Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=tams20
ISSN: 1814-232X (Print) 1814-2338 (Online) Journal homepage: https://www.tandfonline.com/loi/tams20
Spatial description of hake-directed fishing activity off the west coast of South Africa
TP Fairweather , AJ Booth , WHH Sauer & RW Leslie
To cite this article: TP Fairweather , AJ Booth , WHH Sauer & RW Leslie (2006) Spatial description of hake-directed fishing activity off the west coast of South Africa, African Journal of Marine Science, 28:1, 13-24, DOI: 10.2989/18142320609504129
To link to this article: https://doi.org/10.2989/18142320609504129
Published online: 08 Jan 2010.
Submit your article to this journal
Article views: 72
Citing articles: 9 View citing articles
MARINE SCIENCE
EISSN 1814–2338
Spatial description of hake-directed fishing activity off the west coast of South Africa
TP Fairweather1*, AJ Booth2, WHH Sauer2and RW Leslie3
1Department of Ichthyology and Fisheries Science, Rhodes University, PO Box 94, Grahamstown 6140, South Africa;
current address: Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa
2Department of Ichthyology and Fisheries Science, Rhodes University, PO Box 94, Grahamstown 6140, South Africa
3Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa
* Corresponding author, e-mail: [email protected]
Introduction
The South African demersal fishery targets, and is domi- nated by, two Cape hake species. These are shallow-water hake Merluccius capensisCastelnau and deep-water hake M. paradoxus Franca. The majority of the total catch is taken off the West Coast (Payne and Punt 1995), an area falling within the productive Benguela upwelling system (Roel 1987, Roel and Macpherson 1988, Ware 1992). The resource on the West Coast is managed as a ‘single species’ stock, because catch and catch rate have not been species disaggregated and the majority of the catch is M.
paradoxus (Payne and Punt 1995). Fortunately, manage- ment of the resource has been simplified because both species have similar biological parameters (Punt and Leslie 1991) and have historically only been subjected to trawl effort (Payne and Punt 1995).
Since 1994, transformation of the fishing industry to include historically disadvantaged population groups has resulted in an expansion of the demersal fishing industry through the introduction of new participants and the alloca- tion of rights to use longline vessels. The experimental fish- ery was initially formalised in 1998 (Sauer et al.2003) after participating in a scientifically designed fishing experiment (Japp et al. 1995, Japp and Wissema 1999). Longlining differs from trawling with respect to vessel size, areas and
depths fished, and sex and size of fish caught (Japp and Wissema 1999). However, there is no clear understanding of how the combined fishing effort of both sectors will affect the population structure and dynamics of either hake species. The move towards harvesting both species from two different fishing sectors has prompted the development of species-specific stock assessment models for the Cape hakes. Progress in this regard is being addressed by Rademeyer (2003).
It is well known that fish populations are structured in space and time, with distribution and abundance correlated to certain physical parameters such as depth, temperature and substrate type (Swartzman et al. 1994, Le Clus and Roberts 1995, Millar 2000, Sampson 2002). The two Cape hake species have a distribution boundary that appears to be depth related, with M. paradoxus occurring at greater depths than M. capensis, although this could be temperature related because depth and temperature are strongly corre- lated. Roel (1987) reports a boundary between M. para- doxus and M. capensisat 380 ± 45m, whereas Pillar and Barange (1995) note that, whereas juvenile M. capensisare abundant inshore (<150m), a significant overlap occurs between larger M. capensis and smaller M. paradoxus between 150m and 400m. For biological and economic Historically, the two species of Cape hakes Merluccius
capensis and M. paradoxus off South Africa were commercially exploited exclusively by demersal trawl- ing. In 1994, hake-directed demersal longline was intro- duced on an experimental basis, and in 1998 was initi- ated as a commercial fishing sector. The effect of a combined fleet composed of both trawlers and longlin- ers on the Cape hake resource is not fully understood.
Analysis of fishing intensity and catch-rate data revealed that the highest catch rates were found around
the 400m and 500m isobaths for the trawl and longline fisheries respectively. Catch rates from both fishing sectors were also noted to be higher over sediments with a sand component. Differences between areas of the highest fishing intensity and highest catch rates were noted. In addition to other factors, it is suggested that a ‘friction of distance’ effect applies — vessels will trade-off higher catch rates with the increased costs associated with fishing in deeper waters.
Keywords: Cape hake, GIS, longline, Merlucciusspp., trawl, visualisation
Published online 08 Jan 2010
reasons, fishing activity also shows strong spatial patterns;
fishers target those areas that combine high fish abun- dance within the constraint of access to landing and processing facilities (Caddy and Carocci 1999).
Fisheries management is recognising the importance of understanding the use of space by both the resource(s) harvested and the fisheries involved. A better understand- ing of the distribution patterns of both hake species and the areas fished by hake-directed1fishing fleets is therefore required. Given that the data for hake abundance and hake-directed fishing are both spatially referenced, analysis from a spatial perspective is the obvious choice. This paper investigates annual spatial trends in the areas fished, catches, catch rates and fishing intensity by both the trawl and longline sectors, particularly with respect to distances from landing ports, depths and substrate types fished.
Material and Methods Study area
The study area, the South African Western Cape coast, is the continental shelf and upper slope from the Namibian boarder at the Orange River in the north to 20°E. The 1 478km study area was subdivided into three fishing regions,
defined by lines of latitude and, for simplicity, named sequentially after the largest ports found on the West Coast. These were from north to south: Port Nolloth (28.5°S–32°S), Saldanha Bay (32°S–34.5°S) and Cape Town (34.5°S–37°S). Fisheries-dependent data used in the analyses were spatially referenced.
Fisheries data
Commercial trawl data
West Coast catch-and-effort data from commercial trawl v e s s e l s w e r e o b ta i n e d f r o m M a r i n e a n d C o a s ta l Management (MCM) with the permission of the South African Deep-Sea Trawling Association. These data consisted of reports by both wetfish and freezer trawlers, the majority of which operate out of Cape Town or Saldanha Bay. All catch records reported to MCM are referenced to a commercial grid system that is composed of 20’X 20’ blocks extending to the boundary of the Exclusive Economic Zone. All licensed vessels are required to report each fishing event. However, trawlers do not report spatial co-ordinates only the grid block fished. Analysis of the trawl sector was therefore limited to the resolution of 20’X 20’ commercial grid blocks. The position bias was considered small, because trawlers tend to remain within the same area for the day, fishing from dawn until dusk.
The effort (minutes) and catch (tons) over a day by a vessel was summed and the average catch allocated to
1Trawling results in bycatch of other non-target species. In South Africa, some bycatch species may be ‘targeted’ because of their commercial value. All trawls completed with the intent of catching hake must be reported as hake-directed
Table 1: Depth fished by trawlers off the West Coast and fishing regions for the period 1994 –1999
Region Year Mean ± SD (m) Median (m) Range (m) n Effort (1 000h)
Entire West Coast All 393.42 ± 95.680 400 20 – 999 184 909 452.75
1994 385.59 ± 93.900 398 50 – 900 30 718 64.71
1995 386.76 ± 92.580 390 112 – 710 32 745 73.14
1996 390.67 ± 97.640 395 20 – 999 29 557 71.11
1997 402.45 ± 97.120 412 95 – 997 32 865 83.09
1998 398.96 ± 96.310 412 109 – 999 32 100 84.27
1999 396.13 ± 95.210 409 110 – 684 26 924 75.43
Port Nolloth All 410.82 ± 76.320 440 95 – 550 14 215 49.06
1994 384.36 ± 82.820 420 193 – 549 2 397 7.41
1995 392.13 ± 80.770 435 199 – 520 1 380 4.12
1996 426.53 ± 69.870 450 190 – 550 2 595 8.23
1997 413.29 ± 79.790 44 95 – 550 3 042 10.67
1998 405.63 ± 77.870 437 128 – 534 2 247 8.61
1999 431.29 ± 55.350 442 147 – 547 2 554 10.00
Saldanha Bay All 407.70 ± 84.450 411 20 – 999 124 092 298.30
1994 408.43 ± 81.630 411 110 – 682 22 537 45.70
1995 410.87 ± 81.650 413 112 – 710 23 124 52.02
1996 411.11 ± 83.040 415 20 – 900 16 887 41.21
1997 408.91 ± 82.960 413 108 – 997 21 137 52.13
1998 404.95 ± 89.050 414 110 – 999 22 956 58.46
1999 401.25 ± 88.090 403 115 – 682 17 451 48.14
Cape Town All 350.17 ± 114.19 320 50 – 999 46 602 105.39
1994 297.13 ± 90.360 260 50 – 900 5 784 11.60
1995 318.22 ± 89.020 295 115 – 665 8 241 17.01
1996 347.16 ± 110.59 320 100 – 999 10 075 21.67
1997 382.93 ± 127.25 389 135 – 995 8 686 20.28
1998 376.88 ± 119.07 385 109 – 851 6 897 17.20
1999 370.23 ± 115.65 380 110 – 684 6 919 17.30
each position fished during that day. Drag catch per unit effort (cpue) in tons per hour was calculated as
Bottom depth was recorded for all commercial hake- directed trawls. Distribution of fishing intensity and cpue were assessed in depth intervals of 100m. For example, a reported depth of 255m would be classed into the 201–300m stratum. For the purposes of this study, fishing intensity was defined as the number of fishing events.
Observer data
An average of four observer trips per year onboard commercial trawl vessels has been undertaken by MCM staff since 1994. The data collected include species compo- sition of the catch, hake biological information and length frequencies; these data were included in the commercial trawl dataset for analysis.
Commercial longline data
Data from the Hake Longlining Fishing Experiment (Japp and Wissema 1999) included catch, effort, depth and species-specific length frequencies. Longliner vessels reported the latitude and longitude of their start position for each longline set; however, starting depth was only included in the 1999 data. During the experiment, all vessels used the same fishing methods (Japp and Wissema 1999); line setting started at 04:00 and the entire line was soaked from 05:00 to 09:00 when hauling would begin. Although hauling times varied with number of hooks set, the soak times of the various vessels’ gear can be considered comparable and therefore cpue was calculated as tons per 1 000 hooks. Fishing trips lasted between one and three days because vessels needed to ensure high quality of fresh fish for the export market.
Sediment data
Sediment texture off the South African coastline was obtained from the Council for Geoscience (Geological Sur- vey 1986). The primary components of sediment texture were classified as:
Mud — silt and clay-sized sediment particles <0.63mm in diameter;
Sand — mineral particles with a size between 0.6mm and 2.0mm in diameter;
Gravel — unconsolidated sediments composed of rock fragments >2mm;
Rock — compact and consolidated mass of mineral matter, three types are recognised: igneous, sedimen- tary and metamorphic.
The sediment was considered surficial, i.e.non-lithified and unconsolidated. These data are sufficient for the purposes of this investigation because they could be considered repre- sentative of current day sediment deposits (Dyer 1986, F Holzförster, Department of Geology, Rhodes University, pers.
comm.). Areas considered untrawlable by MCM were desig- nated as rock with a surficial sediment covering.
8
CPUE (tons h1) 6 4 2
0 (2 062) (35 625) (54 891) (70 797) (20 163) (923)
200 300 400 500 600 700
STRATA (m)
Figure 2: Box-and-whisker plot of trawl cpue distributed across different depth strata for the period 1994–1999. Number of trawl events in each depth stratum is given in parentheses. The 200-m stratum represents 101–200m, etc.
32°
31°
30°
29°
S
33°
34°
35°
36°
15° 16° 17° 18° 19° 20°E
Orange River
Port Nolloth +
Groen River
Cape Columbine Saldanha Bay +
Cape Town Hout Bay
Gans- baai +
+ Ports Longliners Trawlers +
500m 400m 350m 300m250m 200m
0250 250500
5001 000 1 0001 500
1 500+
Port Nolloth Zone 150m
100m
Saldanha Bay Zone
Cape Town Zone ATLANTIC
OCEAN
Figure 1: Total number of commercial trawls completed in each grid block for the period 1994–1999
Cpue Catch Effort
= × 60
450m
ATLANTIC OCEAN
Orange River Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Gans- baai
500m 400m300m350m250m
Port Nolloth Zone
Saldanha Bay Zone
Cape Town Zone
1994 1995 1996
1997 1998 1999
CPUE (tons h1)
100m 200m150m
02.5
02.5 2.55 5+
Figure 3: Average annual trawl cpue per grid block fished for the period 1994–1999
Table 2: Trawler cpue for all years (1994 –1999) with reference to depth strata. The 100-m stratum represents 1–100m, etc.
Cpue (tons h–1)
Port Nolloth Saldanha Bay Cape Town
Depth strata (m) Mean n Mean n Mean n
100 0.03 ± 0.02 3 0.18 ± 0.26 9 0.67 ± 0.81 2
200 0.51 ± 0.73 137 0.86 ± 1.72 752 0.73 ± 1.00 1 173
300 0.62 ± 0.78 1 706 1.25 ± 1.57 14 270 1.17 ± 1.44 19 649
400 0.67 ± 0.94 1 599 2.00 ± 2.18 42 358 2.19 ± 2.82 10 934
500 1.09 ± 1.49 10 528 1.95 ± 2.25 50 350 2.43 ± 4.00 9 919
600 0.98 ± 1.29 242 1.36 ± 1.64 15 705 1.53 ± 2.04 4 216
700 0.96 ± 0.97 572 1.14 ± 1.25 351
800 0.17 ± 0.25 47 0.36 ± 1.00 94
900 0.11 ± 0.16 27 0.27 ± 1.43 203
999 2.70 ± 3.81 2 0.11 ± 0.38 61
All 0.98 ± 1.38 14 215 1.80 ± 2.10 124 092 1.69 ± 2.63 46 602
Statistical analyses
All data were natural logarithm transformed. The following hypotheses were tested using a one-way ANOVA: cpue distribution does not depend on depth and is independent of substrate type. Possible interactions between year and depth, and year and substrate, were assessed using a two- way ANOVA.
The null hypotheses that the number of fishing events is proportional to depth and to substrate type were tested using a Chi-square goodness-of-fit test.
All data were imported into a TNTmips 6.2 Geographical Information System (GIS; www.microimages.com), where the data were visualised and fishing activity by region and substrate type extracted. Statistical analyses were conducted within S-PLUS 2000 (www.insightful.com).
Results Depth
The null hypotheses that cpue does not depend on depth and the number of fishing events is proportional to depth were rejected (p << 0.001) for both the trawl and longline fisheries.
Trawl fishery
The majority of trawls were conducted between 300m and 500m (Figure 1) with average fishing depths of 350 ± 114m, 408 ± 84m and 411 ± 76m in the Cape Town, Saldanha Bay and Port Nolloth fishing regions respectively (Table 1). Overall, annual fishing intensity (number of trawls) and effort (minutes) were stable throughout the study period.
Mean cpue for the study period (Figure 2) was highest in the same depth range (400–500m), where fishing intensity was highest, albeit with some annual variation (Figure 3).
Fishing intensity was highest in the Cape Town region at 201–300m depth, but mean cpue was highest at 401–500m. The Saldanha Bay region had a marginally higher cpue at 301–400m depth, but with greater fishing intensity at 401–500m (Table 2). An ANOVA indicated an interaction between year and depth (p < 0.01), i.e. the trends in cpue by depth strata were not homogenous over the years investigated.
Longline fishery
Fishing events prior to 1999 were assigned to a depth stra- tum using the GIS. Longline fishing intensity (number of lines set) on the West Coast, illustrated in Figure 4, shows no discernable pattern in depth strata fished. It does, however, illustrate that fishing is concentrated around the two largest ports on the West Coast, namely south of Cape Town and Saldanha Bay, with some fishing excursions into northern waters. Overall, fishing intensity was highest between 201m and 300m, with a marginal shift from the 201–300m stratum to the 301–400m stratum in 1998 that increased in 1999 (Figure 5). The reduction of fishing inten- sity in 1998 is attributable to several logistic and legal issues, which prevented fishing for most of that year. The
Figure 4:Longline cpue for the period 1994–1999. Each point represents one fishing event
FISHING INTENSITY (number of longlines set)
600 400 200
1994 1995
1996 1997
1998 1999
600 400 200
600 400 200
200 300 400 500 200 300 400 500 STRATA (mm)
Figure 5: Distribution of annual longline fishing intensity across each depth stratum for the period 1994–1999
Orange River
Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Hout Bay
Gansbaai
500m 400m 350m300m250m200m
00.1
Port Nolloth Zone 150m 100m
Saldanha Bay Zone
Cape Town Zone
CPUE (tons 1 000 hooks1) 0.10.2
0.20.3 0.30.4 0.40.5 0.5+
450m
ATLANTIC OCEAN
450m
ATLANTIC OCEAN
CPUE (tons 1 000 hooks1)
00.5 0.51 1+
Orange River Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Gans- baai
500m 400m 350m300m250m
Port Nolloth Zone
Saldanha Bay Zone
Cape Town Zone
1994 1995 1996
1997 1998 1999
100m 200m150m
Figure 6: Average annual longline cpue per grid block fished for the period 1994–1999
Table 3: Longline cpue for all years (1994 –1999)
Cpue (tons 1 000 hooks–1)
Year Mean ± SD Median Range n
1994 0.25 ± 0.30 0.13 0 –1.30 473
1995 0.25 ± 0.33 0.05 0 –1.68 207
1996 0.35 ± 0.27 0.30 0 –1.50 962
1997 0.30 ± 0.33 0.24 0 – 7.45 1 375
1998 0.38 ± 0.40 0.32 0 – 4.00 256
1999 0.48 ± 0.34 0.45 0 – 3.55 919
area fished by the longline fleet has expanded annually over the past six years (Figure 6), with average cpue increasing to almost 0.5 tons 1 000 hooks–1 in 1999 (Table 3). There were significant differences between cpue across depth strata for each year (p < 0.03). However a significant inter- action (p < 0.01) was only noted for the Cape Town region.
Substratum
The distribution of surficial sediment on the West Coast is presented in Figure 7. Only around 10% of the West Coast has been found to be untrawlable by MCM (Table 4), the majority of this untrawlable ground is encompassed within areas designated as sand sediment, i.e.a surficial amount of sand is found between bedrock outcrops. The null hypotheses that the number of fishing events is proportional to substrate type and cpue is independent of substrate type were rejected (p < 0.001).
The results of the Chi-square goodness-of-fit test that the number of fishing events (intensity) and substrate type are independent showed highly significant differences (p <
0.001), with 69% of trawling events (χ2 = 2 620 779.96, df = 6) and 72% of longlining events (χ2 = 59 278.05, df = 6) having occurred on sand in the last six years, despite the lower proportion of this sediment type in the study area.
Trawl fishery
Cpue was significantly higher over surficial sediment with a sand component (Figure 8). This trend was evident in both the Port Nolloth and Saldanha Bay regions; however, mean cpue in the Cape Town region was highest on mud (Figure 9). Fishing was concentrated around untrawlable areas in deeper water where cpue was higher (Figure 10) and surfi- cial sediments dominated by a sand component (Figure 7).
Longline fishery
Although trends in cpue by substrate type were not consis- tent over the years investigated, longline cpue was markedly higher on surficial sediment with a sand compo- nent in four of the years (Figure 11). A two-way ANOVA established an interaction between year and substratum types over the full study period (p < 0.01) and within the Cape Town fishing region (p < 0.01).
Whereas 27% (17 973km2) of the area covered by sand is untrawlable (Table 4), owing to rocky outcrops, longline fishing intensity and mean annual longline cpue is highest on sand throughout the study period (Table 5, Figure 12).
Discussion
Many factors may influence where a vessel fishes, includ- ing vessel size, cold-storage or freezer capacity, the type of gear employed, weather conditions and market demands.
Vessel size limits the distance that can be safely traversed from the nearest port to the fishing grounds, together with the duration spent at sea. Larger vessels can withstand rougher seas and with their increased fuel capacity can stay at sea longer than smaller vessels. Similarly, cold- storage or freezer capacity is a constraint because once a
limit is reached the vessels must return to offload. In this regard, smaller vessels are at a greater disadvantage than larger vessels. Gear type acts as a restriction by dictating the type of fishing grounds being sought. For example, within the trawl sector, hard ground and steep slopes must be avoided either to prevent costly net damage or to ensure fishing capabilities. Weather condi- tions will also determine which fishing grounds are targeted. For example, to avoid storms during winter off the Cape west coast, vessels fish further north in the calmer Port Nolloth region. Market requirements are also an important consideration. Prices for fish, particularly for Figure 7: Distribution of surficial sediments off the West Coast (data from Council for Geoscience) and untrawlable (rocky) areas as recorded by Marine and Coastal Management
300m 200m 150m 450m
500m450m
Orange River
Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Gansbaai PortNolloth
Zone
500m 400m 350m300m250m200m
150m100m
Saldanha Bay Zone
Cape Town Zone
SandSandy mud MudMuddy sand Sandy gravel Untrawlable area
Hondeklip Bay
450m
250m 350m 400m 300m
ATLANTIC OCEAN
Table 4:The area of the West Coast covered by different surficial sediment texture classes. Percentages of sediment area class to the total area are shown in parentheses
Sediment class Area covered (km2) Untrawlable area (km2)
Muddy sand 71 904 (31.7%) 3 520
Sand 66 708 (29.4%) 17 973
Mud 54 113 (23.9%) 1 094
Sandy mud 33 685 (14.9%) 1 484
those destined for the export market or domestic demand for varying quantities of different hake products, also vary throughout the year.
Over the six years under study, the trawl sector has fished the same areas consistently, the highest fishing intensity and average cpue being at depths between 300m and 500m. Lower average cpue and lower fishing intensity occur farther north in the Port Nolloth region. Analysis of each fishing region revealed that high fishing intensity did not necessarily coincide with high cpue for either the Cape Town or Saldanha Bay regions. Rough seas, partic- ularly in winter off Cape Point, may explain the higher fishing intensity closer to shore in the Cape Town region.
In contrast, the more sheltered waters of the Saldanha Bay region allow for more fishing in deeper water. In general, higher cpue and a comparatively ‘cleaner’ catch owing to the lower bycatch rate is usually obtained in
deeper water. The ports available to trawl vessels for offloading catch on the West Coast are Saldanha Bay, Cape Town or Hout Bay.
Analysis of the longline sector suggests that fishing intensity is highest between 301m and 500m, concurring with observations by Japp et al. (1995) that the highest catch rates in their pilot study were between 351m and 400m. There has also been a noticeable shift in fishing intensity over time towards fishing in deeper waters and a corresponding increase in cpue. The increase in the number of participants in the longline sector (Japp and Wissema 1999) would also account for the larger area covered by the vessels by the end of 1999 because of the territorial nature of fishers. Of the area fished by longlin- ers, only one-third was on untrawlable grounds. Highest longline cpue was achieved in deep water at around 500m for both the Cape Town and Saldanha Bay regions, over Figure 9: Box-and-whisker plot of trawl cpue distributed across different substrate types for each fishing region for the period 1994–1999.
Number of trawl events is given in parentheses CPUE (tons h1)
1994 1995 1996
1997 1998 1999
SUBSTRATE TYPE Mud
4 6 8
2
0 (14) (7 358) (21 584) (1 762) (6) (8 337) (22 395) (2007) (9) (6 935) (20 652) (1 961)
6 8
4 2
0 (18) (8 227) (21 233) (2 622)
Muddy sand Sand
Sandy mud
(2) (6 553) (18 903) (1 466) (9) (7 591) (23 623) (1 642)
Mud Muddy sand
Sand Sandy mud
Mud Muddy sand
Sand Sandy mud
Figure 8: Box-and-whisker plot of trawl cpue distributed across different substrate types for the period 1994–1999. Number of trawl events is given in parentheses
6 8
4 2
0 (1) (3 972) (10 185) (57) (20) (32 563) (80 591) (10 928) (37) (8 476) (37 614) (475)
Port Nolloth Saldanha Bay Cape Town
Mud Muddy sand
Sand Sandy mud
SUBSTRATE TYPE CPUE (tons h1)
Mud Muddy sand
Sand Sandy mud
Mud Muddy sand
Sand Sandy mud
100m deeper than that of the trawl fishery. These results also confirm reports by Japp et al. (1995) that ‘offshore’
vessels have the highest catch rates. The longline vessels can make use of the smaller ports on the West Coast such as Port Nolloth, in addition to the larger ports avail- able to the trawlers.
The majority of the vessels participating in the longline experiment operated out of Cape Town. It is likely that the significant difference between years in cpue is the result of market forces and the increasing experience of longline fishers, as catch rates tend to be vessel specific, dependent on the skipper’s experience and the individual vessel oper- ating efficiency (Japp et al. 1995).
Trawl vessels that have storage/freezer capacity do not face reduced travel constraints as do the longlining vessels, which must return to port within a couple of days to ensure good quality of the fresh fish destined for export markets.
The position of higher trawl catch rates varied between years, indicating that vessels either fish areas of abun- dance that vary temporally or are more strongly constrained by market forces. Trawl vessels, despite the non-selective nature of their fishing gear, are capable of targeting specific size categories of fish by careful selection of fishing depth
and catch position on the West Coast. Alternatively, DeBlois and Rose (1995) found that shoals of cod Gadus morhua expand and contract in response to foraging activity. Those authors also suggest these aggregation dynamics should be considered when interpreting trawl data. These two scenarios offer some explanation for areas of higher cpue.
Fishing intensity was noted to be highest in shallower water (301–400m) for both sectors. Given that this depth stratum represents a narrow margin on the continental shelf, it can be assumed that the observation is attributable to fishing practices and not to the proportion of the shelf represented by the stratum. This correlation with depth may be the result of several factors. Inclement weather could exclude smaller vessels from deeper waters, particularly off Cape Point, for much of the year. Caddy and Carocci (1999) suggest that fishing vessels face a ‘friction of distance’ dilemma. As discussed by Japp and Wissema (1999), longline vessels are generally restricted to the inshore regions, to maintain catch quality, whereas for trawlers, towing a bottom trawl net in deeper water uses more fuel than towing in shallower water, and the deep- water fishing grounds may be farther from port. In addition, fishing at 201–300m should result in a catch composed of Figure 10: (a) Fishing intensity and (b) average trawl cpue per grid block (1994–1999) considered against the distribution of untrawlable grounds
Orange River
Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Gansbaai Port Nolloth Zone
Saldanha Bay Zone
Cape Town Zone
(a) (b)
FISHING INTENSITY (number of trawls per block)
0250 250500
5001 000 Untrawlable grounds 1 0001 500 1 500+
AVERAGE TRAWL CPUE (tons h1) 0.00.5
0.51.0
1.01.5 Untrawlable grounds 1.52.0
2.03.2 ATLANTIC
OCEAN
large M. capensis and smaller M. paradoxus, which is the ideal size mix for the market. Therefore, although higher cpue may be attained in deeper waters, higher economic yield may be in shallower waters closer to port, despite the lower cpue.
Although the Cape Town region is considered the main hake fishing ground, two-thirds of the fishing that has taken place off the West Coast in the last six years was in the Saldanha Bay region. The continental shelf is much wider in the Saldanha Bay region, with a shallower gradient than that of the Cape Town region. There are better opportunities for ‘easier’ and therefore more economical trawling from Saldanha Bay.
The analyses rejected all four null hypotheses presented.
This suggests that the distribution and abundance of hake, in particular the exploitable proportion of the population, is determined by a combination of depth and substrate type.
Highest cpue was noted in the 301–400m stratum, where surficial sediments had a sand component. Mud is the domi- nant sediment in water deeper than 500m where cpue is lowest. However, in the Cape Town region, sand sediments are restricted to the inshore/coastal waters and the majority of the fishable grounds are mud dominated, a possible explanation for the higher cpue on mud for this region.
Several studies have confirmed that there is a depth- related element in the distribution of marine species. The ontogenetic migration of fish into deeper water has been documented for many fish species, including the Cape hakes (Millar 2000, Sampson 2002). Depth-dependent distribution may be explained by the physiology of the animals. Claireaux et al.(1995a, 1995b) found that Atlantic cod favour colder bottom water layers diurnally and rise into warmer layers at night. Payne (1989) proposed that temper- ature may contribute to the differing relative abundance of M. capensis and M. paradoxus, resulting in aggregations.
The diet of both species varies geographically, but hake can adapt to perturbations in the availability of their prey (Payne et al. 1987).
Although this study shows a strong relationship between substrate type and fishing activity, several mitigating factors should be considered when calculating an abundance esti- mate relative to substrate type. The density and distribution of fish may be affected by turbidity and the concentration of prey items associated with substrate types (Le Clus et al.
1994). This is particularly relevant with respect to soft suspensible sediments such as mud. The physical proper- ties of the sediments themselves, as well as currents and Figure 11: Box-and-whisker plot of longline cpue distributed across different substrate types for the period 1994–1999. Number of longline events is given in parentheses
Table 5:Mean annual longline cpue in untrawlable areas. Number of longlines set each year is given in parentheses and the total number set on each surficial sediment is given in square brackets
Cpue (tons 1 000 hooks–1)
Surficial sediment 1994 1995 1996 1997 1998 1999
Muddy Sand [72] 0.225 (8) 0.056 (5) 0.320 (28) 0.247 (27) 0.062 (2) 0.106 (2) Sand [1 072] 0.282 (126) 0.177 (44) 0.374 (227) 0.268 (275) 0.414 (55) 0.445 (345)
Mud [10] (0) (0) 0.225 (2) 0.270 (6) 0.097 (1) 0.158 (1)
Sandy Mud [10] < 0.001 (1) < 0.001 (1) 0.476 (4) 0.045 (1) 0.240 (1) 0.309 (2)
1994 1995 1996
1997 1998 1999
SUBSTRATE TYPE Mud
Muddy sand Sand
Sandy mud
Mud Muddy sand
Sand Sandy mud
Mud Muddy sand
Sand Sandy mud CPUE (tons 1 000 hooks1)
(7) (83) (346) (37) (11) (45) (134) (17)
1.5 1.0 0.5
0.0 (33) (232) (620) (77)
(56) (298) (910) (111) (3) (33) (208) (12)
1.5 1.0 0.5
0.0 (7) (71) (819) (22)
interaction with biological components, may therefore play an important role (Seiderer and Newell 1999). In defining the importance of habitat variables, Jennings et al. (1996) suggest that substrate should not be used as the sole determinant of biomass.
To achieve a realistic understanding of fishing pressure, the distribution of intensity must be viewed in the context of fishable areas and distances from viable harbours. This is only possible within a spatial analysis framework. When fishing effort is assessed in isolation, for an area as large as the West Coast, several levels of interaction and limiting factors are overlooked. A spatial framework is also impor- tant when identifying where sector clashes are likely to occur and the most economically effective way to harvest the West Coast hake resource.
This study presents a preliminary spatial analysis of hake-directed fishing in South African waters. It has high- lighted the need for further analysis to investigate seasonal variation in commercial catches as well as the relationship between cpue and distance to ‘ideal’ fishing grounds. Such spatial analysis may provide further insight into the South African hake fishery and how best to manage the fishers and the fish.
Acknowledgements— We thank the Council for Geoscience for providing access and permission to use the sediments data. We thank two anony- mous reviewers for their comments on an earlier draft of the manuscript.
Lara Croft is thanked for giving hope when all was thought to be lost.
References
Caddy JF, Carocci F (1999) The spatial allocation of fishing inten- sity by port-based inshore fleets: a GIS application. ICES Journal of Marine Science56: 388–409
Claireaux G, Webber DM, Kerr SR, Boutilier RG (1995a) Physiology and behaviour of free-swimming Atlantic cod (Gadus morhua) facing fluctuating temperature conditions. Journal of Experimental Biology 198: 49–60
Claireaux G, Webber DM, Kerr SR, Boutilier RG (1995b) Physiology and behaviour of free-swimming Atlantic cod (Gadus morhua) facing fluctuating salinity and oxygen conditions.
Journal of Experimental Biology 198: 61–69
DeBlois EM, Rose GA (1995) Effect of foraging activity on the shoal structure of cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 52: 2377–2387
Dyer KR (1986) Coastal and Estuarine Sediment Dynamics.John Wiley, Great Britain, 342pp
Geological Survey (1986) Marine Geoscience Series 3 – Sheet 2:
texture and composition of surficial sediments of the continental Figure 12: (a) Fishing intensity and (b) average longline cpue per grid block (1994–1999) considered against the distribution of untrawlable grounds
ATLANTIC OCEAN
Orange River
Port Nolloth
Groen River
Cape Columbine Saldanha Bay
Cape Town Gans- baai Port Nolloth Zone
Saldanha Bay Zone
Cape Town Zone
(a) (b)
FISHING INTENSITY (number of longline sets)
0200 200600 600800
Untrawlable grounds 8001 200
AVERAGE LONGLINE CPUE (tons 1 000 hooks1)
00.25 0.250.50
0.501.00 Untrawlable grounds 1.002.00
2.002.50
margin of the Republics of South Africa, Transkei and Ciskei.
Government Printer, Pretoria, 1p
Japp DW, Geromont H, Wissema J, Barnes K, Van Zyl J, Leslie R (1995) Report submitted to the Longline Management Committee on the Hake-Directed Longline Pilot Study conducted from 23 May 1994 to 31 May 1995. Unpublished Report WG/10/95/D:H:16, Marine and Coastal Management, South Africa, 49pp
Japp DW, Wissema J (1999) The development of hake-directed longlining in South Africa and a summary of results of the hake- directed longline experiment from 1994 to 1997, 8 March 1999.
Unpublished Report, Marine and Coastal Management, South Africa, 22pp
Jennings S, Boulle DP, Polunin NVC (1996) Habitat correlates of the distribution and biomass of Seychelles’ reef fishes.
Environmental Biology of Fishes 46: 15–25
Le Clus F, Hennig HT-KO, Melo YC, Boyd AJ (1994) Impact of the extent and locality of mud patches on the density and geographic distribution of juvenile Agulhas sole Austroglossus pectoralis (Soleidae). South African Journal of Marine Science 14: 19–36
Le Clus F, Roberts MJ (1995) Topographic and hydrographic effects on catch rates of Austroglossus pectoralis (Soleidae) on the Agulhas Bank. South African Journal of Marine Science16:
321–332
Millar DL (2000) Distribution and abundance of Cape hakes (Merluccius capensis and Merluccius paradoxus) in relation to environmental variation in the southern Benguela. MSc Thesis, University of Cape Town, South Africa, 126pp
Payne AIL (1989) Cape hakes. In: Payne AIL, Crawford RJM (eds) Oceans of Life off Southern Africa.Vlaeberg, Cape Town, pp 136–147
Payne AIL, Punt AE (1995) Biology and fisheries of South African Cape hakes (M. capensisand M. paradoxus). In: Alheit J, Pitcher, TJ (eds) Hake. Biology, Fisheries and Markets.
Chapman & Hall, London, pp 15–47
Payne AIL, Rose B, Leslie RW (1987) Feeding of hake and a first attempt at determining their trophic role in the South African West Coast marine environment. In: Payne AIL, Gulland JA,
Brink KH (eds) The Benguela and Comparable Ecosystems.
South African Journal of Marine Science 5: 471–501
Pillar SC, Barange M (1995) Diel feeding periodicity, daily ration and vertical migration of juvenile Cape hake off the west coast of South Africa. Journal of Fish Biology 47: 753–768
Punt AE, Leslie RW (1991) Estimates of some biological parame- ters for the Cape hakes off the South African west coast. South African Journal of Marine Science10: 271–284
Rademeyer RA (2003) Assessment of and management proce- dures for the hake stocks off southern Africa. MSc Thesis, University of Cape Town, South Africa, 209pp
Roel BA (1987) Demersal communities off the West Coast of South Africa. In: Payne AIL, Gulland JA, Brink KH (eds) The Benguela and Comparable Ecosystems. South African Journal of Marine Science 5: 575–584
Roel BA, Macpherson E (1988) Feeding of Merluccius capensis and M. paradoxus off Namibia. South African Journal of Marine Science 6: 227–243
Sampson MR (2002) Modelling the distribution and abundance of several demersal fish species on the Agulhas Bank, South Africa. MSc Thesis, Rhodes University, South Africa, 109pp Sauer WHH, Hecht T, Britz PJ, Mather D (2003) An Economic and
Sectoral Study of the South African Fishing Industry. Volume 2:
Fishery profiles. Report prepared for Marine and Coastal Management by Rhodes University. Available at www.enviro fishafrica.co.za/projects.php [Accessed 27 October 2004]
Seiderer LT, Newell RC (1999) Analysis of the relationship between sediment composition and benthic community structure in coastal deposits: implications for marine aggregating dredging.
ICES Journal of Marine Science 56: 757–765
Swartzman G, Stuetzle W, Kulman K, Powojowski M (1994) Relating the distribution of pollock schools in the Bering Sea to environmental factors. ICES Journal of Marine Science 51:
481–492
Ware DM (1992) Production characteristics of upwelling systems and the trophodynamic role of hake. In: Payne AIL, Brink KH, Mann KH, Hilborn R (eds) Benguela Trophic Functioning. South African Journal of Marine Science 12: 501–513
Manuscript received November 2004; accepted August 2005
