Catch composition and sustainability of the marine aquarium fishery in Kenya

Fisheries Research 183 (2016) 19–31

Contents lists available at ScienceDirect

Fisheries Research journal homepage: www.elsevier.com/locate/fishres

Catch composition and sustainability of the marine aquarium fishery in Kenya G.M. Okemwa a,b,∗ , B. Kaunda-Arara b , E.N. Kimani a , B. Ogutu c a b c

Kenya Marine and Fisheries Research Institute, P.O. Box 81651, 80100 Mombasa, Kenya Department of Fisheries and Aquatic Sciences, University of Eldoret, P.O. Box 1125, 30100 Eldoret, Kenya State Department of Fisheries, P.O. Box 90423, 80100 Mombasa, Kenya

a r t i c l e

i n f o

Article history: Received 30 January 2015 Received in revised form 15 April 2016 Accepted 30 April 2016 Handled by George A. Rose Keywords: Coral reef fish Aquarium trade Risk assessment Western Indian Ocean

a b s t r a c t Management of the marine aquarium fishery in Kenya and most of the Western Indian Ocean (WIO) region is challenged by a poor understanding of the status and impacts of the fishery on stocks due to lack of long-term species-specific data sets. In this study, we analyzed commercial catch and effort data over a six-year period (2006–2011), provided by a major exporter, constituting over 70% of the total national catch, in order to assess spatial and temporal variations in catch composition for 11 fishing grounds in coastal Kenya. In addition, a semi-quantitative risk assessment based on Productivity Susceptibility Analysis (PSA) was applied on 102 target species to identify species at risk of overexploitation by the fishery. Between 2006 and 2011, approximately 1.54 million aquarium fish were collected constituting 220 species in 36 fish families. The catch was numerically dominated by Labridae (32%), Pomacentridae (14%), Serranidae (9%), Blenniidae (9%), Scorpaenidae (7%), Pomacanthidae (5%) and Acanthuridae (5%). Thirty-two species made up 80% of the catch with the cleaner wrasse, Labroides dimidiatus and the anthias, Pseudanthias squamipinnis being the most collected with a relative abundance of 11% and 7%, respectively. Multivariate nMDS analysis of the catch composition grouped the fishing grounds into three clusters and mostly influenced by the mode of fishing (snorkeling or SCUBA diving). Detrended Correspondence Analysis (DCA) further showed an association of some species to specific fishing grounds. The PSA showed that 91 species (90%) fell in the high productivity and low susceptibility risk categories placing them at a relatively low risk of depletion by the fishery. However, four species: Pomacanthus maculosus, Pomacanthus chrysurus, Amphiprion allardi, and Amphiprion akallopisos were ranked at high risk, and seven species at moderate risk of overexploitation. The findings highlight the need for closer monitoring of the aquarium fishery in Kenya and the WIO; and institution of precautionary management measures such as area closures and species restrictions to ensure sustainability in the fishery. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The trade in live coral reef aquarium fish represents one of the highest value-added products from coral reefs, bringing a higher economic return than most other reef fishery resources (Olivier, 2001). However, the scope of the marine aquarium fishery sector and associated impacts on livelihoods and aquatic communities are often underappreciated or inadequately represented, particularly within the Western Indian Ocean (WIO) region. An estimated 14–30 million live reef fish are collected annually worldwide for the aquarium trade, having an import value of approximately US$ 200–330 million and a retail value of about US $500 million (Wood,

∗ Corresponding author. E-mail address: [email protected] (G.M. Okemwa). http://dx.doi.org/10.1016/j.fishres.2016.04.020 0165-7836/© 2016 Elsevier B.V. All rights reserved.

2001a; Wabnitz et al., 2003; Rhyne et al., 2012). At least 90% of species exploited for the global marine aquarium trade are collected directly from the reef (Wabnitz et al., 2003; Lecchini et al., 2006). The aquarium trade continues to grow in terms of the diversity of fish species. Over 1800 fish species representing 125 families were reported to enter the aquarium fish trade in the United States, the world’s largest importer of live reef fish (Rhyne et al., 2012), reflecting a 22% increase from a previous global estimate of 1471 fish species representing 50 families (Wabnitz et al., 2003). Kenya is among 45 source countries that supply marine aquarium fish to the global trade and is a major supplier among countries of the WIO region (Wood, 2001a; Bruckner, 2005). The aquarium fishery selectively targets juveniles and small-bodied reef fish of specific species, sizes and often sex; with rare species and those that are difficult to collect commanding the highest prices (Wood, 2001a,b; Sadovy et al., 2002). Most of the species have a relatively

20

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

site-attached juvenile benthic phase which can make them particularly vulnerable to localized depletion (Almany et al., 2007). As a result, the sustainability of the global trade has been questioned with concerns about population declines of some target species, loss of biodiversity, habitat damage and ecological changes (Andrews, 1990; Wood, 2001a,b; Rhyne et al., 2012; Dee et al., 2014). A number of quantitative assessments have showed that targeted fisheries for the aquarium trade can lead to localized depletion of some species and changes in the community structure of fish populations (Kolm and Berglund, 2003; Tissot and Hallacher, 2003; Shuman et al., 2005; Jones et al., 2008). In Kenya, resourceuse conflicts associated with the aquarium fishery have been escalating in the recent past, particularly among those associated with the eco-tourism industry resulting in calls by stakeholders for enhanced regulation and management of the fishery. However, setting management measures for the aquarium fishery has been constrained by limited knowledge on the status of exploited stocks due to lack of species-specific catch data at sufficient spatial and temporal scales, and limited resources to undertake rigorous quantitative stock assessment surveys. Alternatively, semi-quantitative assessment approaches (e.g. Zhang et al., 2009; Hobday et al., 2011; Swaleh et al., 2015) can be useful in assessing data-limited fisheries and directing research and management efforts in data poor situations such as in Kenya. The Productivity Susceptibility Analysis (PSA) framework, originally developed to assess the impacts of Australian prawn fisheries on bycatch species (Milton, 2001; Stobutzki et al., 2001; Hobday et al., 2007), has been widely and successfully adapted to assess other fisheries including elasmobranchs (McCully et al., 2013), tuna (Patrick et al., 2010; Arrizabalaga et al., 2011), deepwater trawl fisheries (Dransfeld et al., 2013), groundfish (Cope et al., 2011), reef fisheries (Micheli et al., 2014) and aquarium fisheries (Fujita et al., 2013). In this study, we use fishery dependent data from Kenya’s marine aquarium fishery to quantify spatio-temporal patterns in catch composition, and apply PSA to identify target species that have a high vulnerability risk to collection pressure. Finally, we infer management implications from the analyses, and propose precautionary measures to enhance the sustainability of the aquarium fishery in Kenya and the WIO region.

et al., 2011). Seasonality in fishing activities is strongly influenced by the northeasterly and southeasterly monsoon winds (see McClanahan 1988 for details). Briefly, the northeast monsoon season (NEM, November–March) is a period of calm seas, elevated sea surface temperatures (SSTs) and higher salinities, while the southeast monsoon season (SEM, April–October) is characterized by rough seas, cool weather, lower salinities and higher productivity. 2.2. Description of the fishery Kenya’s marine aquarium fishery targets diverse organisms including reef fish and invertebrates such as soft corals, sea anemones, brittle stars and starfish, crustaceans (e.g., crabs, hermit crabs, cleaner shrimps), mollusks (e.g., snails, clams), and ‘live rock’. The fishery is solely managed by licensing fishing access for aquarium fishers and exporters. There are currently 144 aquarium fishers and 8 aquarium exporters licensed by the State Department of Fisheries (SDF). The level of experience (number of years in the fishery) of most fishers ranges from 3 to 25 years with a mean of 12 years (G. Okemwa, unpublished data). The supply chain is direct from the fisher to the exporter. There are also unlicensed ‘free lance’ or ‘private’ fishers who fish independently and sell directly to the exporters (G. Okemwa, pers. obs.). The main fishing gears used involve a combination of hand-held scoop nets and barrier nets of various mesh sizes and lengths. The barrier nets consist of a float line at the top edge and pieces of lead attached to the bottom line to keep the net floating upright. Slender metal probes or ‘ticklers’ (∼1 cm in diameter) are also used at times to scare fish out of crevices or when amongst coral branches. The fishers either snorkel when in shallow accessible depths (0.5–3 m) along the coastline or SCUBA dive when targeting species that inhabit deeper depths. Mandatory pre-licensing conditions for aquarium exporters introduced in 2009 include: declaration of all contracted fishers, declaration of intended fishing grounds and submission of monthly catch and export returns. Species listed on CITES including seahorses (Hippocampus spp.) are restricted, in addition to handling or trading in all hard corals. There are no species restrictions or restrictions on gears and fishing effort. However, the use of poisons and/or noxious substances is banned. 2.3. Catch and fishing effort data

2. Methods 2.1. Study area This study assesses spatial and temporal dynamics of aquarium fisher catches from 11 fishing grounds within the lagoonal coral reefs along the Kenya coastlnamely: Shimoni, Diani, Shelly, Nyali, Mtwapa, Kanamai, Kilifi, Shariani, Msumarini, Malindi and Lamu (Fig. 1). The Kenya coastline extends to about 600 km, with a relatively narrow (5–10 km wide) continental shelf with reef systems extending from shallow inshore waters to about 20–25 m depths, while shallow lagoonal reefs are typical at depths of 0.5–3 m along the coastline (GOK, 2009). The southern coast from Msambweni to Malindi (approximately 200 km in distance, Fig. 1) consists of an almost continuous fringing reef system (approximately l00 m–3 km in width from the shore). Fore reef slopes interspersed with patchy reefs are typical northwards from Malindi to the Lamu archipelago (approximately 100 km in distance) and southwards from Msambweni to Shimoni (Obura, 1999). The fishing grounds contain a mosaic of substrates including algal mats, seagrass beds, sand, rubble and rocky platforms (GOK, 2009) and are extensively fished by artisanal and commercial fishers using multiple gear types that include basket traps, gillnets, handlines and spearguns (Samoilys

Data on the total catch, species composition and fishing effort from October 2006 to December 2011 were obtained from aquarium fisher logbooks maintained by the main trader in Kenya (representing approximately 70% of the trade, SDF statistics). Exporters maintain daily catch records of their fishers in logbooks for their own purposes. The logbooks detail the date of fishing, the name of the fisher, fishing ground, fishing method used, species and numbers caught. For analysis of annual trends, the data for 2006 was omitted as it represented only 3 months. Export data were extracted from SDF records summarized in annual statistical bulletins for the entire aquarium fishery. The validity of species names used was checked and updated based on Eschmeyer and Fricke (2016). 2.4. Productivity Susceptibility Analysis (PSA) PSA was carried out following methods described by Stobutzki et al. (2001) and Hobday et al. (2007, 2011). Excel worksheets developed by the Marine Stewardship Council (MSC) (see Marine Stewardship Council, 2010) were used for the assessment. For scoring productivity, defined by Stobutzki et al. (2001) as “the capacity of the stock to recover once the population is depleted”,

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

21

Fig. 1. Map of Kenyan coast showing the location of aquarium fishing grounds and marine protected areas.

eight life-history traits were used: average maximum age, minimum population doubling time fecundity, average maximum size (longevity), average size at maturity, von Bertalanffy growth coefficient (K/year), reproductive strategy and trophic level (Table 1, refer to Appendix Table A1 for definitions). Susceptibility (the potential for the stock to be impacted by the fishery) was scored based on five attributes: availability: overlap of fishing effort with a species distribution; encounter ability: the likelihood that a species will be encountered when a fishing gear/method is used within the geographic range of that species; selectivity: the potential of the gear/method to capture or retain species; the desirability or value of the species; and the ecological niche of the species, defined as the ecological connection between a fish species and its habitat (Table 1, refer to Appendix Table A1 for definitions). In order to increase contrast in scoring for the aquarium fishery, the productivity and susceptibility attributes used in the analysis were adapted from Roelofs and Silcock (2008), Patrick et al.

(2009) and Fishbase (Froese and Pauly, 2013), and modified as shown in Table 1. Scoring of the productivity and susceptibility attributes was based on Hobday et al. (2007, 2011) as: 1(low risk), 2 (medium risk) or 3 (high risk). The attributes selected were limited to those where information for scoring was available for most species from various sources including: Fishbase, published literature (see Appendix Table A2) and grey literature. Where species-specific information was lacking, the productivity characteristics of similar species within the same family were used. In addition, the expert opinion of the authors was used with consultative validation when necessary using key informants from industry representatives (two aquarium fishers having twenty and forty years experience and two exporters) as detailed in the Appendix Table A2. Where species-specific information was lacking, the characteristics of similar species within the same family were used. Among the susceptibility attributes, indicative data was sourced online from two firms that import tropical aquarium fish from

22

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

Table 1 Productivity and susceptibility attributes used in the risk assessment of 102 fish species targeted by the aquarium fishery in Kenya based on Hobday et al. (2007, 2011). Productivity attributes:

Low productivity (high risk, score = 3)

Moderate productivity (moderate risk, score = 2)

High productivity (low risk, score = 1)

Minimum population doubling timec Average maximum agea Measured fecunditya Average maximum sizea Average size at maturitya von Bertalanffy (K)b Reproduction strategyb Mean trophic levelb Susceptibility attributes:

>4 years

1.4–4 years

<15 months

>15 years <1000 eggs >60 cm >40 cm <0.15 Live bearer >3.5 Low susceptibility (Low risk = 1) Widespread (Indo-Pacific) Limited accessibility:>30m Generalist: broad range

5–15 years 1000–15,000 eggs 30–60 cm 15–40 cm 0.15–0.25 Demersal egg layer 2.5–3.5 Moderate susceptibility (Moderate Risk = 2) Spread (WIO region)

<5 years >15,000 eggs <30 cm <15 cm >0.25 Broadcast spawner <2.5 High susceptibility (High risk = 3) Restricted (East Africa or local)

Accessible: 10–30m

Readily accessible: 0–10m

USD 0–10

Restricted: three-dimensional habitats associated with coral reefs) USD 10–100

Very restricted: specific microhabitats e.g. branching corals or anemones) >USD 100

<5%

5–10%

>10%

Availability: global distributiona Encounterability: deptha Encounterability: ecological niche (habitat)1 Selectivity: desirability/market value1 Post capture mortality a b c

The scoring bins were adapted from Roelofs and Silcock (2008). The scoring bins were adapted from Patrick et al. (2009). The scoring bins were adapted from Fishbase.

Kenya in order to score for ‘desirability/market value’ attribute. Validation of the market value was also done through consultations with the key informants. In addition, a score of 3 was assigned to the top 2 species collected in each family (based on the catch data in this study) with the premise that they face relatively intense collection pressure irrespective of their market value. Other adjustments made included use of the more qualitative attribute ‘global distribution’ adapted from Hobday et al. (2011) to score ‘availability’ based on information from Fishbase instead of the attribute ‘aerial overlap of species distribution with the fishery’ which could not be accurately quantified. 2.5. Data analysis 2.5.1. Catch and fishing effort Fishing effort was calculated for each fishing ground as the total sum of fisher days (1 fisher day = 1 fisher fishing for one day); while the mean catch per unit effort (CPUE) was calculated as the number of fish per fisher per day (fish/fisher/day ± SE). Simple linear regression (R2) was used to explore the relationship between catch and effort, while seasonal differences in mean CPUE were compared using the non-parametric Mann-Whitney U test. 2.5.2. Species catch composition The catch data was analyzed for each fishing ground separately for two community parameters: Shannon-Weiner’s diversity index (H’) and species richness (SR) (Magurran, 1988), as well as the relative abundance of species. Multivariate ordination techniques were used to explore spatial patterns in catch composition among the fishing grounds; and Detrended Correspondence Analysis (DCA) was used to explore the association of fish species with fishing grounds based on relative abundance. DCA plots were based on species representing >1% of the total catches to allow for easy interpretation. A combination of cluster analysis and non-metric Multidimensional Scaling (nMDS) was performed to identify groupings of fishing grounds that were similar in species catch composition; and to assess the similarity in species catch composition between fishing modes (snorkeling vs. SCUBA diving) based on the% relative abundance of the species in the total catches. The dataset was

pre-treated by square root transformation which has the effect of down-weighing the importance of highly abundant species so that similarities can be more evenly distributed (Clarke and Warwick, 2001). Matrices of similarities were obtained using the Bray-Curtis coefficient (Bray and Curtis, 1957). The nMDS ordination was based on a stress value of 0.1 and the output was superimposed with clusters at 50% level of similarity. Analysis of similarity (ANOSIM) was used to test for similarity in fish composition between the grounds. One-way similarity percentage (SIMPER) analysis was further used to identify the dominant species contributing to similarity and dissimilarity between the fishing grounds. All multivariate statistical analyses were performed using PRIMER v6 software (Clarke and Warwick, 2001; Clarke and Gorley, 2006).

2.5.3. Vulnerability risk scoring In scoring the attributes, each was determined as equally important in influencing overall vulnerability risk and given equal weighting (refer to Appendix Table A3). Simple linear regression was applied to identify the level of correlation among the productivity and susceptibility attributes scores. Hobday et al. (2011) recommend dropping either of two attributes with R values above 0.9 due to redundancy. The strongest correlation among productivity attributes was observed between ‘average size at maturity and “average maximum age” (R = 0.77), while the strongest correlation among susceptibility attributes was between ‘post capture mortality’ and ‘habitat niche’ (R = 0.59) (Table A4). Thus, all selected attributes were included in the analysis as the R values were below 0.9. For each species, the productivity attribute scores were averaged to obtain the risk values, while susceptibility risk values were estimated as the multiplicative product of the attribute scores (Hobday et al., 2011). As an overall measure of relative vulnerability risk, a 2D bivariate plot based on the Euclidean distances of the risk values from the origin was generated with curved lines dividing the plot area into equal thirds representing low, medium and high risk vulnerability categories. Based on Hobday et al. (2007), species having low productivity and high susceptibility scores (with overall risk scores >3.18) were considered as highly vulnerable to the fishery whereas those with high productivity and low susceptibil-

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

Number of fish

3. Results 3.1. Catch trends Export statistics compiled by the State Department of Fisheries indicated that aquarium fish were exported to 27 countries. In 2009, the main export markets were the United States (35%), the European Union Countries (42%), South Africa (9%), Israel (4%) and Japan (2%); whereas, in 2011, minor variations in the volume of fish exported to the different markets were observed although the markets remained the same as: United States (36%), European Union Countries (38%), South Africa (5%), Japan (4%), Turkey (3%) and Israel (2%). Based on the logbook catch data for the 11 fishing grounds, a total of 1.45 million aquarium fish were collected from October 2006 to December2011. This represented an estimated 279,000 fish collected annually ranging from about 235,000 in 2007 to a peak of about 326,700 in 2008 (Fig. 2a). Harvests from six fishing grounds constituted about 94% of the total as follows: Shimoni (33%), Kanamai (20%), Mtwapa (18.5%), Kilifi (12.3%), Nyali (5.6%) and Diani (4.2%). The remaining 6% was obtained from Shariani (3.5%), Shelly (1.7%), Msumarini (0.8%), Malindi (0.3) and Lamu (0.1%) (see Fig. 1 for locations). By fishing mode, snorkel fishers landed a total of 803,010 fish (55.3%) compared to 656,033 (44.5%) landed by SCUBA fishers (Fig. 2b), while the mode of fishing for 4295 fish (0.3%) obtained from independent fishers was not recorded. Shifts in fishing effort among the fishing grounds were observed corresponding with the trend in annual catches (Fig. 3b). Among the fishing grounds, SCUBA diver catches were mainly collected from Shimoni (22% of the total number of fish), Mtwapa (17%) and Kilifi (4%); while catches by snorkel fishers were mainly

Number of fish

120000

2007

10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0

350000 300000 250000 200000 150000 100000 50000 0 2007 Snorkeling

2008

2009

SCUBA

2010

SCUBA effort

2011 Snorkeling effort

100

SCUBA Snorkeling

Lamu

Kilifi

Malindi

Shariani

Msumarini

Mtwapa

Kanamai

Nyali

Shelly

Diani

90 80 70 60 50 40 30 20 10 0 Shimoni

Number of fish (%)

Number of fisher days

ity scores (with overall risk scores <2.64) were considered least vulnerable.

23

Fig. 2. (a) Trends in the total number of aquarium fish collected annually, (b) The spatial variation in composition of fish catches by gear type among the 11 fishing grounds in coastal Kenya.

collected from Kanamai (20%) and Shimoni (11%). A declining trend in total annual catches was observed from 2008 to 2010 in Diani, Shelly, Nyali, and Kilifi, while an increasing trend was observed in Mtwapa and relatively stable or fluctuating trends in the other sites (Fig. 3a).

2008

2009

2010

2011

100000 80000 60000 40000 20000

Lamu

Malindi

Lamu

2011

Malindi

Kilifi

2010

Shariani

Msumarini

Kilifi

Shariani

Msumarini

Kanamai

2009

Kanamai

Mtwapa

2008

Nyali

Shelly

Diani

Mtwapa

Nyali

Shelly

Diani

2007

4000 3500 3000 2500 2000 1500 1000 500 0

Shimoni

Number of fisher days

Shimoni

0

Fig. 3. Variation in (a) the annual total number of aquarium fish collected, and (b) annual total fishing effort (number of fisher days) among 11 fishing grounds from January 2007 to December 2011.

24

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

3.2. Spatial and temporal trends in fishing effort and CPUE

time for snorkel fishers C (R2 = 0.0063, p > 0.05) and SCUBA fishers (R2 = 0.0208, p > 0.05; Fig. 4c). 3.3. Species catch composition The catch from the 11 grounds was comprised of 220 species belonging to 36 families. However, ten families accounted for 94% of the total catch: Labridae (32%, 42 species), Pomacentridae (14%, 14 species), Serranidae (9%, 8 species), Blenniidae (9%, 7 species), Scorpaenidae (7%, 8 species), Pomacanthidae (5%, 10 species), Acanthuridae (5%, 16 species), Microdesmidae (5%, 3 species), Gobidae (5%, 8 species) and Chaetodontidae (3%, 23 species). Labridae were the most collected fish family from all the fishing grounds apart from Lamu, and ranged from 24% of the catch in Shimoni to 53% in Kilifi. In Lamu, angelfishes (Pomacanthidae) were the most collected constituting 57% of the total catch. Thirty-two species constituted 80% of the total catch from the grounds, with the top ten species accounting for over 50% of the catch (Fig. 5). The most commonly collected species included the Bluestreak cleaner wrasse, Labroides dimidiatus (11%), the Sea goldie Pseudanthias squamipinnis (8%), the Fire goby, Nemateleotris magnifica (5%), the sixline wrasse, Pseudocheilinus hexataenia (4.8%), and the Twobar anemonefish, Amphiprion allardi (4.8%). Within the key fish families targeted, L. dimidiatus and P. hexataenia accounted for 34% aEnd 15% of the total Labridae catches, respectively. Among the other key fish families, the catch composition consisted of: Pomacentridae: A. allardi, 36% and Chromis viridis, 30%; Anthiinae: P. squamipinnis, 75%; Acanthuridae: Paracanthurus hepatus, 31% and Acanthurus leucosternon, 16%; Pomacanthidae: Centropyge acanthops, 71%; Blenniidae: Salarias fasciatus, 47% and Ecsenius midas, 36%; and Scorpaenidae: Pterois miles, 60% and P. antennata, 17%.

1600

40

CPUE

Effort

1400 1200

30

1000 800

20

600 400

10

200

Effort (number of fisher days)

CPUE (number of fish.fisher -1 . day -1)

Over the six-year period, a total 60,052 fisher days with snorkeling accounting for 65% of the total fishing effort were recorded in the log books. Approximately 80% of the fishing effort was concentrated in 4 of the 11 fishing grounds: Shimoni (28.5%), Kanamai (21.5%), Mtwapa (15.4%) and Kilifi (14.2%), while the lowest effort was observed in Lamu accounting for 0.03%. The overall mean CPUE (fish/fisher/day ± SE) was 24 ± 0.5 and differed significantly among fishing grounds (Kruskal-Wallis H test, p < 0.05). The SCUBA fishers had a significantly higher mean CPUE (31.9 ± 0.3) compared to snorkel fishers CPUE (20.5 ± 0.1) (Mann-Whitney U test, P < 0.05). Among snorkel fishers, the mean CPUE was highest in Lamu at 48 ± 0.6 (double the overall average) and lowest in Malindi at 13 ± 0.3. The monthly trend showed a three-fold increase in fishing effort in August 2007 with the mean CPUE being highest in March 2007 at 34 ± 0.2 and lowest in September 2011 at 18 ± 0.3 (Fig. 4a). The mean CPUE significantly differed among the years influenced by a significantly lower CPUE in 2007 as indicated by Pair-wise post hoc comparisons (Kruskal-Wallis H test, p < 0.05). The CPUE trend showed strong seasonality (Fig. 4a), demonstrated by a higher mean CPUE (27 ± 0.2) during the NEM months from November to March compared to the CPUE of 22 ± 0.1 during the SEM months of April to September (Mann-Whitney U test, p < 0.05). Results of simple linear regression showed a significant positive relationship between total catch and effort (y = 21.151x + 3091.7, R2 = 0.613; p < 0.05). The linear regression of CPUE with fishing effort showed a weak but negative relationship (R2 = 0.011; P > 0.05; Fig. 4b); while regression of CPUE with time showed no significant change in CPUE over

0

0 OND J FMAM J J A SOND J FMAM J J A SOND J FMAM J J A SOND J FMAM J J A SOND J FMAM J J A SOND 2006

2007

2008

2009

35000 30000 25000 20000 15000 10000 5000 200

2011

36 CPUE (number of fish . fisher-1 . day-1)

Total Catch (number of fish)

40000

2010

400

600

800

1000

1200

Effort (number of fisher days)

1400

1600

34 32 30 28 26 24 22 20 18 16 200

400

600

800

1000

1200

1400

1600

Effort (number of fisher days)

Fig. 4. (a) The monthly trend in the mean CPUE (number of fish/fisher/day), (b) the linear relationship between mean CPUE and fishing effort: y = −0.0018x + 26.085, R2 = 0.011; and (c) the linear trend in mean monthly CPUE for snorkel and SCUBA fishers over time. Snorkel fishers: y = 0.016x + 20.07, R2 = 0.0063; Diving = y = −0.047x + 33.13, R2 = 0.0208). The bounds indicate upper and lower 95% confidence limits.

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

25

Fig. 5. Relative abundance of aquarium fish species making up >1% of the catch for all sites (pooled data), and in selected fishing grounds along the Kenya coast during October 2006 to December 2011.

The species richness and diversity (Shannon-Weiner H’) of the aquarium fisher catches varied among the fishing grounds (Fig. 5). The southernmost site of Shimoni (see Fig. 1) had both the highest number of species collected and the highest diversity (193 species, H’ = 3.59), while the least number of species and lowest diversity of the fisher catches was from the northernmost site of Lamu (38 species, H’ = 2.31). The ANOSIM results showed a significant difference in catch composition between the grounds (P < 0.05); however, the low R value (R = 0.34) suggested that the differences were rather low. Pair-wise tests further revealed strongest differences in catch composition to be between Lamu-Shelly and Lamu-Kanamai grounds with both pairs having a high R value of 0.907. Results of SIMPER analysis (Table 2) indicated that the cleaner wrasse, L. dimidiatus contributed most to dissimilarities between the fishing grounds, while the angelfishes, Pomacanthus chrysurus and P. maculosus contributed to dissimilarity between Lamu and all the other fishing grounds. Three species: N. magnifica, Macropharyngodon bipartitus and H. iridis contributed to similarities in catch composition among SCUBA fishers, while L. dimidiatus and the lionfish, P. miles, contributed most to similarities among snorkel fishers. No significant differences in catch composition were observed between the seasons or between years based on ANOSIM results.

Multivariate DCA of the association of species with the fishing grounds based on% relative abundance showed spatial segregation of some species (Fig. 6). Notably, the orangeback angelfish C. acanthops, the Acanthuridae, P. hepatus and Zebrasoma desjardinii and the Radiant wrasse H. iridis were associated with Shimoni. The Exquisite wrasse, Cirrhilabrus exquisitus and the Midas blenny, E. midas, were associated with Shimoni and Mtwapa; the Sixline wrasse, P. hexataenia was associated with Kanamai and Msumarini, and the damselfish, Chromis nigrura, was associated with Mtwapa and the Blueband goby, Valenciennea strigata, with Kilifi and Mtwapa. nMDS ordination of the catch composition revealed a clustering of fishing grounds into three distinct groups at 50% similarity with a stress value of 0.1. Lamu (group I) was distinctly separated from the other grounds, while Shimoni and Mtwapa (group II) were relatively similar in composition as were sites that formed group III (Fig. 7a, see Fig. 1 for locations). An MDS ordination plot based on the catch composition by fishing mode (SCUBA diving or snorkeling) further showed a clear separation in the species composition of snorkel fishers and divers catches, with snorkeler catches from Lamu still being segregated from the other fishing grounds (Fig. 7b).

26

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

Table 2 One-way SIMPER analysis of aquarium fish catches from 11 fishing grounds along the Kenya coast, showing species contributing to dissimilarity between the three clusters (shown in Fig. 6) differentiated at 50% similarity and a cumulative percentage of 70%: Group I = (Lamu), Group II = (Shimoni, Mtwapa) and Group III = (Diani, Shelly, Nyali, Kanamai, Msumarini, Shariani, Malindi, Kilifi). Species

Mean Abundance

Average Dissimilarity

Ratio

Group I vs. Group II Pomacanthus chrysurus Pomacanthus maculosus Labroides dimidiatus Amphiprion allardi Nemateleotris magnifica Chromis viridis Centropyge acanthops Ecsenius midas Zebrasoma desjardinii Halichoeres iridis Group I vs. Group III Pomacanthus chrysurus Labroides dimidiatus Pomacanthus maculosus Pterois miles Zebrasoma desjardinii

Group I 60.08 11.19 3.11 0 0 0 0 0 6.48 0 Group I 60.08 3.11 11.19 3.99 6.48

Group II 0.28 0.09 11.12 8.25 7.49 6.91 6.76 6.55 2.34 5.74 Group III 0.69 21.3 0.01 11.35 0.42

Average dissimilarity = 92.59 29.9 2.05 5.58 1.23 5.44 0.92 4.13 0.79 3.75 0.91 3.46 0.76 3.38 0.78 3.27 1.01 3.2 1.07 2.87 0.93 Average dissimilarity = 89.03 29.7 2.07 9.39 1.38 5.59 1.25 4.8 1.28 3.18 0.94

Contribution%

Species

Mean Abundance

Average Dissimilarity

Ratio

Chromis viridis Salarias fasciatus Amphiprion allardi Acanthurus leucosternon Group II vs. Group III Labroides dimidiatus Pterois miles Chromis viridis Nemateleotris magnifica Amphiprion allardi Ecsenius midas Centropyge acanthops Macropharyngodon bipartitus Halichoeres iridis Salarias fasciatus Valenciennea strigata Pseudocheilinus hexataenia Paracanthus hepatus Cirrhilabrus exquisitus Pseudanthias squamipinnis

0 0.16 0 4.66 Group II 11.12 2.26 6.91 7.49 8.25 6.55 6.76 3.68 5.74 1.23 4.78 1.52 4.06 2.2 2.78

5.72 5.68 5.39 1.19 Group III 21.3 11.35 5.72 3.59 5.39 0.81 0.47 4.21 1.86 5.68 1.91 4.23 0.19 1.65 0.54

2.86 0.73 2.8 0.87 2.69 1.26 2.17 1.14 Average dissimilarity = 71.36 8.68 1.4 4.82 1.27 4.24 1.02 4.21 1.04 4.17 1.07 3.26 1.04 3.26 0.78 3.03 0.99 2.96 1.05 2.74 0.93 2.52 0.95 2.08 0.88 2.02 0.71 1.58 0.85 1.45 0.79

3.4. Productivity Susceptibility Analysis (PSA) A summary of the risk scores for all the 102 species assessed is presented in Table A3, while the plot of productivity and susceptibility analysis is presented in Fig. 8. The productivity risk values ranged from 1.13 to 2.50 with an average of 1.62, while susceptibility risk values ranged from 1.02 to 3.00 with an average of 1.32. Overall vulnerability risk scores ranged from 1.57 to 3.68 with an average of 2.12. At family level, Pomacanthidae had the highest vulnerability risk having a mean score of 2.68 ± 0.73SD (n = 8); this was followed by Microdesmidae (N. magnifica) with a score of 2.66. The fish families that ranked least vulnerable to overexploitation were Syngnathidae (Doryhamphus excisus) with a score of 1.70 and Blenniidae with mean scores of 1.75 (n = 6). Four species, the Yellowbar angelfish P. maculosus, the Goldtail angelfish P. chrysurus, the Twobar anemonefish A. allardi and the Skunk clownfish A. akallopisos, were ranked as highly vulnerable with vulnerability risk scores ranging from 3.41 and above, while 7 species including the Emperor angelfish Pomacanthus imperator, the Regal angelfish Pygoplites diacanthus, the Clown coris Coris aygula, the Palette surgeonfish P. hepatus, the Radiant wrasse H. iridis, the Fire goby N. magnifica and the Semicircle angelfish Pomacanthus semicirculatus ranked as having moderate vulnerability, with risk scores ranging between 2.66 and 2.99 (Table 3, Fig. 8). Low risk species on the borderline with vulnerability risk scores below 2.64 included the scorpionfish P. miles, the grouper

Contribution%

32.29 6.02 5.88 4.46 4.05 3.73 3.65 3.54 3.46 3.1 33.36 10.54 6.28 5.39 3.57 Species 3.21 3.14 3.03 2.44 12.17 6.75 5.95 5.9 5.85 4.57 4.56 4.24 4.15 3.83 3.53 2.92 2.82 2.22 2.03

Cephalopholis miniata, the carangidae Gnathanodon speciosus and the surgeonfish Z. desjardinii (Table 3).

4. Discussion 4.1. Catch and effort trends The total number of fish collected annually in Kenya is estimated to range from 240,000 to 341,000 individuals. This range takes into consideration an estimated post-capture mortality rate of approximately 5% (Okemwa et al., 2009), and the premise that our data represents about 70% of total landings. The estimate constitutes a two-fold increase in annual catches of ornamental fishes from that reported by Wood (2001a) during the 1990s. The trend is likely associated with increasing demand for marine aquarium resources on the global market (Thornhill, 2012). The significant correlation between fishing effort and total catches and increasing temporal trend in fishing effort depicts a long-term trend of increasing fishing pressure that is a likely response to a reduction in the abundance and availability of some target species. The estimated mean CPUE for Kenya’s aquarium fishers of 24 ± 0.5 fish/fisher/day documented in this study is within previously reported ranges for Australia (20 − 45 fish/man/day) and the Cook Islands (24–36 fish/man/day by Wood (2001a), but lower than that reported by Shuman et al. (2004) for the Philippines (37.5–48 fish/man/day). However, the CPUE estimate for Kenya is likely underestimated taking into account

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31 Acanthuridae (Surgeonfishes)

27

Labridae (Wrasses)

1.0

Pomacanthidae (Angelfishes)

1.0

1.0

0.5

LAMU KANAMAI P. chrysurus P. maculosus 0.5 MALINDI SHIMONI DIANI C. acanthops MSUMARINI SHARIANI C. multispinnis P. semicirculatus 0.0 NYALI SHELLY

KANAMAI A. triostegus MSUMARINI A. lineatus A. xanthopterus N. brevirostris Z. scopas KILIFI SHARIANI

0.5 P. hepatus SHIMONI 0.0

Z. desjardinii 0.0

C. truncatus MALINDI DIANI SHELLY NYALI A. leucosternon LAMU MTWAPA

-0.5

-1.0

-1.0

MALINDI KILIFI C. formosa SHELLY LAMU C. cuvieri NYALI DIANI L. dimidiatus SHARIANI T. hebraicum

-0.5

0.0

0.5

1.0

1.5

2.0

-1.0 -1.5

Dimension 2

-1.0

-0.5

0.0

-2.0 -2.5

0.5

0.0

0.0

0.5

1.0

1.5

2.0

MSUMARINI SHARIANI C. argus DIANI KANAMAI

-10 -9

MTWAPA

0.8

KILIFI N. carberryi P. evansi NYALI MTWAPA P. squamipinnis SHIMONI SHELLY

0.0

-8

-7

-6

-5

-4

-3

NYALI MALINDI

-0.2

P. rhinorhyncus

-0.4

0.0

Taenianotus sp

0.5

1.0

0.5

1.0

1.5

-0.6 -1.2

C. xanthocephalus C. melannotus

MSUMARINI SHARIANI C. trifasciatus SHIMONIC. falcula DIANI C. vagabundus MALINDI C. kleinii C. auriga NYALI H. acuminatus C. guttatissimus KILIFI LAMU

C. lunula KANAMAI

MTWAPA

-2

-1

0

1

2

3

-1.0 4 -1.0

-0.5

0.5

0.0

NYALI P. antennata D. zebra P. radiata KILIFI DIANI SHARIANI KANAMAI P. miles MALINDI MSUMARINI LAMU SHELLY

SHIMONI D. brachypterus

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.0

0.5

1.0

1.5

Gobidae (Gobies)

MTWAPA

0.4

LAMU

-0.5

0.0

C. leucopleura

0.6

SHIMONI

0.0

-1.0

-0.5

-0.5

E. midas

0.2

-1.5

SHELLY

Scorpaenidae (Lionfishes)

KILIFI M. mossambicus

-1.0

-1.0

Species Fishing Grounds

1.0 S. fuscus KANAMAI S. fasciatus SHELLY SHARIANI MSUMARINI DIANI E. brevis

-1.5

Chaetodontidae (Butterflyfishes)

-0.5

D. aruanus -0.5

-2.0

0.5

0.5

-1.5

1.5

LAMU MALINDI

Blennidae (Blennies)

-0.5

1.0

1.0 1.0

P. caeruleus

0.0

0.5

Serranidae (Anthiases and Groupers)

D. carneus A. akallopisos NYALI SHIMONI A. allardi DIANI SHELLY C. dimidiatus MALINDI 0.0 C. viridis C. nigrura SHARIANI MSUMARINI KILIFI MTWAPA -0.5 KANAMAI LAMU D. trimaculatus -1.0

-1.0

A. trimaculatus

-1.5 MSUMARINI

0.5

-1.5

P. imperator MTWAPA

-1.0

P. carpenteri P. mccoskeri

P. hexataenia

KILIFI

-0.5

C. exquisitus H. iridis SHIMONI

KANAMAI -0.5

Pomacentridae (Damselfishes)

-1.5

M. bipartitus A. meleagrides H. cosmetus MTWAPA

0.2

0.4

0.6

-0.5

A. aurora MTWAPA V. puellaris SHIMONI V. helsdingeni V. strigata KILIFI NYALI LAMU MALINDI V. sexguttata

-1.5

-2.0

KANAMAI

MSUMARINI SHELLY

-1.0

SHARIANI DIANI

-1.0

-0.5

0.0

0.5

Dimension 1 Fig. 6. Detrended Correspondence Analysis (DCA) plot of the association of key species (representing >1% of the total catch) with fishing grounds. Black circles = species, open triangles = fishing grounds. Location of fishing grounds indicated on Fig. 1.

Fig. 7. Two dimensional non-metric MDS (nMDS) ordination plots of the (a) total catch, (b) total catch grouped by mode of fishing among 11 aquarium fishing grounds: (SHI = Shimoni, DIA = Diani, SHE = Shelly, NYA = Nyali, MTW = Mtwapa, KAN = Kanamai, MSU = Msumarini, SHA = Shariani, KIL = Kilifi, MAL = Malindi, LAM = Lamu), based on Bray-Curtis similarity of standardised square root transformed species frequency data. Solid contours indicate clusters identified at 50% similarity.

unreported mortalities that may have occurred during fishing as a result of poor handling. As derived in this study, the use of raw or nominal CPUE as an index of relative stock abundance for the aquarium fishery is robust to violation of key assumptions on fishing effort which is influenced by fishing efficiency, the target species, species, population dynamics and environmental variability amongst other factors (see Maunder and Punt, 2004; Maunder et al., 2006). In deriving the CPUE estimates, the efficiency of the fishers was assumed not to differ significantly between individuals; although it is expected that more experienced fishers will be more efficient. In addition,

the target species and number of individuals collected in aquarium fisheries is strongly dictated by exporters and their clients. Some species are routinely collected in large numbers as “filler species”; others are collected opportunistically when encountered due to their rarity or high value, while others are only collected on specific request by a client (Okemwa pers. obs.). Despite these dynamics, CPUE index derived in this study provides a useful indicator which can be used for monitoring the performance of the fishery over time.

28

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

Table 3 Results of Productivity and Susceptibility Analysis (PSA) showing the risk scores of the top 15 species in order of vulnerability risk ranking. The detailed list of the 102 species assessed and scores is shown in the supplementary Table S3. NB: Species with similar scores received the same vulnerability risk ranking. Family

Latin Name

Common Name

Productivity Susceptibility Overall Risk Values Overall Risk Category Vulnerability Risk Ranking

Pomacanthidae Pomacanthidae Pomacentridae Pomacentridae Pomacanthidae Pomacanthidae Labridae Acanthuridae Labridae Microdesmidae Pomacanthidae Scorpaenidae Serranidae Carangidae Acanthuridae

Pomacanthus maculosus Pomacanthus chrysurus Amphiprion akallopisos Amphiprion allardi Pomacanthus imperator Pygoplites diacanthus Coris aygula Paracanthurus hepatus Halichoeres iridis Nemateleotris magnifica Pomacanthus semicirculatus Pterois miles Cephalopholis miniata Gnathanodon speciosus Zebrasoma desjardinii

Yellowbar angelfish Goldtail angelfish Skunk clownfish Twobar anemonefish Emperor angelfish Regal angelfish Clown coris Palette surgeonfish Rainbow wrasse Fire dartfish Semicircle angelfish Devil firefish Coral hind Golden trevally Desjardin’s sailfin tang

2.13 1.88 1.63 1.63 1.88 1.63 2.50 1.50 1.38 1.88 1.88 2.25 2.38 2.38 1.75

Fig. 8. A PSA plot of the risk scores of 102 fish species targeted by the aquarium fishery in Kenya. Curved lines indicate divisions into equal thirds of the plot area showing species at low risk (<2.64), medium risk (2.64–3.18) and high risk (>3.18). The yellow and red highlights indicate species that ranked as having medium to high-risk vulnerability. The species names, detailed scoring and overall vulnerability ranking for all species assessed are presented in the Appendixed Table A3.

4.2. Species catch composition A five-fold increase in the number of aquarium fish species collected from 48 in the 1980s (Samoilys, 1988) to 220 in this study (and 225 as estimated by Rhyne et al., 2012) is a clear indication of increasing fishing pressure. The increase in species targeted for Kenya is in tandem with the global pattern which is fuelled by an increasing demand for new species (Rhyne and Tlusty, 2012). Globally, only 10 fish families account for 83% of the international trade dominated by the Pomacentridae (damselfishes) which account for 42% of the total volume (Green, 2003; Wabnitz et al., 2003). In Kenya, 7 families accounted for 81% of the catches, dominated by the Labridae (wrasses) which made up 32% of the catches. The diversity of wrasses targeted in Kenya was high at 42 species, similar to that for countries such as Sri Lanka (44 wrasse species, Ekaratne, 2000), indicating that Kenya is a key source for wrasse species in the global aquarium fish market. The cleaner wrasse, L. dimidiatus, was the most commonly collected species in most of the fishing grounds. Cleaner wrasses perform an essential ecosystem function of removing ectoparasites from other reef fishes (Grutter, 1996; Wood, 2001a), and in regulating reef fish community structure through increased diversity and

3.00 3.00 3.00 3.00 2.33 2.33 1.30 2.33 2.33 1.89 1.89 1.30 1.02 1.02 1.89

3.68 3.54 3.41 3.41 2.99 2.84 2.82 2.77 2.71 2.66 2.66 2.60 2.59 2.59 2.57

High High High High Medium Medium Medium Medium Medium Medium Medium Low Low Low Low

1 2 3 3 4 5 6 7 8 9 9 10 11 11 12

size of individual fish (Bshary, 2003; Clague et al., 2011; Waldie et al., 2011). Therefore collection of L. dimidiatus in high numbers is likely to have significant ecological implications but these have not been established on most of the reefs in the WIO region including Kenya. The relative composition of catches differed among the fishing grounds with results of the MDS plot separating the grounds into three distinct groups strongly influenced by fishing modes (snorkeling vs. SCUBA diving). There was a distinct segregation of Lamu in the MDS plot due to selective targeting of P. maculosus and P. chrysurus which were the top species collected from the area. DCA further revealed that some species are more associated with specific fishing grounds. The observed temporal shifts in fishing effort between the fishing grounds could be as a result of fishers trying to optimize fishing effort and catches of their target species based on their knowledge of species distributions. Notably, the apparent shift in fishing effort northwards to Lamu, a relatively pristine area, and specific targeting of high-value angelfishes P. maculosus and P. chrysurus could be in response to a reduction in the abundance and availability of these species in other fishing grounds and the need to maximize economic returns. Aquarium fishers in Kenya perceive that these angelfish species have become more difficult to catch over time and as a consequence, they have to travel farther to remote and relatively pristine areas such as Lamu or spend a much longer time searching (G. Okemwa, unpublished data). Furthermore, importation of P. maculosus from Tanzania has been reported (State Department of Fisheries (SDF, 2011), providing more supportive evidence of possible localized depletion on most fishing grounds. Shifts in fishing depth may also be typical to the fishery as documented in Hawaii where aquarium fishers were observed to dive deeper and increase their fishing effort in response to weak recruitment of their target species in shallower areas (Stevenson et al., 2011). Apart from effort and market dynamics, other natural variables influencing the catch composition of aquarium fishers from the different fishing grounds are related to local scale patterns in the abundance and diversity of juvenile reef fish populations driven by recruitment dynamics (Jones, 1990; Doherty, 2002) that are driven by a suite of biological and physical factors such as the presence and abundance of adult conspecifics (Lecchini et al., 2007), larval supply (Kaunda-Arara et al., 2009; Mwaluma et al., 2011; Rankin and Sponaugle, 2014), predation levels (Hixon, 1991), local hydrodynamics and habitat characteristics (McClanahan and Arthur, 2001; Abesamis and Russ, 2010; McClanahan, 2015), and species-specific microhabitat preferences (Lecchini et al., 2007; DeMartini et al., 2010).

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

4.3. Productivity Susceptibility Analysis (PSA) Overall, most species fell in the high productivity and low susceptibility risk categories placing them at a relatively low vulnerability risk to localized depletion by the aquarium fishery in Kenya. Pomacanthidae had the highest overall mean vulnerability risk score of 2.6 ± 0.71SD, with P. maculosus and P. chrysurus being ranked as highly vulnerable. Species of the genus Pomacanthus have life history characteristics that make them susceptible to overexploitation including being relatively long-lived, delayed maturity and low rates of recruitment (Tebua, 2005). The anemonefishes A. allardi and A. akallopisos, which also ranked as having high vulnerability risk, are particularly vulnerable to overfishing due to their high fidelity to specific species of host anemones (Fautin, 1986), and their nest guarding behaviour (Bender et al., 2013). Evidence of the high vulnerability of anemonefishes to the aquarium fishery is documented in quantitative studies by Shuman et al. (2005) in Philippines and Jones et al. (2008) in the Great Barrier Reef, where both anemone and anemonefish densities were observed to be significantly lower in areas exploited by aquarium fishers. The high vulnerability ranking of the four species in this study is in agreement with very early concerns of Lubbock and Polunin (1975) and Samoilys (1988) on their potential for overexploitation. Pomacanthus and Amphiprion species are highly desired on the global market (Sadovy and Vincent, 2002; Shuman et al., 2004). This provides a strong incentive for continued intense collection despite overall abundances being low, which places them at an even higher risk of depletion compared to other target species. Conversely, since these species still provide an important component of Kenya’s aquarium fishery almost 30 years later, it is likely that the populations and catch levels may be sustainable with appropriate management interventions. Seven species ranked as having moderate vulnerability to overexploitation by the fishery including three angelfish species; P. imperator, P. diacanthus and P. semicirculatus; the wrasses, Coris aygula and H. iridis, the surgeonfish P. hepatus and the dartfish N. magnifica. Evidence of declines due to heavy collection by the aquarium fishery have been reported for P. imperator and P. hepatus populations in Philippines and Indonesia (Rubec, 1987) which supports the need to monitor the Kenyan stocks. The moderate risk ranking of N. magnifica concurred with that assessed by Fujita et al. (2013) for Indonesian stocks. The lionfish, P. miles, was ranked among the borderline low-risk species. Darling et al. (2011) report that P. miles populations are generally in low densities and smaller in size in Kenya compared to stocks in the Caribbean. This study also revealed that P. miles is highly targeted in Kenya warranting the need for a quantitative assessment to verify their status.

29

decrease in the MPAs due to an increase in the abundance of predators (Watson et al., 2007). To balance such effects, smaller spatial closures may be beneficial, especially for species that have a high tendency to self-recruit such as anemonefishes (Madduppa, 2012). A number of small community managed areas have been established in Kenya (Rocliffe et al., 2014) which will likely play an important role is sustaining aquarium fish populations. The results highlight the urgent need for closer monitoring and assessment of catches and exports so as to determine the absolute risks of the species. Fishers involved in the aquarium fishery have considerable historical and traditional knowledge of population dynamics of the target species. Tapping into this knowledge, coupled with complementary biology and population studies of target species will further provide more comprehensive information for assessing the status of the fishery. Future research considerations should also focus on conducting fishery independent surveys to assess the status of key target species, as well as an assessment of interactions with artisanal food fisheries to better quantify the cumulative sources of fishing mortality on aquarium fish populations. Concerted efforts are also needed to develop captive breeding and rearing of marine aquarium species particularly those identified to be at risk of localized depletion in this study. In conclusion, this study provides the first quantitative assessment of spatial and temporal trends of the marine aquarium fishery in Kenya and the WIO region. Although the temporal-scale of the datasets was relatively short, the findings address important information gaps and provide an important benchmark for prioritizing future quantitative assessments. An adaptive management approach with full participation of all stakeholders is needed to enhance sustainability of the fishery. Acknowledgements The cooperation of the aquarium fish trade companies and fishers in provision of catch data and information is highly appreciated. The authors thank the Director Kenya Marine and Fisheries Research Institute (KMFRI), KMFRI staff and the State Department of fisheries for logistical support. Funding for this research was supported by grants from WWF Russell E. Train PhD fellowship (grant # RW24) to GMO, the Western Indian Ocean Marine Science Association (WIOMSA)Marine Science for Management (MASMA) programme for catch data collection and analysis. The Kenya Coastal Development Project (KCDP) supported consultative workshops for the risk assessment. We thank three anonymous reviewers for their inputs and comments which greatly improved the quality of the manuscript.

4.4. Management implications and future considerations

Appendix A. Supplementary data

Taking into consideration the existing uncertainties on the real status of target species, precautionary management measures are needed to ensure that they are sustainably exploited. Management interventions should focus on regulating fishing effort by limiting entry to the fishery as well as implementation of area restrictions. There is also a need to consider restricting the collection of high risk species as identified in this study (P. maculosus, P. chrysurus, A. allardi and A. akallopisos). However, we caution that enforcement of such restrictions may be challenging and should be consultatively discussed and agreed upon with the aquarium fishing industry to enhance compliance. The role of Marine protected areas (MPAs) in replenishing fish stocks in Kenya is well documented through spillover of adults (McClanahan and Mangi, 2000; Kaunda-Arara and Rose, 2004) and replenishment of fish larvae (Kaunda-Arara et al., 2009). Though, the abundance of some lower trophic level aquarium species may

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fishres.2016.04. 020. References Abesamis, R.A., Russ, G.R., 2010. Patterns of recruitment of coral reef fishes in a monsoonal environment. Coral Reefs 29, 911–921. Almany, G.R., Berumen, M.L., Thorrold, S.R., Planes, S., Jones, G.P., 2007. Local replenishment of coral reef fish populations in a marine reserve. Science 316, 742–744. Andrews, C., 1990. The ornamental fish trade and fish conservation. J. Fish Biol. 37, 53–59. Arrizabalaga, H., Bruyn, P., Diaz, G.A., Murua, H., Chavance, P., Molina, D.A., Gaertner, D., Ariz, J., Ruiz, J., Kell, L.T., 2011. Productivity and susceptibility analysis for species caught in Atlantic tuna fisheries. Aquat. Living Resour. 24, 1–12. Bender, M.G., Floeter, S.R., Mayer, F.P., Vila-Nova, D.A., Longo, G.O., Hanazaki, H., Carvalo-Filho, A., Ferreira, C.E.L., 2013. Biological attributes and major threats

30

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31

as predictors of the vulnerability of species: a case study with Brazilian reef fishes. Oryx 47, 259–265. Bray, J.R., Curtis, J., 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27, 326–349. Bruckner, A.W., 2005. The importance of the marine ornamental reef fish trade in the wider Caribbean. Rev. Biol. Trop. 53, 127–137. Bshary, R., 2003. The cleaner wrasse, Labroides dimidiatus is a key organism for reef fish diversity at Ras Mohammed National Park, Egypt. J. Anim. Ecol. 72, 169–176. Clague, E.G., Cheney, K.L., Goldizen, A.W., McCormick, M.I., Waldie, P.A., Grutter, A.S., 2011. Long-term cleaner fish presence affects growth of a coral reef fish. Biol. Lett., http://dx.doi.org/10.1098/rsbl.2011.0458. Clarke, K.R., Gorley, R.N., 2006. Primer Version 6: User Manual/Tutorial. Plymoth, PRIMER-E. Clarke, K.R., Warwick, R.M., 2001. Change in Marine Communities: an Approach to Statistical Analysis and Interpretation, vol. 28., 2nd ed. Plymouth-E, Plymoth, UK, pp. 167–181 (Conservation). Cope, J.M., DeVore, J., Dick, E.J., Ames, K., Budrick, J., Erickson, D.L., Grebel, J., Hanshew, G., Jones, R., Mattes, L., Niles, C., Williams, S., 2011. An approach to defining stock complexes for U.S west coast groundfishes using vulnerabilities and ecological distributions. N. Am. J. Fish. Manage. 31 (4), 589–604. Darling, E.S., Green, S.J., O’Leary, J.K., Cote, I.M., 2011. Indo-Pacific lionfish are larger and more abundant on invaded reefs: a comparison of Kenyan and Bahamian lionfish populations. Biol. Invasions 13, 2045–2051. DeMartini, E.E., Anderson, T.W., Kenyon, J.C., Beets, J.P., Friedlander, A.M., 2010. Management implications of juvenile reef fish habitat preferences and coral susceptibility to stressors. Mar. Freshwater Res. 61, 532–540. Dee, L.E., Horii, S.S., Thornhill, D.J., 2014. Conservation and management of ornamental coral reef wildlife: successes, shortcomings, and future directions. Biol. Conserv. 169, 225–237. Doherty, P.J., 2002. Variable replenishment and the dynamics of reef fish populations. In: Sale, P.F. (Ed.), Coral Reef Fishes: Dynamics and Diversity in a Complex Ecosystem. Academic Press, San Diego, CA, pp. 327–355. Dransfeld, L., Gerritsen Hans, D., Hareide Nils, R., Lorance, P., 2013. Assessing the risk of vulnerable species exposure to deepwater trawl fisheries: the case of orange roughy to the west of Ireland and Britain. Aquat. Living Resour. 26, 307–318. Ekaratne S.U.K., 2000. A review of the status and trends of exported ornamental fish resources and their habitats in Sri Lanka. BOBP/REP/88 http://www.fao. org/docrep/007/ad920e/ad920e00.htm. Eschmeyer, W.N., Fricke, R., 2016. Catalog of FISHES: Genera, Species, References (http://researcharchive.calacademy.org/research/ichthyology/catalog/ fishcatmain.asp). Electronic version accessed (25.01.16). Fautin, D.G., 1986. Why do anemonefishes inhabit only some host actinians? Environ. Biol. Fishes 15, 171–180. Froese, R., Pauly, D., 2013. Fishbase. http://www.fishbase.org. (accessed 10.13). Fujita, R., Thornhill, D.J., Karr, K., Cooper, C.H., Dee, L.E., 2013. Assessing and managing data-limited ornamental fisheries in coral reefs. Fish Fish., http://dx. doi.org/10.1111/faf.12040. Government of Kenya (GoK), 2009. State of the Coast Report: Towards Integrated Management of Coastal and Marine Resources. National Environmental Management Authority (NEMA), Nairobi87. Green, E., 2003. International trade in marine aquarium species: using the Global Marine Aquarium Database. In: Cato, J., Brown, C. (Eds.), Marine Ornamental Culture Species: Collection, and Conservation. Iowa State Press, Ames, USA, pp. 31–48. Grutter, A., 1996. Effect of the removal of cleaner fish on the abundance and species composition of reef fish. Oecologia 111, 137–143. Hixon, M.A., 1991. Predation as a process structuring coral reef fish communities. In: Sale, P.F. (Ed.), The Ecology of Fishes on Coral Reefs. Academic Press, San Diego, CA, pp. 475–508. Hobday, A.J., Smith, A.D.M., Webb, H., Daley, R., Wayte, S., Bulman, C., Dowdney, J., Williams, A., Sporcic, M., Dambacher, J., Fuller, M., Walker, T., 2007. Ecological Risk Assessment for the Effects of Fishing: Methodology. Report R04/1072 for the Australian Fisheries Management Authority, Canberra. July 2007. Available from: http://www.afma.gov.au/environment/ecobased/eras/doc/methodology. pdf. Hobday, A.J., Smith, A.D., Stobutzki, I.C., 2011. Ecological risk assessment for the effects of fishing. Fish. Res. 108, 372–384. Jones, A.M., Gardner, S., Sinclair, W., 2008. Losing ‘Nemo’: bleaching and collection appear to reduce inshore populations of anemonefishes. J. Fish Biol. 73, 753–761. Jones, G.P., 1990. The importance of recruitment to the dynamics of a coral reef fish population. Ecology 71, 1691–1698. Kaunda-Arara, B., Rose, G.A., 2004. Effects of marine reef National Parks on fishery CPUE in coastal Kenya. Biol. Conserv. 118, 1–13. Kaunda-Arara, B., Mwaluma, J.M., Locham, G.A., Oresland, V., Osore, M.K., 2009. Temporal variability in fish larval supply to Malindi Marine Park, coastal Kenya. Aquat. Conserv. 19, A10–A18. Kolm, N., Berglund, A., 2003. Wild populations of a reef fish suffer from the nondestructive aquarium trade fishery. Conserv. Biol. 17, 910–914. Lecchini, D., Polty, S., Nakamura, Y., Mosconi, P., Tsuchiya, M., 2006. New perspectives on aquarium fish trade. Fish. Sci. 72, 40–47. Lecchini, D., Planes, S., Galzin, R., 2007. The influence of habitat characteristics and conspecifics on attraction and survival of coral reef fish juveniles. J. Exp. Mar. Biol. Ecol. 341, 85–90.

Lubbock, H.R., Polunin, N.V.C., 1975. Conservation and the tropical marine aquarium trade. Environ. Conserv. 2, 229–232. Madduppa, H.H., 2012. Selt-recruitment in Anemonefish and the Impact of Marine Ornamental Fishery in Spermonde Archipelago, Indonesia : Implications for Management and Conservation PhD Thesis. University of Bremen, pp. 168. Magurran, A.E., 1988. Ecological Diversity and Its Measurement. Chapman and Hall, London. Marine Stewardship Council, 2010. Fisheries Assessment Methodology 6.1 available at www.msc.org. Maunder, M.N., Punt, A.E., 2004. Standardizing catch and effort data: a review of recent approaches. Fish. Res. 70, e141–e159. Maunder, M.N., Sibert, J.R., Fonteneau, A., Hampton, J., Kleiber, P., Harley, S.J., 2006. Interpreting catch per unit effort data to assess the status of individual stocks and communities. ICES J. Mar. Sci. 63, e1373–e1385. McClanahan, T.R., Arthur, R., 2001. The effects of marine reserves and habitat on populations of East African coral reef fishes. Ecol. Appl. 11, 559–569. McClanahan, T.R., Mangi, S., 2000. Spillover of exploitable fishes from a marine park and its effect on the adjacent fishery. Ecol. Appl. 10, 1792–1805. McClanahan, T.R., 1988. Seasonality in east africa’s coastal waters. Mar. Ecol. Prog. Ser. 44, 191–199. McClanahan, T.R., 2015. Biogeography versus resource management: how do they compare when prioritizing the management of coral reef fish in the south-western Indian Ocean? J. Biogeogr. 42, 2414–2426. McCully, S.R., Scott, F., Ellis, J.R., Piling, G.M., 2013. Productivity and Susceptibility Analysis: application and suitability for data poor assessment of elasmobranchs in northern European seas. ICCAT 69, 1679–1698. Micheli, F., De Leo, G., Butner, C., Martone, R.G., Shester, G., 2014. A risk-based framework for assessing the cumulative impact of multiple fisheries. Biol. Conserv. 176, 224–235. Milton, D.A., 2001. Assessing the susceptibility to fish populations of rare trawl bycatch: sea snakes caught by Australia’s northern prawn fishery. Biol. Conserv. 101, 281–290. Mwaluma, J.M., Kaunda-Arara, B., Rasowo, J., Osore, M.K., Oresland, V., 2011. Seasonality in fish larval assemblage structure within marine reef National Parks in coastal Kenya. Environ. Biol. Fishes 90, 393–404. Obura, D., 1999. Status report Kenya. In: Coral Reef Degradation in the Indian Ocean (CORDIO). Status reports and project presentations 32–35. Okemwa, G., Fulanda, B., Ochiewo, J., Kimani, E., 2009. Exploitation of coral reef fishes for the Marine Ornamental Trade in Kenya. In: Hoorweg, J., Muthiga, N.A. (Eds.), Advances in Coastal Ecology. African Studies Centre, Leiden-Netherlands. Olivier, K., 2001. FAO/Globefish Research Programme, vol. 67. United Nations Food and Agriculture Organization, Rome, Italy. Patrick, W.S., Spencer, P., Ormseth, O. et al., 2009. Use of productivity and susceptibility indices to determine stock vulnerability, with example applications to six U.S. fisheries. U.S. Department of Commerce, NOAA Technical Memo, NMFS-F/SPO-101. 90 pp. Available online at http://spo.nmfs. noaa.gov/tm/.process for assessing highly diverse and numerous bycatch. Environmental. Patrick, W.S., Spencer, P., Link, J., Cope, J., Field, J., Kobayashi, D., Lawson, P., Gedamke, T., Cortés, E., Ormseth, O., Begelow, K., Overholtz, W., 2010. Use of productivity and susceptibility indices to assess the vulnerability of United States fish stocks to overfishing. Fish. Bull. 108, 305–322. Rankin, T.L., Sponaugle, S., 2014. Characteristics of settling coral reef fish are related to recruitment timing and success. PLoS One 9 (9), e108871, http://dx. doi.org/10.1371/journal.pone.0108871. Rhyne, A.L., Tlusty, M.F., 2012. Trends in the marine aquarium trade: the influence of global economics and technology. Aquac. Aquar. Conserv. Legis. 5, 99–102. Rhyne, A.L., Tlusty, M.F., Schofield, P.J., Kaufman, L., Morris, J.A. Jr., Bruckner, A.W., 2012. Revealing the appetite of the marine aquarium fish trade: the volume and biodiversity of fish imported into the United States. PLoS One 7 (5), e35808, http://dx.doi.org/10.1371/journal.pone.0035808. Rocliffe, S., Peabody, S., Samoilys, M., Hawkins, J.P., 2014. Towards a network of locally managed marine areas (LMMAs) in the Western Indian Ocean. PLoS One 9 (7), e103000, http://dx.doi.org/10.1371/journal.pone.0103000. Roelofs, A., Silcock, R., 2008. A Sustainability Assessment of Marine Fish Species Collected in the Queensland Marine Aquarium Trade. Department of Primary Industries and Fisheries, Brisbane. Rubec, P., 1987. Fish capture methods and Philippine coral reefs—IMA Philippines visit Part II. Mar. Fish. Monit. 2, 30–31. Sadovy, Y.J., Vincent, A.C.J., 2002. Ecological issues and the trades in live reef fishes. In: Sale, P.F. (Ed.), Coral Reef Fishes: Dynamics and Diversity in a Complex Ecosystem. Academic Press, California, pp. 391–420. Samoilys, M.A., Maina, G.W., Osuka, K., 2011. Artisanal Fishing Gears of the Kenyan Coast. CORDIO/USAID, Mombasa, pp. 36. Samoilys, M.A., 1988. Abundance and species richness of coral reef fish on the Kenyan coast: the effects of protective management and fishing. Proceedings of the 6th Coral Reef Symposium 2, 261–266. Shuman, C.S., Hodgson, G., Ambrose, R.F., 2004. Managing the marine aquarium trade: is eco-certification the answer? Environ. Conserv. 31, 339–348. Shuman, C.S., Hodgson, G., Ambrose, R.F., 2005. Population impacts of collecting sea anemones and anemonefish for the marine aquarium trade in the Philippines. Coral Reefs 24, 564–573. State Department of Fisheries (SDF), 2011. Status report on the marine aquarium fishery 17.

G.M. Okemwa et al. / Fisheries Research 183 (2016) 19–31 Stevenson, T.C., Tissot, B.N., Dierking, J., 2011. Fisher behaviour influences catch productivity and selectivity in West Hawaii’s aquarium fishery. ICES J. Mar. Sci. 68, 813–822. Stobutzki, I., Miller, M., Brewer, D., 2001. Sustainability of fishery bycatch: a. Swaleh, K., Kaunda-Arara, B., Ruwa, R., Raburu, P., 2015. Ecosystem-based assessment of a prawn fishery in coastal Kenya using ecological indicators. Ecol. Indic. 50, 233–241. Tebua, S., 2005. Age-based Demography and Reproductive Ontongeny of Angelfishes Belonging to the Family Pomacanthidae. MSc Thesis. James Cook University, pp. 128. Thornhill, D.J., 2012. Ecological Impacts and Practices of the Coral Reef Wildlife Trade. Defenders of Wildlife, Washington. Tissot, B.N., Hallacher, L.E., 2003. Effects of aquarium collectors on coral reef fishes in Kona Hawaii. Conserv. Biol. 17, 1759–1768. Wabnitz, C., Taylor, M., Green, E., Razak, T., 2003. From Ocean to Aquarium. UNEP World Conservation Monitoring Centre, Cambridge.

31

Waldie, P.A., Blomberg, S.P., Cheney, K.L., Goldizen, A.W., Grutter, A.S., 2011. Long-term effects of the cleaner fish Labroides dimidiatus on coral reef fish communities. PLoS One 6 (6), e21201, http://dx.doi.org/10.1371/journal.pone. 0021201. Watson, D.L., Harvey, E.S., Kendrick, K.N., Anderson, M.J., 2007. Protection from fishing alters the species composition of fish assemblages in a temperate-tropical transition zone. Mar. Biol. 152, 1197–1206. Wood, E.M., 2001a. Collection of Coral Reef Fish for Aquaria: Global Trade, Conservation Issues and Management Strategies. Marine Conservation Society, U.K. Wood, E.M., 2001b. Global advances in the conservation and management of marine ornamental resources. Aquar. Sci. Conserv. 3, 65–77. Zhang, C., Suam, K.A., Donald, G., Richard, M., Jae, B., Park, Hee Won, Lee, Jong Hee, 2009. An ecosystem-based fisheries assessment approach for Korean fisheries. Fish. Res. 100, 26–41.