Commercial fishery no-take zones for African penguins minimize fisheries losses at the expense of conservation gains

Abstract

The African penguin population has declined precipitously in recent decades, and if current rates of decline persist, this species could become extinct in the wild by 2035. Resource extraction of small pelagic fish prey by the purse-seine fishery around African penguin breeding colonies has been identified as a demographically meaningful threat to African penguins. Consequently, long-term, effective no-take zones around breeding colonies have been endorsed by an expert panel of scientists constituted by the South African government. Here, we consider the six largest South African penguin colonies that currently hold 76% of the global population. We evaluate the adequacy of different no-take zone options using a trade-off mechanism recommended by the expert panel. For all six colonies except Bird Island, Algoa Bay, which is subject to the least fishing pressure, the current no-take zone delineations are assessed as having little benefit to the African penguin and little to no cost to the purse-seine fishery. Four of the six current no-take zones include ≤50% of the African penguins’ core foraging areas. Alternative no-take zones that approximate a more balanced trade-off offer more impactful alternatives to the current fisheries restrictions. Given the urgent need to implement evidence-based conservation interventions for the endangered African penguin, we recommend the substitution of the current no-take zones with those proposed herein.

seabirds, penguins, conservation, fisheries, no-take zone, purse-seine, sardine, anchovy

Introduction

Competition between seabirds and industrial-scale fisheries for common prey resources is a significant and growing threat to many seabird species (Croxall et al. 2012, Dias et al. 2019). Resource competition can negatively impact their foraging performance (Boersma and Rebstock 2009, Bertrand et al. 2012), breeding success (Sherley et al. 2018, Sydeman et al. 2021), and survival (Frederiksen et al. 2004), directly affecting demographic processes. Seabird species with specialized diets and whose prey are also targeted by large commercial fisheries are particularly vulnerable to this threat (Furness and Tasker 2000, Searle et al. 2023). The implementation of fisheries no-take zones in marine habitats has shown positive outcomes in alleviating the negative impacts of resource competition for several species (Sydeman et al. 2017, Searle et al. 2023).

African penguins (Spheniscus demersus) are endemic to the Benguela Upwelling Ecosystem off South Africa and Namibia, where they feed on small pelagic prey, predominantly sardine (Sardinops sagax) and anchovy (Engraulis encrasicolus) (Hockey et al. 2005). They are currently listed as globally Endangered (BirdLife International 2024), and there is evidence to support uplisting of this species to Critically Endangered (Sherley et al. 2024). The global population has declined at 7.9% per annum over the last 10 years to ∼9900 breeding pairs; close to just 40% of the ∼24 000 pairs breeding a decade ago. If current trends persist, the African penguin could be extinct in the wild by 2035 (Sherley et al. 2024).

A key driver of the decline in the number of African penguins is regionally poor prey availability; there has been a substantial decrease in the regional biomass of sardine, which has remained low since 2007 (Coetzee et al. 2022a) and has been recognized as a significant factor influencing African penguin survival (Robinson et al. 2015, Crawford et al. 2022). In addition, a spatial redistribution of sardine and to a lesser extent anchovy biomass, likely due to environmental changes in wind and upwelling patterns and disproportionate fishing pressure (Roy et al. 2007, Coetzee et al. 2008), occurred in the late 1990s and has persisted in recent years (Coetzee et al. 2022b). In this context, competition for anchovy and sardine from the commercial purse-seine fishery (Sherley et al. 2018, Sydeman et al. 2021), which is the largest fishery by volume in South Africa (Hutchings et al. 2009), contributes to the overall lack of available food at crucial times during the penguin’s lifecycle (Crawford et al. 2022). The threat posed by a lack of available prey is exacerbated by other threats, including underwater noise from increased shipping traffic (Pichegru et al. 2022), predation by seals (Makhado et al. 2006), oil spills (Wolfaardt et al. 2009a), and disease outbreaks (Roberts et al. 2023).

The need to address the cumulative impact arising from these multiple threats is reflected in various national and international policies and agreements, including the African Penguin Biodiversity Management Plan (DEA 2013) and the International Multi-species Action Plan for the Conservation of Benguela Upwelling System Coastal Seabirds (AEWA 2021). These plans provide guidelines for conservation measures to address these threats, and for the most part, the actions have been successfully implemented (Pichegru 2024). However, there has been a lack of urgent and meaningful long-term interventions to curb the major drivers and threats associated with prey availability, a factor that has significantly contributed to the precipitous decline in African penguin numbers (Crawford et al. 2022).

Between 2008 and 2021, an Island Closure Experiment (ICE) was implemented and included no-take zones for small pelagic fishing around four African penguin colonies in South Africa’s Exclusive Economic Zone (EEZ) (Punt et al. 2023). The experiment aimed to assess the benefit (if any) of no-take zones to African penguins. The ICE included alternating 3-year open and closed regimes in a 20 km radius around the (then) species’ four main breeding sites in two biogeographic regions: Robben and Dassen islands on South Africa’s west coast; and St. Croix and Bird islands on the east coast. The 20 km radius was based on what was known at the time about the average foraging range of African penguins when rearing small chicks (Petersen et al. 2006), an energetically demanding period when foraging range is at its most constrained; these birds forage over much larger areas during other parts of their lifecycle (Carpenter-Kling et al. 2022). The reason for the alternating regime was to control for confounding effects such as natural drivers of prey availability (Pichegru et al. 2012, Punt et al. 2023). Design flaws of the ICE included a mismatch between the duration of no-take zones and important events in the African penguin lifecycle, and between the spatial extent of no-take zones and the birds’ at-sea distribution (Punt et al. 2023). Notwithstanding these limitations, the study demonstrated a demographically meaningful benefit of no-take zones for African penguins (Sydeman et al. 2021, Punt et al. 2023). Further, the impacts of no-take zones during the ICE on African penguin survival during the non-breeding season were not evaluated but could have had a significant impact on overall population-level benefits as African penguins spent a considerable time (e.g. 49% of tracking time during post-moult foraging around Dassen Island) within these no-take zones before and after their catastrophic land-bound moult (Carpenter-Kling et al. 2022).

Given the concerning conservation status of the African penguin and the results obtained during the ICE, the need for purse-seine fishing no-take zones around African penguin colonies was recognized by the South African government’s Department of Forestry, Fisheries and the Environment (DFFE) in 2021 (Coetzee et al. 2021). A set of candidate no-take zones was proposed by a Governance Forum constituted by the DFFE in 2021. However, confirmation on the number, extent, and duration of no-take zones was deferred to engagements between the conservation and purse-seine fisheries stakeholder groups (hereafter referred to as Industry) in an attempt to establish consensus on no-take zone delineations that could minimize costs to the purse-seine fishery and maximize benefits to African penguins. To this end, two stakeholder engagement processes ensued between 2021 and 2022: the Governance Forum Extended Task Team (ETT) and the Consultative Advisory Forum on Marine Living Resources (CAF). Unfortunately, neither process resulted in agreement on mutually acceptable no-take configurations (Punt et al. 2023). This stalemate led the DFFE to institute interim no-take zones around six colonies (out of a total of 12 extant colonies in South Africa) in September 2022; these delineations were based on a mix of previously proposed no-take zone extents, including those originally proposed by the DFFE for Dassen and Robben islands, those proposed by the CAF for Stony Point and Bird Island, those proposed by Industry during the ETT for St. Croix Island, and a new split zone configuration for Dyer Island using delineations originally proposed by the DFFE (vessel size limited zone) and by Industry during the CAF (fully restricted zone). These interim closures (now permanently in place) were not endorsed by the conservation sector but were imposed by the DFFE as a temporary measure to enable a final review of the scientific data by an international panel of experts.

To resolve the impasse, the South African government constituted an international panel of six expert scientists to provide an independent and objective review of the costs and benefits of purse-seine fishery no-take zones. In summary, the panel was tasked with the following:

1. to review the science underpinning benefits of no-take zones to African penguins;

2. to evaluate the evidence supporting long-term no-take zones as a precautionary measure;

3. if benefits of no-take zones are confirmed for African penguins, to recommend a trade-off mechanism to inform no-take zone design;

4. to advise on delineations of no-take zones around six colonies and the durations of closures considering African penguin life-history traits;

5. to provide recommendations on a monitoring and evaluation programme; and

6. to provide short-term research recommendations focusing on the dominant causes of the rapid decline of the African penguin (DFFE 2022).

The panel concluded its assessment in 2023 (International Review Panel’s report, IRPR, Punt et al. 2023), which confirmed the demographic benefits of purse-seine fishery no-take zones for African penguins, including a small yet meaningful improvement in population recovery as determined from the ICE (Punt et al. 2023). The panel noted that, considering the limitations in the design and implementation of the ICE, well designed long-term no-take zones are likely to result in additional benefits to the African penguin population, above and beyond those directly accounted for in the ICE, including benefits associated with survival and recruitment. The panel agreed that preferred penguin foraging areas were best delineated using internationally recognized methods to determine marine Important Bird and Biodiversity Areas based on the Area-restricted Search technique (mIBA-ARS, Lascelles et al. 2016). The panel also noted that the Opportunity Based Model (OBM, Bergh, and Horton 2023), submitted by fisheries scientists contracted by the South African Pelagic Fishing Industry Association to estimate economic costs of different no-take zone options to fisheries, likely overestimated those costs. This is largely due to underpinning assumptions regarding the inability of fishing fleets to replace catches outside the no-take zones. Despite these limitations, the panel indicated that the outputs of the OBM could be used in a relative sense to rank different no-take zone options. In this regard, a precautionary strategy requires that the risk to the resource be considered, and that adequate protection be provided to compensate for the absence of robust economic indicators of cost to the fishery. Precaution also implies agreement on action to avoid a crisis. The panel also recommended a specific trade-off mechanism to identify optimal no-take zone designs that maximize benefits to penguins while minimizing costs to the fisheries.

On 4 August 2023, the principle of no-take zones and some recommendations of the IRPR were endorsed by the South African Minister of Forestry, Fisheries and the Environment (such as year-round no-take zones and a closure duration that accorded with African penguin life history). However, the panel recommendations as to how to determine the extent of these no-take zones were not implemented. Instead, the Minister gave the conservation and fishery sectors five months to reach an agreement on the delineations of no-take zones around the six African penguin breeding colonies in issue. In the absence of an agreement, the interim no-take zones would be instituted from the start of the 2024 small pelagic fishing season until the end of 2033. No consensus was reached between the sectors following the announcement by the Minister; therefore, the interim no-take zones (now the current no-take zones) were reflected in the small pelagic fishing permits issued in January 2024.

In this study, we use the trade-off mechanism recommended by IRPR to identify no-take zones that balance the objectives of minimizing costs to fisheries and maximizing benefits to penguins. We consider a set of candidate no-take zones that have been proposed since the ICE, which includes the current no-take zones and for which fisheries cost data are available. To quantify penguin benefits, we defined a utility score that measures how well the penguins’ preferred foraging areas are represented in each no-take zone. Based on the results of our trade-off analysis, we recommend the implementation of no-take zone(s) that represent an optimal compromise between conservation and the fishing industry at each of the six colonies.

Methods

Calculating foraging ranges and core foraging areas

Marine habitat use by African penguins was assessed using telemetry data. Adult penguins attending chicks <6 weeks old were equipped with GPS loggers following established procedures for the species (e.g. Pichegru et al. 2010, 2012, Campbell et al. 2019). Logger deployments were conducted between 2008 and 2022 at six colonies in South Africa: Dassen and Robben islands on the west coast, Stony Point and Dyer Island on the south coast, and St. Croix and Bird islands on the east coast (Fig. 1). Procedures for validating tracks are explained in detail in the supplementary material.

Figure 1.

Maps showing the location of the six African penguin breeding colonies included in the assessment. The national map includes the South African Exclusive Economic Zone (EEZ), the purse-seine fishery regional catch zones used in this assessment, Z1–Z3 (vertical dashed lines), and the distribution of purse-seine anchovy and sardine catches between 2011 and 2020. The insets (a)–(c) show the current purse-seine no-take zones around each colony, existing fisheries restriction zones in marine protected areas (MPA), and the distribution of anchovy and sardine catches. The current no-take zone around Dyer Island is a split-zone where no fishing is allowed inside the solid line, but vessels <26 m can fish between the dashed and solid lines.

To identify important marine habitat for African penguins, i.e. areas mostly available to penguins around each colony (foraging range) and areas of intense or preferred use around each colony (core foraging area), methods developed to identify mIBAs were used (Lascelles et al. 2016, Beal et al. 2021) using the R package track2KBA (Beal et al. 2021). This approach has been used widely to identify mIBAs for seabirds (e.g. Soldatini et al. 2022, Carr et al. 2023, Correia et al. 2024), including penguins off the Antarctic Peninsula (Dias et al. 2018), and to assess the adequacy of marine protected areas (MPAs) for marine predators, including penguins, in the Southwest Atlantic Ocean (Handley et al. 2020). We used a 90% kernel utilization distribution contour (UD) to calculate foraging ranges (mIBA-UD90) following Börger et al. (2006) and a 54% UD to determine core foraging areas using optimal isopleth value selection (Vander Wal and Rodgers 2012). For the core foraging areas (mIBA-ARS), we adopted the IRPR’s recommended use of ARS for selection of the smoothing parameter (h), and for foraging ranges an h value of 7 km was used following Dias et al. (2018), who adopted this method for Chinstrap (Pygoscelis antarcticus) and Adélie (P. adeliae) penguins in the South Shetland and South Orkney islands. Several individuals had multiple tracks to and from the colony before the GPS loggers were retrieved. For these individuals, we combined multiple tracks into one foraging trip to avoid issues associated with site fidelity and pseudoreplication, following Dias et al. (2018). Details of the procedures followed using these methods are explained in the supplementary material.

The IRPR proposed trade-off mechanism

The IRPR recommended a trade-off mechanism to select purse-seine fishery no-take zones that equitably balance the competing objectives of minimizing fisheries costs and maximizing benefits to African penguins (Punt et al. 2023). The logic underlying the mechanism can be understood by considering how to choose between two zones, call them A and B. If A and B have similar values in one objective (e.g. fisheries costs), but A has substantially better values in the other objective (e.g. penguin benefits), then A is to be chosen. If A has substantially lower fisheries costs than B, but B has substantially higher penguin benefits than A, then the choice depends on the relative magnitude of these differences. The IRPR mechanism determines which scenario should be chosen in these instances, by plotting the fisheries costs against penguin benefits for all the no-take zones considered and fitting a trade-off curve to these points (details below). The zone achieving the most equitable compromise is that closest to the ‘balance point’ on the trade-off curve, where the rate of increase in fisheries costs equals the rate of increase in penguin benefits.

The trade-off mechanism requires that costs to fisheries and benefits to penguins are measured on relative scales with the values for each ranging between zero and one (otherwise, the location of the balance point depends on arbitrary units of measurement). We defined the cost to fisheries as the proportion of total local catch that would be lost through fishing exclusion from each no-take zone; here, a value of 1 represents the maximum catch loss attributed to the largest candidate no-take zone, which in all instances is the foraging range, i.e. mIBA (UD90). The catch loss values we used were the median cost estimates from the OBM, where each fishing set could be re-used five times (Bergh and Horton 2023, Punt et al. 2023). The effects of closures on economic factors, such as job losses, although estimated using a Social Accounting Matrix (Urban Econ Development Economists 2023), were not used in these analyses as they were calculated at a regional scale, were directly derived from the OBM outputs and were assessed as having significant limitations in their modelling assumptions (Punt et al. 2023).

We defined the benefits to penguins as a utility score (⁠|${U}_R)$| for each no-take zone R:

$$\begin{eqnarray} {U}_R = \frac{{\mathop \sum \nolimits_{p \in R} {u}_p}}{{\mathop \sum \nolimits_{p \in FR} {u}_p}}. \end{eqnarray}$$

Here, |${u}_p$| is the number of penguins that regularly forage in pixel |$p$|⁠, estimated using package track2KBA in program R, and |$\mathrm{ FR}\ $| is the total foraging range. The utility score quantifies how well each no-take zone represents the total foraging range at the colony (i.e. a value of 1 on the x-axis in the trade-off space), whilst accounting for the fact that penguins do not utilize all regions of space equally. Lastly, we note that although track2KBA uses the local population size to calculate |${u}_p$|’s, our utility score is independent of this parameter as it is common to both numerator and denominator.

To construct trade-off curves, we plotted benefit-cost points for all no-take zones (at a given colony), as well as points representing the industry-optimized scenario (fisheries cost = penguin benefit = 0) and the penguin-optimized scenario (fisheries cost = penguin benefit = 1). We then find the convex hull of all these points and construct the trade-off curve by fitting a monotone non-decreasing spline to the points on the boundary of the convex hull. The ‘balance point’ can then be found by numerically evaluating where the slope of the trade-off curve equals one, that is, where the rate of increase in penguin benefits equals the rate of increase in fisheries costs.

Trade-off plots are provided for four catch types [anchovy, sardine, sardine bycatch, and redeye (Etrumeus whiteheadii)] for six of the last remaining large African penguin colonies: Dassen and Robben islands off the west coast, Stony Point and Dyer Island off the south coast, and St. Croix and Bird islands off the east coast (Fig. 1). At each colony, we considered the following no-take zones proposed since the ICE was initiated in 2008: (i) African penguins’ foraging range (mIBA-UD90 described above), (ii) African penguins’ core foraging area (mIBA-ARS described above), (iii) 20 km no-take zones used during the ICE, (iv) no-take zones proposed by DFFE in 2021, (v) no-take zones proposed by the CAF, and (vi) the current no-take zones where corresponding fisheries cost data were available.

Lastly, to provide context on the relative catch of different small pelagic stocks taken around each colony, we provide the average yearly catches (DFFE, unpubl. data) of each stock caught between 2011 and 2020 for the entire South African EEZ and three regions: (i) west coast, west of Cape Point (Z1, Fig. 1), (ii) south coast between Cape Point and Cape Agulhas (Z2, Fig. 1), and (iii) east coast east of 24°E (Z3, Fig. 1). For each catch type and for all colonies, we further calculated the estimated catch loss (using the OBM outputs, see above) for the African penguins’ foraging range (mIBA-UD90), and the no-take zone that best approximated the balance point on the trade-off curve or the one that was motivated for as the preferred no-take zone. The distribution of sardine and anchovy catches between 2011 and 2020 using data provided by DFFE is shown in Fig. 1.

Results

Core foraging areas (mIBA-ARS) and foraging ranges (mIBA-UD90) were determined using 1332 complete tracks from 1097 individuals after applying filtering procedures for African penguins from six colonies between 2008 and 2022 (Table 1).

Table 1.

Summary of tracking data from six African penguin colonies showing the total sample size over the specified sampling period.

	Sample period		Sample size (N)
Colony	Year range	Month range	Individuals*	Tracks*
Dassen Island	2008–2019	March–October	136	156
Robben Island	2008, 2010–16, 2018, 2019	April–August	185	207
Stony Point	2017–2022	June–October	60	99
Dyer Island	2009–2012, 2021,2022	May–August	61	64
St. Croix Island	2009–2018	March–July	191	197
Bird Island	2008–2019, 2021,2022	March–July	464	609

*Sample sizes are shown as the number of individual birds tracked and the number of tracks (where there might be more tracks than individual birds if a bird went on more than one foraging trip before the tracking device was removed). These totals represent the sample size after applying filtering procedures (see text for details).

Trade-off assessments

Dassen Island

Penguin utility scores were generally high (⁠|${U}_R$| > 0.83) for all no-take zone types around Dassen Island (Fig. 2). Anchovy constituted the largest catch around this colony although estimated anchovy catch losses for the African penguin foraging range (mIBA-UD90) and the core foraging area (mIBA-ARS), respectively, were ∼4.7% and ∼1.6% of the regional catch and ∼3.7% and ∼1.3% of national catches (Table 2, Fig. 2). The current no-take zone (= DFFE 2021) was most consistently aligned to the balance point in the trade-off space for all stocks, although it omitted 8% of the northern core foraging area (mIBA-ARS) (Fig. 2). There were no catch losses associated with any of the scenarios for sardine bycatch (Table 2).

Figure 2.

Purse-seine fishery no-take zones around Dassen Island proposed since 2008, including African penguin tracks and the trade-off assessments for different small pelagic fish stocks. The trade-off plots include fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curves are shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries.

Table 2.

Estimated national and regional purse-seine catch losses estimated by an OBM (Bergh and Horton 2023) for different no-take zones for six African penguin colonies in South Africa.

				Estimated % catch loss
				Foraging range		Core foraging area		Current no-take zone		Recommended no-take zone
Colony	Catch type	National catch (kt)	Regional catch (kt)	National	Regional	National	Regional	National	Regional	National	Regional
Dassen Island	Anchovy	220	174	3.7	4.7	1.3	1.6	0.5	0.6	1.3	1.6
	Directed sardine	59	4.4	0.4	5.3	0.1	1.9	0.2	2.3	0.1	1.9
	Redeye	47	34	0.9	1.2	0.2	0.2	<0.1	<0.1	0.2	0.2
	Bycatch sardine	10	8.5	0	0	0	0	0	0	0	0
Robben Island	Anchovy	220	174	2.0	2.5	0.6	0.8	0.1	0.1	0.6	0.8
	Directed sardine	59	4.4	0.2	2.8	<0.1	0.2	<0.1	<0.1	0.0	0.2
	Redeye	47	34	0.5	0.7	0.2	0.3	0	0	0.2	0.3
	Bycatch sardine	10	8.5	0	0	0.45	0.53	0	0	0.45	0.53
Stony Point	Anchovy	220	40	1.0	5.3	0.1	0.6	0	0	0.1	0.6
	Directed sardine	59	34	9.3	16.2	1.8	3.1	<0.1	<0.1	1.8	3.1
	Redeye	47	12	1.1	4.2	0.2	0.8	<0.1	<0.1	0.2	0.8
	Bycatch sardine	10	1.4	1.1	8.4	<0.1	<0.1	0	0	<0.1	<0.1
Dyer Island	Anchovy	220	40	13.6	75.4	6.5	36.4	No data	No data	2.4	13.1
	Directed sardine	59	34	32.7	56.5	8.4	14.5	No data	No data	4.2	7.3
	Redeye	47	12	5.1	19.8	2.1	8.1	No data	No data	0.9	3.5
	Bycatch sardine	10	1.4	7.6	56.4	1.8	13.4	No data	No data	1.1	8.1
St. Croix Island	Directed sardine	59	4.2	5.1	70.9	3.0	41.7	No data	No data	2.1	28.7
Bird Island	Directed sardine	59	4.2	0.2	2.5	0.1	0.8	0.1	0.9	<0.1	0.5

Robben Island

The current no-take zone (=DFFE 2021), which is part of an existing MPA declared in 2019, had relatively low penguin utility scores (⁠|${U}_R$|= 0.42) and very low to negligible catch losses (<0.15% of the regional catch) for all stocks around Robben Island (Table 2, Fig. 3). By contrast, the mIBA-ARS no-take zone had high penguin utility scores (⁠|${U}_R$|= 0.86) and the closest alignment to the balance point in the decision space for anchovy and sardine. As at Dassen Island, anchovy constituted the largest catch around this colony but estimated catch losses for the African penguin foraging range (mIBA-UD90) and the core foraging area (mIBA-ARS) were ≤2.5% of the regional catch and <2% of the national catch (Table 2, Fig. 3). There were no clear alignments of any of the no-take options to the balance point in the trade-off space for redeye; redeye catch losses were very low (∼0.7% of the regional catch) for the foraging range (mIBA-UD90) of African penguins from this colony (Table 2, Fig. 3). OBM outputs for sardine bycatch catch loss produced counter-intuitive outputs with the largest (mIBA-UD90) and smallest (current) no-take zones producing no catch losses despite intermediate no-take zones having positive catch loss estimates; these estimates were assessed as being erroneous and were discarded from the trade-off assessment (Table 2).

Figure 3.

Purse-seine fishery no-take zones around Robben Island proposed since 2008, including African penguin tracks and the trade-off assessments for different small pelagic fish stocks. The trade-off plots include fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curves are shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries.

Stony Point

The current no-take zone (=CAF) had low penguin utility scores (⁠|${U}_R$|= 0.34) and mostly negligible catch losses (<0.1% of regional catches) for all stocks fished around Stony Point (Table 2, Fig. 4). Sardine was the largest catch, followed by anchovy (Table 2, Fig. 4). The no-take zone based on the core African penguin foraging area (mIBA-ARS) had a high penguin utility score (⁠|${U}_R$|= 0.91), was closely aligned to the balance point in the decision space for all fish stocks around this colony, and would result in estimated regional catch losses of ∼3.1% for sardine, ∼0.6% for anchovy, and ∼0.8% for redeye (Table 2, Fig. 4).

Figure 4.

Purse-seine fishery no-take zones around Stony Point proposed since 2008, including African penguin tracks and the trade-off assessments for different small pelagic fish stocks. The trade-off plots include fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curves are shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries.

Dyer Island

The DFFE (2021) no-take zone had relatively high penguin utility scores (⁠|${U}_R$|= 0.78) and was most closely aligned to the balance point in the decision space for all stocks around Dyer Island (albeit less so for redeye than the others) (Fig. 5). Anchovy and sardine constituted the largest catches although the relative contribution of the estimated regional catch losses for the DFFE (2021) no-take proposal was ∼13.1% for anchovy and ∼7.3% for sardine around Dyer Island (Table 2, Fig. 5). However, the current fully restricted no-take zone around Dyer Island had a low penguin utility score (⁠|${U}_R$|= 0.32). Although there were no corresponding costs available for this and the partially restricted zone around this colony, sardine and anchovy catch locations were primarily concentrated to the north of the DFFE (2021) no-take proposal with very few catches occurring within the current fully restricted no-take zone (Fig. 1).

Figure 5.

Purse-seine fishery no-take zones around Dyer Island proposed since 2008, including African penguin tracks and the trade-off assessments for different small pelagic fish stocks. The trade-off plots include fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curves are shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries. The current no-take zone is a split zone with the Industry CAF zone fully restricted and the no-take zone aligned to the DFFE (2021) zone is open to fishing to vessels <26 m. The vertical dashed line in the trade-off plots denotes the penguin utility scores for the fully restricted area (Industry CAF); catch loss data was not available for the current no-take zone.

St. Croix Island

The DFFE (2021) no-take zone was most closely aligned to the balance point in the decision space for sardine around St. Croix Island with a relatively high penguin utility score (⁠|${U}_R$|= 0.74) (Fig. 6). The current no-take zone included 50% of the African penguin core foraging area (mIBA-ARS) and had a utility score of 0.61. There were no corresponding fisheries catch loss data available for the current no-take zone. Sardine was the only major stock caught around this colony by the purse-seine fishery with estimated lost catch attributed to the African penguin foraging range (mIBA-UD90) amounting to a substantial proportion (∼70.9%) of the regional catch, while the estimated catch loss associated with the DFFE (2021) no-take option, while still meaningful, was less than half of that at ∼28.7% of the regional catch (Table 2, Fig. 6).

Figure 6.

Purse-seine fishery no-take zones around St. Croix Island proposed since 2008, including African penguin tracks and the trade-off assessment for sardine. The trade-off plot includes fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curve is shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries. The current no-take zone = Industry ETT. The vertical dashed line in the trade-off plot denotes the penguin utility score for the current no-take zone; catch loss data was not available for the current no-take zone.

Bird Island

There was very little purse-seine fishing effort around Bird Island (Fig. 1). The ICE 20 km no-take zone had a high penguin utility score (⁠|${U}_R$|= 0.96) and was closely aligned to the balance point in the decision space for this colony with very little impact on estimated regional sardine catch loss (∼0.5%, Table 2, Fig. 7). The current no-take zone (CAF) has a marginally lower penguin utility score (⁠|${U}_R$|= 0.93) but results in almost double the catch loss (∼0.9% of estimated regional catch) than the ICE 20 km no-take zone (Table 2).

Figure 7.

Purse-seine fishery no-take zones around Bird Island proposed since 2008, including African penguin tracks and the trade-off assessment for sardine. The trade-off plot includes fishery no-take zone options relative to their corresponding benefits to penguins (penguin utility score) and costs to fisheries (catch loss). The trade-off curve is shown with the balance point at which an increase in the rate of penguin benefits is equal to the rate in increase in estimated catch loss to fisheries. The current no-take zone = CAF.

Discussion

Urgency to implement evidence-based conservation interventions

African penguins are recognized as one of three priority penguin species most in need of urgent conservation action (Boersma et al. 2019), and the at-sea distribution of African penguins overlaps with the greatest intensity of purse-seine fishing pressure experienced by any penguin species (Gimeno et al. 2024). The sustainable management of sardine and anchovy stocks has been highlighted as a priority for African penguin conservation efforts (see reviews by Trathan et al. 2015, Boersma et al. 2019). This is partly attributed to the historical correlation between sardine stock and African penguin population collapses following excessive reductions in sardine by purse-seine fishing off Namibia during the 1960s and 1970s (Crawford 2007) with no recovery, and the observed importance of sardine biomass for the survival of adult African penguins in South Africa (Robinson et al. 2015). Sardine and anchovy stocks are also important to several other Benguela predators, including Endangered Cape gannets (Morus capensis) and Cape cormorants (Phalacrocorax capensis) (Crawford et al. 2022), and ecologically and commercially important linefish species, such as silver kob (Argyrosomus inodorus), yellowtail (Seriola lalandi), and geelbek (Atractoscion aequidens) (Parker et al. 2020).

In recent years, the sardine stocks in South African waters have been considerably below the long-term average biomass; spawning biomass was below 800 000 tonnes (∼20% of its maximum recorded value) in 16 of 17 years between 2006 and 2022 (Coetzee et al. 2022a) despite this being a threshold at which ‘recruitment from significantly smaller stock biomasses will be likely to be greatly reduced, resulting in prolonged depletion of the stock with limited potential for recovery’ (Punt et al. 2023). Sardine exploitation rates by the purse-seine fishery have also been high off western South Africa in some years during this period, reaching >70% of the western sardine stock biomass in 2016 (de Moor 2023), which was less than 200 000 tonnes (15% of its maximum recorded value) at the time (Punt et al. 2023). In a global context, this exploitation rate exceeds the exploitation threshold recommended by the Lenfest Forage Fish Task Force for lower trophic level species (Pikitch et al. 2012). Moreover, the availability of both sardine and anchovy to seabirds appears to have declined since the early 2000s (Crawford et al. 2019). This is especially concerning for African penguins, whose breeding success and survival are known to respond negatively to periods where prey biomass falls below 33% of its maximum recorded value (Cury et al. 2011, Robinson et al. 2015). A further consideration is the need for cautious management of the forage fish themselves when dropping below 18% of maximum recorded stock sizes as seabird-induced predation mortality is considerably higher at such low stock levels (Saraux et al. 2020). Given that negative impacts of resource competition are often amplified for seabirds during periods of low prey biomass (Sydeman et al. 2017), it is vital that all efforts to manage the impact of prey depletion are afforded due and timely consideration. Since the inception of the ICE in 2008, South Africa’s African penguin population has decreased by 66%. Following the publication of the results that supported demographically meaningful benefits of no-take purse-seine fishing zones for African penguins in 2018 (Sherley et al. 2018), their population has declined by 44%. Colony extinction probabilities are inversely related to colony population size for African penguins with extinction risk disproportionately elevated for colonies <1000 breeding pairs (Crawford et al. 2001). Populations at all but three of South Africa’s breeding colonies were below that threshold in 2023 (Sherley et al. 2024) highlighting the urgency to prioritize conservation efforts at the last remaining large colonies. If the declines observed over the last decade are replicated over the next 10 years, then there is a credible risk of this species becoming extinct in the wild by 2035 (Sherley et al. 2024). As a result, there is no longer time available to delay implementation of meaningful, science-led, no-take zones.

Small pelagic fishing restrictions around the proposed African penguin colonies are likely to benefit Endangered Cape cormorants who also predominantly prey on small pelagic fish and have similar foraging ranges to African penguins during their breeding season (Hamann et al. 2012, Pichegru et al. 2009). Four of the largest Cape cormorant colonies are situated at four of the six proposed no-take zones at Dassen, Robben, and Dyer islands and at Stony Point (Crawford et al. 2016). Implementation of measures to improve the population status of African penguins around these colonies will also sustain benefits associated with prey facilitation by African penguins to Cape cormorants and other volant seabird species that are known to take advantage of prey elevated from depths (often >33 m) to the surface by African penguins (McInnes and Pistorius 2019).

Trade-off assessments and current no-take zones

Our results highlight shortfalls with the current no-take zones in protecting the preferred foraging areas for African penguins throughout their range. Four of the six current no-take zones include ≤50% of the African penguins’ core foraging areas (Table 3). Although both no-take zones at Dassen and Bird islands cover >90% of their core foraging areas, only the one at Dassen Island sits close to the balance point on the trade-off between protecting penguins and minimizing costs to the fishing industry. Moreover, the current no-take zones around Robben Island and Stony Point significantly underrepresent important foraging areas while having little to negligible costs to the fishing industry. The zone around Robben Island is, in fact, no improvement on the no-take fishing zone implemented when the Robben Island Marine Protected Area was gazetted in 2019, a designation that pre-dated the IRPR (Punt et al. 2023) and was based on a systematic conservation planning process incorporating a diversity of biodiversity features (Kirkman et al. 2021) and not focusing on African penguins in particular. The current no-take zone around Stony Point also pre-dates the IRPR and originated in the CAF process during 2022 as the preferred no-take option of the fishing industry. In contrast to these current inadequate no-take zones, the designs using the mIBA-ARS method recommended by the IRPR represent a substantial improvement in penguin benefits and closely approximate the balance points in the trade-off space for these two colonies.

Table 3.

Proportion (%) overlap of the current purse-seine fishery no-take zones around six African penguin colonies with African penguin core foraging areas (mIBA-ARS) and foraging ranges (mIBA-UD90).

		% overlap
Island	Current area (km2)	Core foraging area (mIBA-ARS)	Foraging range (mIBA-UD90)
Dassen Island	948	92.4	66.2
Robben Island	227	43.4	22.3
Stony Point	84	30.2	10.0
Dyer Island	260	20.8	12.5
St. Croix Island	581	50.0	29.7
Bird Island	826	97.8	59.0

Estimates of catch loss were not available for the current no-take zones around St. Croix and Dyer islands. Nevertheless, based on relative penguin utility scores, the configuration of these no-take zones is inadequate regarding penguin benefits. Penguins from St. Croix Island frequently travel >30 km from this colony, and the 20 km no-take zone in place during the ICE has been shown to be inadequate for these birds (Fig. 6, Pichegru et al. 2012). The current no-take zone around St. Croix Island is even smaller than the 20 km zone, covers only half their core foraging area (Table 3), and is therefore unlikely to provide sufficient benefits to penguins from this colony. For Dyer Island, the current configuration includes a split zone with the total prohibition on purse-seine fishing limited to a relatively small area inshore, where there has been very little fishing in recent years (Fig. 1, Bergh and Horton 2023) and where benefits to penguins are small. This area was the fishing industry’s no-take option proposed during the CAF. The larger partial no-take zone around Dyer Island, which aligns to the DFFE (2021) zone is accessible to vessels <26 m in length. The relative benefits of this split-zone configuration to African penguins could not be assessed, but any benefits could be nullified by the current configuration, given that there is still pressure from the purse-seine fishing fleet. The DFFE (2021) no-take configurations outperform other available configurations for these two colonies (Dyer and St. Croix) with the closest approximation in the trade-off decision space for minimizing costs to industry and maximizing benefits for African penguins.

Currently Bird Island is afforded meaningful protection from purse-seine fishing but, paradoxically, the waters around this colony are least attractive for fishers compared to the other colonies (Fig. 1, Bergh and Horton 2023). Despite the benefits of the current no-take zone, the trade-off assessment identified the 20 km no-take zone as best in terms of balancing costs to the purse-seine fishery while still affording African penguins greater benefits.

The current no-take zone around Dassen Island corresponds with the best approximation of the balance point in the trade-off decision space for anchovy and, based on this assessment alone, is the rational choice for maximizing benefits to penguins whilst minimizing costs to industry. However, the exclusion of the northern parts of the core foraging area in the current no-take zone is significant for several reasons. First, it forms part of the African penguins’ core foraging area proximate to the Dassen Island breeding grounds. Second, small-pelagic fishing concentrated in this northern portion of the mIBA-ARS is likely to have downstream effects (i.e. fishing the line, e.g. Pichegru et al. 2012) on prey availability for African penguins in the remainder of their core foraging area due to the inshore southward movement of anchovy recruits during autumn/winter (Hampton 1987). These months correspond with the most important breeding season for penguins from this colony (Wolfaardt et al. 2009b). Therefore, the model predicting utility scores underperforms in terms of its ability to incorporate the benefits of the northern areas of Dassen Island to African penguins. Considering this discrepancy and that this colony is the only one that still held >2000 breeding pairs in 2023, we would argue that the mIBA-ARS no-take zone is more likely to yield biologically meaningful protection and should be the preference for this colony. Although this could slightly increase costs for the industry, the estimated losses from the OBM are still <2% of regional catches (Table 2) and are likely overestimates compared to actual losses (Punt et al. 2023). Moreover, ignoring an opportunity to optimize the efficacy of this zone (within the trade-off space identified) could mean it fails to yield a conservation benefit, in which case, any losses for the industry would have been needless.

Broader conservation management considerations

There are several broader considerations for conservation management of both African penguins and sardine and anchovy stocks that are relevant for improving the application of the trade-off mechanism and use of no-take fishing zones as a conservation measure. This was recognized by the IRPR, which provided recommendations for a monitoring and evaluation programme to assess the adequacy of no-take zones implemented using the trade-off mechanism as well as recommendations for additional research to improve data inputs (Punt et al. 2023). The South African government has committed to adopt a monitoring programme with a review of no-take zones in 6 years. In terms of the IRPR’s recommended approach, various parameters should be evaluated, including breeding population counts, breeding success (chick condition and chick survival), foraging performance (using telemetry data and automated penguin monitoring system data), recruitment and adult survival. The results of this monitoring and evaluation programme will determine if the no-take zone configuration that is ultimately implemented within this 6-year period is fit-for-purpose in terms of limiting resource competition by the purse-seine fishery.

The IRPR trade-off mechanism aims to reduce subjectivity in selecting balanced closure options. Once specific quantitative measures of cost and benefit are chosen, the method determines which scenario is to be preferred without the need for further judgment. However, subjectivity remains in the choice of these measures of cost and benefit. For example, measuring penguin benefits in terms of a demographic variable e.g. adult survival could result in the selection of different no-take zone scenarios. However, precise estimates of demographic or population-level variables that would result under each no-take zone are not available. Therefore, our choice has been dictated by the available data on fisheries costs and the penguin utility scores based on existing tracking data.

Among specific issues to be considered are those relating to improving the data inputs used in the trade-off mechanism. The first of these relates to ongoing refinements relating to tracking data. The proposed no-take zones in this assessment use tracking data recorded from African penguins attending small chicks. Currently, there is no available tracking data for this species during incubation during which stage several penguin species are known to travel further from their colonies, e.g. for chinstrap (P. antarcticus, Dias et al. 2018), northern rockhopper (Eudyptes moseleyi, Steinfurth et al. 2020), and Magellanic (Spheniscus magellanicus, Dee Boersma and Rebstock 2009) penguins. Moreover, adults attending older chicks during the crèche period are also likely to forage for longer durations than during the brood phase [e.g. for southern rockhopper penguins (Eudyptes chrysocome), Tremblay and Cherel 2005]. As such, the proposed restriction zones are likely conservative in their estimation of preferred foraging areas during the breeding season.

An important data-refinement during the monitoring period is obtaining more accurate measures of fisheries costs. If the IRPR was correct in suspecting an over-estimation of costs, future trade-off assessments using more accurate estimates of fisheries costs and foraging areas aligned to the foraging distributions of African penguins during all breeding stages, including buffer areas to mitigate against fishing-the-line effects (anticipated after a 6-year monitoring period), may reveal a different set of improved priority candidate no-take zones than the ones proposed here. Nevertheless, following the IRPR’s recommended trade-off mechanism using the data currently available demonstrates the inadequacy of the current no-take zone extents in protecting core foraging areas during chick rearing and with little to no cost to fisheries in many instances. The current no-take zones have not been delineated using the IRPR’s recommendations for reasons that are unclear. However, these no-take zones, based on a range of historic delineation proposals are likely ineffective and, at the very least, should be substituted by the proposed no-take zones in this assessment until more representative data on costs to fisheries and benefits to penguins can be achieved. Doing so is an approach consonant with reliance on best available scientific data, application of the principle of precaution, and both international and domestic best practice.

A final consideration is the role of the no-take zones as part of a broader set of conservation interventions to prevent African penguin declines and to consider the population health of sardine and anchovy biomass as a key element of the Benguela Upwelling System as a whole. The demographic benefits attributed to African penguins due to no-take zones during the ICE were during a period (2008–2020) of below-average sardine biomass, which has persisted in recent years (Coetzee et al. 2022b). Notwithstanding the need to limit resource competition around breeding colonies, no-take zones of this kind should not be viewed in isolation of other priority measures linked to the sustainable management of small pelagic fish. There are various potential avenues to improve the latter including those recommended by the IRPR. Firstly, the Operational Management Procedures should be tested to assess if they are adequately sensitive to the sustainable management of sardine stocks (Punt et al. 2023). Secondly, the harvest control rules (HCRs) that form part of the OMP process should be reviewed in terms of their suitability to seabirds that predominantly prey on small pelagic fish and the longer-term harvests of fisheries. Alternative HCRs, as proposed by Koehn et al. (2021), which provide mutually acceptable outcomes for seabirds and fisheries, should be evaluated in the context of the South African small pelagic fishery. Lastly, the spatial management of the small pelagic fishery should be reviewed to incorporate the distribution of African penguins during the non-breeding season to mitigate resource competition during crucial life-history stages such as post-moult.

Conclusion and recommended no-take zones

This study focuses on the critical role and efficacy of no-take zones in tackling one of the most significant contributors to the crisis facing the African penguin, which is the competition between seabirds and industrial-scale fisheries for common prey. The results of our study, using the trade-off mechanism recommended by an international expert panel, have demonstrated that the current no-take zones in place around six of the largest African penguin colonies off the coastline of South Africa, are inadequate for addressing the biological needs of the species they are intended to serve and, in most instances, incur little to no cost to the fishing industry (Table 4). Four African penguin colonies currently have no-take zones that restrict purse-seine fishing in ≤50% of the African penguins’ core foraging areas, and the current no-take zone around Dassen Island excludes a critical area of their core foraging area to its north. The results of this assessment have shown that there are substantially improved fit-for-purpose no-take configurations available that can be implemented in the short-term to urgently address the negative impacts of resource competition on African penguins. Accordingly, we recommend that the design of the no-take zones be based on the best trade-off possible, within the limitations of available data. These are the mIBA-ARS no-take zones for Dassen Island, Robben Island, and Stony Point; the DFFE (2021) no-take zones for Dyer and St. Croix islands; and lastly, the 20 km no-take zone for Bird Island.

Table 4.

Summary of current protection offered to African penguins through purse seine fishing no-take zones, estimated trade-off between protection for penguins and estimated catch loss for purse seine fishery, and recommended no-take zone based on this assessment.

Colony	Currently implemented extent of protection through area closure to purse seine fishing	Source of current no-take zone	Comment on extent of balanced conservation-fishery trade-off	Recommended no-take zone based on this assessment
Dassen Island	More than 90% of core foraging area protected	DFFE (2021)	Balanced overall but note the exclusion of northern area detrimental to penguin foraging, and that the current closure represents no sardine bycatch loss	mIBA-ARS
Robben Island	Less than 50% of core foraging area protected	Robben Island MPA (2019), DFFE (2021)	Limited protection with negligible costs to fishery, existing MPA (pre-dating IRPR)	mIBA-ARS
Stony Point	Less than 33% of core foraging area protected	CAF	Limited protection with negligible loss to fishery	mIBA-ARS
Dyer Island	20% of core foraging area protected	DFFE (2022)	Small benefits to penguins with minimal cost to fishery	DFFE (2021)
St. Croix	50% of core foraging area protected	ETT Industry	Limited protection and uncertain degree of loss to fishery	DFFE (2021)
Bird Island	Almost all core foraging area protected	CAF	Good protection for penguins which exceeds trade-off extent of 20 km radius; uncertain loss to fishery but catches are minimal here	20 km

Acknowledgements

We would like to thank CapeNature, South African National Parks, and Robben Island Museum for allowing us to conduct field work in African penguin colonies and for facilitating vessel transfers to island colonies. We would like to thank all the stakeholders and field assistants who participated in the numerous processes over the years.

Author contributions

Conceptualisation: A.M.M., E.A.W., T.C.-K., M.C., A.K., C.L., L.S., R.B.S., C.S., L.W.; data collection: A.M.M., E.A.W., T.C.-K., P.B., J.S.G., C.H., K.L., A.M., L.P., R.B.S., A.S., L.U., L.W.; data analyses: A.M.M., E.A.W., T.C.-K., M.C.; drafting and review of manuscript: all authors.

Conflict of interest

AMM, EAW, KD, CH, KL, NS are, at the time of publication, involved in litigation against the Minister in the South African Department of Forestry, Fisheries and Environment (DFFE) and other respondents, arguing for the extension of no-take zones around six African Penguin island colonies off the coast of South Africa. MC, RBS have submitted affidavits as expert witnesses in the above-mentioned litigation. LP signed a confirmatory affidavit filed in the above-mentioned case against the Minister of DFFE and others, corroborating the contents of the founding affidavit in this matter insofar as they pertain to LP’s activities and knowledge relating to African Penguins. Nelson Mandela University is not a party to the court application. CS is involved in independent assessments of the ecological sustainability of fisheries for the purposes of the World Wide Fund for Nature – South Africa (WWF) Southern African Sustainable Seafood Initiative (WWF-SASSI). The same author signed a confirmatory affidavit filed in the above-mentioned case against the Minister of DFFE and others, corroborating the contents of the founding affidavit in this matter insofar as they pertain CS’s activities and knowledge relating to African Penguins. WWF is not a party to the court application. LW signed a confirmatory affidavit filed in the above-mentioned case against the Minister of DFFE and others, corroborating the contents of the founding affidavit in this matter insofar as they pertain LW’s activities and knowledge relating to African Penguins. The Endangered Wildlife Trust is not a party to the court application. AM, CL, AK, LU regard the results presented in this article as a product of ongoing collaborative research and declare that the content of the article is not intended to represent the official position or views of the wider institutions/agencies/departments to which the authors belong or are employed by. TC-K, PB, JSG, LS, AS have no conflict of interest to declare.

Funding

Funds for this research were provided by the Charl van der Merwe Trust.

Data availability

All code for this assessment as well as the data for the trade-off analyses can be accessed via Github: https://github.com/amcinnes723/African-Penguin-No-take-zones. Tracking data for inputs into the code for the delineation of core foraging areas (mIBA-ARS) and foraging ranges (mIBA-UD90) can be accessed with permission from data providers for each respective colony at BirdLife International's Seabird Tracking Database: Dassen Island https://data.seabirdtracking.org/dataset/2212, Robben Island https://data.seabirdtracking.org/dataset/2213, Stony Point https://data.seabirdtracking.org/dataset/2214, Dyer Island https://data.seabirdtracking.org/dataset/2217, St Croix Island https://data.seabirdtracking.org/dataset/2215 and Bird Island https://data.seabirdtracking.org/dataset/2216.