Catch per Unit Effort of Decapod Species, C. pagurus and H. gammarus, from a Voluntary Marine Reserve

Abstract

C. pagurus and H. gammarus are deemed to be declining in abundance in the Berwickshire Marine Reserve from personal communications with local inshore fishers. Fisheries data in the form of catch per unit effort (CPUE) were collected for these two commercially important decapods. Other explanatory variables from fishing activity such as the creel and bait type used, the soak time of the fishing gear, and deployment depth were recorded to provide as much detail as possible to describe the effort applied to catch these decapod species. In this study, CPUE was higher for H. gammarus and C. pagurus outside the Berwickshire Marine Reserve. General additive models (GAMs) were used to describe the effects of the explanatory variables and showed that soak time (days) and depth (m) significantly affected CPUE for C. pagurus, not H. gammarus. Sea temperature (°C) showed a negative correlation with the CPUE of both H. gammarus and C. pagurus; however, a positive correlation was found with the number of C. pagurus caught. The data collected in this study provide a foundation in understanding the current abundance of C. pagurus and H. gammarus in a voluntary marine reserve on the east coast of Scotland, which can be used to inform future changes in fisheries management in Berwickshire.

1. Introduction

Pressure is applied to shellfish fisheries, such as C. pagurus and H. gammarus, to compensate for recovering finfish fisheries for food security [1,2]. However, it is suggested that C. pagurus and H. gammarus are deemed “fish to avoid” by the Marine Conservation Society [3]. With contradicting statements, there is a clear need for as much information as possible to fully understand the sustainability status of commercially important species in the Berwickshire region. Personal observations highlighted that many inshore fishers suggest a decline of both H. gammarus and C. pagurus in the region, which these fishers fully depend on for financial support. C. pagurus stocks are currently misunderstood regarding their current level of exploitation [4]. In the region of Berwickshire, based on statistics from Eyemouth Harbour (55.871697, −2.087245), there are around 62 creel fishing vessels under 10 m in length and 3 vessels over 10 m typically fishing for C. pagurus and H. gammarus [5]. These vessels are below 12 m in length and are regarded as part of the inshore fishing fleet [6,7]. In the 10.3 km² Berwickshire Marine Reserve, fishers contribute to 91–97% of the annual value of landings in the U.K., of which in 2019 this contribution was 19% [5,8]. In 2020, the C. pagurus fishery for Berwickshire (Eyemouth Harbour) was valued at GBP 567,000, and the H. gammarus fishery was valued at GBP 2,565,000 [5]. Although generating valuable annual income to the U.K. economy and to the local economy of Berwickshire, it is believed that U.K.-wide C. pagurus and H. gammarus fisheries are overexploited or close to overexploited [2].

Catch per unit effort (CPUE) is a common tool for monitoring and reporting fish populations through analysis of commercial landings to provide an abundance index [2,9,10]. If the resultant CPUE is declining, this may indicate a fishery that cannot support the level of fishing it experiences; likewise, an increasing CPUE can indicate the fished stock is recovering and potential increases in fishing activity may be applied with appropriate management [11]. In addition to landings data, morphometric data (wet weight (g) and carapace length/width (mm)), can be added to understand the growth and health of the fished stock [12]. With all forms of fishing, there are variables that can influence the fishers’ catch. Pots, or creels, are a form of fishing using various bait types. They can be fitted with panels for undersized or unwanted bycatch to escape and can vary structurally with the inclusion or exclusion of one or more “eyes” (the entrance to a fishing creel). Inter- and intra-specific interactions between individuals within the vicinity of the fishing creel can also influence the individuals that enter and leave a creel [13]. In addition, the season [14], soak time [15,16], and environmental changes can potentially lead to a misinformed calculation of CPUE, abundance, and size distribution for a locale [15,17,18]. To understand such a complex system, these variables should be considered in fisheries assessments, although they currently are not; therefore, interpreting CPUE data fully should be undertaken with caution. Furthermore, collating specific information from individual fishing vessels such as the location of creel hauling, the number of creels per fleet, and soak time for each fleet can provide a full description of the CPUE, particularly for mixed trap fisheries [2]. CPUE is generally standardised using statistical methods such as general linear models (GLMs) and general additive models (GAMs), which allow for descriptions of CPUE with greater detail by considering such variables; however, this still does not provide a full descriptor for abundance in real terms as it cannot correlate well with the statistical output of such tests [19,20].

Marine reserves and other protected marine designations are becoming more apparently used to mitigate the declining biodiversity currently observed worldwide [21,22,23]. Most of these designations are backed by legislation; however, in the case of the Berwickshire Marine Reserve, it is a voluntary designation. The Berwickshire Marine Reserve is one of three other statutory reserves on the Berwickshire coastline; of these, one is a special area of conservation, the Berwickshire and North Northumberland SAC, and the other prevents the use of mobile fishing gear, the Static Gear Reserve (Figure 1). Therefore, static gear fishing is the primary form of fishing within and outside the Berwickshire Marine Reserve. At present, the only management protocol for this fishery is the current minimum landing size for both C. pagurus and H. gammarus, which is 150 mm and 90 mm, respectively [24]. The number of creels associated with each inshore fisher is not limited, and therefore, there is an increasing interest in creel limitations, which have started with pilot studies on the west coast of Scotland [25]. Currently, there is no scope for such a limitation in Berwickshire, and anecdotal data from local inshore fishers suggest some vessels are deploying over 1000 creels, with one extreme case of 10,000 creel deployments. Inshore fishing vessels are not required at present to utilise vessel tracking technology, and therefore, monitoring of these vessels’ fishing activity regionally is not common practice either. However, this is soon to change with the current inshore modernisation programme detailed in the “Bute House Agreement” and “fisheries management strategy 2020 to 2030 delivery plan” which aims to enhance monitoring activities of the inshore fishing fleet using onboard vessel technology [26] as this sector is poorly reported and understood. The data collected from these monitoring systems will allow areas of fishing activity to be monitored, which can aid the understanding of fishing effort in a more regional context. Without current context and understanding of the CPUE in Berwickshire, the means of assessing the effects of new management or legislation are limited.

This study aims to provide information for a data-limited fishery in a voluntary marine reserve to understand the local C. pagurus and H. gammarus abundance in the Berwickshire Marine Reserve. We incorporate local inshore fishers that conducted independent onboard assessments from 2018 to 2019 to collate information on CPUE for these species. Included in the assessments are bait type, creel type, and environmental variables, such as sea temperature (°C), dissolved oxygen (%), and salinity (ppt) which cover many of the factors that could influence CPUE.

2. Materials and Methods

2.1. Study Area

The Berwickshire Marine Reserve (55.912714, −2.109902) covers an open sea area of 10.3 km² (Figure 1) which extends out from the Berwickshire coastline to a 50 m depth contour. Within this designation, the habitat is formed from rocky outcrops interspersed with patches of sand (Figure 1). The BMR is situated within the Berwickshire and North Northumberland Special Area of Conservation (SAC). The SAC covers an area of 652 km², from Alnmouth in the south to Fast Castle Head in the north, which includes the St. Abbs and Eyemouth Static Gear Reserve which covers 26 km² and extends one nautical mile offshore from St. Abbs Head in the north to the Scotland–England Border in the south. Onboard independent surveys took place within the Berwickshire Marine Reserve, Southeast Scotland, U.K. (55.912714, −2.109902), and outside this designation (Figure 2). In the local area, the fishing season for both species is typically all year round; however, the intensity of fishing activity is greater in the summer months for H. gammarus and in the winter and spring months for C. pagurus, which can be indicated by the fluctuating annual market value of these decapods.

2.2. Onboard Surveys

In the period between 2018 and 2019, catch per unit effort (CPUE) surveys were conducted using 5 local inshore fishing vessels of sizes 12 m or less. The local fishing vessels used were berthed in either St. Abbs Harbour (55.899190, −2.129224) or Eyemouth Harbour (55.871697, −2.087245). Surveys were conducted throughout the primary fishing season, with the number of surveys decreasing towards the winter months due to lower fishing activity in response to adverse weather. Surveys generally occurred between the times of 0600 h–1500 h, with each survey varying in duration from 1 to 8 h. Multiple creels are fished as fleets. The number of creels deployed in each surveyed fleet reached a maximum number of 45, with some surveys including single pots which are known as single enders in the local area. Fleets were deployed in a random method, as fishers deployed fleets as they would during normal fishing activity. The creels used by the local inshore fishers varied from single hard-eye to double soft-eye parlours (Figure 3). The variation between the creel types used is dependent on the entrance, known as the eye, and the number of eyes per creel. Eyes can be either made from netting, termed soft eyes, or from a hard plastic 6-inch ring termed a hard eye. Typically, the rest of the creel structure is built from a 38-inch frame of either plastic or metal and laced with rope made from a nylon material (Figure 3). Each fleet of creels was deployed in a depth range between 5.2 m and 32 m which varied depending on the fishing vessel used and the month of the year. The length of time creels were deployed, known as soak time, was recorded in days and ranged from 1 to 7 days. Creels used in this study were typically not fitted with escape panels as this is common for the area. All vessel names and locations of creel retrievals were recorded using their onboard GPS or by handheld GPS which all onboard surveyors carried. The recording of all CPUE data was conducted by two members of staff from St. Abbs Marine Station and two staff rangers from the Berwickshire Marine Reserve (BMR).

2.3. CPUE

Catch per unit effort (CPUE) (kg per day⁻¹) was calculated for each onboard survey. For data collected in 2019, landing per unit effort (LPUE) (kg per day⁻¹) and net retrieval per unit effort (NRPUE) (kg per day⁻¹) were also measured. LPUE is the number of individuals of landing size (>85 mm H. gammarus; >140 mm C. pagurus) or above based on the number of creels hauled, and in contrast, the NRPUE is the number of individuals below the landing size.

$$\mathrm{CPUE}=\frac{\mathrm{Total} \mathrm{number} \mathrm{of} \mathrm{target} \mathrm{species} \mathrm{per} \mathrm{fleet}}{\mathrm{Number} \mathrm{of} \mathrm{creels} \mathrm{per} \mathrm{fleet}}$$

$$\mathrm{LPUE}=\frac{\mathrm{Total} \mathrm{number} \mathrm{of} \mathrm{landing} \mathrm{sized} \mathrm{individuals} \mathrm{per} \mathrm{fleet}}{\mathrm{Number} \mathrm{of} \mathrm{creels} \mathrm{per} \mathrm{fleet}}$$

$$\mathrm{NRPUE}=\frac{\mathrm{Total} \mathrm{number} \mathrm{of} \mathrm{undersized} \mathrm{individuals} \mathrm{per} \mathrm{fleet}}{\mathrm{Number} \mathrm{of} \mathrm{creels} \mathrm{per} \mathrm{fleet}}$$

All bycatch was recorded using the common species name and the number of individuals present in each hauled creel. As the number of Necora puber, which are deemed bycatch by the inshore fishers, caught was so high, they were included in the CPUE assessments as they can be collected and sold by the local inshore fishers.

2.4. Environmental Data

Environmental data such as sea surface temperature (°C), dissolved oxygen (mg/L), salinity (ppt), and lunar phase (%) were recorded on each day an onboard survey was conducted. These data, except lunar phase, were collected using a YSI ProSolo digital water quality meter probe deployed at sea before each fleet haul. Data were rounded to two decimal places for analysis. Lunar phase data were collected on each onboard survey day from a meteorological website (Willyweather.co.uk accessed on 1 June 2018).

2.5. Statistical Analyses

All statistical analysis was conducted using R Studio (Version 1.1.463–© 2009–2018 RStudio, Inc., Boston, MA, USA) [27]. Data were inspected for normality of distribution and equal variance using the Shapiro–Wilks test and Bartlett’s test. Histogram plots were used for visual inspection of normality distribution. Unpaired two-sample t-tests were used to compare the number of individuals caught inside and outside the Berwickshire Marine Reserve designation. To find any differences in the number of individuals caught by year and soak time, Kruskal–Wallis tests were conducted. Pairwise Wilcox tests were used to further the analysis of the Kruskal–Wallis tests to include month as a factor. Kendall rank tau correlations were conducted to consider the effects of environmental variables such as sea surface temperature (°C), salinity (ppt), and dissolved oxygen (mg/L) on the number of target species caught and CPUE. Tau values of above 0 were indicative of positive relationships, and values above 1 were regarded as highly positive. Tau values of below 0 were indicative of negative relationships, with values below −1 regarded as highly negative values.

General additive models (GAMs) were used to model the relationships of CPUE for each target species in relation to the explanatory variables, as shown by the following:

$$\mathrm{CPUE} ~ \mathrm{year} + \mathrm{s}\left(\mathrm{depth}\right)+ \mathrm{s}\left(\mathrm{soak} \mathrm{time}\right)+ \mathrm{boundary} + \mathrm{bait} + \mathrm{creel} \mathrm{type}$$

The complexity of the models was reduced by comparing the REML scores (restricted maximum likelihood) and R² values. Models with a low REML score when compared to the other tests were favoured. The models were fitted with a quasi-Poisson function due to overdispersion in the initial models. Soak time (days) and depth (m) were selected as isotropic smooths (s) based on the models used in [20]. Chi-square tests were then used to assign which variables significantly influenced the CPUE of the target species. Significance was assumed when the p-value was <0.05 for all statistical tests.

3. Results

3.1. Catch Composition

From both sampling years collectively, the total number of H. gammarus individuals reached n = 3013, the total number of C. pagurus individuals reached n = 2117, and the total number of N. puber individuals reached n = 3388. These numbers were collected across a total of 23 surveys and 1897 individual creel pots (

Table 1. Total count of the boats, surveys, and creels used in this project. Data are separated by year and show the total count for each group. Total number of H. gammarus, C. pagurus, and N. puber caught over the surveys is shown.

Year	No. of Boats	No. of Surveys	No. of Creels	No. of H. gammarus	No. of C. pagurus	No. of N. puber
2018	1	8	472	389	567	550
2019	5	15	1425	2624 (834 < MLS)	1550 (427 < MLS)	2838

Note: N. puber are not distinguished by a minimum landing size, and therefore, values are not separated similar to H. gammarus and C. pagurus in the table. MLS is the abbreviation for minimum landing size (minimum landing size 150 mm for C. pagurus and 80 mm for H. gammarus).

). In 2019, the number of undersized individuals (<minimum landing size (<MLS)) was 834 for H. gammarus and 427 for C. pagurus. The soak time to catch the greatest number of H. gammarus (>MLS) was 7 days, although this was deemed insignificant (p = 0.07). However, for those H. gammarus <MLS, significance was found for all soak times above 2 days (p ≤ 0.05). A soak time of 5 days showed significance for the highest number of C. pagurus (<MLS) caught (p < 0.05). Contrastingly, a soak time of 2 and 3 days proved to be significant compared to 7 days (p = 0.03) when related to the number of N. puber caught.

The number of H. gammarus (chi-square = 7.36, p = 0.02) and C. pagurus (chi-square = 32.37, p ≤ 0.05) decreased from 2018 to 2019, whereas the number of N. puber (chi-square = 83.60, p ≤ 0.05) increased. The number of H. gammarus increased between July (2018: 1.43 ± 0.07, 2019: 1.26 ± 0.05) and August (2018: 1.67 ± 0.06, 2019: 1.35 ± 0.09) (p = 0.01). For C. pagurus, in 2018, numbers decreased from July (2.28 ± 0.10) to August (1.66 ± 0.10), whereas in 2019, there was not a large discrepancy (July: 1.10 ± 0.05, August: 1.18 ± 0.13) (p ≤ 0.05). For N. puber, the number of individuals caught in August differed from those in July and September (p ≤ 0.05).

3.2. CPUE

CPUE data from 2018 for both C. pagurus and H. gammarus was highest in prawn parlour creel variations (n = 2.92 and 1.85). The lowest values were recorded in double-eye creels for C. pagurus (n = 0.94) and single-eye creels for H. gammarus (n = 0.57) (

Table 2. CPUE for all target species in 2018–2019 based on pot types sampled.

Year	Creel Type	No. of Pots	No. of H. gammarus ≥MLS/≤MLS	No. of C. pagurus ≥MLS/≤MLS	CPUE (L)	CPUE (C)	CPUE (V)
2018	Double-Eye	311	293	291	0.942	0.942	0.836
	Single-Eye	146	70	235	0.570	1.520	1.650
	Prawn Parlour	14	26	41	1.857	2.928	1.571
2019	Double-Eye Parlour	109	101/148	36/93	2.284	1.183	1.12
	Double Soft-Eye Parlour	110	66/73	35/21	1.263	0.509	2.8
	Hard- and Soft-Eye	8	7/16	0/7	2.875	0.875	0.25
	Hard-Eye Parlour	329	143/291	57/123	1.32	0.55	2.29
	Parlour	624	335/988	301/749	2.120	1.682	1.639
	Prawn Parlour	44	21/21	1/5	0.954	0.136	3.659
	Soft-Eye Parlour	144	86/302	25/73	2.69	0.68	2.36

Note: CPUE values for H. gammarus (L), C. pagurus (C), and N. puber (V) are presented for each pot type sampled. Values are calculated by the number of individuals caught divided by the number of pots of the specific creel type. The values shown for total number of H. gammarus and C. pagurus are separated by landing size (≥MLS)/undersized (≤MLS). Numbers of H. gammarus and C. pagurus in 2018 were grouped and not separated by landing size, as represented in the table.

). The CPUE of N. puber was found to be highest in single-eye creels (n = 2.26) and lowest in double-eye variations (n = 0.83). In 2019, CPUE for H. gammarus was highest in hard- and-soft-eye creels and lowest in prawn parlours (

Table 3. CPUE (2018–19) (mean ± standard error mean) of the target species collected from all creels hauled by the respective fishing vessel.

Year	Vessel	BMR Designation	No. of Pots	CPUE (L)	CPUE (C)	CPUE (V)
2018	Vessel 1	In	468	0.82 ± 0.02	1.12 ± 0.04	1.17 ± 0.04
2019	Vessel 1	In	453	1.00 ± 0.02	0.37 ± 0.01	2.86 ± 0.06
	Vessel 2	In	256	1.28 ± 0.05	1.33 ± 0.05	1.53 ± 0.05
	Vessel 3	Out	47	3.72 ± 0.21	0.17 ± 0.03	1.34 ± 0.13
	Vessel 4	Out	370	2.75 ± 0.04	1.31 ± 0.05	1.47 ± 0.05
	Vessel 5	Out	299	2.16 ± 0.04	1.80 ± 0.07	1.80 ± 0.06

Note: Vessels are also separated by their fishing area inside or outside the Berwickshire Marine Reserve designation.

). For C. pagurus, CPUE was highest in parlours and lowest in prawn parlours (Table 3). For N. puber, CPUE was highest in prawn parlours and lowest in hard- and soft-eye creels, respectively (Table 2). In data from 2019, NRPUE was greater than LPUE for all target species. NRPUE was highest for H. gammarus (n = 1.29), whereas LPUE was lowest for C. pagurus (n = 0.31).

Regarding CPUE of the target species inside and outside the Berwickshire Marine Reserve, for H. gammarus, it was greater outside (2.57 ± 0.03) than inside (1.10 ± 0.02) the designation (t = 41.65, p ≤ 0.05). This was also the case for C. pagurus (outside: 1.44 ± 0.04; inside: 0.72 ± 0.02) (t = 11.30, p ≤ 0.05). Only the CPUE for N. puber was greater inside (2.38 ± 0.04) than outside (1.60 ± 0.04). Regarding the number of these target species, both H. gammarus (≥MLS: 1.51 ± 0.04; ≤MLS: 2.39 ± 0.05) and C. pagurus (≥MLS: 1.62 ± 0.07; ≤MLS: 2.91 ± 0.18) were higher outside the designation. The number of N. puber caught was greater inside (3.59 ± 0.11) the designation than outside (3.30 ± 0.12) (t = −5.13, p ≤ 0.05).

In the analysis of GAM models of CPUE of both H. gammarus and C. pagurus, it was found that the simpler model (CPUE ~ year + s(depth) + s(soak time) + boundary) was deemed suitable based on the chi-square test results and REML scores (p ≤ 0.05) (

Table 4. General additive model outputs describing the relationship of CPUE of all target species with depth, soak time, bait, creel type, and area in relation to the Berwickshire Marine Reserve designation.

Species	Model	REML	R2	Deviance	Intercept: t-Value	Intercept: p-Value
H. gammarus	Model A	−1381.5	0.63	63.9%	−0.30	0.76
	Model B	−1575.7	0.72	71.7%	−6.26	<0.05
	Model C	−879.7	0.52	41.7%	−8.76	<0.05
C. pagurus	Model A	−1141.1	0.63	62.6%	16.27	<0.05
	Model B	−838.2	0.46	46.8%	1.66	0.09
	Model C	−580.83	0.33	31.2%	7.27	<0.05
N. puber	Model A	−547.58	0.18	15.8%	−18.12	<0.05
	Model B	−686.55	0.25	26.1%	−10.69	<0.05
	Model C	−895.07	0.42	42.7%	−0.33	0.73

Note: Results include all data from all 5 fishing vessels across both survey years (2018–19). Significance was set at p < 0.05. The models used are as follows: model A: CPUE ~ year + s(depth) + s(soak time) + boundary + bait; model B: CPUE ~ year + s(depth) + s(soak time) + boundary + bait + creel type; model C: CPUE ~ year + s(depth) + s(soak time) + boundary. Best-fitting models were selected based on the REML score (the lower the score, the better the description of CPUE) and R² value (the greater the value, the better the fit to the model). Significance was set at p < 0.05. Chi-square tests were used to compare and test which models were best for the CPUE data.

). For N. puber, the more complex model was deemed more suitable in describing the changes in CPUE (CPUE ~ year + s(depth) + s(soak time) + boundary + bait + pot type) (df = 15.95, p ≤ 0.05). The outputs from all models can be seen in Table 4. Considering each model, soak time (days) was significant for CPUE for all target species (p ≤ 0.05) (Figure 4, Figure 5 and Figure 6). Depth (m) also showed a significant effect on the model and CPUE for all models except for H. gammarus (F = 1.87, p = 0.19) (Figure 5).

3.3. Environmental Data

Between the sampling years of 2018 and 2019, August was the warmest month (sea surface temperature) (2018: 15.38 ± 0.26 °C; 2019: 15.08 ± 0.21 °C). The coldest month with respect to sea surface temperature was March in 2018 (4.77 ± 0.29 °C) and February in 2019 (6.15 ± 0.09 °C). April for both years had the lowest recorded salinity (2018: 33.57 ± 0.06 ppt; 2019: 33.68 ± 0.21 ppt), with September recording the highest in 2018 (34.44 ± 0.04 ppt) and October recording the highest in 2019 (34.39 ± 0.03 ppt). A full breakdown of the environmental variables recorded can be seen in

Table 5. Environmental variables (sea temperature (°C), salinity (ppt), and dissolved oxygen (mg/L)) recorded by month for each sampling year (2018–2019). Data shown as mean ± sem.

	2018		2019
Month	Sea Surface Temperature (°C)	Salinity (ppt)	Sea Surface Temperature (°C)	Salinity (ppt)	Dissolved Oxygen (mg/L)
June	11.80 ± 0.18	34.13 ± 0.02	12.06 ± 0.24	33.91 ± 0.03	9.15 ± 0.06
July	14.80 ± 0.52	34.16 ± 0.05	14.33 ± 0.25	33.88 ± 0.07	8.38 ± 0.06
August	15.38 ± 0.26	33.90 ± 0.30	15.08 ± 0.21	33.86 ± 0.06	8.01 ± 0.25
September	12.75 ± 0.28	34.44 ± 0.04	12.61 ± 0.08	34.22 ± 0.09	8.84 ± 0.17
October	10.88 ± 0.20	34.41 ± 0.04	10.77 ± 0.15	34.39 ± 0.03	8.89 ± 0.09
November	9.99 ± 0.17	34.24 ± 0.04	9.24 ± 0.16	33.78 ± 0.21	9.33 ± 0.07
December	8.40 ± 0.00	34.31 ± 0.00	7.80 ± 0.15	34.00 ± 0.24	9.43 ± 0.05

.

H. gammarus of sizes above the MLS were not positively or negatively related to environmental factors significantly (

Table 6. Kendall rank correlation outputs (p-value and tau value) for abundance counts and CPUE related to environmental variables for C. pagurus and H. gammarus.

Species		Individuals ≥ MLS	Individuals ≤ MLS	CPUE
H. gammarus	Sea Temperature (°C)	p = 0.37 tau = 0.02	p ≤ 0.05 tau = −0.18	p ≤ 0.05 tau = −0.61
	Salinity (ppt)	p = 0.27 tau = 0.03	p ≤ 0.05 tau = 0.15
	DO (mg/L)	p = 0.05 tau = −0.05	p ≤ 0.05 tau = 0.18
	Lunar Phase (%)	p = 0.70 tau = −0.01	p ≤ 0.05 tau = 0.11	p ≤ 0.05 tau = 0.46
C. pagurus	Sea Temperature (°C)	p ≤ 0.05 tau = 0.14	p = 0.00 tau = −0.10	p ≤ 0.05 tau = −0.21
	Salinity (ppt)	p = 0.84 tau = −0.00	p = 0.15 tau = 0.05
	DO (mg/L)	p = 0.00 tau = −0.09	p ≤ 0.05 tau = 0.17
	Lunar Phase (%)	p = 0.00 tau = −0.10	p ≤ 0.05 tau = 0.21	p ≤ 0.05 tau = 0.34

Note: Significance was set at p < 0.05. Negative tau values correspond to negative relationships and positive tau values correspond to positive relationships. The greater the tau value ≥ 1, the greater the positive relationship; and with tau values ≤ −1, the greater the negative relationship. DO denotes dissolved oxygen (mg/L).

). However, those below the MLS were positively correlated with dissolved oxygen (tau = 0.18, p ≤ 0.05), salinity (tau = 0.15, p ≤ 0.05), and lunar phase (tau = 0.11, p ≤ 0.05) and negatively correlated with sea surface temperature (tau = −0.18, p ≤ 0.05) (Table 6). C. pagurus individuals of sizes below the MLS followed the same trend as the undersized H. gammarus; however, there was no significant result in correlation with salinity (tau = 0.05, p = 0.15). Those C. pagurus individuals above the MLS were positively correlated with sea surface temperature (tau = 0.14, p ≤ 0.05) and negatively correlated with dissolved oxygen (tau = −0.09, p ≤ 0.05) and lunar phase (tau = −0.10, p ≤ 0.05). CPUE of both C. pagurus and H. gammarus was tested with lunar phase and sea surface temperature (Figure 7), and it was found that there was a weak negative relationship for both species regarding sea surface temperature (tau ≤ −0.5, p ≤ 0.05) and a weak positive relationship for lunar phase (tau ≥ 0.5, p ≤ 0.05) (Table 6).

3.4. Bait

The type of bait used by the fishers can be seen in Figure 8, with a greater proportion of mackerel (Scomber scombrus) (42.07%) and herring (Clupea harengus) (34.35%) used across all fishing vessels. Bait was included in the GAM models, in which mackerel was significantly related to the number of H. gammarus (p = 0.01) and C. pagurus (p = 0.01) caught. Herring as a bait source was significant for the number of C. pagurus caught (p ≤ 0.05). Contrastingly, scad (Trachurus trachurus) showed significance for the number of N. puber caught (p ≤ 0.05).

4. Discussion

4.1. Catch Composition

In the local area, it is a common theory with local fishers that the stocks for both H. gammarus and C. pagurus are declining (pers. obvs.). Previous studies have implemented the use of logbooks, used by commercial fishers, to collate the information in situ; however, the robustness of this assessment is determined by the number of fishers participating [15,20]. In this study, a low number of fishing vessels (n = 5) were observed independently using onboard observers without the use of fisher logbooks. Logbooks are used to record the annual landings of fishers’ catch every time they fish at sea. In this study we did not observe every fishing trip; therefore, it may be assumed that some detail and nuance have been lost from the results, which may have been provided if logbooks were used. It has been addressed that due to the limited number of vessels fishing used in this study, there is potential that the catch could be more uniform than would be representative of the natural stock, a positive intra-cluster correlation [15,28]. However, by utilising fishers with different fishing grounds and fishing fleets of randomised locations, it is perceived that this positive intra-cluster correlation was reduced. Many factors could influence the C. pagurus populations regionally, such as habitat variations, mating behaviour, and competition for food and shelter [20]. The influence is based on the sex ratio of caught individuals, which differs seasonally and spatially with C. pagurus migrating in the Autumn after emigrating back to inshore grounds to mate and moult [17,20,29]. The habitats outside and inside the marine reserve designation are predominantly rocky outcrops with interspersed sandy patches with the latter increasing in presence outside the marine reserve. Such habitats benefit the female C. pagurus who reside in sandy patches to avoid strong currents and to incubate eggs post-mating [20,30,31] (Figure 1). The rocky outcrops may have led to lower numbers of H. gammarus and C. pagurus caught due to the availability of natural shelter outweighing exploratory behaviour around creels. H. gammarus tend to shelter in the late spring as they moult and harden their shell before re-emerging in the late summer [32,33], which coincides with the increase in CPUE in August and July in 2019 shown in this study. The number of C. pagurus caught has been shown to be related to sediment type following GAM models in [4] which showed that gravel substrate proved to have higher catch rates than softer sediment. This was not included in our models as the locations of creel pots could not be “ground-truthed” to specific substrate types as the habitat map described in Figure 1 was provided after sampling and analysis. The study presented in [4] also used dredges and trawls, a form of mobile fishing, which is not comparable to static gear fishing which involves the need for food-seeking behaviour using odour cues [34,35].

4.2. CPUE

Although independent observations typically are low in replicates, they are useful in understanding CPUE estimates which are location-specific whilst providing additional biological information such as morphometric measurements and condition indices [20,36]. The CPUE for C. pagurus and H. gammarus was greater outside the Berwickshire Marine Reserve designation. This study did not focus on the potential “spill-over” effect associated with the Berwickshire Marine Reserve, but it may be suggestive based on the CPUE trend outside the designation. It can be assumed that the effort required to fish for C. pagurus and H. gammarus is less outside the marine reserve but much greater inside, suggesting overexploitation within the marine reserve. This is contrary to a similar study for the marine protected area (MPA) around the Isle of Arran, Scotland, in which the inverse was found [23]. Inside the reserve, the CPUE for H. gammarus was 123% higher within the designation in 2012; with C. pagurus, the CPUE was similar both inside and outside the designation [23]. However, in 2013 it was found that the CPUE of C. pagurus had declined within the reserve by 49%, equating to a 253% difference between the outside and inside areas of study [23]. Studies outside the U.K. also find increases in abundance inside the designated areas of protection which have also been related to a “spill-over” effect. Spiny lobsters (Panulirus interruptus) in California were 124% higher in abundance inside the reserve and 223% higher on the border of the reserve, suggesting a “spill-over” effect over a 10-year sampling period (2008–18) [37]. Furthermore, in Norway, a 245% increase in CPUE for H. gammarus was located inside MPAs based on pooled results from 2006 to 2010 [38]. It has been stated that warmer sea temperatures are likely to increase catch due to the metabolic demand of the C. pagurus species [20]. In this study, it was found that sea surface temperature correlated with the number of C. pagurus caught above the minimum landing size, which was also similar to that of [20]. This contrasts observations with C. pagurus fishing occurring typically between April and November, with the highest catches observed in October and November [20]. CPUE of H. gammarus also showed a negative correlation with sea surface temperature in this study, but [23] found that temperature had no significant effect on the number of individuals caught. In the Mediterranean Sea, warmer waters (May–August) were associated with higher CPUE for H. gammarus [39]. Soak time ranged from 1 to 7 days, with the optimal soak times differing depending on which target species was being caught. It has been suggested that soak time does not linearly affect the end catch [20], which is apparent in Figure 4, Figure 5 and Figure 6. Other studies suggest that the number of C. pagurus landings positively correlates with soak time [36] but also declines with increasing soak time [39]. In the Northern Taiwan Strait, soak times of <48 h saw a greater catch of target crustacean species, whereas soak times of >48 h saw a greater haul of bycatch [40]. In this study, soak time had a significant effect on the CPUE for all target species as a smoothing factor in the GAM models, although the effects of soak time greatly differed depending on species (Figure 4, Figure 5 and Figure 6). Depth also showed a significant effect on CPUE, except for H. gammarus, which has also been shown in [20] with the number of C. pagurus caught increasing with depth. However, it has been shown in [4] that the number of C. pagurus caught showed a nonlinear decrease with depth (23–84 m); however, this depth range is far deeper than that used in this study (5–37 m) [4]. Creel deployment is typically associated with rocky habitats compared to sand or muddy substrates [6]. Over a 9-year period, the fishing pressure in the rocky inshore area doubled in the Northumberland district, although the creel density was constant [6]. In areas of high fishing density, it has been shown that the number of C. pagurus and H. gammarus caught declined over a 3-year period by 19% and 35% respectively [41]. Future data surveys should integrate the habitat type on which the fishing equipment is deployed, based on the data provided in Figure 1, also known as fisheries habitat interactions [6].

Creels are not standardised fishing equipment; fishers can adapt and change them in response to fishing needs. Such adaptations can include frame material, net gap size, and number and size of entrances [20]. Predominantly, in this area, the creels are not affixed with escape gaps, which likely increases creel saturation. Such escape gaps can increase the efficiency of the creel by decreasing the undersized catch by 34% whilst increasing commercial catch by 125%, which is further exaggerated with the inclusion of two gaps [42]. Many studies record the efficiency of escape gaps and their ability to reduce the catch of undersized individuals whilst maintaining the commercial catch (those at MLS or above) [13,42,43,44]. The results in this study showed that CPUE and the number of target species caught greatly varied among the various creel types used. The CPUE of N. puber in 2019 was greater in the prawn parlour variations (Table 2), CPUE of C. pagurus was greater in parlour creels, and CPUE of H. gammarus was greater in hard- and soft-eye creels. From fisher observations, the hard-eye creel variations are dominated by H. gammarus, and the soft-eye creels are dominated by C. pagurus, although this cannot be confirmed by this study as the number of H. gammarus dominantly outnumbered C. pagurus in all creel types (Table 2). Other adaptations made to creels can include the entrances, with larger and smaller entrances potentially limiting sized individuals. The entrances for the creels used in this study were not measured, and assumptions on catch limitations by entrance size cannot be made and supported. Typically, the size of entrances on creels ranges from 3.5” (89 mm) to 6” (152 mm) in the Berwickshire area, although the industry standard is recorded as 9.8” (250 mm) [29]. A study looking into the influence of entrance size on crayfish noted that the entrance size had no effect on the range of sizes caught [43]. The entrance sizes ranged from 45 to 85 mm in diameter [43], significantly smaller than those required to meet the landing sizes of both C. pagurus and H. gammarus. In contrast, it has been documented that the entrance size, if increased, is likely to increase the mean size of the animal caught, in this case, H. gammarus [13].

4.3. Bait

The dominant bait type used across all fishing vessels was mackerel (Scomber scombrus) (42.07%) and herring (Clupea harengus) (34.36%). Both H. gammarus and C. pagurus are chemosensory species using bimodal sensilla and olfactory systems to respond to chemical stimuli [9,45]. Both herring and mackerel are recognised as oily fish [46] and, it would be assumed, produce a strong chemical stimulus. The chemosensory stimulus is dependent on time, as the degradation and decomposition of the bait will influence the time period at which the fishing gear is most effective, with bait degradation ranging from 4 to 27 days based on values from [47,48]. It could be assumed that bait plumes surrounding the creel pots can vary regionally due to the regional variation in tides and currents [20]. The bait plume influence and the area with which the plume was associated could not be estimated in this study. Strong currents could lead to faster dilution of the bait plume, which could lead to smaller areas of influence that can only be estimated or assumed [9]. When presented with mackerel bait in a closed flume system, scampi (Metanephrops challengeri) showed no alterations in their behaviour but showed greater walking speeds in more turbulent water flow influenced by such bait [49]. Contrastingly, free-ranging H. gammarus exhibited a decrease in walking speed with decreasing distance to deployed creels; however, the behaviour could not be associated with bait influence specifically [9]. This chemosensory stimulus may influence a deterrence behaviour, as crushed crab added to baited creels led to a decrease in caught crabs by 54% [50].

4.4. Models

The best-fitting models for CPUE of H. gammarus and C. pagurus were those excluding the factors creel type and bait type. When additional explanatory factors were applied to each model, the deviance explained increased. However, more variables were significant, which seemed to exaggerate the interpretation of the models. The model used was based on that of [20], with the inclusion of bait type and creel type. Both smoothing parameters, soak time and depth, influenced the CPUE for all target species with varying influences. The CPUE of C. pagurus seemed to increase up to a period of 5 days before steadily decreasing, whereas that of H. gammarus decreased to 4 days before plateauing (Figure 5 and Figure 6). Soak time varied depending on the fishing vessel used and the season. The soak time of static fishing gear can be influenced by the weather, with early termination of fishing periods being dependent on the adversity of the weather. If the creels are heavily saturated, this could likely lead to a decrease in available bait and space, further inhibiting catch. A higher density of H. gammarus in the creels will limit the catch of C. pagurus [2,51]; in contrast, a higher density of C. pagurus would not limit the available space but would reduce the available bait, limiting the attraction of other individuals [52]. The presence of one individual H. gammarus within a creel can decrease the CPUE of C. pagurus and N. puber by factors of 12 and 9, respectively [6]. Similar to the smoothing trend of soak time on the CPUE of H. gammarus, CPUE decreased with increasing depth until around 20 m before plateauing. The CPUE of C. pagurus increased with increasing depth, whilst that of N. puber increased with increasing depth until around 20 m before decreasing. The influence of depth on CPUE is likely linked to bottom sea temperature and sea current, which could be utilised to improve location-specific data [2,20]. C. pagurus are well known to migrate further offshore; in particular, females migrate further offshore during the gestation period to bury themselves in finer sediment such as sand for brooding [42,53]. It is documented that the depth range distribution of C. pagurus is around 6–40 m [54]. The depth range of H. gammarus ideally ranges from 35 to 40 m [40]. H. gammarus have expressed depth diel patterns from September to January as a form of inactivity coinciding with the decrease in water temperature [55,56]. The same study showed that all but one individual remained in depths shallower than 30 m for longer and shorter periods of activity [46]. The CPUE of H. gammarus was influenced by depths below 30 m (Figure 5) which would suggest a similar trend to [53], but also the limited activity in this species is likely to be associated with site fidelity which is common with this species [57], with activity commonly associated with feeding and territorial behaviour [58] within a few (~3.8 km) kilometres from their shelter [39]. Within a 20 m distance of a deployed creel, H. gammarus can remain in this vicinity for up to 17 h, and their presence can reduce the catch of both C. pagurus and N. puber [2,9].

5. Conclusions

Our study suggests that there is an exploitation of the commercially important species C. pagurus and H. gammarus inside the Berwickshire Marine Reserve due to the lower CPUE values recorded. A “spill-over” effect for both species may be a factor contributing to the varying CPUE values recorded in and outside the marine reserve, which would require further investigation. With no creel limitation in the region, it may be suggested that the decline inside the marine reserve may be accounted for by creel density saturating any viable fishing ground and subsequently contributing to regional fishing pressure. This would require further investigation through monitoring of vessel activity within the reserve designation gathered with autonomous tracking devices attached to the inshore fishing fleet. The depth and soak time of deployed creels were associated with the CPUE of the target species. This information could be used in relation to a creel limitation. By reducing the creel saturation on fishing grounds at certain depths and decreasing or increasing the length of species-dependent soak times, more targeted fishing activity can take place without affecting the overall catch for fishers. This is an ideal within a complex system, as other factors such as substrate type and bottom sea temperature would need to be considered, which would highlight fishery habitat interactions, improving the resolution of the findings in this study. It is believed that the information provided in this study will be used in future advice and consideration for changes in fisheries management for Berwickshire.