. The funding body (Department of Primary Industries, Victoria) provided statistical support and interpretation during the development of this manuscript.
