• No results found

Status and trends of coral reef health indicators on Saba (Caribbean Netherlands)

N/A
N/A
Protected

Academic year: 2022

Share "Status and trends of coral reef health indicators on Saba (Caribbean Netherlands)"

Copied!
88
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Status and trends of coral reef health indicators on Saba (Caribbean Netherlands)

Name: Céline van der Vlugt

Reg.nr. 910912900040

MSc Thesis nr. T 1998 THESIS July 2016

AQUACULTURE AND FISHERIES GROUP

LEERSTOELGROEP AQUACULTUUR EN VISSERIJ

Animal Sciences Group Aquaculture and Fisheries Group De Elst 1

6708 WD Wageningen The Netherlands Tel: +31 (0) 317 483307 Fax: +31 (0) 317 483962

(2)

1 Niets uit dit verslag mag worden verveelvoudigd en/of openbaar gemaakt door middel van druk, fotokopie, microfilm of welke andere wijze ook, zonder voorafgaande schriftelijke toestemming van de hoogleraar van de leerstoelgroep Aquacultuur & Visserij van Wageningen Universiteit.

No part of this publication may be reproduced or published in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the head of the Aquaculture & Fisheries Group of Wageningen University, The Netherlands.

(3)

2

T ABLE OF CONTENTS

Abstract ... 4

Introduction ... 6

Literature study ... 9

Anthropogenic factors on coral reef health ... 9

Coral reef indicators ... 12

Material and methods ... 14

Study area ... 14

Sampling methods ... 15

Image analysis ... 18

Data analysis ... 18

Assessment coral reef Health ... 20

Literature study ... 20

Results ... 21

Status fish ... 21

Status Key ecological families ... 23

Status Benthic Indicators ... 28

Reef health index ... 32

Trend in reef health indicators ... 33

Saba compared to the rest of the Caribbean... 34

Discussion ... 36

Current status of reef indicators and effect of the SMP ... 36

Status fish ... 36

Status benthic indicators ... 37

Trends of reef indicators on Saba ... 38

Saba compared to St. Eustatius ... 39

Saba compared to the Caribbean ... 39

Possible anthropogenic factors ... 39

Management options ... 40

Conclusion ... 41

Acknowledgments ... 42

References ... 43

Appendices ... 49

1. Selected fish species and Life history characteristics ... 49

2. Transect info ... 51

3. Fish species table ... 54

(4)

3

4. Output CPCe ... 56

5. Output glmmPQL model ... 59

6. Species benthic indicators ... 65

7. Pictures of corals with disease ... 67

8. R-script ... 75

(5)

4

A BSTRACT

Coral reefs are threatened globally and their trajectories of decline are very similar worldwide. In the Caribbean coral cover has declined by 80% since the 1970s. Global factors, such as ocean warming and acidification, play a role in this decline, but local factors, such as pollution, coastal development and overfishing, are the largest cause of this decline. These factors have caused an overall decline in coral cover and increase in macroalgae cover, demonstrating a shift from coral-dominated reefs to macroalgae-dominated reefs throughout the Caribbean. Understanding how much and why reefs are threatened and furthermore, why some reefs are threatened more than others, is important in providing and improving management options.

This study aimed to investigate the current status and trends of coral reef health around Saba (Caribbean Netherlands). Monitoring surveys were conducted on 20 sites around the island, according to the protocols of the Global Coral Reef Monitoring Network (GCRMN). On each site 5 transects were surveyed for the following coral reef indicators: fish biomass and density, coral and macroalgae cover, coral health, density of coral recruits, density sea urchins and cucumbers and water quality. Fish abundance and lengths were recorded by a Diver Operated Video-system (DOV). Fish biomass was calculated by using length-weight relationships. Coral and macroalgae cover and coral disease prevalence were assessed through photographs of the benthic substrate. Coral recruits, sea urchins and cucumbers were counted along the transect. Water quality was measured as water transparency using a Secchi disk.

Statistical analysis was done using a generalised linear mixed model (GLMM). Each indicator was tested for significant difference between the fished and unfished zone of the Saba Marine Park (SMP).

Furthermore, the effect of habitat complexity on the reef health indicators was taken into account in the model. The Reef Health Index was used to give an overall score on reef health in Saba.

The overall health of the reefs on Saba is ‘fair’, according to the RHI score. When looking at the different zones separately, the unfished zone scores better (fair) than the fished zone (poor). However, when looking at the coral reef indicators separately this difference in status between zones could not be seen so clearly and no statement could be made on whether this was a result of the protection of fishing. Total fish density was significantly higher in the unfished zone. This is in agreement with other studies, where higher fish abundance and biomass are found in areas protected by MPAs due to a decrease in fishing mortality. Acanthuridae (surgeonfish) biomass and density were found to be significantly higher in the fished area, which could be because of a lack of predator control in the fished area. However no other significant effects on fish assemblage were found, which could be explained by the relatively low fishing pressure. Habitat complexity did not differ between zones and did not have an effect on total fish biomass and density, but it did have a significant effect on Acanthuridae density, with higher abundance at higher levels of complexity. Coral cover was significantly higher in the unfished area, although no evidence was found for an indirect effect of fishing through trophic cascades due to a decrease in herbivores and increase in macroalgae. Although macroalgae cover seemed to be higher in the fished zone, no significant difference was found. Zonation had no effect on other benthic indicators.

There was a 20% decrease of coral cover in the last decades and an increase in herbivore biomass and density, with values around 4 times as high in 2015 than in 1991. This is consistent with trajectories in the rest of the Caribbean. A clear increase is seen for commercial fish (groupers and snappers), with a

(6)

5 2-3 fold increase for density and biomass in the fished zone and a 4 and 12 fold increase for density and biomass in the unfished zone. Indicating a decrease of fishing pressure over the years and the allowance for species to grow bigger in unfished areas. Much higher values for commercial fish biomass is seen in Saba than values recorded for other reefs in the Caribbean. Habitat complexity did not show a clear decrease and does not follow the trajectories of reef flattening in the rest of the Caribbean.

Overall, a combination of factors is most likely responsible for the degradation of Saba’s reefs.

Overfishing could have impacted the reefs before the establishment of the SMP in 1987 and these areas might still be recovering from the fishing pressure exerted before the establishment of the SMP.

Run-off and sedimentation increase eutrophication and susceptibility to disease. Together with a decrease in key herbivores, caused by the Caribbean wide die-off of Diadema antillarum (long-spined sea urchin) in 1983, this results in algal blooms, which eventually leads to a phase shift from coral to macroalgae dominated reefs. Reefs damaged by these local threats are more susceptible to the global threats of ocean warming and acidification (e.g. coral bleaching and increase in hurricanes) which will cause further degradation on the long-term.

(7)

6

I NTRODUCTION

Coral reefs are rich in biodiversity and they are of great value for millions of people (Burke et al., 2011).

People benefit from these reef ecosystems, because they are important sources of food and income.

Reefs attract tourists from all over the world and populations along coastlines increase, causing negative effects on the reef ecosystems (Burke et al., 2011). Coral reefs are threatened globally and the trajectories of decline regarding fish abundance and coral cover are very similar worldwide (Pandolfi et al., 2003). There are several factors causing the degradation of coral reefs. Currently, the largest threats are local. More than 60% of the world’s coral reefs are threatened by local human activities, such as overfishing and destructive fishing, coastal development (erosion) and pollution ( e.g. agriculture runoff, solid waste, shipping) (Burke et al., 2011). According to future trajectories, global threats, like ocean warming and acidification, will push the percentage of threatened reefs to more than 90% by 2030 (Hughes et al., 2003; Burke et al., 2011). It is clear that coral reefs need protection, but in order to properly protect and manage them, more specific knowledge is needed to understand which threats affect which reefs.

Coral decline is apparent in the Caribbean as well. Since the 1970s coral cover has declined by 80% in the entire Caribbean region (Gardner et al., 2003). Several factors play a role in the cause of this decline, including both global and local threats, such as climate change, pollution and overfishing. Most studies on coral reef health investigate a part of the Caribbean, of which the results are then projected to the entire Caribbean. There is, however, much variability in the condition of coral reefs in the Caribbean (Schutte et al., 2010). The causes underlying the degradation of coral reefs and why some reefs degrade more than others is poorly understood and needs to be given more attention in order to properly conserve and manage the coral reefs of the Caribbean (Hughes et al., 2010). The status report of the Global Coral Reef Monitoring Network (GCRMN) intended to do this, by documenting the variable condition of Caribbean reefs, in order to understand the factors that are causing the decline and the management options to prevent this decline (Jackson et al., 2014). These authors acquired data for several coral reef indicators from 34 different countries. They found an overall decline in coral cover and increase in macroalgae cover, demonstrating a shift from coral-dominated reefs towards macroalgae-dominated reefs (figure 1).

FIGURE 1: LARGE-SCALE SHIFTS FROM CORAL TO MACROALGAL COMMUNITY DOMINANCE SINCE THE EARLY 1970S AT (A) ALL LOCATIONS AND (B) THE 21 LOCATIONS WHERE LONG-TERM DATA WERE ACQUIRED (REPRODUCED FROM JACKSON ET AL., 2014).

(8)

7 One of the islands in the Caribbean for which information on the status of the coral reefs is not yet known to an adequate extent is Saba, part of the Caribbean Netherlands (figure 2). Saba is located close to the Saba Bank, which is an important nature area in the Caribbean Netherlands. In order to properly manage the Caribbean Netherlands, data is required on the health status of its nature areas, which includes the waters around Saba. Providing a baseline of the current status of reef health on Saba is critical for monitoring reef health in the future. Table 1 gives an overview of events that have impacted Saba’s nature areas since the 1980s (Buchan et al., 2014).

TABLE 1: OVERVIEW OF EVENTS IMPACTING THE BIODIVERSITY AND CONSERVATION OF BIODIVERSITY ON SABA

Timeline

Early 1980s Start of diving business on Saba

Since 1983 Outbreak of White Band Disease causes mortality to Acropora cervicornis (Staghorn coral) and Acropora palmata (Elkhorn coral)

1983-1984 Caribbean wide Diadema antillarum (long-spined sea urchin) die-off 1987 Establishment of the Saba Marine Park

1989 Hurricane Hugo (Category 4)

1995 Hurricane Luis and Marilyn (Category 4 and 2 respectively) 1996 Hurricane Bertha (Category 1)

1998 Hurricane Georges (Category 2)

1999 Hurricane Jose and Lenny (Category 1 and 4 respectively) 2000 Hurricane Debby (Category 1)

2005 Extensive bleaching event 2008 Hurricane Omar (Category 4) 2010 Hurricane Earl (Category 3) 2010 First invasive Lionfish detected

2015 Establishment of Yarari Marine Mammal and Shark Sanctuary 2015 Start fish and reef survey using GCRMN protocol

In an effort to protect coral reefs from the various threats, marine protected areas (MPAs) are established. The establishment of MPAs is widely increasing and is the foremost measure in marine conservation (Molloy et al., 2009; Ortiz & Tissot, 2012; Miller & Russ, 2014). MPAs can enforce different rules depending on its goals, like no fishing or no diving. Many studies have shown the benefits of implementing MPAs, such as increasing fish density and biomass in areas protected from fishing (Polunin & Roberts, 1993; Roberts, 1995; Claudet et al., 2006; Molloy et al., 2009). Studies have furthermore indicated that is important to include the effect of habitat characteristics when investigating the effect of MPAs (Ortiz & Tissot, 2012; Miller & Russ, 2014). Reefs with higher habitat complexity could provide for more species diversity and richness (Rogers et al., 2014). MPAs should provide diversity in habitats for all life stages to ensure proper conservation of the area (Ortiz & Tissot, 2012).

In 1987, a marine park was established around the island of Saba, the Saba Marine Park (SMP). The SMP contains several different zones, including no-take zones where fishing is not allowed (SCF, 2015).

In the other areas of the marine park, some methods of fishing are allowed, like hand lining, spear fishing and lobster traps (Polunin & Roberts, 1993). To assess the effect of the zonation in the marine park, a comparison will be made of the current status of reef health inside and outside the fishing zones. Sites in the unfished zones could harbour healthier reefs due to the protection of the SMP.

Habitat complexity will be taken into account in the analysis in order to differentiate between SMP

(9)

8 and habitat effects. The main hypothesis of this study is that reefs within the no-take zone will be healthier, in terms of higher fish abundance and biomass, higher coral cover and lower macroalgae cover. Moreover, reefs with a higher habitat complexity will have higher fish abundance and biomass.

FIGURE 2: POSITION OF SABA IN THE CARIBBEAN, STUDY AREA OF THIS RESEARCH, INDICATED IN RED. (ADAPTED FROM HOETJES &

CARPENTER, 2010)

(10)

9

L ITERATURE STUDY

The aim of this literature review is to identify the anthropogenic and natural factors that have impacted coral reef health in the Caribbean in the past and are still of concern in the present. Furthermore, several coral reef health indicators are identified, which can be used to assess coral reef health.

A

NTHROPOGENIC FACTORS ON CORAL REEF HEALTH

Coral reefs in the Caribbean have degraded most severely, caused by many factors and often a combination of factors, such as overfishing, pollution, disease outbreaks and global warming (Hughes et al., 2010). Many reefs are showing or have shown a phase-shift, which is a shift in community structure to new dominant species, driven by environmental stressors (Dudgeon et al., 2010). In this case, a shift from coral dominance to reefs dominated by macroalgae, with subsequent effects on other species inhabiting coral reefs. This section will discuss and explain the different factors that have caused the start of degradation of the Caribbean coral reefs and the factors that are currently playing a role in the further degrading of the reefs.

Long before the degradation of reefs through coral bleaching and disease, reefs were already affected by overfishing (Jackson et al., 2001; Pandolfi et al., 2003). Wing et al. (2001) support this finding.

Archaeological evidence from several Caribbean islands indicate a change in reefs after a period of exploitation, with a relation between the magnitude of change and the length of human occupation and intensity of exploitation (Wing et al., 2001). Jackson et al. (2001) recognized that overfishing altered coastal ecosystems as early as 35 000 to 40 000 years ago, however the ecological impact was limited until the 18th century when human disturbance increased. Human and natural disturbances impact the community structure and cause a shift in the dominant species on coral reefs (figure 3).

One of the anthropogenic disturbances is fishing. Fishing affects reefs directly, through the loss of biomass and species diversity, but also indirectly, through the change in community structure and trophic cascades (Jennings et al., 1995; Jackson et al., 2001; Hawkins & Roberts, 2004; Mumby et al., 2006a; Stallings, 2009; Pinca et al., 2012), which can lead to phase shifts. Hawkins and Roberts (2004) demonstrated that besides the known impacts of industrial fishing, artisanal fisheries, which are small- scale fisheries using traditional fishing methods, strongly reduce density and biomass of targeted species on coral reefs as well. They studied six Caribbean islands, with a range in fishing pressure, in order to determine the effect of artisanal fishing on fish assemblages and benthic structure. They calculated the biomass of the most commercially important fish families (groupers, snappers, parrotfish and surgeon fish) and found a decrease in biomass with an increase in fishing pressure, with a higher effect on the large-sized species. Moreover, coral cover and benthic complexity were found to be lower and macroalgae cover to be higher on islands with the highest fishing pressure. Stallings (2009) showed a strong negative relation between human population density and predatory reef fishes, with fishing as the most important driver as well. By implementing marine reserves, the population of these predatory reef fishes may be restored. However, it might be questioned whether an increase in density of predatory reef fishes would limit the population of their prey, including herbivore fishes such as the parrotfish, and with the parrotfish as the main herbivore grazer in the Caribbean, a decrease in their population may limit the levels of grazing, leading to an increase of macroalgae and eventually a phase shift in the ecosystem from coral to macroalgae dominated reefs (figure 3). However, Mumby et al. (2006a) demonstrated that this is not the case. Predation did cause a reduction in grazing by 4 to 8%, however this was overwhelmed by the fact that large-bodied

(11)

10 parrotfish showed an increase in density because they escape the risk of predation, leading to higher levels of grazing and a reduction in macroalgae cover (Mumby et al., 2006a). Furthermore, it is plausible that overfishing leads to outbreaks of disease and eutrophication in coral reef ecosystems.

Because of overfishing, populations of lower trophic levels may become too dense, making them more susceptible for diseases through increased transmission rates (Hochachka & Dhondt, 2000). This may have been the case with the Diadema die-off in the Caribbean (Jackson et al., 2001). Subsequently, altered ecosystems through trophic cascades are much more vulnerable to other disturbances, like eutrophication and climate change. It is clear that fishing affects coral reef ecosystems directly and indirectly. The establishment of marine reserves where fishing is prohibited may therefore be an important tool in preserving healthy reefs.

FIGURE 3: MODEL OF MAJOR FACTORS INVOLVED IN THE CORAL-ALGAL PHASE SHIFT. EXOGENOUS FACTORS ARE IN SHADED ENCLOSURES, WITH ANTHROPOGENIC EFFECTS IN RECTANGLES. DISEASE AFFECTS ALL LIVING COMPONENTS. ARROWS REPERESENT A GAIN TO THE COMPONENT THEY TOUCH, CIRCLES REPRESENT A LOSS. DASHED LINES REPRESENT WEAK LINKS OR THOSE WITH IMPACTS UNDER EXTREME CONDITIONS (REPRODUCED FROM MCMANIUS & POLSENBERG, 2004).

In the 1980s, Caribbean corals have suffered a sudden increase in mortality causing a decline in coral cover of about 80 – 90% (Hughes et al., 2010). The mass die-off of Diadema antillarum in 1983 played an important part in this decline (Lessios et al., 1984). An outbreak of an unknown pathogen was reported to affect the Diadema population in Panama in the beginning of 1983. In a year, this pathogen had spread over the entire Caribbean through surface currents and caused a massive decrease in Diadema populations, reaching levels of near-extinction (Lessios, 1988). Diadema have a profound role in Caribbean coral reefs, they are important herbivores and keep macroalgae levels in check through grazing, ensuring no overgrowth of macroalgae on corals (Lessios, 1988). Therefore, the die-off in 1983 made an impact on the whole ecosystem. Macroalgae cover drastically increased, which resulted in an increase of coral mortality (Lessios, 1988). Despite signs of slow recovery in some places, the Diadema population never fully recovered and remains very low throughout the Caribbean (Hughes et al., 2010;

Bowen, 2015).

In the same time period as the outbreak of the Diadema pathogen, multiple coral diseases were detected throughout the Caribbean; black band disease (BBD), white plague and white band disease (WBD) were among the ones most alarming (Antonius, 1976; Dunstan, 1977; Gladfelter, 1982). A variant of WBD reduced the coral Acropora palmata to 5% of its former cover (Gladfelter, 1982). The various forms of coral diseases spread with a high rate and contributed to the 80-90% decline in coral cover (Bowen, 2015). Of the 66 Caribbean coral species, at least 52 (82%) are susceptible to disease

(12)

11 (Sutherland et al., 2004). Although causes for outbreaks of coral diseases remain questionable, more evidence suggests that human-induced factors are involved. Warmer sea temperatures can deteriorate the effect of disease on coral reefs, it may reduce the immune response in corals and facilitate growth and virulence of pathogens (Selig et al., 2006). Moreover, land-based pollution may introduce pathogens to the marine environment (Sutherland et al., 2004).

Besides outbreaks of disease, global warming has another detrimental effect on coral reefs through the process of coral bleaching. Corals bleach when they are exposed to unusually high or low water temperatures, high UV radiation, reduced salinity, increased sedimentation, disease and pollution (Brown, 1997; Rohwer & Youle, 2010). As the corals lose their symbionts, the zooxanthellae, the polyps become transparent and the white skeleton becomes visible. When corals are bleached, they can survive for a couple of weeks or months at the most, but without the nutrition from the zooxanthellae their growth will stop and they will be more susceptible for diseases and eventually they will die (Knowlton, 2001; Rohwer & Youle, 2010). Rising sea temperature is believed to be the most important factor causing coral bleaching (Brown, 1997; Hoegh-Guldberg, 1999). A rise in sea temperature of only 1% can already cause corals to bleach (Glynn, 1993). This correlation of bleaching with high sea temperature and predicted climate models strongly suggests that bleaching events will be more frequent and could become annual in the future (Hoegh-Guldberg, 1999; Rohwer & Youle, 2010). The most severe bleaching event in the Caribbean occurred in 2005, caused by a high number of degree heating weeks (DHWs) (Wilkinson & Souter, 2008; Eakin et al., 2010). DHWs are the number of weeks where the sea-surface temperature exceeds 1°C above the normal summer temperature. Over 80% of the corals had bleached and over 40% died, with the most severe bleaching occurring in the Lesser Antilles (figure 4).

Underlying the rise in sea temperature is the rise in carbon dioxide levels. Next to ocean warming, higher carbon dioxide levels are also making the oceans more acidic (Rohwer & Youle, 2010; Bowen, 2015). Part of the atmospheric CO2 is absorbed by the ocean, making its waters more acidic and resulting in a decrease in carbonate ions. Coral reef growth depends on the availability of carbonate ions and with the acidification of the ocean, coral skeleton growth will decrease and the existing skeleton will be weakened (Knowlton, 2001; Hoegh-Guldberg et al., 2007). This will result in increased susceptibility to wave action and storms (Hoegh-Guldberg et al., 2007).

An increase in hurricanes in the Caribbean is another concern brought forth by global warming.

Hurricanes bring destruction to coral reefs and furthermore reducing salinity of the water on the reefs because of high rainfall, causing corals to bleach (Gardner et al., 2005; Bowen, 2015).

(13)

12 FIGURE 4: THERMAL STRESS AND BLEACHING DURING THE 2005 CARIBBEAN BLEACHING EVENT. (A) DEGREE HEATING WEEK (DHW) VALUES SHOWING THE HIGHEST THERMAL STRESS RECORDED DURING 2005. (B) % OF CORALS BLEACHED (REPRODUCED FROM EAKIN ET AL., 2010).

Densities of both residents and tourists in the Caribbean are very high and are proven to be harmful to coral reefs. The GCRMN report shows significant negative relation for coral cover with both the density of residents and tourists (Jackson et al., 2014). High densities of residents and tourists cause the coastlines of Caribbean islands to keep developing, impacting reefs either directly, by physical damage, or indirectly, by sediment run-off, sewage and pollution (Burke et al., 2011). These indirect impacts lead to eutrophication, resulting in more nutrient-rich waters. This causes an increase in macroalgae and a decrease in water quality, which in turn results in an increase of coral disease (Burke et al., 2011; Bowen, 2015). Many Caribbean islands have no sewage treatment, an estimated 80 to 90% of wastewater is discharged untreated (UNEP, 2006). Tourism can damage reefs as well, hotel development can pollute and physically damage reefs (Burke et al., 2011). Moreover, divers can damage reefs through trampling and breaking fragile corals (Hawkins & Roberts, 1993). Other threats induced by human activities include watershed-based and marine-based pollution and damage.

Deforestation, agriculture and livestock lead to erosion, sediment run-off and nutrient pollution.

Furthermore, commercial and recreation ships can pollute the water through e.g. leakages, sewage or solid waste, and they can physically damage reefs through e.g. anchors or oil spills (Burke et al., 2011).

C

ORAL REEF INDICATORS

To assess the health of coral reefs several coral reef health indicators can be used. When assessing the health of the entire coral reef ecosystem, important indicators to look at are the fish population, coral and macroalgae cover, habitat complexity, coral disease, presence of key macro-invertebrates (e.g. sea urchins) and water quality (GCRMN, 2014; AGRRA, 2016). All these indicators are, either positively or negatively, interacting. According to the Healthy Reefs Initiative (HRI, 2015), which take into account four of these indicators, a reef is considered unhealthy when it has a low population of herbivorous and commercial fish, low cover of living coral and a high cover of macroalgae.

The abundance of key herbivore fish species (parrotfish and surgeonfish) is important for healthy reefs, because through grazing they decrease the cover of macroalgae. The abundance of key carnivore fish species (snappers, grunts and groupers) is important for healthy reefs in terms of predator control and preventing the occurrence of trophic cascades. Furthermore, the abundance of key carnivore fish gives insight on the effect of fishing since they are commercially important species (Hawkins & Roberts,

(14)

13 2004). The abundance and richness of fish species is influenced by coral cover and habitat complexity.

The percentage of cover and richness in coral species are strongly associated with fish species richness and abundance (Komyakova et al., 2013). Higher levels of habitat complexity increase the availability of shelters and variety of food sources, hereby increasing the abundance and number of species (Hixon

& Menge, 1991; Rogers et al., 2014).

Together with herbivorous fish, sea urchins are needed for healthy reefs to keep macroalgae in check.

The Diadema antillarum die-off in 1983 proved that these sea urchins are key herbivores on Caribbean reefs and are thus important to monitor. Sea cucumbers are important for healthy reefs, because of their significant role on coral reefs in nutrient recycling by converting organic detritus (Purcell et al., 2013).

High macroalgae cover is an indicator of poor reef health because of its negative effect on corals.

Firstly, macroalgae take up the space that coral recruits would settle on. Secondly, macroalgae are able to overgrow coral recruits and small corals of a few years old (Mumby et al., 2006b). And lastly, macroalgae can also partially overgrow larger corals, inflicting damage to the corals and causing the larger corals to separate into smaller patches (Hughes & Tanner, 2000).

Lower coral cover in turn leads to reduced reproduction of coral polyps, because of a lower production of gametes and furthermore because the rate of fertilization of those gametes is reduced since distances between coral colonies increase (Knowlton, 2001). This also results in an increase of budding, leading to a less genetically diverse reef and thus a less stable reef (Knowlton, 2001). Coral cover is furthermore influenced by coral disease. Previous outbreaks of coral disease have proven to cause severe decreases in coral cover and coral health is therefore important to monitor.

Water quality is an indicator of reef health, because it can influence coral and macroalgae growth, coral recruitment and the prevalence of diseases (Jackson et al., 2014). Increased turbidity and reduced light levels are not favourable for corals and may promote macroalgae growth (De’ath & Fabricius, 2010).

(15)

14

M ATERIAL AND METHODS S

TUDY AREA

The study was conducted between September and December 2015 in the waters around Saba. Saba is a volcanic island located in the north-eastern Caribbean. It has an area of 13 km2 with the highest point of 877 meters on Mount Scenery (figure 5). Steep and rocky cliffs form the coastline of Saba. The waters around Saba contain volcanic features, like boulders, pinnacles and lava formations which provide structure for corals to develop (Polunin & Roberts, 1993; Klomp & Kooistra, 2003).

Furthermore, the south-western part of the island harbours some true reefs with a carbonate framework (Polunin & Roberts, 1993).

In 1987, the Saba Marine park (SMP) was established by the Saba Conservation Foundation (SCF). The park circles the entire island, with a range from the high-water mark to a depth of 60 meters (SCF, 2015). The park is divided into four discontinuous zones: a multipurpose zone, a recreational zone, mooring zones and no-take zones (figure 5). The no-take zone is the only area where fishing is not allowed, approximately 33% of the SMP is assigned as no-take area (Noble et al., 2013). Zones 1-3 will be referred as the fished zone and zone 4 as the unfished zone.

FIGURE 5: MAP OF SABA AND THE SMP ZONATION, 1: MULTIPURPOSE ZONE, 2: RECREATIONAL ZONE, 3: MOORING ZONE, 4: NO TAKE ZONE. ALL CIRCLES REPRESENT SITES SURVEYED IN THIS STUDY, WITH BLUE CIRCLES AS SITES SURVEYED BY PREVIOUS STUDIES AS WELL (SEE TABLE 2 FOR SITE NAMES) (ADAPTED FROM DUTCH CARIBBEAN NATURE ALLIANCE, 2014).

4

3 2

1

4

3

4 26

25

24 22

23 20 18

15 14 13

12 11 9 6

Wells Bay

29 19

7

(16)

15

S

AMPLING METHODS

The Global Coral Reef Monitoring Network (GCRMN) protocols were used as sampling methods for this study (GCRMN, 2014). Six elements are described in these methods, 1: the abundance and biomass of key reef fish families, 2: the relative cover of reef-building organisms (corals) and their dominant competitors (macroalgae), 3: the assessment of coral health, 4: the recruitment of reef-building corals, 5: the abundance of key macro-invertebrate species, and 6: water quality (GCRMN, 2014). With these elements an overview of the current condition of the coral reef ecosystem is given. It also provides an indication of possible future trajectories. All these elements were surveyed in this study.

Study sites were chosen from the list of dive sites, making sure the different zones were included, while taking into account the sites that were previously surveyed by other studies (table 2; figure 5). In total 20 sites were surveyed, including the 17 sites that are previously surveyed plus an additional 3 sites in zones 1 and 4, chosen for their contrast in fishing (table 2; figure 5). At each site, the surveys were conducted along 5 transects of 30m each (transect info can be found in Appendix 2).

TABLE 2: INFORMATION ON WHICH SITES ARE SURVEYED BY THE DIFFERENT STUDIES, 1991: POLUNIN & ROBERTS (1993), 1993: ROBERTS (1995), 1993-1995: ROBERTS & HAWKINS (1995), 1999: KLOMP & KOOISTRA (2003), 2008: NOBLE ET AL. (2013).

Zone Nr dive Latitude Longitude Study year

Name dive site SMP site (°'N) (°'W) 1991 1993 1993-1995 1999 2008

Greer Gut 1 20 17°36'42.54 63°14'30.30 x x

Big Rock Market 1 22 17°36'45.06 63°14'10.44 x x x x

Giles Quarter Shallow 1 23 17°36'42.60 63°14'28.80 x

Hole In The Corner 1 24 17°37'3.72 63°13'34.92 x x x x x

Core Gut 1 25 17°37'51.90 63°13'3.54 x

Green Island 1 26 17°38'53.88 63°13'50.16 x x x x

David’s Drop Off 1 29 17°37'6.12 63°13'25.44

Torrens Point 2 9 17°38'35.88 63°15'11.94 x x x x

Wells Bay 3 17°38'30.96 63°15'19.68 x x x

Fort Bay 3 17°36'47.70 63°14'50.16 x x x

Diamond Rock 4 6 17°38'49.80 63°15'24.00 x x x

Man of War Shoals 4 7 17°38'47.94 63°15'19.20

Custom's House 4 11 17°37'54.84 63°15'29.58 x x

Porites Point 4 12 17°37'45.54 63°15'31.98 x x x

Babylon 4 13 17°37'42.66 63°15'34.50 x x x x x

Ladder Labyrinth 4 14 17°37'34.44 63°15'36.24 x x x x

Ladder Labyrinth 2 4 33 17°37'33.60 63°15'37.80 x x

Hot Springs 4 15 17°37'28.68 63°15'34.50 x x x x

Tent Reef 4 18 17°36'58.80 63°15'30.60 x x x x x

Tent Reef Deep 4 19 17°36'59.34 63°15'21.18

(17)

16 The following surveys were conducted along each transect:

1. Abundance and biomass of fish species

In total, the density and size structure of all selected fish species within the survey area were collected (see Appendix 1 for fish species list). The survey provided core information on key families such as snappers (Lutjanidae), groupers (Serranidae), grunts (Haemulidae), parrotfishes (Scaridae), and surgeonfishes (Acanthuridae).

To collect this data a Diver Operated stereo-Video (DOV) system was used. This stereo-video system allows researchers to sample fish in a non-destructive manner and measure fish length relatively accurate (Harvey & Shortis, 1995; Watson et al., 2005). Fish species were recorded by a SCUBA diver, swimming at a slow and continuous speed of about 0.6 km/h along the transects, while holding the DOV (figure 6).

FIGURE 6: DIVER-OPERATED STEREO-VIDEO SYSTEM (DOV) (ADAPTED FROM WATSON ET AL., 2010).

2. Assessment of benthic environment

The goal of this survey was to document the structural complexity of the substratum and the relative cover of reef-building, stony corals and their dominant competitors. As such, data was collected on the estimated complexity using a 6 point scale and the percentage of the reef bottom that is covered by stony corals, gorgonians, sponges, and various types of algae (turf algae, macroalgae, and crustose coralline algae).

The photoquadrat method (GCRMN, 2014) was used to estimate the cover of key taxa on the reef benthos. Photographs were taken of the reef surface in standardized quadrat areas (0.9 m x 0.6 m), along each of the 5 transect lines set for counting fish, with 15 images captured per transect line (i.e., one image taken at every other meter marker on the transect tape). In total, 75 benthic photographs were collected at each site (5 transect lines x 15 photographs per line). The structural complexity of the substratum was estimated by making photographs towards the front along each of the transect lines, with 2 images per transect line (i.e., one image taken at the beginning and one at the end of the transect line). A 6 point scale was used to determine the complexity of the substrate in these photographs: 0 = bare substratum, 1 = low and sparse relief, 2 = low but widespread relief, 3 = moderate complexity with numerous caves and overhangs, 4 = high complexity with cave systems, 5 = extreme complexity with numerous caves and overhangs (Polunin & Roberts, 1993; Noble et al., 2013).

(18)

17 3. Assessment of coral health

Coral health was assessed by documenting the prevalence of disease in stony corals. Disease prevalence is a metric describing the proportion of corals that show signs or pathologies of any disease. The photoquadrats collected for the benthic cover assessment were used to estimate disease prevalence in corals. Data was recorded as the proportion of images collected that contain a coral with any disease pathology. For example, if there are four colonies in a particular photoquadrat and any of these colonies shows signs of disease, this image would be tagged as “with disease”. The number of images that are “with disease” were divided by the total number of images (15 per transect) to generate a proportional estimate of disease prevalence.

4. Coral recruitment and macroalgae height

Data on coral recruitment was collected by estimating the density of young corals that were likely to contribute to the next generation of adult corals on the reef. Coral recruits were defined operationally as any stony coral that is greater than 0.5 cm2 and smaller than 4.0 cm2, and that were visible to the diver in situ. Estimates of coral recruit density were recorded from replicate 25 cm x 25 cm (625 cm2) quadrats. A total of 5 quadrats were surveyed along each of the first 3 transects used for the benthic and fish surveys. These quadrats were placed at 2- meter intervals along the transect line, i.e., with the lower corner of the quadrat placed at the following meter marks: 2, 4, 6, 8, and 10 meter. When possible, each coral recruit was recorded to the species level. Faveo fragum colonies were excluded from the count since adult colonies are often smaller than 4 cm. Furthermore, within each of the quadrats, the canopy height of macroalgae was measured.

5. Abundance of key macro-invertebrate species

The goal of this survey was to estimate densities of biologically and economically important species on the reef. The number and species identity of each echinoid (i.e., sea urchin) and holothurian (i.e., sea cucumber) was recorded along each transect placed for the benthic and fish surveys.

6. Water quality

As an estimate of the integrated water quality, the data that was collected are the distances at which standardized Secchi disks (black and white disk, attached to a measured and marked rope) were visible in the waters at the bottom of the reef. Since surface waters were too clear to vertically deploy the Secchi disk, only horizontal measurements were made. The Secchi disk was placed or held at one location, along with the end of a transect tape. An in-water observer would swim away from the disk, pulling the transect tape and recorded the distance at which the Secchi disk was no longer visible.

(19)

18

I

MAGE ANALYSIS

The videos acquired from the DOV survey were analysed using the software programme EventMeasure (SeaGis, 2012). Species, length and abundance was recorded for the selected fish species (Appendix 1) sighted within 1 meter on each side of the transect line.

The photographs, acquired from the benthic survey, were analysed using the software Coral Point Count with Excel extensions (CPCe) (Kohler & Gill, 2006) (the full output of the CPCe per transect can be found in Appendix 4). On each photograph (covering an area of 60 x 90 cm), 25 points were identified in random locations across the image. The benthic type under each point is classified into a standardized benthic category including key species (and some broader groups) of corals and algae.

Reef building corals were identified to species level, soft corals and macroalgae to genus level.

D

ATA ANALYSIS

Abundance data was directly obtained from the video analysis and the densities of fish species and families were calculated per transect and per site (#/100m2) (info on recorded species can be found in Appendix 3). The weight of each fish with a recorded length was calculated using the length-weight relationship W=a*Lb, where W is the weight of the fish (g), L is its length (cm) and a and b are species- specific parameters (Bohnsack & Harper, 1988). Fish length was directly obtained from the video analysis and parameters a and b were derived from Fishbase (Froese & Pauly, 2016; Appendix 1,).

Lengths could not be measured for ±5% of all recorded individuals. For individuals with missing lengths, the mean length of that species in the same transect was used to obtain biomass. If no length was available from the same transect, the mean length of that species in the same site was used. If the species had not been measured in that site, the mean length of that species measured in another site with a similar habitat was used to obtain biomass. If the species had not been measured at all, a mean length was obtained from Fishbase (Froese & Pauly, 2016). Biomass and density was calculated for every transect and site for five key families: Scaridae (parrotfish), Acanthuridae (surgeonfish), Haemulidae (grunts), Lutjanidae (snappers) and Serranidae (groupers).

Statistical analysis was done in an R environment (R Core Team, 2015). To investigate the effect of the zonation system of the SMP, sites in the fished and unfished zone were compared (with 10 sites in each zone). Every indicator was tested for significant difference between the fished and unfished area.

However, zonation is not the only factor that can contribute to this difference. Differences in habitat can also explain the possible differences in reef health indicators. Therefore, habitat complexity was also taken into account when checking for differences.

As there are 5 transects per site, the values of reef health indicators at these 5 transects are likely to be more related to each other than to the indicator values from transects on different sites. The nested structure of the data is visualised in figure 7. Since a linear regression model does not take this relatedness into account, a mixed effect model was used for statistical analysis (Zuur et al., 2009). The mixed effect model has two components that contain explanatory variables, the fixed and the random component. Blocks in observational studies replicated across sites or times are among the most familiar types of random effect (Bolker et al., 2009). In this case, location (site), is used as a random effect.

Since the data are not normally distributed and it includes random effects, the best tool to analyse this data set is a generalized linear mixed model (GLMM). The GLMM combines the properties of two

(20)

19 statistical frameworks, linear mixed models (which includes random effects) and generalized linear models (which handles non-normal data by using link functions and family distributions) (Bolker et al., 2009). In R, the glmmPQL (pseudo- and penalized quasilikelihood) model was used to test for significance. The R-script including the MASS-package (Venables & Ripley, 2002) can be found in Appendix 8. PQL is a flexible technique and is the most widely used GLMM approximation (Bolker et al., 2009). By specifying the family distribution, link function and structure of random effects, all indicators can be analysed using this model (table 3). For density data the Poisson distribution is selected with a logarithmic link, this is typically used for count data (Bolker et al., 2009). For proportion data a logit transformation was used, since the previously widely-used arcsine-transformation is no longer thought to be effective (Warton & Hui, 2011).

To check if the data set was balanced, the number of transects per Polunin score were counted for the fished and unfished zone (table 4). To create a more balanced data set, containing roughly equal numbers of transects per fishing zone and Polunin score, transects with a Polunin score of 1 were removed from the analysis, because there was only one transect of this score in the unfished zone.

The indicators were tested for the main fixed factors zonation and complexity, and the interaction between the main factors. The structure of the model is as follows:

If the interaction turned out to be non-significant, the indicators were tested without the interaction and with the main factors separately. Afterwards, the model was tested for goodness of fit by testing the residuals. If no significance was found, only the results for the model with the main fixed factors tested separately were reported.

The estimates and standard errors were obtained from the model and were transformed back from the log or logit scale. The means and confidence limits were calculated per zone. If the interaction was significant, the means and confidence limits were calculated per zone, per complexity score.

FIGURE 7: SET UP OF THE GCRMN DATA. MEASUREMENTS WERE TAKEN ON 20 SITES (10 IN THE FISHED ZONE AND 10 IN THE UNFISHED ZONE), AND ON EACH SITE 5 TRANSECTS WERE SAMPLED.

GCRMN data Fished

Site 1

Tr 1 Tr 2 Tr 3 Tr 4 Tr 5

Site 10

Tr 1 Tr 2 Tr 3 Tr 4 Tr 5 ...

Unfished

Site 11

Tr 1 Tr 2 Tr 3 Tr 4 Tr 5

Site 20 Tr 1 Tr 2 Tr 3 Tr 4 Tr 5 ...

Fixed Random

Indicator value = Zonation + Complexity + Zonation*Complexity + Location

(21)

20 TABLE 3: TYPES OF INDICATORS AND THEIR ASSOCIATED DISTRIBUTIONS AND LINK FUNCTIONS.

What Type Family Link Random effect

Fish biomass Weight Gaussian Log Location

Fish density Count Poisson Log Location

Coral and

macroalgae cover

Proportion Binomial Logit Location

TABLE 4: BALANCE BETWEEN THE FISHED AND THE UNFISHED ZONE IN TERMS OF HABITAT COMPLEXITY. NUMBERS INDICATE THE AMOUNT OF TRANSECTS IN THE FISHED AND UNFISHED ZONE PER POLUNIN SCORE.

Polunin score

1 2 3 4 5

F 4 17 24 5 0

UF 1 6 38 5 0

A

SSESSMENT CORAL REEF

H

EALTH

Reef health was evaluated using the Reef Health Index (RHI). The RHI evaluates the ecological condition of the reefs around Saba according to four key indicators that are important to the structure and functioning of healthy coral reef ecosystems (Healthy Reef Initiative (HRI), 2015). The mean value of each indicator was compared to the thresholds established by the HRI (figure 8), these indicator values were then given a grade from 1 (critical) to 5 (very good). The mean of the four grades was calculated in order to obtain a RHI score for each site.

FIGURE 8: THRESHOLDS FOR THE FOUR INDICATORS OF THE REEF HEALTH INDEX (RHI) (REPRODUCED FROM HEALTHY REEF IINITIATIVE, 2015).

L

ITERATURE STUDY

In order to assess the trends of coral reef health indicators on Saba, literature was searched for studies on coral reef health in Saba since 1990. The Global Coral Reef Monitoring Network status report of the Caribbean from 1970 – 2012 (Jackson et al., 2014) was consulted for data on trends of coral reef health indicators in the wider Caribbean. The trends of the indicators in the Caribbean were then compared to the trends of the indicators in Saba.

(22)

21

R ESULTS S

TATUS FISH

Total biomass and density of all fish was recorded per site (figure 9). Sites in the unfished zone appeared to show higher values of fish biomass, but there was also more variability between sites (figure 9). The same goes for total fish density, with higher densities in the unfished zone, also with much variability between sites. When comparing fish biomass with density, there were a few peaks in fish density where no peaks were found at the same sites for fish biomass. This could be explained by differently-sized fish at different sites. The results of the generalised linear mixed model (GLMM), testing the effect of zonation and complexity, are shown in table 5 (the full output of the GLMM can be found in Appendix 5, here only the significant and marginally significant results are shown). The interaction between zonation and complexity had no significant effect on total fish biomass and was left out of the model. The main factors zonation and complexity showed no significant effect on total fish biomass at the 0.05-level, but the difference between the fished and unfished zone was marginally significant (p=0.083; table 5; figure 10b), with a mean biomass of 5.06 (kg/100m2) in the fished and a biomass of 11.6 (kg/100m2) in the unfished zone. For total fish density, the interaction between zonation and complexity did have a significant effect (p=0.018) and was therefore included in the model. Furthermore, the difference between the fished and unfished zone was significant (p=0.025).

The mean density in the fished zone had a mean value of 30.9 (#/100m2), irrespective of habitat complexity, while the density in the unfished zone ranged from a mean of 125.4 (#/100m2) at complexity 2 to a mean of 32.2 (#/100m2) at complexity 4.

FIGURE 9: CURRENT STATUS OF THE BIOMASS (KG/100M2) AND DENSITY (#/100M2) OF ALL SELECTED FISH SPECIES PER SITE, WITH A DISTINCTION BETWEEN THE FISHED AND THE UNFISHED ZONE.

(23)

22 FIGURE 10: TOTAL FISH BIOMASS (KG/100M2) AND DENSITY (#/100M2) PER ZONE AND PER COMPLEXITY SCORE (A AND C RESPECTIVELY), AND TOTAL FISH BIOMASS (KG/100M2) AND DENSITY (#/100M2) CALCULATED FROM THE GLMM WITH 95% CONFIDENCE LIMITS (SHOWN FOR SIGNIFICANT EFFECTS OF THE FIXED FACTORS) (B AND D RESPECTIVELY). GRAPHS MARKED WITH * INDICATE SIGNIFANT DIFFERENCE.

TABLE 5: GENERALISED LINEAR MIXED MODEL RESULTS FOR TOTAL FISH BIOMASS AND DENSITY (F = FISHED ZONE, UF = UNFISHED ZONE, C2/3/4 = POLUNIN COMPLEXITY SCORES 2, 3 & 4).

Total fish biomass Estimate Std. Error t-value p-value

Main factors separate

Intercept 1.62 0.39 4.16 <0.00005

Fished vs Unfished 0.83 0.45 1.84 0.083

Total fish density With

interaction

Intercept 3.16 0.62 5.12 <0.00005

Fished vs Unfished 1.67 0.68 2.45 0.025*

Complexity 2 vs 3 0.64 0.65 0.98 0.33

Complexity 2 vs 4 0.04 0.74 0.05 0.96

Interaction F/UF:C2/C3 -1.65 0.68 -2.43 0.018*

Interaction F/UF:C2/C4 -1.40 1.03 -1.36 0.18

a b

c d

*

(24)

23

STATUS KEY ECOLOGICAL FAMILIES

The current status of fish biomass and density per key ecological family were recorded per site (figure 11).

SCARIDAE

Scaridae biomass and density showed much variability between sites in both zones (figure 11). No significant effects of zonation or habitat complexity were found for Scaridae biomass and density (table 6). Mean biomass was 661 (kg/100m2) in the fished and 907 (kg/100m2) in the unfished zone (figure 13b). Mean density was 3.8 (#/100m2) in the fished and 4.7 (#/100m2) in the unfished zone (figure 14b).

ACANTHURDIAE

Sites in the fished zone appeared to show higher values of Acanthuridae biomass and density (figure 11). Zonation indeed had a significant effect on both biomass and density (p=0.0085, p=0.0067), with a mean biomass of 771 (kg/100m2) in the unfished and 1701 (kg/100m2) in the fished zone (table 6;

figure 13d). Mean density was 14.4 (#/100m2) in the fished and 7.2 (#/100m2) in the unfished zone (figure 14d). Also, complexity was marginally significant for Acanthuridae biomass (p=0.051) and significant for density (p=0.026) (table 6; figure 13c & 14c).

LUTJANIDAE

Sites in the unfished zone showed higher values of Lutjanidae biomass and density, although there was much variability between sites (figure 11). The peaks in biomass were caused by a few large sized snappers recorded in those sites. No significant effects of zonation or habitat complexity were found for Lutjanidae biomass and density (table 6). Mean biomass was 70 (kg/100m2) in the fished and 985 (kg/100m2) in the unfished zone (figure 13f). Mean density was 1.3 (#/100m2) in the fished and 3.4 (#/100m2) in the unfished zone (figure 14f).

HAEMULIDAE

Sites in the unfished zone showed higher values of Haemulidae biomass and density, although there was much variability between sites (figure 11). No significant effects of zonation or habitat complexity were found for Haemulidae biomass and density (table 6). Mean biomass was 149 (kg/100m2) in the fished and 572 (kg/100m2) in the unfished zone (figure 13h). Mean density was 1.9 (#/100m2) in the fished and 4.6 (#/100m2) in the unfished zone (figure 14h).

SERRANIDAE

Sites in the unfished zone showed higher values of Serranidae biomass, although there was much variability between sites (figure 11). However, the difference in density was not so clear as was seen for biomass. The peaks in biomass were caused by a few large sized groupers recorded in those sites (figure 12). Serranidae biomass and density were non-significant for all terms (table 6), with mean biomass of 396 (kg/100m2) in the fished and 899 (kg/100m2) in the unfished zone (figure 13j). Mean density was 3 (#/100m2) in the fished and 3.8 (#/100m2) in the unfished zone (figure 14j).

(25)

24 FIGURE 11: CURRENT STATUS OF FISH BIOMASS (G/100M2) AND DENSITY (#/100M2) PER FAMILY PER SITE (SITES BRM TO WB ARE IN THE FISHED ZONE, SITES BA TO TRD ARE IN THE UNFISHED ZONE).

FIGURE 12: DISTRIBUTION OF GROUPER DENSITY (#/100M2) AND BIOMASS (G/100M2) PER SITE. SMALL: GRAYSBY & CONEY; MEDIUM:

RED HIND & ROCK HIND; LARGE: MARBLED, YELLOWFIN & BLACK.

Fished

Fished Unfished

Unfished

Fished Unfished

Unfished Fished

(26)

25 TABLE 6: GENERALISED LINEAR MIXED MODEL RESULTS FOR ACANTHURIDAE BIOMASS AND DENSITY (F = FISHED ZONE, UF = UNFISHED ZONE, C2/3/4 = POLUNIN COMPLEXITY SCORES 2, 3 & 4). WITHOUT INTERACTION: THE MAIN FACTORS, ZONATION AND COMPLEXITY, ARE TESTED TOGETHER IN THE MODEL, BUT WITHOUT THE INTERACTION BETWEEN THE TWO. MAIN FACTORS SEPARATE: THE MAIN FACTORS ARE TESTED SEPERATELY; A MODEL FOR ZONATION ALONE, AND A MODEL FOR COMPLEXITY ALONE.

Scaridae biomass Estimate Std. Error t-value p-value

Main factors separate Intercept 6.49 0.28 23.10 <0.00005

Fished vs Unfished 0.32 0.34 0.92 0.37

Acanthuridae biomass

Without interaction Intercept 6.98 0.37 18.72 <0.00005

Fished vs Unfished -1.00 0.34 -2.95 0.0085*

Complexity 2 vs 3 0.79 0.40 1.98 0.051

Complexity 2 vs 4 -0.41 1.06 -0.39 0.70

Main factors separate Intercept 7.44 0.15 49.08 <0.00005

Fished vs Unfished -0.79 0.36 -2.21 0.04*

Lutjanidae biomass

Main factors separate Intercept 4.25 4.67 0.91 0.37

Fished vs Unfished 2.64 4.68 0.56 0.58

Haemulidae biomass

Main factors separate Intercept 5.01 2.24 2.24 0.028

Fished vs Unfished 1.35 2.31 0.58 0.57

Serranidae biomass

Main factors separate Intercept 5.98 0.73 8.15 <0.00005

Fished vs Unfished 0.82 0.80 1.03 0.32

Scaridae density

Main factors separate Intercept 1.35 0.17 8.07 <0.00005

Fished vs Unfished 0.21 0.22 0.94 0.36

Acanthuridae density

Without interaction Intercept 2.33 0.30 7.80 <0.00005

Fished vs Unfished -0.92 0.30 -3.06 0.0067*

Complexity 2 vs 3 0.69 0.30 2.27 0.026*

Complexity 2 vs 4 -1.01 0.44 -1.38 0.17

Main factors separate Intercept 2.67 0.13 20.71 <0.00005

Fished vs Unfished -0.69 0.28 -2.46 0.024*

Lutjanidae density

Main factors separate Intercept 0.26 0.65 0.40 0.69

Fished vs Unfished 0.97 0.72 1.34 0.19

Haemulidae density

Main factors separate Intercept 0.67 1.07 0.63 0.53

Fished vs Unfished 0.86 1.19 0.72 0.48

Serranidae density

Main factors separate Intercept 1.11 0.17 6.58 <0.00005

Fished vs Unfished 0.23 0.22 1.07 0.30

(27)

26 FIGURE 13: FISH BIOMASS (G/100M2) PER FAMILY, PER ZONE AND PER COMPLEXITY SCORE (LEFT), AND FISH BIOMASS (G/100M2) PER FAMILY CALCULATED FROM THE ESTIMATES WITH 95% CL ERRORBARS (RIGHT) (ERRORBARS FOR LUTJANIDAE AND HAEMULIDAE IN FISHED ZONE TOO LARGE FOR FIGURE). GRAPHS MARKED WITH * INDICATE SIGNIFANT DIFFERENCE.

*

a b

c d

e f

g h

i j

(28)

27 FIGURE 14: FISH DENSITY (#/100M2) PER FAMILY, PER ZONE AND PER COMPLEXITY SCORE (LEFT), AND FISH DENSITY (#/100M2) PER FAMILY CALCULATED FROM THE ESTIMATES WITH 95% CL ERRORBARS (RIGHT). GRAPHS MARKED WITH * INDICATE SIGNIFANT DIFFERENCE.

*

a b

c d

e f

g h

i j

(29)

28

S

TATUS

B

ENTHIC

I

NDICATORS CORAL COVER

Overall, coral cover appeared higher in the unfished zone (figure 15). Zonation indeed had a significant effect on coral cover (p=0.019), with a mean of 4.2 (%) coral cover in the fished and 6.6 (%) in the unfished zone (table 7; figure 18). Habitat complexity was marginally significant for coral cover (p=0.062) (table 7; figure 18). A table of recorded coral species can be found in Appendix 6.

MACROALGAE

Macroalgae cover appeared lower in the unfished zone, however there was a lot of variability between sites (figure 15). Macroalgae and cyanobacteria together accounted for the majority of benthic cover for a mean of 28% on most sites. No significant effects of zonation or habitat complexity were found for macroalgae cover (table 7). Mean cover was 16.1 (%) in the fished and 13.6 (%) in the unfished zone (figure 18). A table of recorded macroalgae species can be found in Appendix 6.

Mean macroalgae height was higher in the fished zone (figure 16). Zonation indeed had a significant effect on macroalgae height (p=0.047), with a mean height of 2.11 (cm) in the fished and 1.15 (cm) in the unfished zone (table 7; figure 18). Also, complexity was significant for macroalgae height (p<0.0005) (table 7; figure 18).

CORAL RECRUITS

Density of coral recruits seemed to be higher in the fished zone, however at one site no coral recruits were found (figure 16). No significant effects of zonation or habitat complexity were found for density of coral recruits (table 7). Mean density was 5.3 (#/100m2) in the fished and 4.1 (#/100m2) in the unfished zone (figure 18). A table of recorded coral recruit species can be found in Appendix 6.

WATER QUALITY

Water quality (horizontal transparency) seemed to be the same over all sites (figure 16). No significant effects of zonation or habitat complexity were found for water quality (table 7). Mean horizontal transparency was 26.1 (m) in the fished and 25.7 (m) in the unfished zone (figure 18).

KEY INVERTEBRATES

Sea urchins and sea cucumbers were only found on 6 sites and densities were very low (figure 16). The most common species was Diadema antillarum. The number of observations were too low to make an analysis between zones. A table of recorded key invertebrate species can be found in Appendix 6.

HABITAT COMPLEXITY

Habitat complexity was recorded per site (figure 18). All transects were taken into account when analysing for the effect of zonation. No significant effects of zonation were found for habitat complexity, with a mean value of 2.6 (Polunin score) in the fished and 2.9 in the unfished zone (table 7; figure 17).

CORAL HEALTH

38 of the 1500 (2.53%) photoquadrats contained one or more diseased corals. The most common disease was the dark spot disease, other occurring diseases were red band disease, aspergillosis and white plague. Photos of the diseased corals can be found in Appendix 7.

(30)

29 FIGURE 15: CURRENT STATUS OF BENTHIC COVER (%) PER SITE (SITES BRM TO WB ARE IN THE FISHED ZONE, SITES BA TO TRD ARE IN THE UNFISHED ZONE).

TABLE 7: GENERALISED LINEAR MIXED MODEL RESULTS FOR CORAL COVER AND MACROALGAE HEIGHT (F = FISHED ZONE, UF = UNFISHED ZONE, C2/3/4 = POLUNIN COMPLEXITY SCORES 2, 3 & 4).

Coral cover Estimate Std. Error t-value p-value

Without interaction

Intercept -3.27 0.15 -21.58 <0.00005

Fished vs Unfished 0.41 0.16 2.55 0.019*

Complexity 2 vs 3 0.22 0.14 1.54 0.13

Complexity 2 vs 4 0.40 0.21 1.89 0.062

Main factors seperate

Intercept -3.11 0.13 -24.67 <0.00005

Fished vs Unfished 0.47 0.17 2.74 0.014*

Macroalgae cover Main factors separate

Intercept -1.65 0.22 -7.53 <0.00005

Fished vs Unfished -0.20 0.31 -0.64 0.53

Macroalgae height With

interaction

Intercept 1.14 0.20 5.65 <0.00005

Fished vs Unfished -0.94 0.44 -2.13 0.047*

Complexity 2 vs 3 -1.08 0.21 -5.12 <0.0005*

Complexity 2 vs 4 -0.09 0.48 -0.18 0.86

F/UF:C2/C3 1.07 0.45 2.38 0.023*

F/UF:C2/C4 -0.82 1.44 -0.57 0.57

Density coral recruits Main factors

separate

Intercept 1.66 0.14 12.02 <0.00005

Fished vs Unfished -0.25 0.21 -1.20 0.25

Water quality Main factors separate

Intercept 3.24 0.06 54.80 <0.00005

Fished vs Unfished 0.03 0.08 0.36 0.72

Habitat complexity Main factors separate

Intercept 0.95 0.06 15.40 <0.00005

Fished vs Unfished 0.13 0.08 1.50 0.15

Fished Unfished

(31)

30 FIGURE 16: CURRENT STATUS OF MACROALGAE HEIGHT (CM), DENSITY OF CORAL RECRUITS (#/100M2), HORIZONTAL TRANSPARENCY (M) AND DENSITY OF KEY INVERTEBRATES (SEA URCHINS AND CUCUMBERS) (#/100M2).

FIGURE 17: CURRENT STATUS OF HABITAT COMPLEXITY PER SITE (LEFT) AND ESTIMATES OF MEAN HABITAT COMPLEXITY PER ZONE WITH 95% CL ERRORBARS (RIGHT).

(32)

31 FIGURE 18: BENTHIC INDICATORS PER ZONE AND PER COMPLEXITY SCORE (LEFT), AND MEANS PER INDICATOR CALCULATED FROM THE ESTIMATES WITH 95% CL ERRORBARS (RIGHT). GRAPHS MARKED WITH * INDICATE SIGNIFANT DIFFERENCE.

*

*

Referenties

Outline

GERELATEERDE DOCUMENTEN

The number of samples equal to the required power (p=0.8) was calculated from a change of 50% and 100% on the observed average of the total number of individuals, recorded over

In two natural populations with extra hand pollination of Epilobium angustifolium, also an ovule clearing technique has been used (Wiens et al. A fertilization rate of 97% and

The aims of this thesis are to (1) better understand the effect of nutrient limitation on macroalgae and seagrasses, (2) study the impact of short-term

Keywords: Caribbean Netherlands, monitoring, nature, biodiversity, ecosystems, species, treaties This research project was carried out by Alterra and IMARES Wageningen UR at

In order to update the 1985 atlas of Bonaire’s coral reefs (Van Duyl, 1985), a hyperspectral mapping campaign was performed in October 2013 using the Wageningen UR

The assessment survey tool captured information for each site on the current level of capacity and needs to improve capacity in the following 24 thematic assessment areas:

• priority sites have higher relative resilience, or lower relative vulnerability, are greater relative sources of fish and coral larvae and not weak sinks, and are exposed to

Motivated by the need for regular data monitoring and for quantification of the state and change of benthic and pelagic organisms, the Global Coral Reef Monitoring Network