This study was conducted along the Garden Route coastline of South Africa between the Groot River in the east and the western border of the Kaaimans River mouth (Figure 1). The study area consisted of predominantly rocky coastline (78%) and sandy beaches (22%) and was 248 km long. The area falls under the jurisdiction of four local municipalities, with several towns and villages scattered throughout the adjacent interior. Coastal access points are plentiful. The majority of people living in the area are situated in or near the towns of George, Knysna and Plettenberg Bay (Table 1).

Three MPAs are situated within the study area: The Tsitsikamma MPA consists of 68 km of closed (no resource use) coastline and forms part of the Garden Route National Park, managed by South African National Parks, whilst both the Robberg and Goukamma MPAs, managed by CapeNature, are open to shore angling, but no boat angling is permitted.

Monthly aerial surveys covering the survey area were conducted between December 2008 and November 2009 in order to obtain ‘instantaneous’ counts of shore angling effort. Survey days and flight times were randomly selected, but depended on the weather, and pilot and observer availability. All flights were conducted during daylight hours, with the earliest flight taking place at 09:10 and the latest at 15:55. All counts were done in an east–west direction, as the observers were seated on the starboard side of the plane.

The first four surveys were undertaken in a four-seater Robinson helicopter, after which a four-seater Cessna light aircraft was used. Flight parameters, including air speed and altitude, were standardised, but depended on weather conditions and the pilot's discretion. Flight times ranged between 70 min and 80 min, with an average air speed of 189 km.hr⁻¹ for both aircraft. Altitude ranged between 30 m a.s.l. and 80 m a.s.l. for the Robinson and between 150 m a.s.l. and 250 m a.s.l. for the Cessna.

On each flight, two observers counted shore anglers and one scribe entered the data into a global positioning system-enabled pocket PC. Each spotter was equipped with a pair of binoculars (10 x 42 magnification) and continuously scanned the coastline, verifying the other spotter's counts. When large groups of anglers were encountered, a second flyby was made and anglers were re-counted by both spotters. Information recorded during the aerial surveys included the date, time, weather, sea conditions, number and location of anglers or boats and habitat where shore anglers were fishing (rock or sand).

Data were also obtained from three aerial surveys conducted in December 2008 as part of a study researching the ecology, value and management of the Garden Route (Chalmers et al. 2009). Methods were similar, with the same Cessna plane and pilot being used, two observers and similar standardised flight speed and altitude. This information was used in temporal, but not spatial (coastal effort maps), analyses.

Angling effort showed a heterogeneous distribution throughout the study area and although variability in count data was high, a significant difference in angling effort occurred between MPAs and non-MPAs (Kruskal–Wallis one-way ANOVA, df = 6, p-value < 0.05) (Figure 2) and equal area breaks (Kruskal–Wallis one-way ANOVA, df = 24, p-value < 0.05) (Figure 3).

The majority of anglers (72.50%) fished in open access areas outside formal conservation borders (including no-take and controlled zones) and a Bonferroni post-hoc test showed that the Tsitsikamma MPA had significantly less fishing effort than all other sections (p-value < 0.05) with the exception of zone 9, in which no anglers were encountered. Although the Tsitsikamma MPA is zoned for no-take (i.e. no consumptive use), a total of 32 illegal fishermen were counted during all the surveys, with an overall angler density of 0.03 anglers.km⁻¹ (Figure 3).

Outside of formal protected areas, the highest density of anglers occurred in section one (Wilderness) with 1.35 anglers.km⁻¹, followed by section 15 (Plettenberg Bay) and 17 (Nature's Valley) with overall angler densities of 0.78 anglers.km⁻¹ and 0.58 anglers.km⁻¹ respectively (Figure 2). These three sections contributed to the overall angler densities being highest (0.73 anglers.km⁻¹) along the larger section of coastline between the Kaaimans River mouth and the western border of the Goukamma MPA. This was followed by the Plettenberg Bay coastline (Robberg to De Vasselot) with 0.65 anglers.km⁻¹ and the De Vasselot section of the Garden Route National Park with the third highest density of 0.51 anglers.km⁻¹ (Figure 3). The Goukamma and Robberg MPAs had overall angler densities of 0.41 anglers.km⁻¹ and 0.14 anglers.km⁻¹ respectively (Figure 3), whilst the coastal section between the Goukamma and Robberg MPAs, consisting of large stretches of exposed rocky coastline with few access points, had a relatively low overall angler density of 0.16 anglers.km⁻¹.

Angling effort occurred on both rocky and sandy substrates throughout the study area with both high densities (2.23 anglers.km⁻¹ – 3.94 anglers.km⁻¹) occurring over rocky (Figure 4a) and sandy substrates (Figure 4b). Although total angler density on rocky substrates was lower (0.31 anglers.km⁻¹) than on sandy substrates (0.47 anglers.km⁻¹), fishing effort between substrates was not significantly different (Mann–Whitney U test, p-value 0.221).

Although more anglers were encountered during the autumn (0.37 anglers.km⁻¹) and winter (0.38 anglers.km⁻¹) months (Figure 5a), with fewer anglers fishing in spring (0.20 anglers.km⁻¹) and summer (0.21 anglers.km⁻¹), no significant difference was observed (Kruskal–Wallis one-way ANOVA, df = 3, p-value 0.266). This was due to the high variability within the data, exacerbated by the low sample size. Fishing pressure was significantly higher during weekends and public holidays (0.38 anglers.km⁻¹) than on week days (0.18 anglers.km⁻¹) (Mann–Whitney U test, df = 1, p-value 0.028) (Figure 5b).

Total annual shore angling effort was estimated at 49 812 angler outings. Recent angler interviews conducted along portions of this coastline indicate that anglers spend on average between 5 hours (Smith 2012) and 6 hours (Van Zyl 2011) fishing per outing. Using these figures, a total annual fishing effort estimate was estimated to be between 226 645 angler hours and 298 872 angler hours.

Angler distribution is unlikely to be uniformly distributed along any coastline and may be impacted by access points and infrastructure that have a clustering effect (Smallwood et al. 2012). Similarly, habitat preferences of target fish species can influence the spatial distribution of anglers attracted to particular sites and different species (Smallwood et al. 2013).

Recreational fishing along the Garden Route coastline showed a heterogeneous spatial distribution, with some localised areas having high angler densities. These areas were generally associated with ease of access and proximity to the more heavily populated urban areas of George and Plettenberg Bay. A similar pattern was shown along the KwaZulu-Natal coast of South Africa (Mann et al. 2008), where angler effort was largely concentrated around access points, with increased effort along the more developed and highly populated sections of the coastline. The concentration of anglers around easy access points along South Africa's coastline has more than likely been heightened by the 2002 ban (Schedule 44 of the National Environmental Management Act [Act No. 107 of 1998]) on beach driving, which limits anglers’ access to the more remote sections of coastline (Mann & Tyldesley 2012; Mann et al. 2008).

In fisheries where angling effort is widely dispersed (e.g. coastlines), instantaneous estimates of total fishing effort obtained from aerial surveys are more accurate and preferable to land-based roving creel surveys (Pollock et al. 1994). Although no previous aerial surveys have been undertaken along this section of coastline for comparative purposes, a number of land-based roving creel surveys covering sections of the coastline have been completed, including Plettenberg Bay (King 2006), the Goukamma MPA (Van Zyl 2011) and the Wilderness to Sedgefield coastline (Smith 2012). These projects, conducted at finer spatial scales, show similar patterns and general trends to those recorded during the aerial surveys conducted in this study.

Greater angling effort (1.03 anglers.km⁻¹) was recorded along the Wilderness to Sedgefield (Smith 2012) and Plettenberg Bay coastlines (0.71 anglers.km⁻¹ [King 2006]) during land-based surveys, in comparison to 0.84 anglers.km⁻¹ and 0.69 anglers.km⁻¹ recorded during the aerial surveys of this study. Due to the generally shorter distance and slower speed of land-based creel surveys, a higher estimation of effort in comparison to aerial surveys would be expected (Pollock et al. 1994). The Goukamma study did not report on angler densities in terms of anglers.km⁻¹ of coastline, but estimated total effort relating to angler hours (21 428 [Van Zyl 2011]), approximately half the 48 167 angler hours estimated in the Wilderness study (Smith 2012). This is similar to the trend seen in the effort estimations of the aerial surveys, with an overall angler density of 0.46 anglers.km⁻¹ recorded for the Goukamma MPA.

The distribution and density of recreational users in general is known to be impacted by large temporal factors such as seasons and holiday periods (Hingham & Hinch 2002). Large-scale patterns in recreational shore fishing along the South African east coast have also been shown to fluctuate seasonally, with an increase in summer and a decrease in winter (Brouwer et al. 1997). Both King (2006) and Smith (2012) have shown that recreational shore fishing within Plettenberg Bay and the Wilderness coastline followed this trend, attributing the increase in effort to an influx of visitors during school holiday periods. However, results from the aerial surveys are more consistent with those from Goukamma (Van Zyl 2011) and the KwaZulu-Natal shore fishery (Dunlop & Mann 2012; Mann et al. 2008), which both showed an increase in shore angling density over the winter months.

Two popular targeted shore angling species, galjoen (Dichistius capensis) and elf (Pomatomus saltatrix), show strong seasonal abundance and have closed seasons between 15 October and the last day in February, and 01 October and 30 November respectively. It is thus likely that those anglers who specifically target these species are temporarily absent from the fishery during these closed periods (Mann et al. 2008; Van Zyl 2011).

Given that no previous aerial surveys have been conducted along the Garden Route coastline and shore-based surveys have only covered portions of the coastline, it is impossible to say whether total fishing effort is increasing or not. A decrease in shore-based angling effort along the KwaZulu-Natal coastline (Dunlop & Mann 2012) has been attributed to security concerns, declining catches, an increase in the cost of fishing and stricter linefish regulations. As catch rates within the shore fishery have decreased (Brouwer & Buxton 2002), the duration and frequency of fishing may have altered, which in turn would affect the turnover rate used to calculate the total estimated effort. However, the current average fishing duration of between 5 hours (Smith 2012) and 6 hours (Van Zyl 2011) is very similar to the 5 hours calculated 15 years earlier (Brouwer 1997). Independent effort calculations are seen as a crucial part of the management and monitoring of the South African shore fishery (Dunlop & Mann 2012) and the total angling effort estimated from this study provides a baseline for future work.

With greater biodiversity and abundance of fish, MPAs where fishing is allowed are attractive locations for recreational anglers (Cooke et al. 2006; Götz, Cowley & Winker 2008; Parker, Booth & Mann 2013). However, non-compliance by anglers fishing in closed, no-take areas can undermine conservation and fishery management objectives. Along the Garden Route, shore fishing is allowed in all MPAs except Tsitsikamma. However, illegal shore fishing (poaching) was observed on several occasions during this study within the no-take Tsitsikamma MPA. Although fishing levels inside the park were low compared to outside adjacent areas, illegal fishing within this important MPA is cause for concern.

Within the Ningaloo Marine Park in North West Australia, anglers were found to be concentrated in localised areas with easy access, and limited non-compliance was observed within no-take zones (Smallwood et al. 2011). The low levels of non-compliance were attributed to a number of factors, including (1) multiple educational tools (signboards and brochures), (2) visual zonation reference points, (3) various enforcement patrols (boat- and land-based) and (4) a willingness by anglers to fish at alternative sites (Smallwood & Beckley 2012). More research is required regarding the illegal fishing taking place within the Tsitsikamma MPA to gain a greater understanding of the extent of this activity and the underlying drivers, which could then be used by management to address and resolve this issue.

Although the Goukamma and Robberg MPAs are open to shore-based fishing, these areas had comparatively lower angler densities than other sections of coastline outside the formal conservation areas. Fishing locations within Robberg are limited due to a large seal colony on the eastern side and the exposed rocky coastline on the west, restricting fishing to a few well-known locations (King 2005). The Goukamma MPA has a combination of habitat types and land-based roving creel surveys conducted within this reserve (Van Zyl 2011) showed that anglers were clustered around easily accessible areas, favouring rocky and mixed shores over sandy beaches. Similar to the aerial survey data presented here, angler densities were greater around the Buffalo Bay section of the reserve, which may be due to ease of access, with anglers in the Goukamma section having to walk up to 6 km to reach some fishing locations.