ArticlePDF Available

Impacts of human activities on marine animal life in the Benguela: A historical overview


Abstract and Figures

This review provides a historical overview of human activities in the Benguela and documents their effects on marine animal life. Considered are the activities of conventional industrial and inshore fisheries but also nonfishery activities, such as mariculture, regulation of river flow, introduction of marine invasive species, marine contruction and mining, pollution and climate change. Human influences may conveniently be divided into four epochs: aboriginal (c. 10,000 before present (BP)–c. 1652), preindustrial (c. 1652–c. 1910), industrial (c. 1910–c. 1975) and postindustrial (c. 1975–present). The aboriginal epoch is characterised by low levels of mainly intertidal exploitation; the preindustrial epoch by intense exploitation of few large, accessible species; the industrial epoch by technological development and a subsequent massive escalation in catches; and the postindustial epoch by improved resource management and stabilisation of catches, but increasing nonfishery impacts on the system. Over 50 million t of biomass has been extracted from the system over the past 200 yr, resulting in significant changes in community structure. Extraction rates peaked at over 1.3 million t yr–1 in the 1960s and have subsequently declined by over 50%. Populations of whales, seals and pelagic and demersal fishes are recovering from historical overexploitation, while those of inshore stocks, particularly abalone, rock lobster and inshore linefishes, remain severely depressed.
Content may be subject to copyright.
Oceanography and Marine Biology: An Annual Review
, 303–392
© R. N. Gibson, R. J. A. Atkinson, and J. D. M. Gordon, Editors
Zoology Department, University of Cape Town, Rondebosch 7700, South Africa
Anchor Environmental Consultants, Zoology Department, University of Cape Town,
Rondebosch 7700, South Africa
History Department, University of Cape Town, Rondebosch 7700, South Africa
University of Pretoria, c/o South African Museum, Box 61,
Cape Town, South Africa
Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa
Department of Archaeology, University of Cape Town, Rondebosch 7700, South Africa
Western Australian Marine Research Laboratories, P.O. Box 20, North Beach, WA6020 Australia
This review provides a historical overview of human activities in the Benguela and
documents their effects on marine animal life. Considered are the activities of conventional indus-
trial and inshore sheries but also nonshery activities, such as mariculture, regulation of river
ow, introduction of marine invasive species, marine contruction and mining, pollution and climate
change. Human inuences may conveniently be divided into four epochs: aboriginal (c. 10,000
before present (BP)–c. 1652), preindustrial (c. 1652–c. 1910), industrial (c. 1910–c. 1975) and
postindustrial (c. 1975–present). The aboriginal epoch is characterised by low levels of mainly
intertidal exploitation; the preindustrial epoch by intense exploitation of few large, accessible
species; the industrial epoch by technological development and a subsequent massive escalation in
catches; and the postindustial epoch by improved resource management and stabilisation of catches,
but increasing nonshery impacts on the system. Over 50 million t of biomass has been extracted
from the system over the past 200 yr, resulting in signicant changes in community structure.
Extraction rates peaked at over 1.3 million t yr
in the 1960s and have subsequently declined by
over 50%. Populations of whales, seals and pelagic and demersal shes are recovering from
historical overexploitation, while those of inshore stocks, particularly abalone, rock lobster and
inshore lineshes, remain severely depressed.
This review is a product of the History of Marine Animal Populations (HMAP) project, the historical
component of the Census of Marine Life Programme (, a decade-long
multinational project funded largely through the Alfred P. Sloan Foundation and Consortium for
Oceanographic Research and Education (CORE). The initial objective of the HMAP project has
been to identify a series of large marine ecosystems (or global sheries), for which good ecological Page 303 Wednesday, June 30, 2004 2:19 PM
304 C. L. Grifths et al.
and historical catch data exist, and to document the effects of human activities on the structure and
functioning of these systems. The Southwest African Shelf, termed the Benguela here, is one of
seven such case studies being investigated.
For the purposes of this review the Benguela region (Figure 1) is dened as extending from Cape
Agulhas in the south to the Namibian–Angolan border (17˚S) in the north, a distance of some 2500
km. These boundaries also mark the approximate biogeographical transition zones between the
cool–temperate biota of the Benguela and those of the warm–temperate South Coast Province of
South Africa to the east and the more subtropical Angolan region to the north (Emanuel et al. 1992,
Branch & Grifths 1988). The northern part of this coastline (N of about 32˚S) is extremely arid and
virtually linear, the only signicant embayments being at Luderitz, Sandwich Harbour and Walvis
Bay, and the only river of note the Gariep (Orange), which forms the border between South Africa
and Namibia. In the South the coastline becomes more irregular, with several prominent capes (Cape
Columbine, Cape Peninsula, Cape Hangklip) and larger bays (St. Helena Bay, Saldanha Bay/Lange-
baan Lagoon, Table Bay, False Bay, Walker Bay). The seaward boundary of the region is considered
to be that of the exclusive economic zones (EEZs) of South Africa and Namibia.
Figure 1
Map of the Benguela region, showing place-names mentioned in the text.
Cunene River
Langebaan Lagoon
Cross Swakopmund
Walvis Bay
Gariep (Orange)
Marcus Island
Saldanha Bay
Cape Columbine
St. Helena Bay
Dassen Island
Robben Island
Table Bay
Cape Peninsula
Cape Point
False Bay
Dyer Island
Cape Agulhas
Walker Bay
Quoin Point
Gans Bay
Muizenberg Betty’s Bay
Cape Town
Berg River
Elands Bay
Olifants River
Lamberts Bay
Algoa Bay
South Africa
Atlantic Ocean
1: 9500000
50 0 50100 Kilometres
Sandwich Harbour Page 304 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 305
Evidence for a human presence on the shores of the Benguela dates from the Early Stone Age
(1–0.5 million yr before present (BP)), but systematic exploitation of marine resources appears
only to have commenced during the last interglacial period (120,000 yr BP) (Parkington 2001a;
see also below). Marine resources soon became integrated into a hunter-gathering economy. Indeed,
the fatty acids contained in the marine food chain are thought to have been important in human
evolutionary development (Crawford et al
1999, Parkington 2001a,b, Broadhurst et al. 2002). Low
population levels and rudimentary technology essentially limited the impacts of hunter-gatherers
to the intertidal. The establishment of pastoralism (1900–1400 yr BP) ultimately altered human
use of marine resources, curtailing human access to the coast to occasional visits determined by
the annual movements of their livestock (Smith 1992).
The rst European seafarers entered the Benguela in the late 15th century, en route to Asia,
but in the mid-17th century the Dutch East India Company (DEIC) established a permanent
settlement at Table Bay. The new colony expanded steadily up the west coast to the Berg (c. 1700),
Oliphants (1750), and Buffels (1798) Rivers. The DEIC also annexed ve bays north of the Orange
River (the current Namibia), including Walvis Bay and Luderitz, in 1793. The DEIC’s discourage-
ment of private enterprise and the low rainfall in the region limited settlement, and hence the impact
of European colonisation on marine resources (Van Duin & Ross 1987) during this period. When
the British supplanted the DEIC in 1806 they extended their jurisdiction to the Orange River in
1847 and subsequently annexed all the Namibian islands (1866) and Walvis Bay (1879), while
allowing Germany to seize the mainland between the Orange and Cunene Rivers as its colony in
1884. The British allocated land for agriculture and mining and leased out use rights to seabirds,
seals and shes to facilitate settlement and trade on the west coast. This was further encouraged
by railway construction and the introduction of steam shipping to the coastal trade in the nal
quarter of the 19th century (Van Sittert 1992).
British rule ended in 1910, with the amalgamation of its colonies and the Boer republics
into the Union of South Africa. South Africa brought the Benguela under a single political
administration for the rst time in 1915, when it conquered German Southwest Africa during
the First World War. Southwest Africa was subsequently administered as a South African colony
until its independence as the Republic of Namibia in 1990, and the return of Walvis Bay, a
South African enclave, to Namibian administration in 1994. During the 20th century the human
population along the Northern Benguela coastline remained small and restricted largely to a
series of factory-cum-holiday towns. The economy in most of this region remains based largely
on natural resources, including alluvial diamonds, rock lobster, pelagic shes, and beachfront
property. The only major population centre – Cape Town and its satellite settlements – lies in
the extreme south. This region has a vibrant and diverse economy quite different from that of
the more arid west coast and incorporates substantial industrial, commercial, agricultural and
tourism components.
A number of previous reviews have synthesised existing information on various marine com-
ponents of the Benguela ecosystem. These include articles on physical features and processes
(Shannon 1985), chemical processes (Chapman & Shannon 1985), plankton (Shannon & Pillar
1986), the major shes and invertebrate resources (Crawford et al. 1987), the coastal zone (Branch
& Grifths 1988) and marine geological aspects (Rogers & Bremner 1991). Some of these reviews
remain valuable but those dealing with exploitation of biological resources have dated rapidly,
because management policies and the status of many marine living resources have undergone radical
transformation in recent years. This review aims to provide an updated historical overview of the
status of exploited marine stocks in the Benguela region and to consider other, nonexploitative
anthropogenic inuences that may affect marine animal life in the region. Thus, for the rst time,
it is possible to see, in a single source, the interactive effects of all types of human impact on the
Benguela. These inuences are discussed under separate headings below, beginning with the earliest
forms of exploitation and the more conventional sheries and proceeding to other, more indirect
environmental inuences. Page 305 Wednesday, June 30, 2004 2:19 PM
306 C. L. Grifths et al.
Precolonial exploitation
South Africa’s 3000 km coastline is dotted by many thousands of archaeological sites (shell middens
and caves) that bear witness to the long-term exploitation of marine resources (shellsh, crustaceans,
shes, seabirds, and marine mammals). The earliest evidence for marine exploitation by people in
southern Africa dates to the Middle Stone Age, around 120,000 yr BP, and is found in fossilised
open shell middens along the west coast and cave sequences on the south and east coasts (Volman
1978, Klein 1999, Henshilwood et al. 2001, Marean & Nilssen 2002). Much of what is known
about prehistoric exploitation of marine animals in southern Africa, however, derives from the far
more numerous Later Stone Age (LSA) sites, dating to the last 12,000 yr. The majority of coastal
sites dating before that time became submerged along the coastal shelf as a result of rising sea
levels from –120 m since the end of the last Glacial Maximum, around 18,000 yr BP (Van Andel
A diverse range of observations is available for the LSA sites, namely, which species were
exploited, their relative abundance in the archaeological record, the technology used to exploit the
species and the seasonality of their exploitation (Avery 1987, Buchanan 1988, Jerardino 1996,
1997, Jerardino & Parkington 1993, Jerardino & Yates 1997, Inskeep 1987, Noli & Avery 1988,
Parkington et al. 1988, Poggenpoel 1996, Schweitzer 1979, Smith et al. 1992). Palaeoenvironmental
conditions prevalent during coastal visits by hunter-gatherer groups are also derived from archae-
ological sources, along with those obtained from conventional sources, such as geological proles
and cores (Jerardino 1995, Parkington et al. 2000, Compton 2001).
Given the current need to control patterns of resource exploitation by our technologically
advanced society, it is interesting to speculate as to the impact prehistoric groups had on marine
resources using their simple technology. Broadly speaking, the precolonial exploitation of marine
animals consisted mainly of the collection of at least 15 species of molluscs and crustaceans
(Buchanan 1988, Jerardino 1997, Jerardino & Navarro 2002) combined with some hunting, but
mostly scavenging. Species scavenged included washed up seabirds of about 10 species, Cape fur
seals (
Arctocephalus pusillus pusillus
) and cetaceans (Avery 1987, Smith et al. 1992, Jerardino &
Parkington 1993). Fishing of at least 10 species was also practiced, with the aid of simple technology
such as gorges and shhooks made out of bone, wooden spears, reed baskets, and nets, as well as
stone traps (Avery 1975, Poggenpoel 1996). Judging from the amount of visible archaeological
debris and available lists of identied species, the harvesting pressure exerted on marine resources
varied in intensity with both locality and time.
A west coast study
Among the small number of coastal projects that have focused on precolonial settlement and
subsistence in South Africa, archaeological investigations in the Lambert’s Bay and Elands Bay
areas (32–32˚ 35'S) have yielded the best data for evaluating the impact of prehistoric inhabitants
on marine resources. Aside from the high density of observations generated for this particular
stretch of coastline, there is a relatively good palaeoenvironmental record for the area, and a
good level of understanding of present marine animal communities (Branch & Grifths 1988;
marine mammal sections, p. 308–317). These factors make this area the best candidate for the
study of precolonial exploitation of marine animal populations in South Africa.
The vast majority of the excavated sites in this study area consist of deposits dominated by
marine shell remains and varying quantities of marine and terrestrial vertebrates. The rate at which
these deposits were accumulated has clearly changed since the present shoreline was established
about 8000 yr BP (Figure 2). The rst signs of relatively fast accumulation of shell midden deposits
and intensive shellsh collection date to around 3500 yr BP (Jerardino 1996, Jerardino & Yates
1996). Subsequently, between 3000 and 2000 yr BP, shellsh exploitation was greater than at any
other time during the Holocene period. During this millennium, enormous shell middens Page 306 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 307
(megamiddens) containing tons of black mussel shells and relatively few bone and cultural remains
accumulated immediately behind rocky platforms (Jerardino & Yates 1997) (Figure 2). The overall
dietary mix of people (hunter-gatherers), as reconstructed from isotopic measurements on skeletons
buried along stretches of the west coast and from archaeological food waste, was more marine
between 3000 and 2000 yr BP than either before or after (Lee-Thorp et al. 1989, Jerardino 1996).
The scale of shellsh exploitation was dramatically reduced after 2000 yr BP, a period coincident
with the arrival of pastoralism to the west coast of South Africa. During the last 2000 yr, the
precolonial diet was derived predominantly from terrestrial resources.
Studies focusing specically on prehistoric shellsh exploitation have shown that human impact
on rocky shore molluscs seems likely to have uctuated in response to a succession of different
settlement patterns (from more mobile to more sedentary), demography and palaeoenvironmental
conditions at the time of resource exploitation (Jerardino 1997, Klein 1999). Although comparative
contemporary data for exploitation levels at these same sites are not available, it is clear that these
prehistoric levels of exploitation were very low and almost always sustainable. This is in marked
contrast to the extremely high and unsustainable levels of subsistence exploitation currently occur-
ring on the east coast of South Africa (Grifths & Branch 1997).
The marine bird and mammal records are still insufciently studied to derive meaningful
conclusions as to the impact of prehistoric people on these species. The same applies to the
archaeological record of Cape rock lobster (
Jasus lalandii
). The study of this species is particularly
important, as it is well known to inuence directly and indirectly the abundance and population
structure of its prey and other interacting species (Castilla et al. 1994; see also rock lobster section,
p. 340–345). Groundwork for the study of the exploitation of Cape rock lobster in the study area
was recently laid out (Jerardino et al. 2001, Jerardino & Navarro 2002) and preliminary observations
point to an intricate combination of variables such as sea level change, possibly resulting in shrinking
Figure 2
Summary of temporal changes in the volumes of shell middens in the vicinity of Elands Bay and
Lambert’s Bay, with notes on changes in prehistoric human use of marine resources. (Data from Jerardino &
Yates 1996.)
15 7 5 4 3
Thousands of years B. P.
based at
by drying
Low-level harvest
Pastoral life begins.
Density reduced.
End to coast-
based dwelling.
Hiatus in
Volume of shell middens (m × 1000)
3 Page 307 Wednesday, June 30, 2004 2:19 PM
308 C. L. Grifths et al.
availability of suitable hideouts for lobsters, and greater exploitation pressures as a result of
increasing hunter-gatherer populations together causing uctuations in the mean sizes of rock
Clearly, much remains to be done to improve our understanding of ecosystem change and the
role of people in changing marine ecosystems in the precolonial past. In particular, work needs to
be done to build the necessary databases with observations already obtained; to generate more
detailed observations of sea level change, shoreline conguration, marine productivity and sea
surface temperatures; and to carry out additional eldwork at sites presenting longer and well-
resolved sequences. These topics are the subject of ongoing research effort.
This account is conned to those cetacean species that mainly occur over the continental shelf in
the Benguela region, which in the case of large whales restricts the coverage to southern right
whales (
Eubalaena australis
), humpback whales (
Megaptera novaeangliae
), and the inshore stock
of Bryde’s whales (
Balaenoptera edeni
). Despite this geographical restriction, it must be appreciated
that at least the rst two species are highly migratory, and so subject to human impact quite removed
from the Benguela region. Whaling in the Southern Ocean is perhaps the prime example, but
because of continuing uncertainty about linkages between breeding and feeding areas, it is difcult
to assign pelagic catches to a stock inhabiting a particular coastal region. As a consequence, this
review mainly concerns impacts (including catches) that occurred directly within the Benguela.
Prior to the arrival of European settlers, recorded exploitation of cetaceans in the region is
conned to reports of the utilisation of stranded whales and dolphins for food and other materials.
A few coastal dolphins (probably mostly bottlenose dolphins,
Tursiops truncatus
) were killed by
native peoples wading out from the shore (Budack 1977) but it is unlikely that any of these activities
adversely impacted the populations.
European colonisation at the Cape in 1652 resulted in an immediate interest in the commercial
exploitation of large whales that abounded in the neighbouring bays. The earliest attempts at their
capture, however, were desultory and largely ineffective. This changed dramatically with the advent
of visiting pelagic whaleships, largely from the U.S., but also from France and Britain, in the late
18th century. Although there was a brief episode of whaling by the Dutch West Indian Company
in the Walvis Bay region early in the 18th century, the number of voyages involved was small and
unlikely to have signicantly affected abundance. The main onslaught began around 1780, when
Figure 3
Catches of southern right whales off South Africa by decade, 1790–1940. (Redrawn after Best et
al. 1997.)
Number Page 308 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 309
whaleships from New England began to overwinter at the Cape. Their principal quarry was the
southern right whale, which yielded large quantities of oil and whalebone. Despite relatively
primitive equipment (open boats and hand harpoons), the size of the eet (up to 30 whaleships in
one bay at a time) and the predictability of right whale behaviour led to a rapid decline in whale
abundance. By the 1840s the pursuit was largely abandoned by visiting whaleships. Estimates of
the landed catch in the South Atlantic by U.S. whaleships from 1805–1909 range from
28,500–32,200 individuals, with the bulk of the catch (24,500–27,000) being taken prior to 1840
(Best 1987). Richards & du Pasquier (1989) have made an independent estimate of the number of
right whales taken by all eets (not just the U.S.) on the coast of southern Africa (including
Mozambique) between 1785 and 1812 as 12,000. From their Table 1 it can be estimated that about
34% of these were taken by vessels travelling to Delagoa (Maputo) Bay or the east coast of Africa,
and the rest by vessels visiting Walvis Bay or the Cape of Good Hope.
Meanwhile, shore-based whaling for right whales nally began on an organised basis at the
Cape in 1792. Despite whaling stations springing up at a number of locations in the Western and
Eastern Cape, their catches never reached the levels of those of the visiting whalers. Catches peaked
at between 200 and 400 whales decade
in the early 1880s (Figure 3), whereas the total catch
landed from 1792–1912 is estimated at only 1580 whales (Best & Ross 1986). Unlike their foreign
counterparts, however, the colonial whalers were able to continue catching because their costs were
much lower and the activity could be pursued in association with other shing enterprises, such as
beach-seining. In this way, the catch of only one or two whales a season could still be highly
protable, with almost all the products being exported.
Other species, notably humpback, bottlenose, sperm, blue, nback (probably Bryde’s), pygmy
right, and killer whales, were taken in this shore-based open-boat shery (Best & Ross 1986), but
the numbers recorded were too low to have been of population signicance. As right whales declined
in abundance, the pelagic whalers of other nations began to capture other species on the southern
African coast, notably humpback whales. Although there are no published estimates of the number
of humpback whales taken by the shery in this region, U.S. catches worldwide between 1815 and
1905 have been estimated at 14,000–18,000 animals, with the peak catch (11,000–15,000) being
taken between 1855 and 1889 (Best 1987). Plots of catch positions given by Townsend (1935)
indicate a substantial concentration between Gabon and central Angola from June to September.
Many of the humpback whales taken in this shery may therefore have been from the population
that is believed to migrate through the Benguela region to its breeding grounds off equatorial West
Africa in winter. (A rough estimate from Townsend’s chart is that between a third and half of all
catches were from West Africa, suggesting total kills of between 4500 and 9000 animals.) This
shery was also characterised by a much higher struck-and-lost rate than for southern right whales,
and as several of these animals would have been dead (sunk) or died later, the landed catch is
probably an underestimate of the total removals from the population.
No assessment of the effect of these catches on the population has been undertaken, but within
20 yr of the end of the peak catch, humpback whales were again the target of a shery but this
time one potentially far more destructive. In 1909 modern whaling began on the west coast of
southern Africa. Instead of open boats powered by sail or oars, with hand harpoons as the principal
weapon, whales were pursued by steel-hulled steam-driven catchers of 100 t or more, with the
harpoon red from a mounted cannon and carrying an explosive grenade at its tip. Methods of
processing the whale were initially not very different from those of the earlier shery, with utilisation
being largely conned to the blubber and tongue and the rest of the carcass being jettisoned. The
escalation in catching effort was enormous, so that by 1913 at least 16 land stations or moored
factory ships were whaling between Cap Lopez in Gabon and Hangklip in South Africa. Catches
soared accordingly, from about 600 in 1909 to nearly 6000 whales in 1913 (Best 1994). Such a
whaling intensity was clearly nonsustainable, and by 1915 catches had crashed to less than 200
(Figure 4). Thereafter the industry largely switched to other species (blue, n, and sei whales
especially), and by 1963 (when humpback whales were nally given protection by the International Page 309 Wednesday, June 30, 2004 2:19 PM
310 C. L. Grifths et al.
Whaling Commission (IWC)) only a handful were being taken annually by the sole surviving land
station at Donkergat in Saldanha Bay (Figure 1). Curiously, episodic whaling off Gabon (1934–37,
1949–52) was reasonably successful, suggesting that the humpback whales passing Saldanha Bay
may represent a different component of the population, possibly one feeding to the east of the
continent, off Queen Maud Land.
Modern whaling also affected right whales. Despite their rarity, right whales were valued
by the industry as highly as sperm (and considerably more than blue, humpback, or sei) whales,
and this must have encouraged their continued exploitation. It is an indication of just how scarce
right whales must have been at the beginning of the 20th century, that only 100 were taken in
modern whaling on the South African coast between 1908 and 1937 (Best & Ross 1986, Figure
3). Some projections have indicated that at its lowest point (about 1937), the South African
right whale population might have contained as few as 30–68 mature females (Tormosov et al.
1998). Since 1935 the species has been internationally protected but this has not prevented
some illegal catching, particularly by pelagic eets from the Soviet Union, which took at least
3368 southern right whales between 1951–52 and 1970–71 (Tormosov et al. 1998). Such
poaching ceased with the introduction of the International Observer Scheme and since 1971
the South African population of right whales has been increasing steadily at 7% a year (Best
et al
2001). In 1997 the population stood at an estimated 659 adult females, equivalent to a
total population of some 3100 animals (IWC 2001). This compares to an estimated original
population size for southern Africa (both east and west coasts) of 20,000 right whales (Richards
& du Pasquier 1989). The latter estimate, however, is difcult to interpret. It includes whales
from three widely separated grounds (Walvis Bay, Cape of Good Hope and Delagoa Bay),
whose relationship to each other is still unknown, and for which only the Cape of Good Hope
can be considered as equivalent to the current South African population. Their estimate also
ignores the effect of recruitment during exploitation (resulting in an overestimation of original
Figure 4
Annual catch of humpback whales off the west coast of South Africa.
Catch (number)
SW Cape
Total catch including
Angola, Gabon Page 310 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 311
population size) and only includes catches from coastal waters. Substantial catches of right
whales also occurred between Cape Town and Tristan da Cunha in the mid-Atlantic (Townsend
1935). It is now known that these catches are highly likely to have included numbers of right
whales that also visit the coast of South Africa (Best et al. 1993; unpublished satellite tagging
data). Model projections have shown that overall, the southern right whale population is about
10–14% of its original abundance (Figure 5).
The current status of humpback whales on the west coast of southern Africa is unknown.
Incidental sightings (and a preliminary shore-based survey at Cape Columbine in 1993; Best et al.
1995) would suggest that some increase must have occurred since protection, given the size of the
catch in the last few years of exploitation and the number of incidental sightings currently being
made. However, at present there are no estimates of population size or trend.
The third species of large whale occurring over the continental shelf, the Bryde’s whale, was
only “discovered” when modern whaling started on the west coast of South Africa. The rst
published description of its external appearance was based on animals examined at the Donkergat
whaling station in Saldanha Bay. Hence one can only speculate that the species was among those
rarely taken by open-boat whalers and declared as “nbacks.” Unfortunately, publication of the
external description was not enough to ensure that the species was always correctly identied in
catches thereafter. Confusion with sei or n whales persisted until well into the 1960s (Best 1994),
so it is difcult to reconstruct a reliable catch series
for the species. To add further complication,
two separate populations of Bryde’s whales have been described from the west coast of South
Africa, one inshore over the continental shelf (largely nonmigratory) and one offshore, which
appears to migrate between equatorial regions in winter and waters off southern Namibia in summer
(Best 2001). Both populations feature in the catches, so although there are morphological differences
between the two, unless the whales are examined by trained personnel, it is impossible to separate
them. In January–February 1983 a shipboard survey was undertaken of the continental shelf of
Figure 5
Model projections of the total population size of southern right whales, 1770–1997, using high,
low, and base estimates of historic catch scenarios and a 1997 population size of 7571 whales. (From IWC
Number of Whales
High Base Low Page 311 Wednesday, June 30, 2004 2:19 PM
312 C. L. Grifths et al.
South Africa between East London and St. Helena Bay, corresponding to the known range of the
inshore stock of Bryde’s whales. This resulted in a line transect estimate of 582 ± 184 whales. This
is likely to have been an underestimate, as the survey was carried out in closing mode (Best et al.
1984). Although the status of this stock is unknown, because of its relatively restricted range it is
unlikely that it was ever very large.
Cape fur seals
The Cape fur seal
Arctocephalus pusillus pusillus
is the only indigenous pinniped inhabiting the
shores of southern Africa. It breeds at 25 colonies, 15 of which are in Namibia and 10 in South
Africa (Figure 6). Of these, seven are on the mainland and 18 on islands. There are an additional
nine sites where seals haul out, but little or no breeding has been recorded. Cape fur seals
preferentially choose nearshore rocky islands on which to breed, which are cooler than the mainland
and afford them protection from land predators. However, overcrowding on the small, coastal islands
has caused them to overow onto the nearby mainland and establish new colonies there.
The colonies are distributed around 3000 km of coastline from Algoa Bay in southeast South
Africa to Cape Frio in northern Namibia (Figure 6). Although there is no evidence that they breed
there, seals have been recorded in Angolan waters up to about 650 km north of the Cunene River.
About 90% of the population is found on the west coast, taking advantage of the rich sheries of
the Benguela ecosystem, whereas only about 10% occurs on the south coast, where food resources
are less abundant (Rand 1959, Shaughnessy 1979, 1982, David 1987, 1989).
Figure 6
Map showing breeding and nonbreeding colonies of Cape fur seals around the coast of South Africa
and Namibia.
Möwe Bay
Mercury Is.
Breeding colonies
Nonbreeding colonies
Seal Is.
Partridge Pt
False Bay
Black Rocks
Subarea 2
Bird Is. Lambert’s Bay
Strandfontein Pt
Port Nolloth
Conception Bay
Sandwich H.
Pelican Pt. WALVIS BAY
Cape Cross
Coast Park
Ichaboe Is. Klein Ichaboe Is.
Wolf Bay
Atlas Bay
Van Reenen
Bay Page 312 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 313
Exploitation of seals is one of the oldest commercial “sheries” in southern Africa. Sealers
aboard the old sailing vessels plundered as many colonies as they could nd, without any form of
control, though there are few early records of numbers taken (Rand 1972, Best 1973). The rst
known sealers were Dutch and they killed about 45,000 seals near the Cape of Good Hope in 1610.
As a consequence, early Dutch sealing destroyed most of the colonies close to Cape Town. Before
the arrival of Dutch settlers at the Cape in 1652, French sealers were sometimes active on the
islands in and around Saldanha Bay. After the colonisation of the Cape, ships of the English East
India Company based in Table Bay, while waiting to resupply the homeward eet, may have raided
local seal islands. There was little further sealing until the late 18th and early 19th centuries, when
British and American sealers were active on the west coast of southern Africa (Shaughnessy 1984,
David 1989).
In those days there were no legal controls and sealing was completely indiscriminate. Rookeries
were invaded even during the breeding season and all age classes were taken, including pregnant
cows and small black pups only a few weeks old. As a result, the seal population was reduced to
low levels by the end of the 19th century, by which time at least 23 colonies had become extinct
in South Africa and Namibia (Rand 1972, Shaughnessy 1984). Subsequently, several of these islands
were colonised by seabirds and permanent accommodation was built on some of them for the
protection of the guano harvest. This effectively prevented recolonisation by seals, which are
sensitive to human presence.
Legislation and harvesting
The desire by the government to improve the control over seal harvesting and to end private sealing
prompted it to promulgate the rst legal protection under the Cape Fish Protection Act of 1893,
which stipulated that no seals might be harvested without a government permit. In 1909 the sealing
season was limited to prevent disturbance during the breeding season and was further amended
and curtailed in 1936. The union government controlled sealing in Namibia by means of the Sealing
and Fisheries Proclamation of 1922 and the Sealing and Fisheries Ordinance of 1949. Both acts
prohibited sealing without a licence. Sealing is currently managed under the Sea Birds and Seals
Protection Act of 1973, which prohibits landing on any island and the capture or killing of any
seal or seabird without a permit. Under this Act the minister is empowered to prescribe the age,
size and sex of seals killed, as well as the season and localities where sealing may take place
(Shaughnessy 1984).
Despite the low seal population at the beginning of the century, harvesting continued at certain
colonies almost every year from 1900 (Wickens et al.
1991). The harvest of pups was small initially
but grew progressively as seal numbers increased. From 1900–10 the total annual harvest was
2600–9300 pups. During the economic depression of the 1930s it was between zero and 16,800
pups and by the 1950s it had increased to 27,200–45,000 pups. By the 1970s the harvest had grown
to 62,400–81,200 pups and the industry reached its zenith in the 10 yr preceding 1983, when the
average harvest was about 75,000 pups (Figure 7).
Until 1965 the bulk of the harvesting was carried out by government sealers of the Guano
Islands Division of the Department of Commerce and Industries. From that date the government
began handing over concessions for individual colonies to private enterprise by inviting public
tenders for the sole right to seal at each colony. By 1979 all concessions were in private hands and
government sealing had ceased (Shaughnessy 1984).
No total allowable catch (TAC) was in operation until 1974, because there was inadequate
knowledge of the size of each colony. Sealers, therefore, took as many pups as they could, given
the weather conditions and the time available. However, this gap in knowledge was lled with the
commencement of regular seal research in 1971 and all concessions awarded after 1974 included
a TAC. Page 313 Wednesday, June 30, 2004 2:19 PM
314 C. L. Grifths et al.
In addition to the harvest of pups, some adult bulls were taken during the summer breeding
season in November–December, when the bulls congregate on the colonies. Prior to 1983 the bull
harvest usually totalled 1000–3000 animals but over 5000 were killed during some years in the
1940s and 1960s (Figure 7). However, the bull harvest increased substantially after 1983 and a
record number of over 20,000 bulls was harvested in 1984 (see below).
Figure 7
Harvests of pups, bulls, and cows from 1900–2001 at all colonies of the Cape fur seal combined
(A) and at South African (B) and Namibian (C) colonies separately.
Pups Bulls Cows
Number (103)Number (103)
Pups Bulls Cows
Pups Bulls Cows
Number (103)
C Page 314 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 315
The total harvest of pups and bulls from 1900–2001 was over 3 million (approximate total
mass of 100,000 t), made up of about 2.8 million pups (approximate mass of 63,560 t at a mean
mass of 22.7 kg pup
; David 1987) and 244,000 bulls (approximate mass of 36,600 t at a mean
mass of 150 kg bull
). The largest number of colonies harvested in a single year was 13 in
1975. The sealing industry thrived until 1983, when the market collapsed as a result of political
developments in North America and Europe. In that year militant conservation organisations
put pressure on the Canadian government to stop the Canadian harvest of harp seals. The lobby
spread to Europe and resulted in the European Parliament requesting member nations to place
a voluntary embargo on the import of all seal products. This move effectively caused the collapse
of the South African industry because the bulk of the skins were sold in Europe. Sealing in
South Africa ceased altogether in 1990, following a ban imposed by the minister of environ-
mental affairs, but continues in Namibia at a reduced level at the Cape Cross and Luderitz
colonies (David 1989).
The most valuable product traditionally was always the skins of the pups 7–10 months
old, which formed the bulk of the harvest. However, there are some other by-products produced
by the industry, the most important of which is seal oil obtained by rendering down blubber
scraped from the skin and carcass. This is done in large drums or pots on site. It is also possible
to render the stripped carcasses into meat meal and bone meal, provided there is a factory on
site, but the value of these products is not high and barely covers production costs. Seal meat
has also been processed into pet food but marketing the product was not successful (David
Another product of considerable value is the genitalia of the bulls, which are dried and sold in
the Far East as a supposed aphrodisiac. After 1983 the concessionaires attempted to compensate
economically for the loss of the skin market by harvesting more bulls. This form of utilisation is
very wasteful because at some colonies no other part of the bull was used and not even blubber
oil is produced.
The method of killing pups in the harvest is controversial and has been the subject of
sporadic conict with conservationists and animal rights groups for many years. It is basically
the stun-and-stick method used in abattoirs, which is primitive and unaesthetic. The thin skull
of the pup is shattered by a blow from a heavy club, which has much the same effect as a bullet.
The chest is then immediately opened with a large knife and all the major blood vessels round
the heart severed, so that the carcass is rapidly exsanguinated. Although the method is not pretty,
it is effective and efcient and no better method has been found. Bull seals are always shot in
the side of the head but to shoot thousands of crowded, jostling pups would not be practical or
safe (David 1989).
Disturbance programmes
Another impact of human activities on the seal population has been through deliberate disturbance,
to cause seals to vacate a particular location. For example, when the permanent human occupation
of Mercury Island ceased in the 1980s, seals began to recolonise the island and quickly built up
their population to a level where it became established as a new breeding colony. The seals spread
over a large part of the island and had a very negative impact on breeding gannets (
Morus capensis
and African penguins (
Spheniscus demersus
), when they began to overrun the bird colonies. This
was considered to be unacceptable because both seabird species are Red Data Book species. A
decision was therefore taken to chase the seals and encourage them to move to the adjacent
mainland. The island was then reinhabited and the seals were systematically disturbed during three
successive breeding seasons. The desired result was eventually obtained when the seals deserted
the island completely and moved to the nearby mainland (Crawford et al. 1989). Page 315 Wednesday, June 30, 2004 2:19 PM
316 C. L. Grifths et al.
The population size of seals
No accurate gures of population size were available until census techniques were developed in
1971. The total population in 1997 was estimated to number some 1.5–2 million animals, of which
about 60% were in Namibia (Figure 8). It grew at an average rate of about 2.8% yr
from 1971–92.
It seems clear that the great increase in numbers during the 20th century was the normal response
of a species recovering from overexploitation. However, the growth in population was not uniform.
For example, there was the sharp decline in 1986 due to the decision by the minister to permit
extensive bull sealing to extend into December, which caused major disturbance of the colonies at
the height of the breeding season. This resulted in signicant mortality of pups and disruption of
mating. Another major decline in 1995 was caused by the death by starvation of tens of thousands
of pups and adults in the Luderitz area, which in turn was due to a warm water oceanographic
event that caused the death or emigration of all the local sh populations (Roux 1998). Thus the
seal population was unable to nd food and the lactating females abandoned their pups, which died
of starvation. Many thousands of adults also died.
Only very young black pups (3–6 wk old) are censused from the air. This is because, up to
that age, they cannot swim and must therefore stay on land. Furthermore, because of their jet-black
natal coat and small size they can be distinguished clearly from adults. Two techniques are used
to estimate numbers of pups: aerial photography and tag–recapture. The photography is timed to
coincide with the peak of the pupping season, when maximum numbers are expected to be present,
and takes place in mid-December. The black pups are counted on large monochrome prints of high
Whereas all colonies can be photographed within a few days, it is not possible to cover all
colonies with a tagging programme. Tagging is labour-intensive and takes much longer to accom-
plish, so normally only one colony can be tagged per year. The tagging programme is carried out
Figure 8
Estimated trend (with 95% condence limits) in the population size of Cape fur seals, 1900–2001.
(After Best et al. 1997).
Number (×10 )
Lower 5th Median Upper 5th
3 Page 316 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 317
in mid-January, when the breeding season is over and all bulls have left the colony. The pups are
then aged about 6–8 wk and are sufciently robust to be handled. Tagging is used as a backup to
aerial photography and also to study seal movements.
The question naturally arises as to how long the population will continue to expand. This is
difcult to answer because the population size before exploitation is unknown. The large size of
four of the more recently established mainland colonies (Kleinsee, Cape Cross, Wolf Bay, and Atlas
Bay) may more than compensate for the 23 extinct colonies but this is uncertain. In pre-exploitation
times it may have been that breeding space was limiting, due to the relatively small size of most
of the islands. Now, however, the formation of large mainland colonies, which can spread, has
altered this situation. Therefore, human activities, in the form of removal of large predators from,
and restriction of human access to, mainland areas, have probably assisted in the recovery of the
seal population. The fact that seals have continued to increase in recent years, in parallel with a
burgeoning shing industry, is an interesting observation and may well be indicative of a sufciency
of food to support both the seals and the shing industry.
There is a rich diversity of seabirds in the Benguela system, including 15 species that breed in the
region (8 endemic) and about 60 species that visit it (Ryan & Rose 1985).
Humans have affected seabirds in the region in several ways, including through exploitation
of some birds and their products, habitat modication and by-catch mortality. Seabird products that
have been harvested include eggs, feathers and guano. Nesting habitat has been greatly modied
by human activities, notably guano collection and the provision of alternative nesting sites, such
as guano platforms constructed along the northern coast of Namibia. Other anthropogenic changes
that have affected seabirds include altered supply of food and pollution, especially oil spills.
Seabirds are an incidental by-catch in several sheries, most importantly longline sheries. More
recently tourism to seabird colonies has developed to the extent where it has the potential to
inuence populations. Additionally, changes to the structure and functioning of the Benguela
system, in particular variations in the abundance of seals, as a result of man’s activities (see above),
have inuenced seabird populations.
The rst men from Western civilisations to see African penguins were Bartholomew Diaz and
his crew in 1487. Diaz described them as “birds as large as ducks, they do not y because they do
not have feathers on their wings. We killed as many of them as we desired and they bray like asses”
(Shelton et al. 1984). This heralded an era in which islands around southern Africa were regularly
plundered to provision ships with the products of seabirds and other marine life. The settlement of
the Cape in 1652 saw an increase in the frequency of visits to islands for this purpose. Although
penguins were killed for food, for fuel to supply ships’ boilers and to be rendered down for their
fat, the primary attraction was their eggs (Randall 1995).
By the late 1700s exploitation and disturbance by humans had led to the cessation of breeding
by African penguins
Spheniscus demersus
at Robben Island and the island was only recolonised
by penguins in 1983. Breeding by African penguins stopped, and has not recommenced, at 10 other
colonies. At three of these sites the likely cause was again exploitation and disturbance by humans;
at ve colonies penguins were displaced by Cape fur seals
Arctocephalus pusillus pusillus
; while
at one colony scarcity of food was the probable reason for cessation of breeding. At the 10th site
breeding was temporary (Crawford et al. 1995b).
Unsustainable harvests of penguin eggs were taken at several islands in the 19th and 20th
centuries, leading to large decreases in the sizes of some penguin colonies. Almost 600,000 eggs
were collected at Dassen Island in 1919, after which harvests decreased steadily (Figure 9). On
the assumption that the decreasing egg harvests reected the trend in the population of adult
penguins at the island, it was estimated that 48% of all eggs produced were harvested and that the
population of penguins aged 2 yr or older was at least 1.45 million in 1910 (Shannon & Crawford Page 317 Wednesday, June 30, 2004 2:19 PM
318 C. L. Grifths et al.
1999). By the early 1990s, this had decreased to 30,000 (Crawford et al. 1995c). Sanctioned
collections of penguin eggs in southern Africa stopped in 1967 (Shelton et al. 1984). Feathers were
collected as a form of down in the early part of the 20th century but this activity probably did not
have a great inuence on seabird populations.
Collection of guano in southern Africa commenced in March 1843 at Ichaboe Island. Manage-
ment of this industry is documented by Hutchinson (1950), Shaughnessy (1984) and Best et al.
(1997). In 1845, licences were issued to vessels collecting guano. From 1847 at Ichaboe Island
and 1869 at certain other islands, concessions were granted for the harvesting of guano. These
concessions terminated in the 1890s. From the 1890s until 1975, management of the islands and
the collection of guano were primarily undertaken by the South African government. In 1976, the
collection of guano was again leased out to private enterprise. Concessions were not renewed in
the 1990s, when the harvesting of guano at islands lapsed. In the 1930s, seawalls were built around
some islands to protect nesting birds from high seas and to reduce losses of guano into the ocean.
For the period 1844–95 it has proved possible to reconstruct the annual harvest of guano from
Namibia (but not from South Africa) using records of guano imported to and exported from several
countries (Van Sittert & Crawford 2003). Accurate records of guano harvests are available since
Figure 9
Harvests of the eggs of African penguins (top) off southern Africa and (bottom) at Dassen Island,
1871–1970. (Redrawn from Best et al. 1997.)
Number (10 )
Number (10 )
3 3 Page 318 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 319
1892 for South Africa and 1896 for Namibia (Figure 10). The guano harvest was decreased in some
years by heavy, unseasonable rains and in others by the periodic scarcity of sh (Hutchinson 1950).
After commercial purse-seine sheries became established off South Africa and Namibia, the
amount of food available to seabirds was reduced and the production of guano decreased (Crawford
& Shelton 1978). Commencing in 1931, platforms to attract seabirds to breed and to deposit guano
were constructed along the coast of northern Namibia between Walvis Bay and Cape Cross. Guano
is still collected annually at these platforms (Figure 10).
The main producers of seabird guano in the Benguela system are Cape gannets
Morus capensis
at six islands and Cape cormorants
Phalacrocorax capensis
at the platforms and most islands. To
a lesser extent, guano is also produced by Bank cormorants
Phalacrocorax neglectus
, other cor-
morants, and African penguins. Guano was collected from April or May, at the conclusion of the
breeding seasons of Cape gannets and Cape cormorants, but at Malgas Island in 1977, commence-
ment of collecting at too early a date caused the deaths of 800 chicks after they had been displaced
from nests. About the same time unrecorded numbers of chicks at Ichaboe and Possession Islands
suffered a similar fate (Crawford et al. 1983).
Figure 10
Collections of seabird guano at (top) islands off Namibia and South Africa and (bottom) platforms
off Namibia, 1892–1995. Sardine stocks collapsed off Namibia in the 1970s and off South Africa in the 1960s.
(Redrawn from Best et al
Tonnes (10 × 3)
SA Islands Namibian Islands
Tonnes (10 × 3) Page 319 Wednesday, June 30, 2004 2:19 PM
320 C. L. Grifths et al.
As a result of guano collecting Cape gannets had little material at islands with which to construct
nests. This caused them to breed later and sometimes to nest on at ground. Although there were
attempts to offset this by placing phosphatic sand (collected from penguin colonies) at gannet
breeding areas (Ross & Randall 1990), eggs were probably lost from nests constructed with too
little guano (Jarvis 1970). Parts of some gannet and penguin colonies also became basin-shaped
as a result of guano collection, allowing rainwater to accumulate, ooding nests and decreasing
breeding success (Randall & Ross 1979). Collection of guano during the breeding season of
penguins caused considerable disturbance at breeding sites. The removal of accumulated deposits
of guano from islands also forced penguins to nest on the surface (Frost et al. 1976). This reduces
breeding success, since burrows have a more constant microclimate than surface nests and shield
birds from direct insolation, which may cause penguins to abandon nests (Randall 1983).
Populations of seabirds, especially kelp gulls
Larus dominicanus vetula
and great white pelicans
Pelecanus onocrotalus
, that ate the eggs or young of, or disturbed, those species that produced
guano were reduced in the 20th century through deliberate destruction of eggs and chicks, shooting,
and harassment (Crawford et al. 1982, 1995a). Control of kelp gulls (mainly the destruction of
eggs and chicks) commenced in 1937, was discontinued as a policy in the early 1960s, but continued
at some localities until 1978 (Crawford et al. 1982). It was recently reintroduced at two Namibian
islands and at Bird Island in Algoa Bay.
The Western Cape population of great white pelicans bred at Robben Island in the early 1600s.
At some stage in the next 250 yr it abandoned this island. As with penguins, exploitation and
disturbance by humans were the most likely reasons. From at least 1869 until at least 1919, pelicans
bred at Dyer Island, sometimes being persecuted by island staff to decrease their predation on
guano-producing birds. Between 1894 and 1904, some bred at Quoin Rock, from where they were
eventually displaced by Cape fur seals. From 1930–54, they bred at Seal Island (False Bay), where
human disturbance included riemen shooting at seals, commercial harvesting of seals, and use of
the island for naval target practice. From 1956, they bred at Dassen Island, where they were left
relatively undisturbed. The population in the Western Cape had been reduced to about 30 pairs
between 1930 and 1956, but following cessation of persecution, it increased to 504 pairs in 1993
(Crawford et al. 1995a) and 603 pairs in 2000 (Figure 11).
The construction of the guano platforms off northern Namibia between 1930 and 1971 provided
alternative nesting space for Cape cormorants and great white pelicans after sand islands in Cape
Cross Lagoon and Sandwich Harbour, where they had previously bred, became joined to the
mainland (Cooper et al. 1982). It is not known to what extent, if any, the construction of these
platforms increased the populations of these two species off northern Namibia. However, construc-
tion of Bird Rock Platform north of Walvis Bay and the grounding of
Meisho Maru No. 8
Cape Agulhas enabled extensions of the breeding range of crowned cormorants
415 km to the north and 16 km to the east. Various man-made structures, including
jetties, disused boats, articial islands in sewage or salt works, and roofs of buildings, have provided
some seabirds with safe nesting habitat. Other natural sites have been rendered unsuitable for some
species through, for example, islands being joined to the mainland to provide safe anchorage for
shing boats but thereby allowing access to mainland predators.
Up until the mid-20th century, factors inuencing trends in seabird populations had mainly
been those operating at breeding localities, such as harvesting of eggs and the collection of guano.
Tow ard the end of the 20th century, seabirds were increasingly inuenced by factors at sea, the
most important of which were an altered supply of food, marine pollution, and incidental mortality
of seabirds caused by sheries (e.g., Crawford et al. 1995c). For some species the food supply
decreased and brought about large decreases in population sizes. For others additional sources of
food were made available.
The seabirds most affected by a decreased supply of food have been those that feed mainly on
sh exploited by commercial sheries. Colonies of African penguin between Lüderitz and Dassen
Island and of Cape gannet in Namibia underwent massive decreases after collapses of sardine Page 320 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 321
Figure 11
Trends in populations of (A) African penguins, (B) great white pelicans, and (C) kelp gulls in
Western Cape in the 20th century.
Breeding pairs (×10 ) Western SA
Breeding pairs Western Cape
Pairs (×10 ) on ten major islands
33 Page 321 Wednesday, June 30, 2004 2:19 PM
322 C. L. Grifths et al.
Sardinops sagax
resources off South Africa during the 1960s and Namibia during the 1970s. Off
South Africa, sardine was replaced by anchovy
Engraulis encrasicolus
Engraulis cap-
; Grant & Bowen 1998). Cape gannets switched their diet to this species. However, anchovy
stocks on the Agulhas Bank were beyond the foraging range of African penguins breeding at
colonies to the north of Table Bay, causing many young penguins to emigrate from the west coast
to breeding localities on the Agulhas Bank. More recently, as South Africa’s sardine stock has
recovered, there has been a reversal of this trend. In Namibia, there was no replacement of sardine
in the epipelagic zone. The population of African penguins at Namibian localities to the south of
Lüderitz decreased from more than 40,000 pairs in 1956 to only about 1000 pairs in 2000. The
area occupied by breeding Cape gannets at Namibian islands fell from 6.24 ha in 1956 to 0.63 ha
in 1996. There was some emigration of young gannets from Namibia to islands off the Western
Cape (Crawford et al. 1985, 2001, Crawford 1998, 1999). Bank cormorants decreased off the
Western Cape following a decrease in production of Cape rock lobster
Jasus lalandii
(Crawford et
al. 1999).
Additional food that has been made available to seabirds includes unwanted sh and offal
discarded by shing boats, food at coastal rubbish tips and abattoirs and sh present in freshwater
impoundments that have been built in proximity to the coast (Berruti et al. 1993, Crawford et al.
1995a). These sources of food have been used especially by great white pelicans, Cape gannets,
kelp gulls, Hartlaub’s gulls
Larus hartlaubii
, and some of the nonbreeding visitors to the Benguela
system (Ryan & Rose 1985). The populations of great white pelicans and kelp gulls increased in
the Western Cape in the past two decades (Figure 11). This was probably a response to the lifting
of controls on these populations. However, the overall supply of food for these two species may
be greater now than previously, so their populations may rise above pristine levels. Because both
species feed on the eggs and chicks of other seabirds (Crawford et al. 1995a, 1997), the reproductive
success of prey species may be reduced below levels that would pertain in an undisturbed system.
The eggs and chicks of penguins, which at many islands can no longer burrow into guano, are now
more accessible to predators than previously. Availability of an articial supply of food may have
led to an extension in the breeding season of Hartlaub’s gulls (Ryan 1987).
Oiling is a major threat to several seabirds. It causes feathers to clump, leading to a breakdown
in their insulating properties. As a result, birds become hypothermic and are forced to leave cold
waters. They dehydrate, mobilise stored energy reserves and may lose up to 13% of their body
mass within a week (Morant et al. 1981). Unless rescued, they will eventually starve. Oil ingested
by preening can also cause ulceration of the mouth, oesophagus, and stomach and, in severe cases,
can lead to substantial blood loss. Oil absorbed into the system can cause red blood cells to rupture,
leading to anaemia (Birrel 1994). Furthermore, an immunosuppressant effect makes birds more
susceptible to diseases (Morant et al. 1981) such as pneumonia and aspergillosis. Ingested oil may
produce a greater diversity of pathogenic bacteria. If a bird gets oil in its eye, it can lead to ulceration
of the cornea and blindness unless treated (Crawford et al. 2000). From 1970–2001, more than
46,000 African penguins and 5000 Cape gannets were oiled (Morant et al.
1981, Adams 1994,
Underhill et al. 1999, Crawford et al. 2000). Many have been successfully cleaned and returned to
the wild, although there has also been heavy mortality. Some of the rehabilitated birds breed again
in the wild (Randall et al
There is limited mortality of African penguins from entanglement in shing nets, discarded
line and other material. There are unconrmed reports of penguins being used as bait in rock lobster
traps (Ellis et al. 1998). Off Namibia, Cape gannets are killed as accidental by-catch in longline
sheries. Tuna sheries off southern Africa also kill albatrosses and petrels (Ryan & Boix-Hinzen
1998). Tourism to seabird colonies may decrease breeding success through disturbance.
The population of Cape fur seals in the Benguela system was reduced to low levels at the start
of the 20th century, but has recovered during the 20th century (see seal section, p. 312). Increasing
seal numbers adversely inuence specialist seabirds in three ways: they compete for breeding space,
compete for food and inict mortality (Crawford et al. 1989, 1992, 2001). Page 322 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 323
In summary, humans have caused large decreases in the populations of abundant seabirds in
the Benguela system, initially through unsustainable exploitation and later by competing with them
for food. Seabirds have also been killed by oil spills and as incidental by-catch in sheries. However,
for some seabirds and seals, humans have created additional breeding habitat and additional sources
of food. Populations of these opportunistic species increased during the 20th century, following
cessation of uncontrolled harvesting or the removal of population controls and they in turn are now
adversely inuencing populations of some of those seabirds that have more specialised requirements
for feeding and breeding.
Pelagic sheries
The pelagic shery in the Benguela region uses purse-seine nets to target visible, near-surface
shoals of small pelagic shes, which are generally caught at night. Once a shoal has been located,
the net is set in a circle around it and the bottom of the net pulled closed by means of a foot rope.
The net is then brought alongside the vessel and shes are sucked out with a suction pump and
transferred to the hold. Depending on the species targeted, the shes are kept in refrigerated seawater
or a strong brine solution. Once its hold is full, the vessel returns to harbour, where the catch is
processed. Most pelagic shing trips last a single night and landings of up to 200 t trip
are not
uncommon. Because of the high degree of targeting and selectivity of purse-seine operations, by-
catch of other species is generally low.
The anchovy
Engraulis encrasicolus
and sardine
Sardinops sagax
have been the dominant
species of the Benguela pelagic shery, with round herring
Etrumeus whiteheadi
and horse mackerel
spp. occasionally important. Off South Africa, where anchovy has dominated pelagic
landings for the past three decades, over two thirds of anchovy caught are juveniles of around 6
months old, migrating from the west coast nursery grounds to the spawning grounds off the south
coast. Anchovy, round herring, and juvenile sardine and horse mackerel that shoal together with
anchovy or round herring are reduced to sh meal and sh oil. Adult sardine are either canned or
frozen for human consumption or frozen for use as bait in other sheries. Currently, the pelagic
shing industry operates primarily from ports along South Africa’s west and southwest coasts,
although commercial pelagic shing also takes place from Mossel Bay and Port Elizabeth. In
Namibia, the purse-seine eet is based in Walvis Bay. As is the case for pelagic sheries worldwide,
that of the Benguela is characterised by high volumes and relatively low values compared to other
Development of the South African pelagic shery took place after World War II. Fishing for
sardine and horse mackerel started in St. Helena Bay and soon expanded along South Africa’s
entire west coast. Rapidly increasing effort from the late 1940s led to a peak catch of almost
500,000 t in 1962, over 80% of which was sardine (Crawford et al
1987). Catches of both species
declined rapidly after 1962, the sardine stock collapse being ascribed to overshing, expansion of
shing grounds to the south and variable recruitment (Beckley & van der Lingen 1999). Reduced
horse mackerel catches were ascribed to the depletion in the population of the exceptionally strong
year classes of 1946, 1947, and 1948 (Crawford et al. 1987). Since their collapse horse mackerel
landings have consisted primarily of juvenile shes and have remained low, generally less than
10,000 t yr
. Annual sardine catches were low and relatively stable at around 50,000 t during the
period 1967–94 but have since shown a steady increase to 190,000 t in 2001. In 1964 the South
African pelagic shery switched to smaller-meshed nets to target anchovy, which then replaced
sardine as the mainstay of the South African pelagic shery and has dominated purse-seine catches
since. Average catches of anchovy off South Africa have been in the region of 230,000 t yr
have varied between 40,000 t in 1996 and 596,000 t in 1987. A time series of pelagic landings by
major species is given in Figure 12.
Sardine was also the dominant species initially targeted by Namibian purse-seine vessels. The
collapse of the South African sardine stock in the early 1960s resulted in increased shing effort Page 323 Wednesday, June 30, 2004 2:19 PM
324 C. L. Grifths et al.
off Namibia, and annual pelagic catches rose rapidly from around 200,000 t (composed entirely of
sardine) in the 1950s to a maximum of 1.55 million t (1.4 million t of which was sardine) in 1968
(Boyer and Hampton 2001). In addition to the local purse-seine eet operating out of Walvis Bay,
factory ships from the eastern European midwater eet, shing outside territorial waters, also
targeted sardine in the late 1960s. Poor catch records from this eet indicate that the sardine catches
reported during the 1960s must be regarded as a minimum (Boyer & Hampton 2001). After 1968,
the Namibian pelagic shery experienced a decline in landings to around 625,000 t in 1972, followed
by slightly increased catches for a few years, before another abrupt collapse in the early 1980s
(Figure 12). These collapses have been primarily attributed to overshing, although poor sardine
recruitment resulting from adverse environmental conditions exacerbated the decline. As in South
Africa, smaller-meshed nets were used to target anchovy after the collapse of the sardine stock,
with annual anchovy catches uctuating around 200,000 t during the 1970s and 1980s. A pro-
nounced peak of around 350,000 t in anchovy landings was recorded in 1987, the same year the
highest anchovy catches were recorded off South Africa. Since then catches have rarely exceeded
Figure 12
Pelagic sheries catches in the Benguela ecoregion, 1950–2001. Top, landings of anchovy, sardine,
and other species (horse mackerel, chub mackerel
Scomber japonicus
, round herring, and lanternsh
panyctodes hectoris
) by the South African shery. Bottom, landings of anchovy, sardine, and other species
(horse mackerel only) made by the Namibian shery. In both panels the total pelagic catch from the entire
Benguela region is shown by the solid line. (Data for South Africa updated from Crawford et al. 1987; and
for Namibia from Boyer and Hampton 2001, and A. Kreiner, Ministry of Fisheries and Marine Resources,
Namibia, personal communication.)
Catch (10 tonnes)
SA Other
SA Anchovy
SA Sardine
Benguela Total
Catch (10 tonnes)
Namibian Other
Namibian Anchovy
Namibian Sardine
Benguela Total
33 Page 324 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 325
50,000 t and have declined to zero following the anomalous environmental event recorded in the
Overall, it can be seen that landings by pelagic sheries in the Benguela ecoregion have
uctuated 10-fold over the period 1950–2001, between 185,000 t and 1.9 million t, with a long-
term average of 770,000 t (Figure 12). Peak landings were made in the decade 1965–75, primarily
by the Namibian purse-seine eet (80% of total pelagic landings in 1968 and 1969). Whereas
landings by South African purse-seine vessels have remained relatively constant over the past 50
yr, those by the Namibian shery have exhibited a 40-fold variation. Pelagic landings in the
Benguela region are no longer dominated by Namibia. Since 1979, over 50% of pelagic landings
have been taken by the South African purse-seine eet and this percentage has increased steadily.
By 2001, 94% of total pelagic catches from the Benguela were made off South Africa.
Fluctuations in stock size
The large-scale uctuations in population size and changes in the relative abundance of anchovy
and sardine indicated by pelagic shery catch data are typical of upwelling ecosystems. Analysis
of scale deposits in sediments from upwelling systems suggests that these population uctuations,
or regime shifts as they have become called, are a natural phenomenon that occurred in the absence
of shing (Schwartzlose et al. 1999). Estimates of anchovy and sardine stock size in the Southern
Benguela, derived initially from shery-dependent data, but more recently from shery-independent
surveys, corroborate the catch data (Figure 13). Off South Africa, anchovy replaced sardine as the
dominant small pelagic species following the collapse of the sardine population. This dominance
lasted from 1965 until the mid-1980s, after which a steady increase in the sardine population
resulted in populations of approximately equal size during the latter half of the past decade.
Estimates of Namibian sardine biomass indicate that the stock collapsed from more than 11 million
t in 1964 to well below 1 million t by the mid-1970s, and the population has not attained more
than 500,000 t since (Figure 13). Unfortunately, biomass estimates for Namibian anchovy are not
available. Unlike the Southern Benguela, however, anchovy in the Northern Benguela did not
replace sardine to a great degree following the latter’s collapse. After sustaining moderate catches
from the 1970s to the mid-1990s, the anchovy population became severely depleted following the
Benguela Niño of 1994–95 (Boyer & Hampton 2001).
Ecological impacts of the shery
Because of their intermediate trophic level and massive population sizes, small pelagic shes in
upwelling ecosystems exert top-down control of zooplankton and also bottom-up control of
predators such as other shes and marine birds. Small pelagic shes therefore constitute a crucial
link in mediating energy ows between lower and upper trophic levels and, because of the low
numbers of species comprising this group, have been termed wasp-waist populations (Cury et al.
2000). Their critical position means that changes in the abundance of these small pelagic shes
have substantial impacts on the ecosystem. For example, overshing and collapse of the sardine
stock during the 1960s were followed by collapses of colonies of African penguin along the west
coast of southern Africa (Crawford 1998 and see above). Whereas the recovery of the Southern
Benguela sardine population has resulted in a stabilisation of penguin colonies between Lüderitz
and Table Bay, this has not arrested the overall decline in penguin numbers. The inability of the
penguins to cope with recent shifts in dominance in prey species may have resulted from increased
competition for food with shermen during the 20th century (Crawford 1998). Similarly, the decline
of the sardine stock off Namibia resulted in severe decreases in numbers of Cape gannets, which
were not able to successfully exploit the mesopelagic horse mackerel and the goby Sufogobius
bibarbatus that replaced sardine (Schwartzlose et al. 1999). Page 325 Wednesday, June 30, 2004 2:19 PM
326 C. L. Grifths et al.
Explicit analyses of the ecosystem effects of pelagic shing have been achieved through
computer simulation in which a dynamic, mass-balanced trophic model (called Ecosim) of the
Southern Benguela was subjected to various types of shing pressure (Shannon et al. 2000). These
simulations have suggested that increased shing on small pelagic shes in systems where wasp-
waist (both top-down and bottom-up) control dominates results in major perturbations propagating
through the ecosystem. Competing species are favoured through enhanced availability of zooplank-
ton prey, and the increased abundance of these competitors delays recovery of the target species
once shing pressure is reduced.
While intense shing on small pelagic shes in the Benguela has had large and obvious
ecosystem impacts, more subtle effects on the sh populations have been suggested, principally
the erosion of intraspecic diversity (Cury et al. 2000). For example, the composition of pelagic
sh schools has been shown to reect the relative species abundance within the pelagic community.
When anchovy or sardine are abundant they tend to form relatively pure schools, whereas when
their abundance is low, they join schools of other more abundant species. These variations in average
percentages of a pelagic species contained in schools track the overall relative population abundance
remarkably well (see Figure 7 of Cury et al. 2000 for an example from the Southern Benguela)
and have led to the development of the school trap hypothesis (Bakun & Cury 1999). This hypothesis
postulates that “a sh species driven to school together with a more abundant species must
effectively subordinate its specic needs and preferences to the ‘corporate volition’ of a school
largely driven by a different set of needs and preferences.” The school trap thus constitutes a
mechanism for adverse population interaction between co-occurring small pelagic shes and could
maintain a collapsed population in a depleted state for lengthy periods, as well as affecting spatial
dynamics such as migrations. Possible evidence for the school trap hypothesis has been suggested
from a study of changes in the spawning habitats of small pelagic shes off South Africa over the
past two decades. During the 1980s and early 1990s, sardine and anchovy in the Southern Benguela
showed a broad-scale overlap in their spawning habitats. From 1994 onward, however, their
spawning habitats became markedly distinct, with sardine spawning principally off the west coast
and anchovy spawning predominantly off the south and east coasts (van der Lingen et al. 2001).
Figure 13 Estimates of biomass of some important species of pelagic sh in the Benguela ecoregion,
1950–2000. Estimates of South African sardine and anchovy biomass up to 1982 derived from virtual
population analysis (VPA) and from 1984 onward from hydroacoustic surveys. Estimates for Namibian sardine
were VPA-derived from 1952–85, and from hydroacoustic surveys from 1990–2000. Note that the VPA-derived
estimates are considered to reect only large-scale trends in abundance because of several limitations in the
data (see Schwartzlose et al. 1999). (Data updated from Beckley and van der Lingen 1999, for South African
and Namibian sardine; from Butterworth 1983, for anchovy VPA estimates; and from Barange et al. 1999.)
Biomass (×10 tonnes)
Namibian Sardine
SA Anchovy
SA Sardine
6 Page 326 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 327
This was also the rst year of the acoustically estimated biomass time series when the relative
biomass of sardine was higher than that of anchovy (Barange et al. 1999).
The school trap may not only operate between species, but intraspecically as well. From the
point of view of a sh population, a shery represents nothing more or less than a massive source
of predation pressure, and sustained pressure on a particular component of the population could
result in removal at the population level of the tendencies expressed by the targeted component.
Off Namibia, sardine primarily spawned near the coast between two intense upwelling centres (near
Cape Frio and Luderitz) that appeared to provide the best available reproductive habitat, some
spawning also taking place offshore, in the region of the Angola–Benguela front (Bakun 2001).
The development of the pelagic shery occurred directly within the inshore, primary reproductive
habitat, and the inated predation pressure arising from shing may have resulted in a steady
preferential removal from the sardine population of individuals with strong afnities for the primary
reproductive habitat. As a result, essentially all of the Namibian sardine’s reproductive output in
recent years has been concentrated in the Angola–Benguela frontal zone (Bakun 2001). These two
examples indicate that overshing not only can alter the abundance of small pelagic sh populations,
and hence have a substantial impact on ecosystem structure, but also can affect ecosystem func-
tioning by altering the composition and spatial distribution of such populations.
Demersal and midwater trawl sheries
The demersal shery off southern African started in the early 1900s (Lees 1969). Initially the prime
target species were Agulhas sole Austroglossus pectoralis and west coast sole Austroglossus microl-
epis (Payne & Badenhorst 1989), but this changed with the discovery of the vast hake resource off
the west coast during the First World War (Payne & Punt 1995). From these humble beginnings,
the demersal shery, based on the Cape hakes Merluccius capensis and Merluccius paradoxus, has
developed into the most valuable shery in the region. The historical background to the hake shery
of southern Africa and the management of the stocks have been described in detail by Andrew
(1986), Payne (1989), Punt (1991), Payne & Punt (1995), Gordoa et al. (1995), Boyer & Hampton
(2001), and van der Westhuizen (2001), and hence only a brief summary is provided below.
Initially the demersal shery was concentrated close to Cape Town and total annual landings
of hake were around 1000 t (Lees 1969). The expansion of the shery was initially slow, and hake
landings were less than 50,000 t yr–1 by the end of the Second World War. Improved technology
enabled the expansion of the shing grounds and the annual hake landings began to increase more
rapidly, reaching 50,000 t by 1950 and 160,000 t by 1960 (Payne & Punt 1995). Exploratory shing
by Japan and Spain in the early 1960s showed that catch rates were higher off Namibia than off
South Africa (Gordoa et al. 1995). This resulted in the expansion of the shery into Namibia waters,
an increase in the foreign distant-water eets, and a rapid escalation of hake landings, reaching
over 1 million t by 1972 (Figure 14).
This rapid escalation of shing effort prompted the establishment of the International Com-
mission for the Southeast Atlantic Fisheries (ICSEAF) in 1972. Over the next few years ICSEAF
introduced a minimum mesh size of 110 mm, a system of international inspection, and allocated
quotas to member countries. South Africa declared a 200-nautical mile exclusive economic zone
on November 1, 1977, and excluded all but a small amount of foreign effort, thereby reducing the
hake catches off South Africa (Figure 14). South Africa then embarked on a rebuilding strategy
for the Cape hake resource by setting conservative annual catch limits.
At that time, South Africa governed the ex-German colony of Southwest Africa (as Namibia was
then known). In the absence of an internationally recognised government, an EEZ could not be
instituted in the Northern Benguela. Nonetheless, some foreign shing effort was withdrawn due to
the falling hake catch rates. The international shery off Namibia was managed by ICSEAF until
independence in 1990. The newly recognised nation then declared a 200-nautical mile EEZ and Page 327 Wednesday, June 30, 2004 2:19 PM
328 C. L. Grifths et al.
excluded all foreign shing effort, resulting in a dramatic drop in hake catches (Figure 14). Since
1990 there has been a perceived gradual recovery of the hake resource and an increase in hake catches.
The wider ecological effects of the hake shery are difcult to estimate. There are three main
aspects to be considered: discarding of hake, catches of other species, and substratum damage. The
minimum marketable size for hake was certainly greater in the early days of the shery than today,
resulting in a higher proportion of discards. The “ofcial” hake catch statistics (Figure 14) prior to
1972 were thus increased by 39% in accordance with a decision made by ICSEAF in 1978 (Andrew
1986). In the 1980s the hake shing industry was forced to develop a market for small hake because
of the depleted nature of the resource. A consequence of the rebuilding strategy has been an increase
in the availability of large hake, which has allowed the industry to diversify the market. There has
been a gradual shift to deeper water as the industry has started targeting larger shes (Glazer &
Butterworth 2002). Walmsley (2004) estimated that 17,000–25,000 t of sh (including 7000–12,000 t
of hake) were discarded in 1997 by the demersal trawl eet off the west coast of South Africa.
However, there are consistent rumours of high grading off South Africa, particularly by small quota
holders trying to maximise the nancial returns on their limited hake quotas.
Some by-catch species such as Agulhas sole, kingklip (Genypterus capensis), and monksh
(Lophius vomerinus) are sought after because they fetch a higher price per unit mass than hake.
Other species (e.g., horse mackerel Trachurus trachurus capensis and ribbon sh Lepidopus
caudatus) are retained or discarded depending on market demand, whereas a third group (e.g.,
grenadiers Caelorinchus symorhincus and Malacocephalus laevis) are unwanted incidental by-catch
and are routinely discarded.
For a brief period in the early 1970s there was a directed shery for west coast sole in the
vicinity of the Orange River. The shery was closed in the late 1970s due to a stock collapse and
remains closed in South Africa, but there has been a small directed shery off Namibia since 1994
(Boyer & Hampton 2001). In 1983 an experimental kingklip-directed longline shery was started
off South Africa (Japp 1988, 1989). Catches increased rapidly from 1042 t in 1983 to a peak of
8684 t in 1986; the shery was closed due to collapse of the stock in 1990. Subsequent modelling
showed that the kingklip resource was already under pressure as by-catch in the hake shery before
the longline shery commenced (Punt & Japp 1994). There is a directed shery for monksh off
Namibia (Boyer & Hampton 2001, Maartens & Booth 2001a,b) and some targeted shing off South
Africa (Walmsley et al. 2004).
Horse mackerel are semipelagic and are exploited by purse-seine nets and by both midwater
and bottom trawls. After the peak hake landings off Namibia in the early 1970s, some effort was
diverted to midwater trawl shing for horse mackerel in response to declining hake catch rates
(Payne & Crawford 1989). The midwater trawl shery for horse mackerel subsequently became
Figure 14 Catches taken by the demersal shery in the Benguela, 1917–2000.
Tonnes ( 10 )
Horse Mackerel
3 Page 328 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 329
the largest shery by volume off Namibia, with catches of around 500,000 t yr–1 between 1978 and
1987 (Boyer & Hampton 2001). Adult horse mackerel were exploited by purse-seine off the west
coast of South Africa in the 1950s and 1960s, with catches peaking at 102,600 t in 1952 (Johnston
& Butterworth 2001), but this resource has largely disappeared. Small quantities of adults are taken
in bottom trawls and adults in spawning condition have been observed during research cruises to
the west coast. (Note that the horse mackerel landings shown in Figure 14 include catches taken
by midwater trawl off the south coast of South Africa.) It is generally assumed that the horse
mackerel resource off the west coast of South Africa was a southern extension of the large Namibian
resource. With the collapse of the sardine shery off South Africa, some purse-seine effort was
redirected toward juvenile horse mackerel, but catches have generally been small, except for 1989
and 1996, when the purse-seine catch of juvenile horse mackerel exceeded 25,000 t. The purse-
seine catch of juvenile horse mackerel of the South African west coast is limited to 5000 t in terms
of current regulations. However, in some years substantial quantities of juvenile horse mackerel
(up to 100,000 t) are taken in pelagic purse-seine nets off Namibia (Boyer & Hampton 2001).
The historical proportion of hake in the catches is not known because only data on the landings
(i.e., retained catch) are available. Walmsley (2004) found that the proportion of hake in the catches
taken off South Africa increased with increasing depth. Therefore, one would expect that the
proportion of unwanted by-catch was higher during the early stages of the shery, when vessels
were small and shed in shallow waters close to Cape Town. Payne & Punt (1995) show that the
hake contribution to the landings off South Africa decreased from around 90% in the early 1960s
to around 60% in the early 1990s. They attribute this decline in the hake proportion to a combination
of channeling effort into mixed species sheries and to landing a greater proportion of the by-catch.
Bottom trawls are unselective, and there is a growing body of literature on damage they cause
to the seabed (see, for example, Hollingworth 2000). As a result, there is increasing pressure to
develop less destructive shing methods. A widely used alternative is longlines. These are lines of
over a kilometre in length with hooked branch lines called snoods placed at regular intervals. An
experimental longline shery for hake was started off South Africa in 1983 and currently some
5000 t of hake are caught by this method off South Africa and 5000–10,000 t off Namibia. Although
longlines are more selective than bottom trawls and do not damage the seabed, they are not without
their ecological problems. In particular, longlines are responsible for a substantial mortality of
seabirds (Barnes et al. 1997). Due to their very low reproductive rate and delayed onset of
reproductive maturity, these k-selected species cannot withstand the observed levels of mortality
(Ryan et al. 2002 and references therein). In addition, there is substantial loss of catch to predators
such as Cape fur seal and through sh breaking off when the line is hauled in slightly rough weather.
Inshore net sheries
Both beach-seine and gill nets are used by the Benguela inshore net shery. Beach-seines are mobile
nets, usually rowed out into the surf zone under the directions of a spotter, to encircle a shoal of
sh (most commonly mullet Liza richardsonii). A crew of 6–30, depending on the size of the net
(up to 275 m in length) and the length of the haul, then haul the head ropes shoreward. As the net
approaches the shore, the ends, or wings, of the net are brought together, and the trapped sh driven
into the bag in the middle of the net. Occasionally no spotter is used and a “blind seine” is made
in areas or at times when shes are likely to occur. Smaller 50- to 100-m beach-seine nets may
also be deployed by walking out into the surf to encircle sh and are locally referred to as foot
nets. Beach-seine shing has remained essentially unchanged since the technique was introduced
to South Africa during the mid-1600s. The only technological improvements relate to the use of
woven nylon rather than cotton nets, glass bre as opposed to wooden boats and four-wheel drive
vehicles to transport the rigs on sandy beaches.
Gill netting is normally a passive form of shing in which nets are deployed, usually from a
boat, in the hope that shes will swim into them and become entangled. Two main types of gill Page 329 Wednesday, June 30, 2004 2:19 PM
330 C. L. Grifths et al.
nets have been used in the region: positively buoyant, 44-mm stretch-mesh drift nets (used for
catching Liza richardsonii) and negatively buoyant set nets set along the seaoor and buoyed and
anchored at both ends. In the past, set nets of various mesh sizes (44–178 mm) were used, but
since 1982 the only legal mesh size has been 178 mm, used to target St. Joseph shark Callorhinchus
capensis. Although cotton and multilament braided nylon mesh were used in the past, most modern
gill nets are made from monolament nylon mesh. Technological advances (such as outboard
motors, echo sounders, and spotlights) have also allowed commercial gill netters to employ a more
active type of gill netting. Shoals may be completely encircled, or the nets set in a semicircle in
the path of the shoal. The sher then scares the shes into the net by revving the outboard motor
and completing the circle behind the shoal.
Origins and history of inshore net shing
The origins of inshore net shing in South Africa date back to the use of beach-seine nets by
English and Dutch seafarers during the 17th century (De Villiers 1987). By the end of that century,
the Dutch had established beach-seine shing outposts as far aeld as Langebaan Lagoon (Poggen-
poel 1996). Most of the catch was salt cured, dried, and used to supply the Dutch East India
Company trading boats, troops, and slaves at the castle in Cape Town. With the expanding colonial
settlement in the Cape, beach-seine shing became widespread and was the most frequently
employed shing method in most areas. Thompson (1913) writes:
Nearly every sea-board hamlet and coast farm have a net or two, whilst at Struis Bay, Keimouth and
similar stations “trekking” is practically the only mode of shing carried on. The St. Helena Bay and
Saldanha Bay areas and False Bay are the chief centers of the seining industry, which in these localities
has long been the means employed for dealing with the large shoals of harders, white stumpnose, elft,
albacore and young white steenbras that are periodically on the move close inshore.
By this time the link between agriculture and beach-seine shing was well established. Coastal
farmers either worked their own nets or purchased sh from shing settlements on the coast. The
shes were still processed into sun-dried “bokkoms” and given to farm labourers as “rantsoen vis”
in partial payment for their work.
The rst major threat to the supremacy of beach-seine nets as the gear of choice for inshore
shing, particularly along the west coast, was the introduction of gill nets by Italian immigrants
during the late 1800s (Thompson 1913). User group conict between the traditional beach-seine
shers and the new, more technologically advanced gill net shers began to occur. Despite their
admiration of the gill net, the state was also not keen to alienate the beach-seine shers or farmers
and drafted various regulations aimed at minimising conict between the two sectors (Thompson
1913, Van Sittert 1992). During the early 1900s many of the Italian gill net shers began to work
in the expanding rock lobster shery and competition for sh and the rantsoen vis market between
the two groups decreased (Van Sittert 1992). By 1912, however, over 300 gill nets were in use in
the Berg River and in the sea to the south of the river mouth (Van Sittert 1992). This indicates the
extent to which these nets had been incorporated into the inshore sheries along the west coast
and beach-seining was no longer the dominant net shing method.
The importance of inshore gill and beach-seine shing along the west coast was further reduced
with the start of the purse-seine shery during the 1930s. Purse-seine shers were more efcient
at catching large quantities of harders Liza richardsonii and maasbanker Trachurus trachurus than
beach-seine and gill net shers. The large quantities of purse-seine-caught shes thus ooded the
rantsoen vis market, causing prices to collapse and considerable hardship for gill and beach-seine
net shers (Van Sittert 1992). This prompted many traditional beach-seine and gill net shers to
either join the purse-seine shery or seek alternative employment. Increased demand for sh during
the Second World War led to the construction of numerous canning factories and the purse-seine Page 330 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 331
eet began targeting the more abundant pelagic pilchard Sardinops sagax and anchovy Engraulis
encrasicolus. Smaller purse-seine vessels did, however, continue to catch considerable quantities
of L. richardsonii. Despite numerous complaints by gill and beach-seine shers, management was
slow to respond, ofcially requesting purse-seines not to target Liza richardsonii in 1973 and
banning the practice in 1984.
Management of inshore net sheries
After the early attempts to resolve conict between gill and beach-seine shers around the turn of
the last century, the inshore net sheries received little management attention for almost 50 yr. Any
person could sh anywhere along the coast with gill or beach-seine nets, the only restriction being
a minimum mesh size of 44 mm, aimed at preventing catches of juvenile Liza richardsonii. In
1967, the old conict between beach-seine and gill net shers again threatened to develop. An
investigation by the then Sea Fisheries Research Institute revealed that the numbers of beach-seine
and gill nets in use was increasing rapidly. This was possibly due to the collapse of the west coast
sardine stock in the late 1960s and the subsequent need for purse-seine shermen to once again
supplement their incomes by inshore net shing. With the growth of recreational line shing, anglers
and conservation-minded members of the public also became increasingly opposed to inshore net
shing, which they alleged threatened stocks of numerous sh species.
As a result, numerous control measures were introduced, including compulsory registration of
all nets in 1974 (with permit holders required to submit daily catch returns on a monthly basis),
gear and catch restrictions, and the restriction of net shing to certain areas. A reduction of inshore
net shing effort became ofcial policy and permits were only awarded to so-called bona de
shers or pensioners with a history of participation in the shery. The number of beach-seine
permits issued for the popular angling areas of False Bay and Walker Bay were drastically reduced
and gill netting was conned to the west coast (north of Cape Town). With the exception of False
Bay, the inshore net shery has received negligible management action since the early 1980s,
although along with other South African shing sectors, net shing rights are currently under review
and a substantial reduction in effort is likely.
Spatial distribution of effort
Although limited gill net sheries targeting Liza richardsonii and other species appear to have
developed around Cape Town and along the southwest coast, management measures aimed at
restricting their use, in favour of beach-seines, were proposed as early as 1926. Gill net and beach-
seine shing were probably introduced to Namibian waters by South African pelagic shers during
the 1950s and 1960s, and a limited market probably constrained the expansion of the shery beyond
the ports of Walvis Bay and Luderitz. The area between Elands Bay and Langebaan on the west
coast and False Bay and Walker Bay on the southwest coast were the main areas of net shing
activity for much of the last century.
It was only with the compulsory registration of gill and beach-seine nets in 1974 that information
on the full extent of this net shing activity became available. A total of 1303 net permits were
issued to 593 permit holders, who operated between Cape Agulhas and Walvis Bay. In Namibia,
only 17 net permits were issued initially, but this increased rapidly to 185 permits issued to 34
permit holders by 1978. By 1985, however, only six of these permit holders were reporting catches.
Approximately 70% of all permits issued were for set or drift gill nets, the remainder being beach-
seine permits. By far the majority (77%) were issued to shers operating between Elands Bay and
Langebaan on the west coast. Along the southwest coast 191 beach-seine permits were issued in
1974. In line with ofcial policy to reduce the impact of beach-seine netting on so-called linesh
species, the number of permits issued for the southwest coast was reduced by 66 to 64% by 1985. Page 331 Wednesday, June 30, 2004 2:19 PM
332 C. L. Grifths et al.
There has been only a moderate reduction (23%) in the number of net permits issued for the west
coast. Potential netting effort, however, actually increased by allowing all gill net permit holders
north of Saldanha the option of using 178-mm set nets, in addition to their small-mesh harder nets.
The numbers of permits issued also only reect potential netting effort. A recent study has shown
that a large number of permit holders (up to 40% in some areas) are inactive, or only operate a
few times per year in an apparent recreational fashion (Hutchings 2000).
Long-term trends in reported catches
Anecdotal evidence indicates that historically a large portion of the inshore net catch comprised
species other than Liza richardsonii. Linesh species such as white steenbras Lithognathus lithog-
nathus, elf Pomatomus saltatrix, kob Argyrosomus inodorus, white stumpnose Rhabdosargus glo-
biceps, yellowtail Seriola lalandi and galjoen Dichistius capensis contributed over 90% to the total
beach-seine catch at Muizenberg (False Bay) around the turn of the century (Lamberth 1994). The
contribution of these species to net catches has, however, decreased substantially since then, and
Liza richardsonii made up over 77% of the total mass landed by the period 1977–87 (Figure 15).
Despite legislation introduced in 1974 preventing the targeting of linesh by net shers, illegal net
shers still land in the region of 150 t yr–1. A recent survey (Hutchings 2000) indicated that at least
29 species are caught as by-catch in the legal gill net shery along the west coast and contribute
approximately 5% to the total catch. Lamberth (1994) recorded 65 by-catch species in False Bay
beach-seine hauls, and beach-seines along the rest of the southwest coast probably have a similar
catch composition
Although a variety of species were taken by inshore net sheries over much of their history,
it is impossible to separate these net-caught fractions of the catches from those made by other
shing sectors. Reporting of by-catch by net shers has also been shown to be grossly inaccurate
(Lamberth 1994, Hutchings 2000). It is therefore not possible to determine the impact of net shing
on the stock size of these species. However, Liza richardsonii species were caught almost solely
by net shing and some historical catch data for this species do exist.
Gilchrist (1914a) provides the earliest record of Liza richardsonii catches in the St. Helena
Bay area. In this report, evidence of decreasing catches in St. Helena Bay is provided in the form
of annual catches of adult and juvenile Liza richardsonii by Messrs. Stephan Bros. for the years
1880–1913. Assuming these are accurate (gures are not rounded off), the recorded catches for a
33-yr period provide a valuable insight into the net shery in St. Helena Bay at the time. A drastic
reduction in annual catches occurred, with an average annual catch prior to 1900 of approximately
102 t (calculated from a conversion ratio of ve adults kg–1 and eight juveniles kg–1), declining to
only 16 t thereafter, a 85% decrease (Figure 16). It appears likely that the observed decline was
due to the high shing effort by both estuarine and marine net shers. The particularly noticeable
Figure 15 Beach-seine catches in False Bay for three different time periods. (After Lamberth 1994.)
1896—1906 1950—1968 1977—1987
Tonnes (10 )
Linefish Harders
3 Page 332 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 333
decrease in the number of juveniles caught suggests that a degree of recruitment overshing had
Despite the decline in the St. Helena Bay catch for this period, the total Liza richardsonii
landings for the region as a whole (Figure 17) rose to 1775 t for the period 1927–31 (Marine and
Coastal Management, unpublished data), substantially more than the 259–323 t recorded for
1898–1900 (Gilchrist 1899, 1900, 1901). The total catch for the period 1974–99 appears to show
a sustained decline since 1987–99 (Figure 17). These data, however, are based on compulsory catch
returns by net shers. Such returns have since been shown to be inaccurate, with many permit
holders under- or overreporting catches or failing to submit returns. Indeed, as little as 8% of the
total catch is reported in some areas, and the actual total catch of Liza richardsonii was estimated
at around 5000 t for 1998 (Hutchings 2000). It is therefore unlikely that these statistics are accurate
reections of the real trends and data for this period should be treated with extreme caution.
Figure 16 Recorded net catches of adult and juvenile mullet (Liza richardsonii) by Messrs. Stephan Bros.
of St. Helena Bay, 1880–1913. (After Gilchrist 1914a.)
Figure 17 Total reported mullet (Liza richardsonii) catch over time.
Catch (tonnes)
Juveniles Adults
Catch (tonnes)
Southwest coast
West coast
Namibia Page 333 Wednesday, June 30, 2004 2:19 PM
334 C. L. Grifths et al.
Given the paucity of catch data for most of the net sheries history and the inaccuracy of more
recent catch statistics, it is impossible to make informed comments on the stock status of Liza
richardsonii in the Benguela. Undoubtedly over 300 yr of inshore net shing has substantially
reduced the biomass of Liza richardsonii and of other species in the region. Indeed, there is
compelling circumstantial evidence that the Liza richardsonii stock is overexploited in some areas,
but there is simply insufcient data to quantify this.
Commercial lineshing is one of the oldest shing industries in the Benguela system. Fleets of
row- and sailboats were used for handline shing off South Africa from the beginning of the
19th century and off Namibia about a half-century later (Kinahan 1991, Grifths 2000). Although
snoek (Thyrsites atun) comprised the mainstay of this shery, warm-temperate species, chiey
croakers (Sciaenidae) and seabreams (Sparidae), have played substantial roles in warmer waters,
both north of Walvis Bay and east of Cape Point. While a total of 15 species are targeted by
commercial line sheries in the Benguela, this review focuses on the six most important (based
on annual catch). As with other line sheries, catch and effort data are unevenly distributed in
space and time. In order to accommodate this limitation, but also allow for the presentation of
data for separate stocks of the same species, the study area was divided into the following
regions: Southwestern Cape (Cape Point to Cape Agulhas), Western Cape (Cape Point to Orange
River), and Namibia. The Namibian coastline was not subdivided because commercial vessels
are based only at Swakopmund and Walvis Bay, operating between Meob Bay in the south and
Rocky Point in the north.
Fishing effort in the Southern Benguela rose dramatically during the 20th century, numbers of
commercial vessels increased 10-fold in the Southwestern Cape and 20-fold in the Western Cape.
This trend was not, however, reected in the northern part of the system. During the 1990s the
average numbers of active commercial vessels were 577 in the Southwestern Cape, 986 in the
Western Cape, and 10 in Namibia, equating to 1.2, 2.1, and a mere 0.006 vessels km–1 coastline,
respectively (Grifths 2000, Holtzhausen, personal communication).
Overall catches of the six main target species over three periods for which adequate data are
available are shown in Figure 18. Catches in the Western Cape, which are dominated by snoek,
have increased dramatically over the past century, whereas those on the Southwestern Cape, which
are made up of a more diverse mixture of species, have declined over the same period. The reasons
for this become more evident when examining the biology of the component species individually
(see below).
Snoek (Thyrsites atun)
Snoek is a medium-size, pelagic predator attaining a weight of 9 kg (Nepgen 1979) inhabiting the
coastal waters of the temperate Southern Hemisphere. It is found off southern Africa, Australia,
New Zealand, the east and west coasts of southern South America, Tristan da Cunha, and the islands
of Amsterdam and St. Paul (Nakamura & Parin 1993). Southern African snoek have been recorded
from northern Angola to Algoa Bay, but are mostly found between the Cunene River and Cape
Agulhas, i.e., in the Benguela ecosystem. This snake mackerel is a valuable commercial species
and an important predator of small pelagic shes. It apparently has two separate subpopulations
in the Benguela ecosystem, one off Namibia and the other off South Africa, with medium-term (c.
5 yr) exchange in response to environmental events and food availability (Grifths 2003). Taken
with handlines since the early 1800s, Thyrsites atun became a signicant by-catch of the hake-
directed trawl shery that developed during the 20th century. Initially discarded, snoek were retained
by trawlers only after 1972. Page 334 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 335
The catches of snoek can be analysed in a number of ways: in terms of total domestic and
foreign catch, capture location, or method of capture. Total annual catches from the Benguela
(Figure 19A) increased markedly from 5000–19,000 t in 1960–77 to 26,000–82,000 t in 1978–90
with the inux of foreign trawlers in 1978, and then dropped again to 12,000–25,000 t following
their withdrawal in 1991. The proportion of the catch made in Namibian waters (Northern Benguela)
rose commensurately from 10–20% to 40–80% during the period of foreign involvement, and then
declined to 5–9% after they were excluded (Figure 19B). Looking at the method of capture (Figure
20), handline catches of snoek in the Southern Benguela during the 20th century, despite high
interannual variation, demonstrate some noteworthy trends. Annual catches increased concomitantly
with the increasing effort between 1896 and 1950. Between 1950 and 1978 they uctuated between
2500–13,000 t, and between 1979 and 1999 they showed a steady increase from a low of 2000 t
to around 18,000 t. Factors such as the abolishment of a 5-month closed season (1960s–1981) in
1981 and increased targeting of snoek following the demise of other lineshes (Grifths 2000)
may have played a role in the latter. Annual trawled by-catch of snoek (Figure 20A) made by South
African vessels in the Southern Benguela also uctuated considerably, but increased generally from
around 1500 t in 1972 to 15,600 t in 1991. This peak was probably caused by the redirection of
Figure 18 Mean annual commercial handline catches during three periods of the 20th century of the six most
important lineshes of the Southern Benguela: carpenter (Argyrozona argyrozona), yellowtail (Seriola lalandi),
hottentot (Pac hymetopon blochii), silver kob (Argyrosomus inodorus), geelbek (Atracoscion aequidens), and
snoek (Thyrsites atun). (Constructed from data presented by Grifths 2002.)
1897–1908 1927–1931 1986–1998
Catch (tonnes)
1897–1908 1927–1931 1986–1998
Catch (tonnes)
Western Cape
Southwestern Cape Page 335 Wednesday, June 30, 2004 2:19 PM
336 C. L. Grifths et al.
excess trawl effort at snoek and horse mackerel when South African vessels were excluded from
Namibia (the South African hake catch is limited by a TAC and therefore could not accommodate
the extra effort). A steady decline in the annual trawl catch of snoek since 1991 is attributed to a
general increase in operational depth, associated with a recent focus on the production of prime
quality hake for overseas markets, rather than with declines in snoek abundance. Given the inu-
ences of effort levels and sher behaviour on the trawl and handline catches of snoek, it is impossible
to draw any conclusions on the long-term abundance of the resource. However, handline catch per
unit of effort (CPUE) during the 1990s (available for the periods 1897–1906, 1927–31, and
1986–98) was only 42% lower than the historical maximum (1927–31) (Table 1) suggesting that
the stock is currently not overexploited (Grifths 2000). Handline catches off Namibia, on the other
Figure 19 (A) Domestic and foreign snoek (Thyrsites atun) catches in the Benguela ecosystem. (B) The
division of annual catch between northern and southern subregions (1973–1987). Data were obtained from
International Commission for the Southeast Atlantic Fisheries statistical bulletins (ICSEAF 2 & 17) and FAO
Yearbooks of Fishery Statistics (FAO 46, 52, 70, 88/1). Catch for the (ICSEAF) division 1.5 (1973–1987) was
divided according to the ratio of Namibian and South African coastlines comprising the region. Because FAO
yearbooks do not report catch by subregion within the Benguela, country totals for the period 1988–90 were
assigned to the Northern and Southern Benguela based on previous shing patterns (i.e., Israeli and Portuguese
catches assigned to the south and all other foreign catches to the north); foreign involvement ceased after 1990.
Catch (tonnes)
Foreign catch
Domestic catch
Proportion of total catch
Southern Benguela
Northern Benguela
B Page 336 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 337
Figure 20 Annual (A) trawl and (B) handline catches of snoek made in South African and Namibian waters
by domestic vessels.
Table 1 Vital statistics available for linesh stocks of the Benguela ecosystem
Species Area % CPUE
assessment Reference
Snoek Namibia
South Africa 57.6
Grifths 2000
Silver kob Namibia
Southwestern Cape 9.2
B = 40%
SB/R = 8%
Grifths 1997b, 2000, Kirchner 2001
Geelbek Southwestern Cape 2.5 SB = 5% Grifths 2000, Hutton et al. 2001
Carpenter Southwestern Cape 3.2 ? Grifths 2000
Hottentot West Coast
Southwestern Cape
Grifths 2000
Yellowtail South Africa 59.8 SB = 45% Grifths 2000, Penney 2000
Note: Data include current catch per unit of effort expressed as a proportion of historical peak (% CPUE) and stock
assessment details. Remaining biomass (B), spawner biomass (SB), and spawner biomass per recruit ratios (SB/R) are
given as percentages of pristine under the stock assessment column.
Catch (tonnes)
Northern Benguela
Southern Benguela
Catch (tonnes)
Northern Benguela
Southern Benguela
B Page 337 Wednesday, June 30, 2004 2:19 PM
338 C. L. Grifths et al.
hand, dropped dramatically from 1500–4700 t (1950–83) to <1000 t in 1984, following the period
of high foreign catches in the region. Lack of any indication of recovery may well have been the
result of the persistent low levels of primary prey biomass, particularly clupeoids, since 1980 (Boyer
& Hampton 2001).
Silver kob (Argyrosomus inodorus)
Silver kob is a medium to large sciaenid attaining a weight of 37 kg (Grifths & Heemstra 1995).
Two stocks have been identied within the Benguela ecosystem: Meob Bay to Cape Frio (Kirchner
& Holtzhousen 2001) and Cape Point to Cape Agulhas (Grifths 1997a). Although annual yield
from the Southwestern Cape stock declined by 40% during the 20th century, CPUE dropped by at
least 90% (Table 1 and Figure 21). This evidence of severe overexploitation is further supported
by an estimated spawner biomass per recruit ratio of as little as 8% of pristine and an instantaneous
shing mortality rate threefold that of natural mortality (Grifths 1997b). Owing largely to the
lower commercial effort (above), and also to the fact that around 80% of the Namibian coastline
is inaccessible to shore-based recreational activity (due to mining operations), this silver kob stock
is considerably more healthy. Annual commercial handline catches of Namibian silver kob uctu-
ated between 78 and 2800 t between 1964 and 2000, depicting no clear trend. Remaining biomass
is estimated at around 9000 t, or 40% of pristine (Kirchner 2001). It should, however, be noted
that a commercial ski boat shery established in the Swakopmund area in the early 1980s has since
collapsed (Holtzhausen, personal communication), indicating a case of localised depletion. Recre-
ational shers catch approximately 30% of the estimated 1200-t annual catch (1995) of silver kob
in Namibia (Holtzhausen et al. 2001). Most of the annual catch from the Southwestern Cape stock
is taken by the commercial handline eet (90.8%), with the remainder divided between recreational
anglers (7%) and the beach-seine shery (3.8%).
Geelbek (Atractoscion aequidens)
The geelbek is a predatory sciaenid that attains a weight of 15 kg and feeds almost exclusively on
sardines and anchovy (Grifths and Hecht 1995). Once the most important linesh of the South-
western Cape (Figure 18), catch rates in this region declined by more than 95% during the 20th
century. Although geelbek caught in the Southwestern Cape form part of a single South African
stock, similar declines in CPUE throughout its range conrm the severe stock depletion (Grifths
2000). Stock assessment indicates that spawner biomass had declined by at least 95% in 1996 due
to overshing (Hutton et al. 2001).
Carpenter (Argyrozona argyrozona)
Endemic to the warm-temperate waters of South Africa, this sparid reef sh is the second most
important linesh of the Southwestern Cape (Figure 18). Recent tagging studies suggest that
carpenter of the western and central Agulhas Bank comprise a single stock (Grifths and Wilke
2002). Despite its current importance, catch rates of carpenter on traditional linesh grounds of
the Southwestern Cape and Southern Cape (Cape Agulhas to Plettenberg Bay) have declined by
as much as 97 and 95%, respectively (Table 1; Grifths 2000). Stock assessment indicates that the
reproductive potential (egg per recruit) of the western carpenter stock has been reduced to between
6 and 14% of pristine levels (Brouwer and Grifths, unpublished data). Page 338 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 339
Hottentot (Pachymetopon blochii)
Hottentot is an omnivorous cool-temperate sparid attaining a weight of 3 kg and inhabiting kelp
forests and subtidal reefs to a depth of about 40 m. It is most abundant between Cape Agulhas and
Luderitz in southern Namibia. Although targeted when larger and more lucrative species are less
abundant, catch rates of this resident endemic sh have declined by 77 and 62% in the Southwestern
Cape and Western Cape, respectively (Grifths 2000). A quantitative assessment of this resource
in 1985 (Punt et al. 1996) indicated that, with the exception of Gans Bay, hottentot were not
overexploited. However, the growth rate used in this analysis (Pulfrich & Grifths 1988) was
overestimated, as it was based on whole (vs. sectioned) otoliths (maximum age being underesti-
mated by 30%), and further assessment is required.
Yellowtail (Seriola lalandi)
This circumglobally distributed carangid, which attains a maximum weight of 34.2 kg, comprises a
number of disjunct populations. The southern African population occurs along the entire South
African seaboard and is found from the shore to the edge of the continental shelf. It is, however,
Figure 21 Commercial handline catch and catch per unit of effort for silver kob (Argyrosomus inodorus)
stocks in (A) Namibia and (B) the Southwestern Cape.
Catch (tonnes)
Catch (tonnes)
CPUE (kg/boat/year)
Catch CPUE
B Page 339 Wednesday, June 30, 2004 2:19 PM
340 C. L. Grifths et al.
most abundant in the area extending from the central Agulhas Bank (within the 200-m contour) to
the central Western Cape. Tagging studies reveal that sh move freely within this range and that
they comprise a single stock (Wilke & Grifths 1999). Although a traditional linesh in South Africa,
purse-seining for yellowtail was introduced in 1972. Commercial handline catch rates around Cape
Agulhas dropped by two orders of magnitude by the early 1980s, but then recovered after netting
was banned in 1982 (Penney 2000). Stock assessment (virtual population analysis (VPA) with
extended survivor analyses) indicates that the stock is presently optimally exploited (Penney 2000).
Overview of linesh resources
Stock assessments and trends in CPUE reveal that, with the exception of snoek and yellowtail,
other linesh species targeted in the Southern Benguela were heavily overexploited during the last
century. Factors contributing to their demise have included unchecked commercial effort and
biological factors such as predictable locality (coastal migrants and resident reef shes), longevity
(12–25 yr), and late maturity (7–55% of maximum age) (Grifths 2000). Plans to rebuild South
African linesh stocks include drastic reductions in commercial effort and the introduction of more
stringent bag limits for recreational anglers.
Apart from the loss of long-term yield, overshing of teleost predators, particularly in the
Southwestern Cape, is thought to have had other effects on the Benguela ecosystem. Sardine
Sardinops sagax and anchovy Engraulis encrasicolus are important food of pelagic and several
reef-associated lineshes. These two prey species spawn on the Agulhas Bank and utilise the
Benguela current to transport eggs and larvae onto nutrient-rich west coast nursery areas (see
pelagic sh section, p. 323). Thereafter, the juvenile sardine and anchovy return to the Agulhas
Bank, where they are spawned, thereby biologically pumping energy/carbon from the highly
productive west coast upwelling system onto the eastern seaboard. This process presumably allows
warm-temperate reef ecosystems to support larger shoals of piscivores than the reefs themselves
could sustain. It is therefore no accident that the most important reef-associated species have been
those that feed on clupeoids, i.e., carpenter, silver kob, and geelbek. The extent to which these
lineshes may have fertilised reefs (nitrogen excretion and faecal pellets) is unclear, but it is possible
that their demise has concomitantly reduced the links of reef ecosystems with the pelagic food
web. Ecosystem modelling is regarded as imperative if the impact of overshing of teleost predators
is to be fully understood.
Rock lobsters
Historically, the commercial shery for west coast rock lobster, Jasus lalandii, has extended from
Sylvia Hill, north of Luderitz, to Cape Hangklip in False Bay (Figure 1). The species occurs over
a wide depth range, from intertidal rock pools to a few hundred metres depth, the depth of occurrence
extending deeper in the southern than in the central and northern parts of the range. However, there
is some uncertainty as to whether these differences have always been as extreme as they are today.
Because of their shallow habitat, rock lobsters have long been targeted by coastal dwellers.
Archaeological records show that they formed a component of the diet of early indigenous inhab-
itants on the west coast (Grindley 1967, Parkington 1976; see above, p. 307). Exploitation of
lobsters for personal use and sale continued after colonisation, especially among the poor in the
community; indeed, an active shery for rock lobsters by licenced recreational shers persists to
this day.
The rst commercial processing plant was established in South Africa in 1875 to can rock
lobster tails, but teething problems got in the way of the new industry and exports only developed
continuity in the 1890s. At that stage, and for the next 70 yr, before traps were introduced into the
shery in the 1960s, the entire west coast rock lobster catch was made using baited hoop nets Page 340 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 341
(otherwise known as ring nets or drop nets), mostly shed from wooden dinghies, with one person
rowing and another setting and hauling the nets. Between 1890 and the early 1920s the industry
expanded with the building of rock lobster canning factories in other towns along the west coast,
as far north as Luderitz. Expansion was particularly rapid after the First World War (Figure 22),
with landings in both the South African and Namibian sheries showing rapid growth to the start
of the Second World War, at which point exports declined. This downturn was short-lived, and the
industry soon took off with renewed vigor to reach peak landings of over 20,000 t yr–1 in the early
and mid-1950s (Figure 22). Since the mid-1950s, rock lobster production in both countries has
declined to a small fraction of its peak. The downward spiral has tended to be by a series of steps,
which in most cases has been due to adjustments to production/TAC quotas and minimum sizes.
These need to treated separately for the two countries.
A production quota was rst imposed on the South African west coast rock lobster shery in
1946. Landings for the shery derived from export production, and in later years, landed mass
gures show that the quota did little to constrain catches between 1946 and the mid-1950s. Very
large catches were recorded between these years, particularly in 1950 and 1951, when they exceeded
16,000 t (Melville-Smith & van Sittert, in press).
Industry production quotas were introduced in 1946. In 1970 these were superseded by indi-
vidual company quotas and, from 1980 onward, by regional TACs for eight regional shing zones
(Pollock 1986). Periodic, generally downward adjustments to quotas and TACs have been a feature
of the shery since the 1950s, as it became clear that the high landings of earlier years were not
sustainable. The shery experienced signicant changes in the 1990s resulting from a sudden
decrease in growth increments that led to poor recruitment into the shery. The low catches of
legal-size animals led to radically revised management arrangements for the shery in the early
1990s. Since 1993–94 the minimum size has been xed at 75 mm carapace length (CL), which is
14 mm smaller than the 89-mm minimum size rst introduced 60 yr earlier (Figure 23). The TAC
has also been decreased to around half of what it was in the 1980s.
From its inception in the early 1920s, the Namibian shery has been subject to management
restrictions of one form or another. The rst legislation governing rock lobster shing in Namibia
Figure 22 Annual commercial landings of west coast rock lobster (Jasus lalandii) in the Benguela ecoregion,
1890–2001. RSA, Republic of South Africa.
Catch (tonnes)
RSA Namibia Page 341 Wednesday, June 30, 2004 2:19 PM
342 C. L. Grifths et al.
was introduced in 1922 in the form of a legal minimum size of 4 inches (101.6 mm) and a closed
shing season. Both the minimum size limit and shing season were changed the following year,
but the fact remains that conservation measures were recognised and imposed early on in the shery.
The proclamation of sanctuary areas in 1924, legislation limiting the number of factories licenced
for processing lobsters, and production quotas in 1940 followed. Unlike in South Africa, where
quotas limited landings, in Namibia the minimum size limit and number of processing factories
appear to have had the most signicant effects on historical landings, rather than production quotas
(which, as shown by Beyers & Wilke 1990, were seldom attained).
The increase in landings through the early and mid-1950s followed an extra factory coming
onstream in 1949, and several changes to the legal minimum size in the early 1950s (Figure 23).
These included a period in the 1951 and 1952 seasons during which shers were required to land
all animals caught, the sublegal-size ones being sold on the local market, in effect eliminating the
legal size limit (Beyers & Wilke 1990).
Further increases in landings in the mid- to late 1960s followed the granting of another factory
licence in 1964, thereby increasing shing effort, and a reduction in the minimum size. When
catches declined further, the minimum legal size limit was abolished altogether for 1968 and 1969
(Beyers & Wilke 1990). The marketing of small tails proved problematic and further declining
catches led to the implementation of a new size limit and a big reduction in the production quota
in 1970. Despite these new arrangements, landings remained depressed compared to earlier years
and slowly declined to the beginning of the 1990s when, over the space of two seasons, catches
plummeted (Figure 22) as a result of very low catch rates. Since the 1992–93 season catches have
been constrained by small TACs of under 300 t.
Over the years there has always been an unquantiable portion of the landed catch that has
gone unreported. These unreported landings, which are probably a signicant portion of the reported
landings, have taken diverse forms:
Figure 23 Legal minimum sizes for west coast rock lobsters (Jasus lalandii) in Namibia and South Africa,
Minimum size (mm CL)
South Africa Namibia Page 342 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 343
1. Rock lobster caught for “shermen’s bait” was exempt from regulations imposed on
catches made for export until 1920.
2. Canned lobster, which dominated landing gures up to the 1950s, was (particularly in
the early years) prone to high product rejection rates due to swelled or blown cans (Von
Bonde & Marchand 1935).
3. Illegal shing has been an important informal supplier of west coast rock lobster to
domestic markets.
4. Illegal catches have traditionally been used by unscrupulous inshore shers for bait or
consumption at sea.
5. There have been eras when, at least in certain regions of the shery, regulations were
inadequately enforced (Pollock 1982).
Since 1983 it has been necessary for South African recreational rock lobster shers to purchase
licences, and this database has been used to conduct telephone surveys of recreational shers
(Cockcroft & Mackenzie 1997). Since the introduction of these surveys in 1991 to the 1995–96
season, the recreational catch in South Africa has been around 400 t season–1. That in Namibia was
at least 9 t in 1993–94 (Melville-Smith et al. 2000). The lack of continuity in these estimates has
led them to be excluded from Figure 22.
Impact of shing on the stock
West coast rock lobsters have a larval life of about 9 months (Silberbauer 1971), during which the
larvae are found as far as 300–460 km offshore (Lazarus 1967, Pollock 1986). This lengthy life
cycle provides ample time for mixing of the larval pool, and it is therefore reasonable to assume
that Namibian and South African west coast rock lobster form a single stock.
The rise and fall of rock lobster landings in the two countries are remarkably similar (Figure
22). Both show rapid growth between the two World Wars. The decrease during the Second World
War was followed by even larger landings, which peaked in the early 1950s. There is little doubt
that by this time the shery was being grossly overexploited, with catches in both South Africa
(Melville-Smith & van Sittert, in press) and Namibia (Beyers & Wilke 1990) well over the respective
quotas, and efforts to maintain production in Namibia were desperate enough to reduce the minimum
size (Figure 23). Although CPUE in the central grounds of the South African west coast rock lobster
shery was stable between the mid-1950s to mid-1960s, catch rates had fallen from much higher
levels in the early 1950s and effort had increased (Heydorn et al. 1968). A particularly dramatic
decline in catches occurred in both Namibian and South African sheries between the mid-1960s
and early 1970s, although in the case of the Namibian shery this decline was temporally masked
by the abolition of a minimum size in 1968 and 1969.
There are important environmental differences between the northern (north of the Olifants
River) and southern parts of the west coast rock lobster shery. These differences are mainly due
to the presence of oxygen-decient bottom water, which at times extends close inshore, in the more
northerly regions (Pollock & Shannon 1987, Grobler & Noli-Peard 1997). This results in a tendency
for Jasus lalandii to increasingly be conned to shallower waters as one moves northward (Pollock
& Beyers 1981). Pollock & Shannon (1987) have hypothesised that the lower production of the
northern regions of the shery between 1960 and 1970 was the result of environmental change
over this period, resulting in a progressive expansion of oxygen-decient shelf water. They suggest
that this forced the animals into shallower water, where reduced habitat and consequent competition
for food resulted in lower growth rates and higher natural mortality, which owed through to the
shery as reduced surplus production.
Traps were introduced in the South African and Namibian sheries in the 1960s. The intro-
duction of this new technology affected the shery in a number of ways. Because traps were hauled
with mechanical winches, they opened up deepwater shing grounds unavailable to hoop net shing. Page 343 Wednesday, June 30, 2004 2:19 PM
344 C. L. Grifths et al.
Also, because trap shing gear is unattended, it permitted shing at night and in weather conditions
unsuitable for operating dinghies. Finally, the large volumes of lobsters taken in each trap made
sorting the legal- from undersize catch difcult and probably resulted in large-scale discard mor-
tality. Thus it would seem likely that this technology helped to maintain landings at unsustainable
levels and led to a false impression of the availability of lobsters in ensuing years.
After the declines in landings during the late 1960s and to a lesser extent the 1970s, the 1980s
were a period of relative stability in both the South African and Namibian sheries, with seasonal
landings of around 3500–4000 and 1000–1900 t, respectively. Both countries experienced a major
decline in CPUE and landings in the 1989–90 and subsequent seasons. These changes have been
direct and indirect results of a reduction in the annual growth increment (Melville-Smith et al.
1995, Goosen & Cockcroft 1995, Cockcroft 1997, Pollock et al. 1997). The 1990s have also been
a period in which there have been more mass strandings of lobsters caused by hypoxic conditions
associated with high biomass dinoagellate blooms (Cockcroft 2001) than in any previous decade.
Of particular concern has been the loss in the 1990s of 5–12% of the Jasus lalandii spawner biomass
as a result of mass stranding mortalities (Cockcroft 2001).
There is no question that the size structures of the west coast rock lobster populations in South
Africa and Namibia today bear no resemblance to those of the pristine stock. It is clear from the
minimum sizes that were in place (Figure 23) and the quantities of lobsters that were being landed
each season (Figure 22) that either there must have been a very large initial biomass of animals
(>89 mm CL), which was depleted over time, or productivity in earlier years must have been
considerably higher than in the modern era. Historical size composition data are scarce, but where
available (e.g., Gilchrist 1913, 1914b, Anonymous 1935, 1939) they conrm that animals of the
modal size classes of prewar years were of a size seldom ever recorded in modern catches.
Pollock et al. (2000) gives the harvestable biomass (i.e., biomass of animals of >75 mm CL)
for the South African shery as ~5% of pre-exploitation levels, and the spawning biomass as ~20%
of pristine levels. These results, backed by model results, suggest that the depressed state of the
spawning biomass may have resulted in a decline in recruitment in recent decades, and in particular
since the mid-1980s, when the biomass underwent substantial change. With this fact in mind, a
stock rebuilding strategy has been developed for the South African shery in recent years (Pollock
et al. 2000). The Namibian shery is being managed by setting conservative TACs and has the
stated long-term objective of rebuilding stock levels to those of the 1980s (Grobler & Noli-Peard
Wider ecological effects of Jasus lalandii exploitation
West coast rock lobsters are known to feed on a wide range of food items, including molluscs,
polychaetes, shes, sea urchins, other crustaceans (including their own species), algae and sponges.
They thus play an important role in inuencing the structure of shallow subtidal benthos on hard
substrata in some (Branch & Grifths 1988), but not all, rock lobster grounds (Tomalin 1993). The
depleted stocks, and in particular removal of large size classes, which are able to consume prey
that is not available to small lobsters (Pollock 1979, Grifths & Seiderer 1980), have presumably
had ecological consequences to benthic communities on the lobster grounds. However, the extent
and implications of these changes remain largely unknown.
In contrast to the depleted state of the rock lobster stock on the west coast, there has been a
substantial increase in the abundance of west coast rock lobsters on the South African southeast
coast since the late 1980s (Tarr et al. 1996, Mayeld & Branch 2000). These increased densities
of lobster, which lie outside of commercial shing grounds, are believed to have depleted sea urchin
populations and, in so doing, have disturbed the ecological balance between sea urchins and juvenile
abalone, thereby contributing to the decline in the abalone shery (see next section).
In summary, there is little doubt that whereas the primary reason for the present state of both
the Namibian and South African rock lobster sheries is overexploitation, there have been multiple Page 344 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 345
factors contributing to this situation. While most of these factors have an obvious link to human
impacts, for example, advancements in gear technology or illegal catches by the nonindustrial
sector, there are others, such as the hypotheses of large-scale environmental change and the
incompletely explained reasons for changes in lobster growth rates, for which there may be little
or no relationship with human activity.
Abalone, Haliotis midae, known locally as perlemoen, have been utilised by coastal shers for at
least 125,000 yr, as indicated by Middle Stone Age cave deposits and until recently, the species
remained readily accessible to shore pickers in the intertidal area. There are a number of other,
smaller abalone species in southern Africa, but only Haliotis midae has been the target of com-
mercial shing.
The present-day shery covers the coastline between Quoin Point and Cape Columbine, with
the most productive region being from Cape Hangklip to Quoin Point. Fishing is carried out by
licenced commercial divers who operate from twin-engined ski boats launched from the shore.
Except for the introduction of satellite navigation and more seaworthy shing boats, the method
of shing is essentially the same as it has been since the inception of the commercial shery in
the 1950s. A “hookah” system is used, consisting of an onboard compressor that supplies air to
the diver through a hose. The minimum size limit of 114-mm shell breadth has been unchanged
since 1955. Being collected by hand, the harvest method has no secondary impact on the kelp
forest ecosystem.
Records of commercial catches are available from 1953 (Figure 24) and were calculated from
production gures (amount marketed, canned, or frozen annually). Initially, catches were around
500 t yr–1 and uctuated due to varying market demand. However, after 1960, catches escalated
dramatically and, in the absence of any limiting quota, reached a peak of around 2800 t in 1965.
During this period shing was essentially a mining operation, removing pristine accumulated beds
of abalone, where up to 20,000 abalone were taken from a single bed (Newman 1965). These
aggregations occurred at densities of 15–20 abalone m–2, permitting individual catches as high as
3000 abalone day–1. As these accumulations were mined out, catches, as well as catch rates, declined
annually from 1965–70, and although production quotas were imposed for the rst time in 1968
and 1969, they were not lled. Only in 1970 did the production quota rst limit catches. Thereafter,
catches have been quota limited, and the stabilising effect of this measure is apparent in Figure 24.
From 1983 the quota system was changed to one based on the whole mass of abalone delivered
to the factories, because investigations showed that the production quota was being misused, in
that abalone were being cut up into pieces and not declared as quota. It is likely that the quota was
exceeded by at least 15% during the 1970s and up to 1982. Area-bound, annually revised TACs
were introduced in 1985. From 1992–3 catches were sealed at the slipway, before transport to the
factories, as there were allegations that abalone were being landed but not delivered to the factories.
In a further renement from 2000 the weight of the catch was recorded using mobile scales, directly
at the landing sites.
Participation in the shery has changed over the years, initially being free access to all, with
around 100 divers active by the early 1960s. These divers delivered on a catch-as-catch-can basis
to the four or ve licenced abalone processing factories. These factories were allocated a xed
percentage of the quota from 1968. In an effort to limit the shery to bona de shers, diver
numbers were reduced gradually to around 50 by 1989. From 1984 the divers became legally
obligated to deliver to specic factories and were also each allocated a xed percentage of the
quota, the value of which was derived from past performance. This placed divers in a position of
greater nancial security, giving them a stake in the shery that was saleable and heritable. From
1998–99, the balance of power, which had previously rested with the factories, which “owned” the
quota, was reallocated when divers and quota holders were amalgamated into one group called Page 345 Wednesday, June 30, 2004 2:19 PM
346 C. L. Grifths et al.
right holders. These right holders could harvest or market their abalone on an equal basis, subject
to certain legal constraints.
A new subsistence sher category was created in 1997 and continued for 3 yr. This sector,
however, proved virtually unmanageable due to the shore-based snorkel shing method permitted,
and the associated lack of controls, with a consequence that the amount of abalone landed for the
rst two seasons is unknown. In the third year, the shing method was changed to a boat-based
operation with individual allocations of 300 kg, enabling better control to be exercised over the
landings and the catch for the season to be accurately recorded. This sector was then converted to
a small-scale type of shery called Limited Commercial, from the 2001–2 season, which follows
normal commercial shing practices.
Recreational shing for abalone is also an important component of the shery dynamics. The
number of recreational shers is not limited, but they must purchase an annual permit that limits
the daily bag limit (presently) to three abalone, and they may operate from the shore only, using
snorkel equipment. Monitoring of recreational shing catches started in the late 1980s, with catches
increasing to equal 89% of the commercial catch in 1993–94. A series of management measures
were imposed over the years to reduce recreational catch. The most effective of these was a reduction
of the length of the shing season, from the original 270 days to only 11 days (specied weekend
days) for the 2001–2 season. Nonetheless, this only resulted in a reduction of recreational take to
28% of the combined commercial sectors for 2001–2, due to greater recreational shing effort
being applied during the very limited season.
Status of the shery
The abalone resource is presently facing a severe crisis. This is a result of two developments,
poaching and ecological change, both of which coincidentally affected the resource from around
Poaching has always been a factor; however, since around 1994 it has become increasingly
important, to the extent that recent data indicate that the shery is unlikely to remain sustainable
unless major improvements in compliance occur. Conscation records obtained from the various
branches of compliance services have been collected since 1994, and these reect a great increase
in activity over the last 3 yr (Figure 25), a trend that is conrmed by anecdotal information from
a variety of sources. Some of the recent increases in conscations may be attributable to increased
effectiveness of compliance staff, but this increased efciency is not believed to be equivalent to
more than a factor of approximately 25–30%. This is because as compliance efciency improves,
Figure 24 Commercial catches of abalone from 1953 showing the stabilising effect of quotas from 1970 and
the declines evident from 1996.
Tonnes (10 )
Commercial Subsistence
3 Page 346 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 347
so do the evasive skills of the highly organised gangs involved. However, there is no doubt that
the damage caused to the resource by poaching is serious. The volume of abalone removed illegally
must be considerable, with most estimates assuming that conscations represent only around 10%
of actual illegal catch. Around 55% of the abalone conscated are below minimum legal size. Given
that maturity is reached about 1 yr before attaining the shery size limit, poaching is removing
large numbers of abalone from the population before they have an opportunity to reproduce.
In addition to the problems caused by overshing, a dramatic environmental change has recently
occurred in the centre of the historically most productive region of the abalone population, Cape
Hangklip to Hermanus, and this has severely curtailed the reproductive effectiveness of the popu-
lation. The change became noticeable from around 1994, when a large-scale incursion of rock
lobsters Jasus lalandii into the area rst became apparent. The rock lobsters, which previously
occurred in low densities in this area, have consumed the majority of small invertebrates in this
area, such as sea urchins Parechinus angulosus and gastropod molluscs. The disappearance of the
sea urchin population in particular has negatively affected the reproductive efciency of the abalone,
since juvenile abalone derive important shelter from predators under the spine canopy of the sea
urchins. In the absence of urchins, abalone are very vulnerable to predation, including that from
lobsters. As a result, recruitment of abalone in this area is now estimated to be only 10% or less
of normal. For this reason, any signicant removal of abalone from the area between Cape Hangklip
and Hermanus must now be seen as an essentially unsustainable practice.
The natural tendency of pristine Haliotis midae populations is to develop into high-density
aggregations of adults in the shallow inshore zone. This adaptation facilitates reproductive efciency
in these broadcast spawners. Furthermore, abalone have developed an efcient drift-weed trapping
feeding behaviour, which is essential for effective herbivory in such high-density, essentially
nonmobile populations. Fishing, poaching, and ecosystem changes have resulted in severe declines
in density. Extensive diving surveys in the 1980s showed that those areas that supported large
populations of abalone demonstrated average densities in the shallow inshore area of around 0.8–1.3
abalone m–2. These densities are similar to the those recorded in shery-independent abalone surveys
(FIAS) of the Betty’s Bay marine reserve, initiated in 1995, which averaged around 1.2 abalone
m–2 from 1995–98. In contrast, as a result primarily of poaching, the average density in three of
the four main abalone commercial grounds has now declined to below 0.3 abalone m–2.
It is not known whether these declines in abundance have reached the point where nearest-
neighbour distances are such that fertilisation efciency of the population is negatively affected.
Recruitment studies carried out in the late 1990s indicated that abundance of juvenile abalone less
Figure 25 Records of numbers of abalone conscated from 1994. Data for 2002 represent conscations for
the rst 4.5 months only.
Number (10 )
3 Page 347 Wednesday, June 30, 2004 2:19 PM
348 C. L. Grifths et al.
than 2 yr old was not noticeably different in commercially shed grounds compared to refugia.
However, at present, in certain areas that have been heavily denuded through poaching, it is likely
that recruitment success has been compromised, with further negative implications for this already
threatened stock.
Freshwater inows and estuaries
Rivers and their associated estuarine systems impact on the marine environment of the Benguela
in a number of important ways, most notably as nursery areas for marine shes, as habitat for
wading birds, and as sources of sediment and nutrients and other materials (including pollutants)
entering coastal seas. As an arid to hyperarid region the Benguela coastline supports very few
estuaries. Indeed, no perennial rivers reach the sea within the entire Namibian coastline, although
the Cunene and Orange (Gariep) Rivers form its northern and southern borders, respectively. The
only major river systems within the entire Benguela region are thus the Orange, Olifants, and Berg
Rivers. All have been highly manipulated and, indeed, the attitude of 20th-century South African
engineers to rivers may be summed up by the following quotation:
There can be hardly a single true South African, and certainly no irrigation engineer, with soul so dead
that he can contemplate our greatest river tearing down to the ocean through a vast area of country,
which is thirsting for water, without feeling that some great effort should be made to design and carry
out irrigation works for the Orange River, which would rival those famous works of other great rivers
of the world.
— Dr. A.D. Lewis, Director of Irrigation, October 12, 1928, Graaf-Reinet
This anthropocentric view of rivers, that water running down them to the sea is wasted, has
led to overmanipulation for water supply of all west- and southwest-owing rivers of the subcon-
tinent. As a result, rivers that historically contributed to estuarine and inshore coastal processes,
especially the Orange-Vaal, the Cape Olifants, the Great Berg, the Palmiet and the Breede (Figure
1), either no longer ow or ow at greatly reduced volumes and in a greatly regulated manner, due
to overabstraction and the construction of water supply reservoirs and interbasin water transfer
schemes (IBTs). The situation has become so serious that a number of recent Global International
Water Assessment (GIWA) workshops have identied human manipulation of water resources and
overutilisation as the single most important threat to the ecological functioning of the region
(Prochazka et al. 2001, Davies 2002).
This part of the review focuses on two of these catchments, the Orange-Vaal and the Great
Berg. However, these systems epitomise the severity of anthropogenic ow modication wrought
on other systems throughout the region.
The Orange-Vaal system
The Orange-Vaal is the largest watershed in southern Africa south of the Zambezi (17˚S), draining
approximately 47% of South Africa (South African Department of Water Affairs (DWA) 1986),
providing between 5.5 ¥ 109 m3 (Keulder 1979) and 11.9 ¥ 109 m3 of water yr–1 (Cambray et al.
1986) and contributing 22.1% of the total mean annual runoff (MAR) from South Africa (Noble
and Hemens 1978).
Twenty-three major dams have been constructed within the Orange-Vaal catchment since 1872
(Benade 1988), and it contains nine of South Africa’s 30 largest dams. Nine of the 23 major dams
were completed in the 1970s, while only one, Katse (187 m), has been constructed since 1978.
Together with Mohale Dam (under construction), this dam forms Phase 1A of the massive Lesotho Page 348 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 349
Highlands Water Project (LHWP), an IBT transferring water from the headwaters of the Orange
in Lesotho to the Vaal for consumption in Gauteng (Davies & Day 1998).
The ow regulation of the Orange River can be traced back to 1877, the start of irrigation
schemes in the Great Fish and Sundays Valleys (Eastern Cape), as well as those in the middle
Orange River. Irrigation in the middle Orange seems to have begun c. 1883, although according
to Alexander (1974), surveying for Boegoeberg Dam had already started in 1872. Boegoeberg was
eventually built in 1931, but in the intervening period over a dozen private weirs had already been
built in the main river channel.
With the advent of the Union of South Africa in 1910, non-conservation orientated agricul-
tural activities in the Great Fish and Sundays Valleys led to large-scale soil erosion, ash
ooding, and siltation, and attention turned toward the diversion of water from the Orange River
to the Eastern Cape (the Orange River Development Project (ORDP); Alexander 1974). With
the announcement of the ORDP in the early 1960s, many scientists remarked that it was
inappropriate to develop dams on the Orange without understanding the basic biological pro-
cesses in the system. As the ORDP continued, calls for the urgent prioritisation of baseline
ecological data were made, notably by Chutter (1973). Similar calls were made in 1988, with
the advent of the LHWP (Petitjean and Davies 1988), but both calls were ignored until the late
1990s. Rapid population increase in Gauteng, as well as the development of the Orange Free
State gold elds and the Vaalhartz Irrigation Scheme, negatively affected the Vaal and thus also
the middle and lower Orange, between the late 1940s and early 1950s (Alexander 1974). With
emphasis switching from irrigation supply to industrial and domestic potable use, the single-
purpose diversion of Orange River water to the Great Fish and Sundays systems grew into a
multipurpose scheme, comprising three reservoirs, the Gariep, van der Kloof, and Torquay dams;
hydroelectric power stations; irrigation systems; and the 83-km-long Orange-Fish Tunnel IBT
(Benade 1988).
Presently the Vaal catchment contains 16 of the 23 major storages of the Orange-Vaal Basin.
Their existence has led to an extreme cumulative case of overabstraction and river regulation (Figure
26 and Figure 27). Indeed, the Orange-Vaal system is probably the most regulated river system in
Africa (e.g., Cambray et al. 1986) and one of the top three overregulated systems in the world,
ranking with the Colorado (U.S.) and the Murray-Darling (Australia).
The Orange per se had a more reliable preregulation runoff than the Vaal, but with pronounced
seasonal variation and widely uctuating interannual variation; erratic high ows transported silt
and adsorbed nutrients to the coastal zone (Cambray et al. 1986; Figure 26 and Figure 27). For
instance, postriver regulation, Bremner et al. (1990) reported that the oods of 1998 delivered 24.3
km3 of water and 64.2 ¥ 106 t of suspended sediments to the sea through the mouth at Alexander
Bay. The same paper noted the source of material as bed scour and bank erosion below the major
dams, as well as erosion of the estuary mouth. The resulting sediment transport led to the formation
of a delta extending 1.2 km offshore with a mass of approximately 3.6 million t. Interestingly, prior
to ow regulation by dams, sediments of the Orange were derived from the Middle Orange
Interannual ow variation has also been exceptionally large. The maximum Orange River ood
probably reached 31,200 m3 s–1, although the largest recorded ood occurred at Hopetown in 1874
(11,330 m3 s–1). Palmer (1996) reports that “prior to the building of dams in the [1970s], the [Orange]
river often ceased owing in winter [1862–3, 1903, 1912, 1933, and 1949] … and was reduced to
isolated pools.” At the coast, the estuary received aperiodic, large discharges, resulting in limited
tidal exchange and irregular scouring, such that permanent sediment deposition was reduced. The
mouth remained fresh for several months every year and according to Brown (1959), experienced
extreme seasonal ooding, high silt loads and turbidity. In this context, Branch et al. (1990) report
that the postregulation oods of 1988 led to abnormal dilution of coastal waters, with major effects
on intertidal and shallow subtidal organisms. Mass mortalities of patellid limpets were recorded
along the shore within 10 km of the mouth. The loss of these grazers led to dense growths of Ulva Page 349 Wednesday, June 30, 2004 2:19 PM
350 C. L. Grifths et al.
Figure 26 Natural and regulated ow regimes of the Orange/Gariep River, 1978–1980.
Figure 27 Seasonal variations in the monthly runoff at the Orange/Gariep River mouth under natural ow
conditions (solid line, average ow for 1935–69) and after construction of the Gariep and van der Kloof Dams
(dotted line, average for 1971–83). (After Cambray et al. 1986.)
Flow m × 103
Natural Regulated
Monthly Maximum Daily Average Flow Rate (m3s1)
Average 1935/36 to 1968/69 Average 1971/72—1982/83
JFMAMJJASOND Page 350 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 351
and Enteromorpha in mid- to low-shore zones and Porphyra along the high shore. Clearly similar
temporary shifts must have occurred along the intertidal and subtidal zones during preregulation
high-ow events discharged by the historically unmodied Orange.
Lake Gariep, completed in 1970, is the main regulator, providing ood control, hydroelectric
power and acting as a silt trap (Kriel 1972, 1978). The modications of ow (Figure 26 and Figure
27) comprise changes in the (1) annual runoff, (2) interannual variation in runoff, and (3) the
marked seasonality of the preregulation ow regime. Most oods from the catchment above Gariep
are contained by Gariep and van der Kloof, completed in 1978, cutting the maximum ood (31,000
m3 s–1) by 65% (Kriel 1972). The lower van der Kloof Reservoir supplies a stable, continuous water
supply for irrigation farmers and the regime is in direct conict with the variable demand required
by the aquatic environment in as much as estuarine, mouth, and inshore coastal characteristics are
affected. Suspensoid loads in the Orange River are very high and constitute 0.46% of the ow
volume. In perspective, a ood peak of 8500 m3 s–1 in 1967 was estimated to carry some 250,000
t ha–1 of silt (South African Department of Information 1971), while the average annual inow of
silt to Lake Gariep is c. 32 ¥ 106 m3 yr–1 (Kriel 1972). This silt is now mainly trapped by Gariep.
The total natural, or virgin, annual river ow at the mouth is estimated to have been some
10.670 ¥ 109 m3 (Prins 1990). In 1991 the ow was measured at 5.340 ¥ 109 m3. This reduction
may have serious implications for the rich avifauna (e.g., Frost & Johnson 1977, Siegfried &
Johnson 1977), where sandbars provide ideal roosting and nesting sites for pelicans, cormorants,
gulls, and terns. Courtenay-Latimer (1963) listed 172 species from Holgat to the mouth; however,
the inuence of the regulated river on the avifauna is complex. For instance, previous oods would
have destroyed nests and inundated roosting sites (e.g., Morant 1990, reporting on the oods of
1988), while reduced seasonal ooding and altered drought conditions may make the mouth more
stable for aquatic birds. On the other hand, permanent loss of silt and nutrients can only be
detrimental to the system and certainly reduced runoff contributes to salinity problems due to an
effective increase in concentration of available salt in the smaller volumes of water.
As a direct result of the South African National Water Act (no. 34 of 1998), environmental
ows or in-stream ow requirements (IFRs) from dams are mandated, but although an IFR for the
Orange River has been calculated (Venter & van Veelen 1996), it has yet to be implemented and
little if any cognisance has been taken of coastal and inshore marine requirements, particularly
sediment and nutrient discharges through the estuary.
The Great Berg River
The Great Berg (Berg) River rises on the northern slopes of the Franschoek Mountains in the
Western Cape near Cape Town. Flowing north and west, it forms a complex oodplain and estuarine
system before entering the Atlantic at St. Helena Bay. Like the Orange, it is allogenic, owing
through progressively more arid areas toward the sea where the MAP is <360 mm yr–1 (Ninham
Shand Inc. 1992). Extensive viticulture, grain and stock agriculture are practiced, with water demand
reaching a peak in the low-ow summer months. Historically the ow of the Berg (1928–88; Berg
1993) has been reduced such that ood-ow return periods of >75% are now the norm. The virgin
MAR to the mouth was 903 ¥ 106 m3 yr–1, while present-day annual runoff is 23–30% lower at
693 ¥ 106 m3 yr–1 (Berg 1993; Figure 28).
Irrigation demands and the growing Cape Town metropolitan region have led to construction
of a number of dams and IBTs. Only 1.6% of Cape Town’s water supply is generated within its
boundaries. Present water developments include the Riviersondereind (a tributary of the Breede)-
Berg River (RSB) IBT (Snaddon & Davies 1998, 1999, Snaddon et al. 1998, 1999). Irrigation
releases increase upper river ows in summer by between 560 and 4000% and deliver 15 ¥ 106 m3
yr–1. This, coupled to overabstraction within the Breede catchment itself, has led to signicant
reductions in ow to the estuary of the Breede, such that changes in the ood tide delta since 1942
have annihilated the once dense Zostera beds at the mouth (de Villiers 1988). Page 351 Wednesday, June 30, 2004 2:19 PM
352 C. L. Grifths et al.
The uncontrolled growth of Cape Town and the development of industry at Saldanha near
Langebaan Lagoon (recognised by the Ramsar Convention as a wetland of international importance)
have led to the development of the controversial Skuifraam Dam in the upper Berg catchment near
Franschoek (Cape Town supply). Skuifraam will harvest all upper river ows and all small to
medium ood ows in the system (Berg 1993) and will decrease peak discharges in upper Berg
by 40%, and in the lower system by 20%, above present abstraction rates (Basson and Beck 2001).
Multiple ood ows in any year cannot be maintained for Skuifraam; return ows of 1:2 yr will
be reduced to 1:3 yr (simulation data). With the construction of Skuifraam, ows in the Berg will
drop from 914 ¥ 106 m3 MAR to 572 ¥ 106 m3 MAR (62% of virgin). Once additional water supply
developments are executed, ows will fall further to 272 ¥ 106 m3 MAR (25% of virgin) (Little
1993), with severe implications for the continued functioning of the estuary, the oodplain, and
the adjacent coastal zone.
To place the Berg River estuary and oodplain in context, O’Keeffe et al. (1992) note that 76%
of the estuaries of the Western Cape are degraded. Of these, the Berg is undoubtedly the most
Figure 28 Virgin (natural) and regulated (present day, 1993) seasonal distribution of runoff to the Berg River
estuary. (After Berg 1993.)
Flow (Millions of Cubic Metres per Month)
Natural Present Day
JFMAMJJASOND Page 352 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 353
important. The once mobile estuary mouth was articially entrained in 1966, changing tidal
amplitude and causing higher seawater intrusion to the oodplain (Huizinga et al. 1993). Upstream
impoundments now prolong periods of low ow in rivers, increasing salinity intrusion and reducing
sediment dynamics in the estuary and at the mouth. The development of Skuifraam will exacerbate
these features, reducing ushing and scouring, and may ultimately lead to mouth closure (Huizinga
et al. 1993), with severe consequences for inshore sheries and coastal processes.
The nearly 6000-ha oodplain–estuarine complex supports 10 different vegetation communities,
making it the most structurally complex system of its type in southern Africa (McDowell 1993).
Among other features, the oodplain–estuary has the third largest area of salt marsh along the Cape
coast, second only to Langebaan Lagoon, within the Benguela region. Mouth entrainment and ow
regulation have signicantly impacted the vegetation (McDowell 1993).
Since 1975, 127 bird species have been recorded utilising 5 of the 10 wetland communities
(Hockey 1993a), and despite the fact that the system is not a Ramsar site, in terms of its avifauna
it is far more signicant than the adjacent Langebaan Lagoon, the Orange River Mouth, and even
the St. Lucia Wetlands of KwaZulu–Natal (all of which are Ramsar sites). The system ranks second
in importance only to Walvis Bay in Namibia (also a Ramsar site). In 1992, 46,000 individuals
recorded represented regionally signicant populations of 31 bird species and nationally signicant
populations of 25 species, with ve Red Data species present (Hockey 1993a). In order to maintain
these populations Hockey (1993a) recommended a minimum average ooding frequency of 1.9
yr–1. Given the development of Skuifraam, this is unlikely to happen. Further, reduced ows over
the past 40 yr, coupled to desiccation, increased salinities, and possibly pollution, have had severe
impacts on the density of benthic invertebrates, thereby adding to concerns for the maintenance of
the avifauna (Hockey 1993b).
The system is also vital in terms of its role as a coastal zone sh nursery. Bennett (1993) has
remarked that the shes recorded from the Berg River estuary–oodplain system represent 77% of
the total coastal species, compared with between 49 and 52% for other estuaries. Indeed, the Berg
has higher percentages of resident species (23% vs. between 4 and 18% for other estuaries in South
Africa), dependent species (27% vs. between 9 and 25%), and partially dependent species (30%
vs. between 18 and 27%) (Bennett 1993). Accordingly, ow reductions in the Berg have had far
more impact on coastal sheries than anywhere else in South Africa. Additionally, the estuary is
one of only two permanently open estuarine nursery areas in the region. It is thus clear that further
ow modication of the Berg will have serious implications for the maintenance of inshore sheries.
In this context, Schrauwen (1993) notes that although a local sport shery is developing, commercial
sheries are already overexploited and that Berg River estuary salinity increases (reduced ows)
render parts of the industry nonviable.
An IFR has been calculated for the Berg, but again, as in the Orange, it has not been
implemented. Thus, like all other river systems owing to the Benguela, the natural river
ecosystem is at risk, with disturbing future implications for estuarine and coastal processes in
the region.
Although certain forms of sh and shellsh farming have been practised in South Africa for many
years, it is only since the 1980s that aquaculture has grown into a commercialised and economically
viable industry. Freshwater aquaculture in South Africa is, however, severely limited by water
availability, and the only real potential for the development of the industry in South Africa therefore
lies in the sea (Cook 2000).
In the Benguela region, marine aquaculture is at present practiced mainly in Saldanha Bay,
with smaller operations at various other sites. Mussels and oysters have traditionally formed
the bulk of production, but seaweeds and abalone have become more important in the past few Page 353 Wednesday, June 30, 2004 2:19 PM
354 C. L. Grifths et al.
years. Although prawns are produced in brackish water ponds in KwaZulu–Natal on the east
coast, they are not cultured in the Benguela region because of low water temperatures. Over
the past 10 yr, total mariculture production in South Africa has remained relatively static, but
some important changes have occurred with respect to individual species, the production of
some species having been discontinued, while others have seen signicant production increases.
Total production in the Benguela region in metric tons per year (animals only) is shown in
Figure 29. Some of these values were obtained directly from South African farms, while others,
particularly for earlier years, were obtained from published Food and Agriculture Organisation
(FAO) statistics (Grainger & Garcia 1996). Further details with respect to each of the major
species listed are provided below.
Mussel production in South Africa uses either the raft or longline system, the former proving to
be the more popular. Although between 1988 and 1992 attempts were made to farm some indigenous
species (e.g., Choromytilus meridionalis and Perna perna), the industry has now settled on the
introduced Meditteranean mussel Mytilus galloprovincialis as the most economically viable species.
In Saldanha Bay, the main centre of the industry, spat are collected locally on ropes. Since the rst
farmed mussels were produced in the early 1980s, annual production rose rapidly, reaching 2300
t in 1994. After a slight decrease in production in 1995 and 1996, production again rose in 1997
as a new small-farmer cooperative began to come onstream. Production reached a peak of 2600 t
in 1998, with a total value of a little over U.S.$2.5 million. Problems associated with harmful algal
blooms (HABs) in the Benguela region have severely curtailed mussel production over the past
few years, and in 2000 production fell to the very low value of 160 t. It appears, however, that
production is now beginning to increase again.
As in many other countries, oyster production in South Africa and Namibia utilises the fast-growing
Pacic oyster, Crassostrea gigas. Although an exotic species in South Africa, importation of spat
is permitted because it has never been thought to show any signs of becoming invasive (although
there are recent reports of wild populations in several Southern Cape estuaries). At present, the
vast majority of production is from imported spat, but recently a small local hatchery started
operating. Several farming methods are used and each has been developed in response to the
Figure 29 Mariculture production in the Benguela region from 1988–2001.
Tonnes (10 )
Pacific oyster
Mediterranean mussel
3 Page 354 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 355
demands of local conditions. The most common method uses racks or oyster baskets in shallow
coastal lagoons. Total annual production in South Africa reached over 600 t in the early 1990s but
has since dropped considerably. In 1998, about 480 t was produced, the total value being about
U.S.$2.26 million. By 2001, production had dropped to 147 t.
Research into seaweed farming has been conducted over the past few years, concentrating mainly
on Gracilaria and Ulva. Gracilaria is used mainly for agar extraction, whereas both Gracilaria
and Ulva are used as important food items in the emerging abalone farming industry. Most of the
farming effort in South Africa has been concentrated in the vicinity of Saldanha Bay, where a
thriving Gracilaria harvesting industry existed before harbour construction altered the natural
habitat of the seaweed and rendered the harvesting industry uneconomic. Both tank cultivation and
suspended cultivation from rafts are being used, but at this stage, the raft method seems to be
preferred for commercial production. Raft design is based on that already in use in Namibia, where
Gracilaria farming is already an established industry, incorporating both actual culture and some
limited factory processing where semirened agar is produced. Production gures have uctuated
over the past few years as new producers have experimented with Gracilaria production, only a
few of whom have survived to form fully commercial production units. A signicant increase in
1998 occurred as abalone farmers began producing macroalgae as an abalone feed. Estimated
production in South Africa in 1998 was 25 t, the total value of which was a little over U.S.$108,000.
It is predicted that production in 2002 may reach 105 t.
Perhaps the greatest mariculture potential in South Africa is for abalone farming. The local species,
Haliotis midae, is prized among markets in the East and the market price has encouraged local
companies to invest in this industry. All farms are land-based and water is pumped ashore to them.
The coastline supports huge kelp beds and most farms use kelp as food, although an articial food
is being produced locally, which may replace or supplement kelp on some farms. About 10 farms
are now under construction, with the most advanced almost at the full commercial production phase.
Farmed abalone from South Africa began to appear on the international market in small
quantities in about 1993, but at that stage, less than 1 t yr–1 was being produced. By 1997, annual
export production was about 10 t, and since then, production has gradually increased to about 317
t in 2001. It is predicted that this production level will continue to increase over the next few years.
Although prices have uctuated recently, the 1998 production was valued at about
U.S.$950,000. Because abalone is sold as an international commodity and is priced in U.S. dollars,
the value of the industry has increased signicantly over the past year or so, and it is estimated
that export production is now valued at least at 95 million rands. All South African farmed abalone
is exported, mainly to Japan, as it is illegal to sell farmed South African abalone on the local market.
Although there is probably potential in South Africa for that production level to be increased,
market demand will obviously dictate the eventual size of the industry.
Marine nshes
The farming of marine nshes is a relatively new activity in the Benguela region, with turbot
(Psetta maxima) being the only species currently produced commercially. Although there are several
indigenous species of marine shes that appear to have great mariculture potential, the technology
for farming them is only just being developed. It seems likely, however, that this is an area of
mariculture where signicant expansion can be expected in the future. Several regions of South Page 355 Wednesday, June 30, 2004 2:19 PM
356 C. L. Grifths et al.
Africa seem to have potential for salmon or turbot farming, and several applications to start farms
are being considered at present. One of the main problems is that since nshes are exotic species
to South Africa, the potential hazards of introducing them are not yet known and conservation
authorities have been expressing severe reservations about possible problems. In addition, it is
doubtful whether there is a sufciently large market for these shes in South Africa and export
potential thus needs careful investigation. However, it appears to be only a matter of time before
the rst salmon farms will be commercialised.
Production losses
In recent years, there have been production losses in most of the major forms of mariculture. In
mussel farming, the major loss was caused by adverse weather conditions leading to raft and
longline destruction. In some cases, HABs have interrupted mussel harvesting, but this cannot
really be classied as a loss, because mussels were again harvested after the bloom had ended. In
the case of oysters, however, major permanent production losses have been caused by HABs,
particularly associated with the organism Aureococcus anophagefferens. This organism began to
affect oyster production toward the end of 1998, and its effect was severe in 1999. HABs have also
affected abalone production, but only to a very small extent.
Historical and future production trends
Two trends appear in the historical mariculture production gures for the Benguela region (Figure
29). First, a general stabilisation of the industry has occurred where a number of species that were
cultured on an experimental or pilot plant basis (e.g., red bait, Mactra, and mullet) have now been
discontinued and replaced by species that are more economically viable. For example, a few years
ago, three species of mussels were cultivated, but now all mussel farms have settled on the one
species, Mytilus galloprovincialis, which produces the best return on investment.
The second trend has been where species have proved to be economically viable, leading to
increased production levels. This applies to Gracilaria, the Mediterranean mussel, and in particular
to abalone. Although the Pacic oyster has proved an economically viable mariculture species,
total recent production has not increased. This, however, has resulted from environmental factors
and does not reect the viability of the species. A number of species have been identied as having
good mariculture potential and numerous experiments are being conducted to assess this. Marine
nshes, particularly species of indigenous origin, appear to have very good potential for future
development. A future trend is therefore likely to include development and expansion of marine
nsh culture.
Environmental monitoring of the waters used for mariculture production has become extremely
important in order that products conform to international standards. In the past, such monitoring
has been inadequate in South Africa, and this will have to be improved if farmers wish to continue
to export their products.
Because the Benguela region of South Africa has relatively few areas of sheltered water,
mariculture activities are concentrated into a few prime sites. Research has been directed at
establishing the carrying capacity of such sites, and the results are likely to limit the future expansion
of mariculture activities (Cook & Grant 1998). In addition, the monitoring and control of HABs
in the future will become increasingly important as population growth dictates that increasing
numbers of people will be living close to the coast.
Mariculture in southern Africa is seen as an industry with a great deal of potential for devel-
opment and expansion. The recently adopted South African Marine Living Resources Act recognises
mariculture as a growth industry and the directorate of Marine and Coastal Management has been
charged with the duty of developing and regulating the industry. It is therefore likely that mariculture Page 356 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 357
will expand rapidly in the near future. The main areas in which short-term future growth is likely
to occur are in abalone, seaweed and marine nsh farming. All these industries are relatively new
to southern Africa. Although the marine nsh industry is still at an experimental level, the others
have been expanding rapidly over the past few years. One of the principal objectives of management
will be to ensure that expansion occurs in an environmentally acceptable and sustainable manner.
Environmental monitoring of mariculture activities will be essential to ensure this sustainable
Marine invasive aliens
Marine organisms have intentionally, or more often accidentally, been transported across the world’s
oceans since the earliest human attempts at exploration, colonisation, and commerce (Carlton 1989,
1999). On an intraoceanic scale these movements date back thousands of years, but interoceanic
movements only really became signicant in the age of European global exploration, which began
in the 1400s.
Early marine introductions took place via two main mechanisms. Semiterrestrial and littoral
marine species were transported along with rocks, sand or other dry ballast placed inside the holds
of ships. A much more diverse group of attached fouling species, particularly algae, barnacles,
hydroids, ascidians and molluscs, were transported on the outside of the hull, or bored into the
timber of the vessels. How many such species survived voyages through the tropics to southern
temperate waters is unknown, but Carlton (1999) has estimated that early wooden vessels could
have each carried 150 or more species per voyage. Many of these vessels sank en route or remained
in their ports of destination, thus inoculating substantial and concentrated populations of potentially
invasive species at these sites.
The proportion of early introductions that resulted in naturalised populations is extremely hard
to estimate in retrospect, since these introductions predate any kind of ecological survey, sometimes
by hundreds of years. What seems certain, however, is that many species now regarded as wide-
spread or cosmopolitan were in reality introduced by early explorers. Such species are now referred
to as cryptogenic (of unknown origin) and may include some of the most common and ecologically
important species in the area of introduction. Potential examples of cryptogenic forms abound and
are discussed by Carlton (1999). Several such species occur in the Benguela region, where they
have been considered as naturally occurring. Examples include a variety of fouling species, includ-
ing hydroids (Obelia), bryozoans (Membranipora), ascidians (Botryllus, Botrylloides) and amphi-
pods (Corophium, Jassa). Shell and wood-boring cryptogenic species are also represented in the
Benguela region, for example, the sponges Cliona spp. and the amphipod Chelura terebrans (the
better-known wood-boring shipworm Banksia carinata and isopod Spheroma terebrans are found
only on the Indian Ocean coast of South Africa).
Over the past 100 yr the mechanisms by which marine species have been transported have
undergone considerable change. Boring species are no longer able to penetrate the steel hulls of
modern vessels, and although some fouling forms still cling to the outside of vessels, antifouling
paints and increased ship speeds have progressively reduced the importance of this as a transport
mechanism. Other vectors, however, have increased in importance. These include the deliberate
introduction of marine species for mariculture, research or aquarium display purposes (sometimes
along with their parasites or pathogens). Far more important, however, is the transport of species
in ballast water, which is taken on board cargo vessels to improve their stability and trim. The scale
of this mechanism is unprecedented. Carlton (1999) estimates that if only 10% of the 35,000 vessels
at sea at any given time are carrying ballast, and if just two unique species are carried in each
vessel, more than 7000 species are “in motion” on the world’s oceans on any given day. Ballast
water, with its fauna, is moreover probably turning over at a rate of 5–10% d–1, giving a much Page 357 Wednesday, June 30, 2004 2:19 PM
358 C. L. Grifths et al.
higher total if longer time periods are considered. The net result has been a huge increase in rates
of introduction with each succeeding decade.
Because of the increased speed of vessels, ballast water is capable of transporting both adult
planktonic and larval benthic forms. The trend has thus been one away from transport of sessile
adult invertebrates, toward introduction of larval and planktonic forms, which can include pathogens
and toxic dinoagellate spores. The construction of new harbour areas, especially in regions of
high conservation value, also establishes new foci for introduction. A prime example here has been
the construction of the new harbour in Saldanha Bay, an ecologically sensitive and previously
relatively pristine bay, in the 1970s. The volume of ballast water transported into this site was
estimated at 6.8 million t yr–1 by Carter (1996).
What, then, is the situation with respect to alien marine species along the west coast of
southern Africa? Grifths (2000) lists 22 marine species introduced to South African waters,
while one additional form (the alga Schimmelmannia elegans from Tristan da Cunha) has
subsequently been reported from Table Bay docks (De Clerck et al. 2002). However, several of
the species listed by Grifths (2000) are known to have become locally extinct, are dubious
records (e.g., observations of dead mollusc shells), are of uncertain introduced status, or occur
only along the Indian Ocean coast. Only eight introduced species are conrmed and extant in
the Benguela region. One of these, the oyster Crassostrea gigas, is an intentional introduction
and is farmed at a variety of sites (there have been recent reports of wild populations of this
species in estuaries along the Southern Cape coast, but this is just outside the region covered
by this review). The remaining seven are accidental introducions, most of which are conned
to harbours or lagoonal sites, where their impact is limited. Only one such species, the ascidian
Ciona intestinalis, is known to have environmental or ecological impact, as it fouls the ropes
of mussel culture rafts and smothers the mussel spat. Only two introduced species have become
invasive on the open coast of the Benguela. These are the Mediterranean mussel Mytilus
galloprovincialis and the European shore crab Carcinus maenas.
Carcinus maenas was rst recorded from Table Bay docks in 1983 (Joska & Branch 1986) and
by 1990 had spread from Camps Bay to Saldanha Bay (Le Roux et al. 1990), a distance of some
100 km. No signicant subsequent expansion has been noted; indeed, no further specimens have
been reported from Saldanha Bay since the single observation reported by Le Roux et al. (1990)
and one dead carapace collected in 2002 (C.L. Grifths, personal observation). Notably all estab-
lished populations are from shores sheltered from waves, suggesting that Carcinus maenas has
difculty establishing on the exposed open coastline of the Benguela region, and hence in reaching
suitable sites distant from its point of origin. There is, however, considerable concern that this
species will cause severe damage if it does establish a population in the Saldanha Bay–Langebaan
Lagoon complex, which has an abundance of suitable sheltered habitat and is also an important
conservation and mariculture site.
Mytilus galloprovincialis was rst recorded in South Africa by Grant et al. (1984). By that
stage it had already established extensive populations along the entire west coast between Cape
Point and Luderitz. The exact date and site of introduction are unknown, but circumstantial evidence
suggests that it was recent, probably in the late 1970s, and mediated by man (Hockey & van Erkom
Schurink 1992, Grifths et al. 1992). By the early 1990s Mytilus galloprovincialis had spread as
far east as Port Alfred, in the Southeastern Cape, and was the dominant intertidal organism along
the entire west coast, with an estimated biomass of some 194 t wet mass km–1 of rocky coast (van
Erkom Schurink & Grifths 1990). Unpublished data suggest that standing stocks on the west coast
have subsequently increased substantially and that the species continues to spread eastward at a
rate of about 10–20 km yr–1 (N. Steffani & C.D. McQuaid, personal communication). There has
also been further northward expansion into northern Namibia, although densities remain lower in
this region than in South Africa.
Relative to indigenous mussel species, Mytilus galloprovincialis has a rapid growth rate, high
fecundity, and enhanced tolerance to desiccation (van Erkom Schurink & Grifths 1990). Its spread Page 358 Wednesday, June 30, 2004 2:19 PM
Impacts of Human Activities on Marine Animal Life in the Benguela: A Historical Overview 359
has therefore greatly increased the overall biomass and vertical extent of mussel beds along the
entire west and (to a lesser extent) south coasts of South Africa. This has had several implications
for the wider intertidal community.
Before the arrival of Mytilus galloprovincialis, the dominant space-occupying invertebrates
in the mid- to low intertidal of the Cape west coast were the slow-growing, indigenous mussel
Aulacomya ater and limpets, notably Scutellastra (formerly Patella) granularis and Scutellastra
argenvillei. Much of the shore was open space, kept clear by the intense grazing activities of
these and other large limpets. Because Mytilus galloprovincialis grows much faster and extends
higher into the intertidal zone than Aulacomya ater (van Erkom Schurink & Grifths 1990,
1993), it is now the dominant space-occupying intertidal species at most sites. The net result
has been a massive increase in both mussel cover and biomass and a movement of both the
upper limit and centre of gravity of the mussel beds upshore. This has been accompanied by a
decline in the overall biomass of Aulacomya ater, although paradoxically this species now
occurs higher up the shore than before, since it can nd protection within the dense mats of
Mytilus galloprovincialis.
As well as outcompeting indigenous mussel species, Mytilus galloprovincialis competes suc-
cessfully for primary rock space against adult limpets (Hockey & van Erkom Schurink 1992,
Grifths et al. 1992). However, the shells of large mussels also offer a favoured settlement and
recruitment site for juvenile limpets. Thus, as mussels encroach, adult limpets initially become
spatially constrained and then eventually eliminated. However, at the same time, the densities of
smaller limpets on the mussels increase enormously. The net result is an initial increase in limpet
biomass (while the adults on the rock and the juveniles on the mussels are both present), followed
by a decline, when the adults are nally eliminated. Effects on limpet fecundity differ between
species. In the midshore, small Scutellastra granularis living on mussel shells attain sexual maturity
and the overall mass of gametes released by the dense population of small individuals in areas of
100% mussel cover actually exceeds that of the few large animals previously found on the bare
rock (Grifths et al. 1992). By contrast, the larger, low-shore limpet Scutellastra argenvillei also
recruits onto mussel shells, but it is unable to attain sexual maturity in this conned habitat. As a
result, reproductive output cannot be maintained in areas invaded by mussels (N. Steffani, personal
Mussel beds are also structurally complex habitats that provide refuge to a diverse commu-
nity of associated organisms. Grifths et al. (1992) showed that the infaunal communities
colonising Mytilus galloprovincialis and Aulacomya ater beds were similar in both species
richness and composition (69 and 68 species, respectively, with 70% shared). However, because
Mytilus galloprovincialis grow to a larger size, attain a greater biomass, and develop thicker,
more structurally complex beds, they support a much denser invertebrate fauna than Aulacomya
ater beds (76,600 vs. 34,000 individuals m–2). They also tend to provide refuge for larger
infaunal individuals.
A third implication of the increase in overall mussel biomass is that mussels form an
important component in the diets of a wide range of predatory species, including both aquatic
forms, such as shes, rock lobsters, starshes, predatory whelks and octopuses, and terrestrial
ones, including shorebirds and humans (Grifths & Hockey 1987). The introduction of Mytilus
galloprovincalis has led to a massive increase in mussel standing stock, and hence enhanced
food availability to natural predators. The benets to terrestrial species may be particularly
marked because they are now able to gain access to mussel stocks in the upper shore, where
none occurred previously. Hockey & van Erkom Schurink (1992) and Grifths et al. (1992)
both compared the contents of oystercatcher middens in Saldanha Bay before and after the
Mytilus galloprovincialis invasion. They show a dramatic switch in oystercatcher diet away
from limpets (31 to 19%) and Aulacomya ater (36 to 3%) in favour of Mytilus galloprovincialis,
which comprised over 66% of oystercatcher diet between 1987 and 1991. The increased
availability and accessibility of food to oystercatchers resulting from the Mytilus galloprovin- Page 359 Wednesday, June 30, 2004 2:19 PM
360 C. L. Grifths et al.
cialis invasion have simultaneously reduced the length of time oystercatchers have to forage
on each tide (and hence increased their resilience to storms, human disturbance, etc.). They
have also increased the proportion of pairs successfully raising two chicks (Hockey & van
Erkom Schurink 1992).
As shown above, the Mytilus galloprovincialis invasion has had profound effects on the appear-
ance and ecological processes occurring within rocky shores over much of the South African
coastline. While the ecological costs of this invasion may be signi