Effects of Community- and Government-Managed Marine Protected Areas on Tropical Seagrass and Coral Communities

Angelica Chirico Effects of community- and government-managed marine

Effects of community- and government-managed marine protected areas on tropical seagrass and coral communities protected areas on tropical seagrass and coral communities

Angelica Chirico

ISBN 978-91-7911-220-2

Department of Ecology, Environment and Plant Sciences

Doctoral Thesis in Marine Ecotoxicology at Stockholm University, Sweden 2020

Effects of community- and government-managed marine protected areas on tropical seagrass and coral communities Angelica Chirico Academic dissertation for the Degree of Doctor of Philosophy in Marine Ecotoxicology at Stockholm University to be publicly defended on Friday 11 September 2020 at 13.30 in Vivi Täckholmsalen (Q-salen), NPQ-huset, Svante Arrhenius väg 20.

Abstract Tropical seagrass beds and coral reefs are among the most productive and diverse ecosystems on Earth and provide ecosystem services, such as fish production and coastal protection, and support livelihoods of millions of people. At the same time, these ecosystems are threatened globally by anthropogenic disturbances, such as overfishing, pollution and global warming. Implementation of marine protected areas (MPAs) is one of the main strategy to achieve conservation goals and has proven to restore biodiversity and fish stocks, at least on coral reefs. However, studies assessing protection effects on seagrass communities are scarce. Moreover, many MPAs are government-managed and increasingly criticized for excluding and marginalizing local communities. Therefore, MPAs that are managed by the communities themselves, i.e. community-managed MPAs, constitute a promising yet poorly studied alternative. The aim of this thesis was to investigate ecological effects of government- and community-managed MPAs on seagrasses, corals, and their associated benthic and fish communities in the tropical seascape. We used a space-for-time replacement approach and surveyed coral and seagrass communities in fished areas, recently established community MPAs (1-6 years of protection) and old government MPAs (20-44 years) in coastal Kenya, East Africa. Results suggest that only a few years of protection in community MPAs can increase diversity of benthic communities (Paper I), and also protect economically valuable fish stocks (Paper II). Protection also appeared to induce a community shift, from dominance of pioneering and stress-tolerant coral and seagrass species in fished areas, to structurally complex climax species in old government MPAs (Paper I). Additionally, effects of protection on seagrass communities seems to be most apparent in the mid-lagoon by favoring seagrass species with high shoot density; an effect that was mostly caused by species turnover but also phenotypic plasticity. Meanwhile, effects in the shallow intertidal and reef zones were weak or non- existing (Paper III). Finally, a two-year field experiment suggests that a community MPA speeds up seagrass recovery and decrease sediment erosion following experimental disturbance, most likely by reducing additional disturbances (e.g. fishing practices) on recovering plants and sediments (Paper IV). Based on these results I make three conclusions. First, MPAs seem to protect seagrasses in a similar way as they protect corals, suggesting that MPAs can aid local seagrass conservation. Seagrass beds should therefore be actively incorporated in marine spatial planning. Second, even though recently established community MPAs were not as effective as the old government MPAs, they appear to benefit both seagrass and coral communities (Paper I, II, IV). Given that previous studies show that they can also fulfill socio-economic community level-values (e.g. involvement in MPA design and enforcement), our findings emphasize their potential as a complement to government MPAs. Third, MPAs are an effective tool to protect seagrass and coral communities from local disturbances, particularly in mid-lagoon and reef areas, but they do not appear to protect the shallow intertidal seagrass beds (Paper III), possibly because of MPA-related tourism activities. This highlights the need for more detailed MPA evaluations, but also the need for more holistic conservation approaches, like integrated coastal zone management.

Keywords: coastal ecosystems, conservation, marine spatial planning, locally managed, fisheries closure, benthic communities, fish, foundation species, macroalgae, secondary succession, life-history, trait variability, plasticity, tourism, human disturbance, fishing, experimental research, Western Indian Ocean, causal modelling, structural equation model, permanova, multivariate data.

Stockholm 2020 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-182771

ISBN 978-91-7911-220-2 ISBN 978-91-7911-221-9

Department of Ecology, Environment and Plant Sciences

Stockholm University, 106 91 Stockholm

EFFECTS OF COMMUNITY- AND GOVERNMENT-MANAGED MARINE PROTECTED AREAS ON TROPICAL SEAGRASS AND CORAL COMMUNITIES

Angelica Chirico

Effects of community- and government-managed marine protected areas on tropical seagrass and coral communities

Angelica Chirico ©Angelica Chirico, Stockholm University 2020

ISBN print 978-91-7911-220-2 ISBN PDF 978-91-7911-221-9

Cover photo: A thornback cowfish (Lactoria fornasini) in a Thalassia hemprichii seagrass bed in Kenya (Photographer A. Chirico)

Printed in Sweden by Universitetsservice US-AB, Stockholm 2020 Till mamma, mormor och farmor

There is no force more powerful than a woman determined to rise.

Unknown

List of Papers

This thesis is based on the four following papers and referred to in the text by their roman numerals:

I. Chirico AAD, Uku JN, Eklöf JS. Marine protected areas increase diversity and alter composition of benthic communities across a tropical seascape gradient. Manuscript.

II. Chirico AAD, McClanahan TR, Eklöf JS. (2017) Community- and government-managed marine protected areas increase fish size, biomass and potential value. PLoS ONE 12(8): e0182342. doi: 10.1371/journal.pone.0182342

III. Chirico AAD, Alonso Aller E, Eklöf JS. Effects of marine protected areas on inter- and intraspecific trait variability in tropical seagrass assemblages. Manuscript.

IV. Chirico AAD, Zuma M, Kautsky N, Eklöf JS. A marine protected area speeds up seagrass recovery and decreases sediment erosion following experimental disturbance. Manuscript.

Paper II is reprinted with kind permission from PLoS ONE (open access journal).

My contributions to the four papers are: Paper I – planning and design of study, data collection, processing and analyzing data and main responsibility of writing. Paper II – planning and design of study, data collection, processing and analyzing data and main responsibility of writing. Paper III – planning and design of study, data collection and main responsibility of writing. Paper IV – planning and design of experimental setup, partly responsible in data collection, processing and analyzing data and main responsibility of writing.

1 Contents

Introduction ...... 4 The tropical seascape – importance and threats ...... 4 Tropical marine management ...... 7 Marine protected areas ...... 7 Management of seagrasses ...... 8 Limitations of marine protected areas ...... 9 Community resilience and disturbances ...... 11 Trait-based ecology and phenotypic plasticity ...... 11

Scope of thesis ...... 13

Methods ...... 15 Description of the study area ...... 15 The Kenyan seascape ...... 15 Coastal fisheries ...... 15 The first phase of Kenyan marine protected areas ...... 17 The birth of the tengefu movement ...... 18 Approaches for assessing effects of MPAs ...... 18 Field surveys ...... 19 Benthic habitat surveys ...... 19 Fish surveys and fish data preparation ...... 20 Seagrass morphological traits...... 22 Laboratory analyses ...... 22 Intra- and interspecific trait variability ...... 23 Seagrass recovery experiment ...... 23 Statistical analyses ...... 25 Univariate analyses ...... 25 Multivariate analyses ...... 25 Structural equation modeling ...... 26 Assessment of intra- vs. interspecific trait variability ...... 26

Synthesis of results and discussion ...... 28 Effects of MPAs on benthic communities in seagrass beds and on coral reefs ...... 28 Effects of MPAs on fish assemblages in seagrass beds and on coral reefs ...... 31 Effects of MPAs on seagrasses species and trait composition ...... 33 Effects of a MPA and size of disturbance on seagrass recovery and sediment erosion ...... 36 Protection effects differ depending on metric ...... 39

2 Effects of protection on seagrass beds versus coral reefs ...... 40 Community-managed MPAs as an alternative or complementary protection tool ...... 41

Conclusions ...... 44 MPAs are effective but not effective enough ...... 44 Community-managed MPAs as promising complements to state-run MPAs ...... 45 Include seagrass beds in management plans ...... 45

Future perspectives ...... 47

Populärvetenskaplig sammanfattning (svenska) ...... 50

Muhtasari (Kiswahili) ...... 52

Acknowledgements ...... 55

Financial support...... 57

References ...... 58

3 Introduction

The tropical seascape – importance and threats Tropical marine ecosystems including mangrove forests, sand and mud flats, algal and seagrass beds, rocky beaches, and coral reefs form a complex landscape mosaic of interconnected systems, termed the tropical seascape (Ogden and Gladfelter 1983, Ogden 1988). The complex physical structure provided by foundation species such as mangroves, perennial macroalgae, seagrasses and corals create shelter, nursery and feeding grounds for thousands of species and are therefore considered to be ‘hotspots’ of marine diversity and productivity (Roberts et al. 2002, UNEP 2006, Graham et al. 2007). Particularly mangroves, seagrasses and hard corals, provide vital regulating and supporting ecosystem services such as oxygen production, coastal protection and nutrient cycling, as well as spiritual and recreational values to society (e.g. Daily et al. 2000, Millennium Ecosystem Assessment 2005, Martínez et al. 2007, Isbell et al. 2017). These ecosystems also deliver provisioning services such as firewood, construction material, medicine, and seafood (Moberg and Folke 1999, Moberg and Rönnbäck 2003, Nordlund et al. 2016, Nordlund et al. 2018, Mitra 2020). Humans in tropical coastal regions often depend heavily on these services for income and food requirements (Martínez et al. 2007). Despite the relatively low distribution of corals and seagrasses (cover <3% of the worldwide seafloor), they are essential for human well-being and among the most valuable habitats on earth (Millennium Ecosystem Assessment 2005, Barbier et al. 2011, Costanza et al. 2014, Selig et al. 2019). For example, coral reef tourism globally is calculated to be worth US$36 billion dollar per year (Spalding et al. 2017) (Figure 1).

Figure 1. Pristine coral reefs generate billions of dollars every year. Photography from Mafia Island, Tanzania (left) and the Great Barrier Reef, Australia (right). (Photographer A. Chirico)

Mangroves and scleractinian corals (hereafter ‘corals’) have long been known for their importance to deliver ecosystem services to humans, e.g. fish production, buffer waves, export and absorb nutrients, and store carbon (reviews; Moberg and Folke 1999, Mitra 2020). Seagrass ecosystems, on the other hand, have only more recently been acknowledged for their ecological and economical importance (Costanza et al. 1997, Duarte et al. 2008, Cullen-Unsworth et al. 2014, Nordlund et al. 2016). Seagrasses are marine, clonal flowering plants anchored in mud or sand bottoms in shallow waters that can form large meadows (Green and Short 2003). They flourish at the interface between the land and the sea, improving water quality and protecting shorelines (Bos et al. 2007, Duarte et al. 2008, Nordlund et al. 2016), as well as exchange energy, material and living biomass with nearby and distant marine habitats (Heck et al. 2008). Seagrasses are keystone species that have a disproportional strong influence on connected species by ecosystem engineering effects, and are some of the most productive ecosystems on earth (Hemminga and Duarte 2000). They serve as vital feeding grounds for iconic vertebrates like sea cows, sea horses, turtles and sharks (Short et al. 2007, Duarte et al. 2008). Addition- ally, seagrass ecosystems are estimated to provide valuable nursery habitats to fish involved in >20% of the world’s 25 largest fisheries (Unsworth et al. 2018b), and support fisheries and human livelihoods on a local and global scale (de la Torre-Castro and Rönnbäck 2004, Unsworth et al. 2010, Cullen- Unsworth et al. 2014, Nordlund et al. 2018, Unsworth et al. 2018b). Addition- ally, even though seagrass beds only occupy ca. 0.1% of the ocean floor they are considered one of the largest carbon sinks globally (Duarte et al. 2005, Duarte et al. 2010, McLeod et al. 2011). Seagrasses can store carbon in the seabed 35 times faster than tropical rainforests and have the capacity to mitigate climate change (McLeod et al. 2011).

5 Figure 2. A multi-specific seagrass bed in Mombasa MPA, Kenya. (Photographer A. Chirico)

While many humans in tropical coastal regions are highly dependent on marine ecosystems for their livelihoods (Martínez et al. 2007, Selig et al. 2019), human-induced environmental changes and population growth have resulted in increased demand and pressure on marine resources, and are now threatening many of the world’s marine ecosystems (Burke et al. 2001, Lotze et al. 2006, Worm et al. 2006, Arkema et al. 2013). Sea level rise and warming, as well as increased frequency of storms, runoff from land, eutrophication and poor water quality have resulted in extensive loss of coral reefs (Hoegh- Guldberg 1999, Hoegh-Guldberg et al. 2007, Sweet and Brown 2016), seagrass beds (Orth et al. 2006, Waycott et al. 2009, Quiros et al. 2017, Unsworth et al. 2018a), and mangroves (Alongi 2002). Loss of these ecosystems can have severe consequences for the environment as well as for human well-being. For example, extensive harvesting of mangroves for firewood and aquaculture can result in reduction of coastal protection from wind and storms (Mitra 2020), and removing seagrasses have been shown to increase beach erosion and turbidity (Daby 2003, Lamb et al. 2017). More generally, marine biodiversity loss is threatening the ocean's capacity to provide food, maintain water quality, and recover from disturbances (Worm et al. 2006).

At the same time as fish production and fishery catches are considered among the most important ecosystem services from the tropical seascape, overfishing is one of the major threats to coral reefs and seagrass beds (e.g. Jackson et al.

6 2001). Fishing gears dragged on the seabed, such as trawls and beach seines, as well as boat anchors and trampling, are reducing the structural complexity of reef-building corals (Austin et al. 1997, Alvarez-Filip et al. 2009) and can form fragmented seagrass areas or complete loss of seagrass beds (Short and Wyllie-Echeverria 1996, Orth et al. 2006). Also indirect effects of fishing threaten these habitats. For example, overexploitation of triggerfish, a family of fish that feed on sea urchins and other invertebrates, can increase abundances of sea urchins, that in turn consume and erode corals (Bak 1994, McClanahan 1995, Eakin 1996, Dumont et al. 2013) and overgraze seagrasses (Eklöf et al. 2008). Additionally, removal of herbivorous fishes, such as surgeonfish and parrotfish, can increase abundance of macroalgae that compete with corals for space, light and nutrients and contribute to severe coral decline (Hughes 1994). The loss of coral and seagrass habitats results in further decrease of fish biomass, abundance and biodiversity since many fish species rely on coral reefs and seagrass beds as foraging and nursery grounds (Jones et al. 2004, Rogers et al. 2014, Unsworth et al. 2018b).

Tropical marine management

Marine protected areas To secure future delivery of ecosystem services and counteract overexploitation and further decline in biodiversity, implementation of marine protected areas (MPAs) has become a widely used conservation strategy (Olsen et al. 2011, Spalding et al. 2013). MPAs are a form of place-based management where a geographical area is set aside to be managed in a way that that fulfill conservation goals, usually aiming to prevent or limit human fishing pressure (National Research Council 2001, Emslie et al. 2015). Restrictions can include the prohibition of certain types of fishing gear (e.g. fine-mesh nets), periodic closures (e.g. the area is closed for fishing during certain periods), or complete no-take zones that exclude all forms of extraction of natural resources (Kaiser et al. 2000, Nemeth et al. 2006, Lester and Halpern 2008). Previous studies show that MPAs, primarily permanent no-take areas, can be an effective tool for reducing local threats and successfully protect and restore marine ecosystems, habitat-forming organisms, biodiversity and fish stocks (e.g. Roberts and Hawkins 2000, McClanahan et al. 2008, Lester et al. 2009, Barbier et al. 2011, Graham et al. 2011, Hilborn et al. 2020). Additionally, it has been suggested that protection can strengthen ecosystem resilience to climate change (Roberts et al. 2017), increase coral reef functioning (Topor et al. 2019) and enhance capacity to withstand disturbances, e.g. flood impacts (Olds et al. 2014).

7 Protection effects often strengthen with time of protection (number of years since protection started) due to succession dynamics and changes in population and community structure over time (Roberts and Hawkins 2000, McClanahan et al. 2007b, Claudet et al. 2008, McClanahan et al. 2009, Mörk et al. 2009, Grime and Pierce 2012). These temporal effects can also differ between organism groups. For example, a few years can be enough for small fishery target fish species that migrate into MPAs in seek of shelter (e.g. Gell and Roberts 2002, Halpern and Warner 2002), whereas >18 years of protection might be needed for targeted and highly mobile large predatory fish (Russ and Alcala 2004). Effectiveness of MPAs also depend on the enforcement and size of the protected area, with large (at least 100 km2), well-enforced and old (>10 years) no-take areas being the most effective (Walters and Holling 1990, Edgar et al. 2014).

Management of seagrasses Marine coastal management and ecological research has traditionally focused on hard-bottom ecosystems such as tropical coral reefs, temperate macroalgal reefs, and rocky shores, while less attention has been paid to shallow, coastal soft-bottom seagrass ecosystems (Orth et al. 2006, Wells et al. 2007, Duarte et al. 2008). In 2006, 60% of published research on threatened coastal habitats (reported in Web of Science as of May 2006) was carried out on coral reefs, compared to only ca. 5% in seagrass beds (Duarte et al. 2008). A more recent study show that this imbalance has worsened and that seagrass research effort (proportion of yearly ecology and ecosystem publications on ISI Web of Sci- ence) has leveled off between the mid-2000s to 2016, whereas coral and mangrove research effort have increased continuously (Unsworth et al. 2019). To secure future delivery of ecosystem services derived from seagrass ecosystems, seagrasses should be of increased priority and more often included in research and conservation plans. This is especially important given the global ecological and economic importance of seagrasses (Waycott et al. 2009, Nordlund et al. 2016, Nordlund et al. 2018, Unsworth et al. 2018b), which are experiencing extensive losses worldwide (Jackson et al. 2001, Eklöf et al. 2009, Waycott et al. 2009, Unsworth and Cullen 2010, Baden et al. 2012, Unsworth et al. 2018a). Temperate seagrass ecosystems are often included in management strategies in developed countries (Kenworthy et al. 2006, Orth et al. 2006, Montefalcone et al. 2009, Appolloni et al. 2018), but are often ignored by managers in tropical developing countries (Unsworth and Cullen 2010). However, since tropical seagrass beds are often found in close proxim- ity to coral reefs, they frequently occur within MPAs designed to conserve coral reefs. Even though protection effects on seagrass ecosystems have not been as thoroughly tested as on coral reef research, seagrasses could at least theoretically benefit from protected areas, since fishing activities, like boat moorings, can damage seagrass beds (Short and Wyllie-Echeverria 1996, Orth

8 et al. 2006). To the best of my knowledge, only a few studies have tested effects of MPAs on seagrasses, and most of these suggest that MPAs can protect seagrass ecosystems (but see limitations below). Protection effects include increases in seagrass shoot density and leaf production (Marbà et al. 2003, González-Correa et al. 2007, Ferrari et al. 2008) and benefits on seagrass-associated fish stocks (Valentine et al. 2008, Valle and Bayle-Sempere 2009, Unsworth et al. 2010, Fraschetti et al. 2013, Seytre and Francour 2014, Alonso Aller et al. 2017). Of these, only one study is conducted in the Western Indian Ocean (Alonso Aller et al. 2017), urging the need for studies investigating MPA effects on tropical seagrass communities.

Figure 3. A shoal of yellowfin surgeonfish (Acanthurus xanthopterus) swimming over a seagrass bed in Mombasa MPA, Kenya. (Photographer A. Chirico)

Limitations of marine protected areas Even though MPAs can protect coral reefs, and to some extent seagrasses (although studies are relatively few), there are also limitations with many MPAs. A recent study from East Africa show decline and partial recovery in seagrass cover over a 10-year period, despite the presence of a MPA (suggesting large influences by other factors such as water currents) (Alonso Aller et al. 2019). There are many types of disturbances that can intrude MPAs and compromise protection effects. For example ‘invasion’ by overgrazing sea urchins (Eklöf et al. 2009) and runoff from land (Quiros et al. 2017) may diminish protection effects on seagrasses, emphasizing that effects of MPAs on seagrasses needs

9 to be further examined. Moreover, even though protection can increase recovery rates after disturbance due to greater biodiversity and higher levels of op- erating ecological processes (Roberts et al. 2017), protected areas can, by promoting more stress-sensitive coral species in the absence of destructive fishing, be more vulnerable to climate change and heat waves than unprotected reefs (Darling et al. 2010, 2013). The role of MPAs in selecting for more fragile species that have increased sensitivity to some disturbances is called the ‘protection paradox’, and has recently been suggested to limit the benefit of protected areas to ecosystem resilience (Bates et al. 2019).

Locally banning all type of resource extraction, including fishing, can also have great socio-economic costs (Klein et al. 2008, Adams et al. 2011), especially in development countries where local communities can depend on near- shore fisheries for income and daily food requirements (Roberts and Hawkins 2000, Burke et al. 2001, UNEP 2006, Martínez et al. 2007). Loss of fishing grounds can create poverty traps and increase food insecurity for local communities (Smith et al. 2010). Implementation of MPAs that exclude local communities has resulted in low acceptance, due to limited involvement in establishment and reduced fishing grounds and/or income, and hence conflicts (Christie 2004, Pomeroy and Douvere 2008, Agardy et al. 2011). Low compliance to regulations that ban fishing has resulted in illegal fishing (poaching) that jeopardize management goals around the world (McClanahan 1999, Christie 2004, Hilborn 2007, Cinner et al. 2011, Giakoumi et al. 2018). Addi- tionally, strong opposition makes implementation politically difficult (Lester and Halpern 2008). On the other hand, strong community involvement - i.e. to engage local communities in planning, decision making, implementation and/or enforcement of protected areas - has led to fewer conflicts and high compliance to rules and regulations (e.g. Pollnac et al. 2001, Roberts et al. 2001). Together with declining fish catches for local fishers and opportunities for restoring fisheries and/or generate alternative income through MPA en- trance fees, this has resulted in an increase in “bottom-up” driven MPAs worldwide (e.g. White and Vogt 2000, Pollnac et al. 2001, Govan 2009, Pascal 2011, McClanahan et al. 2016a, Villaseñor-Derbez et al. 2019). These MPAs are often situated close to the villages (community-based) and/or managed by the locals (community-managed), and are also referred to as ‘locally managed marine areas’ (LMMAs). Community-managed MPAs can also maintain social values within the community, such as cultural and existence values, involvement in community organizations and confidence in management of resources (Hicks et al. 2009). Protection effects of the usually small community- managed MPAs is less studied than those of large government-managed protected areas like marine national parks, particularly in the Western Indian Ocean. Previous research on community-managed MPAs demonstrate effects like increased hard coral cover and coral reef fish diversity, density and biomass (White and Vogt 2000, McClanahan et al. 2006, Samoilys et al. 2007,

10 Aburto-Oropeza et al. 2011, Clements et al. 2012). If these biological effects are general, and additionally protect other habitats than coral reefs, these bottom-up MPAs can constitute a good complement or alternative to government place-based protection.

Community resilience and disturbances Humans are altering the capacity of ecosystems to resist and recover from disturbances (Nyström et al. 2000). Reducing disturbances from human activities, such as fishing, can increase seagrass resilience (Orth et al. 2006). Areas that are already disturbed are often more vulnerable to further damage (Nyström et al. 2000, Whitfield et al. 2002), even though disturbances also can select for less vulnerable stress-sensitive species (see e.g. Darling et al. 2013, and the 'protection paradox' above). It is therefore advantageous if a community recovers fast following a disturbance. For example, coral reefs with heavy fishing pressure on grazing fishes promote growth of algae that suppress coral growth (e.g. Tanner 1995, Foster et al. 2008). When corals are exposed to heat waves and bleach, recovery is very slow due to low grazing rates (e.g. Rasher and Hay 2010). This can further result in regime shifts from coral to macroalgal dominance (e.g. Cheal et al. 2010, Graham et al. 2015).

Disturbances in the tropical seascape occur frequently and differ in intensity and at temporal and spatial scales. A local disturbance can be the physical damage caused by a bottom trawl or dragnet dragged along the seabed (Walker et al. 1989, Eckrich and Holmquist 2000), whereas other disturbances can be more diffuse and derive from distant sources like land (pollution) or heat waves (Cesar 2011). The larger the area a disturbance impacts, the lower is often the rate of recovery, and therefore also greater risk that the ecological community can shift into a different state, i.e. scale-dependent recovery (Pet- raitis and Latham 1999). An experimental study of macroalgae removal showed that in small plots the community recovered to pre-disturbance state, whereas large plots (ca. 8 m2) did not recover and turned into an alternative community state with different species composition (Petraitis and Dudgeon 2005). Since changes in community composition are likely to have consequences for ecosystem functioning it is important to study how different sizes of disturbance affect recovery of and community composition (Jones et al. 1994).

Trait-based ecology and phenotypic plasticity Traits are morphological, behavioral, physiological and phenological characteristics measurable at the level of individual organisms (Violle et al. 2007)

11 and can be used to quantify changes and status of ecological communities (Barnett et al. 2019). Habitat-forming organisms, such as trees and seagrasses, modify, maintain and create habitats to other species by their structures and thus strongly affect associated communities (Jones et al. 1994). Traditional assessments in community ecology have focused on classifications based on species identity or broad functional groups (e.g. Littler and Littler 1980, Bruno et al. 2005). However, to understand how and why organisms and communities respond to environmental variability in space and time, trait-based approaches have become increasingly common (Lavorel and Garnier 2002, Díaz et al. 2007, Madin et al. 2016). A recent study of macroalgal communities showed that morphological traits, e.g. frond leaf length, were more important than species identity in predicting epifaunal densities (Stelling-Wood et al. 2020). Similarly, Parker et al. (2001) found that seagrass leaf surface area was more important than host identity for epifaunal biomass. These and other studies (e.g. Diaz et al. 2004, Fortunel et al. 2009, Pakeman et al. 2009) indicate that trait-based approaches may be more useful than traditional species-based approaches. Therefore, trait-based approaches that measure traits at the level of individual organisms, rather than using proxy values for the species, can contribute to knowledge that is not constrained to taxonomic groups (Stelling- Wood et al. 2020).

Many organisms can change their morphology to cope with different environmental conditions, so called phenotypic plasticity (Bradshaw 1965). Seagrasses exhibit high levels of intraspecific variability in morphological traits (Abal et al. 1994, Udy and Dennison 1997), which can help them survive in unfavorable conditions. For example, seagrasses subjected to poor water quality and suffer from light deprivation have been found to have increased leaf length (Bulthuis 1983) or increased leaf width (Lee and Dunton 1997) as a morphological response to maximize photosynthetic capacity. At the same time they can also have decreased leaf density (Ruiz Fernandez and Romero 2004) as a response of energy allocation. Such morphological changes are likely to alter ecological functions maintained and services delivered by the organisms (Miner et al. 2005). For example, large-sized seagrasses accumu- late more carbon than smaller (Duarte and Chiscano 1999, Gullström et al. 2018) and also form more structurally complex seagrass beds that harbor more fish (Gullström et al. 2008). Therefore, trait diversity and variability in habitat-building organisms is particularly interesting from the perspective of ecosystem engineers such as seagrasses, since their traits can strongly affect ecosystem functions and structure (Jones et al. 1994, Gagic et al. 2015).

12 Scope of thesis

The aim of this thesis was to investigate ecological effects of government- and community-managed MPAs on seagrasses, corals and their associated benthic and fish communities in the tropical seascape (Figure 4), using coastal Kenya as a case study area. More specifically, the objectives were to:

• assess the influence of protection, and time of protection, on bottom cover, diversity and taxonomic composition of benthic habitat-forming species by conducting field surveys in MPAs and fished areas (Paper I).

• assess effects of protection on traditional fish metrics – density, size and biomass – by surveying fish communities in MPAs and fished areas, and estimate effects of protection on the more rare metric, monetary value of fish stocks, using data from a 12-year fish market survey (Paper II).

• assess to what extent protection influence seagrass species composition and trait composition, and whether effects vary along the tidal gradient, by extractive seagrass sampling in MPAs and fished areas (Paper III).

• assess the influence of disturbance size and presence of a MPA on seagrass recovery from disturbance by experimentally removing seagrasses in two different sizes of plots in a MPA and three fished areas and follow the benthic communities during two years (Paper IV).

Figure 4. Conceptual figure presenting the studies of the four papers (I-IV), highlight- ing the assessed relationships between MPAs, experimental disturbance and marine organisms. (Symbols: courtesy of the Integration and Application Network)

14 Methods

Description of the study area

The Kenyan seascape Coastal Kenya was selected as the study site since the coast is relatively ho- mogenous and straight with sandy beaches, rich in seagrass and coral areas and has several marine government- and community-managed protected areas, providing opportunities to address the aim of the thesis. Kenya has a 600 km shoreline in the Western Indian Ocean and a fringing reef system that runs parallel to the coast, and form lagoons as part of the tropical seascape. Ap- proximately 20% of the world’s seagrasses (14 species) are found in the West- ern Indian Ocean (Green and Short 2003), of which 12 species have been found in Kenya (Ochieng and Erftemeijer 1993).

Studies were undertaken in intertidal and subtidal seagrass beds and fringing and patchy coral reefs. The intertidal zone is typically dominated by seagrasses such as Thalassia hemprichii and Halodule spp., and the mid subtidal lagoon with seagrass species like Thalassodendron ciliatum and patches of corals, typical of that of other shallow coasts in the Western Indian Ocean (Gullström et al. 2002). Out at the reef crest, corals and macroalgae are usually dominating but are often interspersed with seagrass and sandy patches (McClanahan and Muthiga 2016). There are about 45 common hard coral genera, that include branching, plating and encrusting species from the genera Porites, Acropora, Pavona and Montipora (McClanahan et al. 2007a).

Coastal fisheries A majority of households in coastal Kenya depend on marine resources for food and income (UNESCO 1997, Martínez et al. 2007). Coastal economic activities include artisanal fisheries, tourism, coral mining (for cement facto- ries) and mangrove cutting to provide firewood. Fisheries in the Western In- dian Ocean, including Kenya, are often small-scale and conducted close to the mainland in seagrass beds and on coral reefs (e.g. Jiddawi and Öhman 2002, de la Torre-Castro and Rönnbäck 2004). Fishing is typically a subsistent livelihood and traditional fishing gear target multiple species and include traps,

15 gill nets and handlines (McClanahan and Kaunda-Arara 1996, Okemwa 2017) (Figure 5). Beach seines and spear guns are frequently used, even though they are illegal due to their detrimental effects on marine resources (Obura 2001, Evans et al. 2011, Okemwa 2017).

Figure 5. Fisheries in Kenya a) Fishermen in a canoe with a fish basket trap (Swahili madema), b) fish trap in a shallow seagrass bed, c) jacks (Carangidae) caught by handline, and d) a mix of fish species drying in the sun. (Photographer A. Chirico)

Fish catches in the artisanal fisheries are comprised of up to ca. 160 species, but usually dominated by only a few herbivorous species, such as the marbled parrotfish (Leptoscarus vaigiensis) and the African white-spotted rabbitfish (Siganus sutor), that can be sold on local markets (Okemwa 2017). A recent study suggests that nearly 40% of coral reef fish stocks in East Africa are below sustainable levels, which calls for more careful regulations of fishing (McClanahan 2019). At the same time, the marine resources are heavily impacted by other anthropogenic stressors, such as deforestation, overcrowding of people and poor sanitation that have resulted in erosion of the coastline, runoff from land and decline in water quality (UNESCO 1997, Obura 2001, Green and Short 2003, McClanahan et al. 2005b). Together, these threats have resulted in significant loss of fish biomass and important coral and seagrass habitats, which in turn has reduced fishery yields for local people (e.g. McClanahan and Mangi 2000). This is disquieting and can generate serious

16 socio-economic consequences, since half of the population in the coastal prov- ince in Kenya lives below the poverty line (Hoorweg et al. 2009).

The first phase of Kenyan marine protected areas Before colonial times and up to the 1920s, marine resources were regulated by informal institutions that were led by community elders that gave permission for fishing and solved resource use conflicts (McClanahan et al. 1997, Glaesel 2000). After colonial times, population growth and heavily exploitation of marine resources has caused increased degradation of marine ecosystems and resulted in that several formal MPAs have been established along the Kenyan coast. The first generation of formal protected areas were established in the late 1960s and aimed to conserve and manage the unique habitats and variety of marine biodiversity, mainly for fisheries and/or tourism purposes (Francis et al. 2002, McClanahan et al. 2005b). Malindi, Watamu, Kisite and Mombasa Marine National Parks were established in 1968, 1972, 1978 and 1991, respectively, and are managed by the Kenya Wildlife Service (gov- ernmental corporation), with minimal involvement of local communities (Francis et al. 2002, McClanahan et al. 2005b). These four MPAs are located close to the shoreline, except Kisite that is located around offshore islands. Together, these protected areas cover ca. 50 km2. The four government-managed MPAs do not allow any extracting activities (no-take areas) and have been shown to successfully restore and protect coral reefs and fish stocks (e.g. McClanahan 2000, McClanahan et al. 2005b, McClanahan et al. 2007b, McClanahan 2014). The beaches around these protected areas are surrounded by hotels and the MPAs are open to tourism-related activities, such as snorkeling and diving, and thousands of visitors every year generate great tourism revenues (McClanahan 1999, Muthiga 2009). However, even though protection goals aim to contribute to the socio-economic welfare of nearby local communities, locals adjacent to the government-managed MPAs are only included in informal activities, such as boat operators, and at the same time experience reduced fish catches, suggesting that these objectives are not met (McClanahan et al. 2005a, Muthiga 2009). The exclusion of local communities in MPA decision-making and loss of fishing grounds has often generated conflicts between fishing communities and authorities during and after MPA establishment (Obura 2001, McClanahan et al. 2005b, Hicks et al. 2009). For example, Mombasa Marine National Park was declared protected from fishing in 1987, but poaching continued until 1991 and the size of the protected area had to be greatly reduced for it to be accepted by locals (McClanahan and Kaunda-Arara 1996). The lack of involvement of local fishing communities in the MPA establishment, as well as interest conflicts between fishing livelihoods and conservation incentives, are suggested to be the main reasons for low compliance that in turn jeopardized management goals (Abunge et al. 2013).

17 The birth of the tengefu movement Since the late 20th century local communities have become aware of the value of coastal resources and are empowered to manage them, resulting in that a new generation of MPAs have developed in East Africa, as well as other parts of the world (Francis et al 2002). In Kenya, several “bottom-up” driven and community-managed MPAs (locally called tengefu, a Swahili word roughly meaning something that is set aside) have been established with the aid of NGOs and private financers, to protect fish stocks and habitats from overfishing (McClanahan et al. 2016a). The first tengefu (Kuruwitu closure) was established in 2006. Since then, several more have been implemented, for example Wasini, Kanamai and Kibuyuni (established in 2008, 2009 and 2009, respectively). Some of these community-managed MPAs allow some types of resource use; for example, Wasini is a periodic closure open to fishing approx- imately 3 months per year, so that fisheries can still contribute to food and income for fishery communities. Community-managed MPAs in Kenya are smaller (<0.5 km2) than the government MPAs (between 6-28 km2). In terms of protection effects, Kuruwitu closure has been shown to enhance coral cover and increase fish biomass on coral reefs (McClanahan et al. 2016a). To what extent the other community-managed MPAs affect marine resources is yet to be evaluated, and on the basis of my knowledge no studies have so far inves- tigated seagrass communities within these areas. The MPAs managed by the communities themselves can also generate small tourism revenues that benefit the local villages (McClanahan et al. 2016a), in contrast to government-managed protected areas that instead primarily benefit government corporations and operators in the highly developed tourism industry (e.g. large hotels).

Approaches for assessing effects of MPAs The ideal way to assess effects of MPAs is to use data collected both before and after establishment of several MPAs, and additionally sample regularly over time both inside and outside the protected areas; a so called Multiple Before-After Control-Impact (MBACI) study design (Underwood 1993). However, such designs are quite rare due to lack of long-term monitoring data and, particularly, sampling prior to implementation of protected areas (Ojeda- Martínez et al. 2011). Moreover, historical data on seagrass cover is more or less missing from the Western Indian Ocean (Waycott et al. 2009), just as baseline data from coral reefs prior to establishment of MPAs (Francis et al. 2002). In Kenya, long-term data are available for hard coral cover and coral reef fish, but data from seagrass beds are generally lacking (but see Harcourt et al. 2018), and effects of MPAs on seagrass ecosystem has therefore rarely been assessed.

18 Another option is to sample and compare established protected and unprotected areas regularly over time, but since tropical regions experience strong seasonal changes (e.g. monsoon rains), comparative field surveys should preferable be done at the same time of a year for several years, making such as design difficult to use, especially during the relatively short amount of time available during a PhD. To assess ecological effects of MPAs we therefore used a space-for-time field survey approach. Space-for-time substitution as- sumes that spatial and temporal variation are equivalent and is therefore an alternative to long-term before vs. after studies (Pickett 1989). A space-for- time approach is plausible and the most commonly used method to empirically evaluate the ecological effects of MPAs, by comparing biological status inside and outside protected areas (Pickett 1989, Micheli et al. 2012, Fraschetti et al. 2013, Strain et al. 2019). Moreover, we took advantage of the three different types of management (fished areas, and government- and community-managed MPAs) by assessing the effects of protection per se (differences between management types), as well as the different ‘ages’ (1-44 years of protection) and hence assessed effects of time since protection (years of protection) on different response variables. This type of temporal approach has been successfully used to demonstrate temporal recovery of coral reef fish communities in Kenya and elsewhere (e.g. McClanahan and Graham 2005, McClanahan et al. 2016b). The reference sites were carefully selected to avoid environmental differences confounding the results, e.g. we avoided rocky areas and areas dominated with mangroves, and sites were instead primarily sheltered lagoons with sandy and/or muddy coral and seagrass areas.

Field surveys

Benthic habitat surveys To test effects of protection on composition and diversity of benthic habitat- forming organisms (Paper I), we surveyed 12 sites along the Kenyan coast: four government-managed MPAs, four community-managed MPAs (tengefus) and four fished reference areas (Figure 6). At each site three habitat zones (intertidal- and subtidal seagrass beds, and coral reefs) were surveyed. We used a stratified sampling method that is more suitable than random sampling, when investigating organisms along a gradient (Duarte and Kirkman 2001). In each habitat zone 10 sampling stations (55 m from each other) placed parallel to the shoreline along a 500 m transect where surveyed (N=30 per site). Cover of all benthic organisms (>2 cm) was estimated visually within three randomly placed quadrats (0.5 m2), ca. 5 m from each other at each station. To be able to test whether protection effects differed between seagrass beds and coral reefs, we had to use one single survey method in both habitats.

19 The quadrat-based method (most commonly used in seagrass studies; Short et al. 2006) was chosen because most sampling stations (66%) were in seagrass beds. The line-intercept method is more common on coral reefs (Hill and Wilkinson 2004) and may capture more species by sampling a larger area. Therefore, our data may underestimate the actual taxonomic diversity in the reef areas, but on the other hand, using the same method in two different habitats gives a unique opportunity to compare how they are influenced by protection.

Figure 6. Map over the study sites in Kenya. Each of the study sites are subjected to one of the three management types; government-managed MPA, community-managed MPA or fished area (see figure legend). The parenthesis including roman numerals show which of the four papers (I-IV) the site was used in.

Fish surveys and fish data preparation

Fish survey design We assessed effects of protection and habitat on fish assemblages (Paper II) by sampling within six sites; two fished areas and four MPAs (two community-managed and two government-managed) (Figure 6). At each site we used

20 underwater visual census (point transects) in 7-9 replicates >30 m apart on coral reefs and in monospecific seagrass beds (T. hemprichii) (N = 92).

All fish survey methods have their pros and cons. One group of methods in- volve destructive sampling with nets or trawls, which sample many but not all fish species. They are also detrimental to fish and can destroy fragile benthic habitats (e.g. Austin et al. 1997, Alvarez-Filip et al. 2009). Additionally, such destructive methods are illegal in most MPAs, and are not preferable in unprotected areas either. Underwater visual census (UVC) is therefore much more favorable and a commonly used method to survey fish; it is cost-effective, quick and nondestructive, and generates fish data that can be compared between studies (Brock 1954, Samoilys and Carlos 1991, Hill and Wilkinson 2004). The technique has also been carefully evaluated in terms of precision, bias and accuracy (Sale and Sharp 1983, Bell et al. 1985, Thresher and Gunn 1986, Stjohn et al. 1990, Samoilys 1997, Samoilys and Carlos 2000).

Point transects are commonly used to visually survey fish in seagrass beds (Nakamura and Sano 2004, Dorenbosch et al. 2005, Dorenbosch et al. 2006, Unsworth et al. 2007), whereas belt transects are more common on coral reefs (Brock 1954, Hill and Wilkinson 2004). Since we aimed to compare fish variables between seagrass and coral habitats, using different methods was not an option. Since many seagrass fish species hide under the canopy we considered point transects to be more suitable, to not miss hiding fishes and hence underestimate their density. Also, point counts have been successfully used in the past when comparing fish in coral and seagrass habitats (e.g. Dorenbosch et al. 2005). Finally, this method is less time-consuming and requires minimal equipment compared to belt transects (Hill and Wilkinson 2004). In one comparative study, point transects were shown to yield higher fish density than belt transects (Bortone et al. 1989). Therefore, our estimates of absolute fish density, biomass and value, may overestimate actual values.

Fish surveys and data preparation Quadratic point transects (5×5m) were marked with thin ropes that were at- tached to wooden sticks driven into the bottom, and left for at least 20 minutes before fish surveys began, to minimize fish disturbance. All fishes observed within the quadrat were counted and identified to species level (based on Lieske and Myers 2002), and fish sizes were estimated visually to the nearest 5 cm. The surveys were conducted by a single snorkeler (A. Chirico) to avoid observer bias. The observer trained to estimate fish sizes using fish-like plastic sheets with known lengths until standard errors were below 5%. Counting fish by snorkeling can, in contrast to SCUBA diving, underestimate numbers of certain fish families, e.g. some “shy” Acanthuridae spp. (Dearden et al. 2010). However, snorkeling is a cost-effective method using minimal equipment and results are still comparable across different treatments, especially when using

21 standardized methods (e.g. using a single observer) (Thompson and Mapstone 1997, Williams et al. 2006). Additionally, the snorkeler was positioned about one meter from the shoreward corner of the transect and moved as little as possible to avoid disturbing fish. Since fish counts were conducted when water movement was minimal (mid-neap tides), body movements (swimming) from the snorkeler were also minimal, and hence fish disturbance. Point counts were carried out between 09:00 and 15:00, since tropical fish behavior differs between day, night and dusk/dawn (Piet 1996). All fish that entered the quadrat were counted for 12 minutes with the observer positioned in the corner of the transect, and then additionally for 3 minutes with the observer actively searching for more cryptic fishes hiding under the canopy or on the bottom (following Dorenbosch et al. 2006, Unsworth et al. 2007).

Count and size data was used to calculate biomass of fish using species-specific length-weight relationship from FishBase (Froese and Pauly 2013). Us- ing 12-years of Kenyan fish market data from ten busy landing sites (areas where fishers convene to sell their catch) (McClanahan 2010), we then calculated the potential monetary value of the fish assemblages (Kenyan shilling per transect). Data was based on five value groups generally sold on Kenyan fish markets; scavengers (including emperors [Lethrinidae], snappers [Lutjan- idae] and grunts [Haemulidae]), goatfish (Mullidae), rabbitfish (Siganidae), parrotfish (Scaridae), and 'rest of catch' (fish with lower value) (McClanahan 2010).

To link the fish point transect data to seagrass and coral cover (Paper II), each 25 m2 fish transect was surveyed using 5 randomly placed 0.5 m2 quadrats. Cover of all benthic substrate and organisms was estimated in each transect and mean cover of seagrass and coral (% cover per transect) was calculated and used in the analyses.

Seagrass morphological traits

Laboratory analyses At each sampling station (see Benthic habitat surveys above), seagrass samples were collected in one of the three quadrats (if seagrass was present) using a steel corer, 10 cm in diameter, to a sediment depth of 20 cm. Care was taken to include whole seagrass shoots growing within the area of the core. Samples were sieved and rinsed in the field and stored in freezer boxes, and then brought to the lab, where they were frozen. Later, the samples were thawed, rinsed with freshwater and cleaned from organisms living on the surface of

22 seagrasses. To assess effects of protection on seagrass traits composition (Pa- per III), we measured five typical morphological traits on all sampled seagrass shoots; shoot density, number of leaves per shoot, maximum leaf length (mm), leaf width (mm), and above:below-ground biomass ratio (g, dry weight). All plant material (above- and below-ground separated) was then dried at 80° C and weighted.

Intra- and interspecific trait variability Since variability and diversity of traits within a community have consequences for ecological functions and delivery of ecosystem services, it is important to address how the environment affects organism trait composition (Jones et al. 1994, Gagic et al. 2015). Previous studies have found that morphological traits can be more important than species identity (Parker et al. 2001, Stelling-Wood et al. 2020). Therefore we used a trait-based approach (Paper III) and assessed to what extent protection influence seagrass species and trait composition. Trait-based studies often assess environmental effects on traits on a community level and/or assess interspecific trait variability (e.g. mean trait value to describe entire species from literature) (Albert et al. 2010). However, most organisms, including seagrasses, can change their morphological traits in response to the environment, i.e. plasticity (Udy and Dennison 1997, Violle et al. 2007, Maxwell et al. 2014). Therefore, environmental conditions that induce changes in trait composition on a community level, e.g. average plant height, can include both changes through species turnover (interspecific variability) and plasticity (intraspecific variability) (Lepš et al. 2011). Only assessing interspecific variation may therefore result in over- or underestimating effects on environmental conditions. By measuring seagrass morphological traits at the level of individual samples (Paper III), we were able assess how both inter- and intraspecific trait variability varied with protection.

Seagrass recovery experiment To assess if protection and size of disturbance affect seagrass recovery from physical disturbance (Paper IV), manual removal of seagrasses (clearings of two sizes: 0.25 and 1 m2) were conducted within seagrass beds dominated by T. hemprichii, in one MPA and three fished reference sites. Clearings were circular and created by removing all seagrass including the roots and rhi- zomes, using gardening rakes (based on Eklof et al. 2011) (Figure 7). Five small clearings, five large clearings and five controls (1m2 area with no disturbance) were randomly distributed (ca. 7 m apart from each other) at each site in three rows parallel to the shoreline (N = 60). The plots were regularly

23 surveyed during two years (2012-2014), estimating percentage cover of benthic organisms, primarily seagrass (species level) and algae, and the depth of the clearing (as a measure of erosion).

Figure 7. Photography of a small experimental plot two days after seagrass removal. (Photographer A. Chirico)

An experiment design with one protected site is not ideal for this study. To be able to draw general conclusions about protection effects on seagrass recovery, more than one protected area should be included. The original plan was to include two sites inside and two sites outside the MPAs. However, since the MPA in one site (Kanamai) did not encompass the intertidal seagrass zone, this was not possible. We then considered several other alternatives. First, we tried to include one of the nearby government MPAs, but the Kenyan Wildlife Service did not allow this type of destructive method. Second, we considered to conduct the experiment in another community-managed area further south, but that would have been logistically difficult since it was far away, and additionally, environmental differences could have confounded potential effects. Regardless, this study is, to my knowledge, the first to experimentally test effects of protection on the recovery from disturbance on seagrass, and therefore brings new insights regarding protection effects on seagrass recovery and can constitute an important start for further studies.

24 Statistical analyses

Univariate analyses Effects of management type (categorical factor with three levels; fished, community-managed and government-managed MPA), and habitat type (categorical with two or three levels) on diversity of benthic organisms, reef topographic complexity, total hard coral and seagrass cover (Paper I) and on fish density, size, biomass and monetary value (Paper II), were examined using different types of general linear mixed models in R (v. 3.0.1) (R Core Team 2017). In some analyses we also used time since protection (age of MPA), as a continuous factor, to better identify temporal trends (Paper I and II).

For the experimental data (Paper IV) we used general linear mixed models (‘nlme’ package, Pinheiro et al. 2018) in R v. 3.5.3. (R Core Team 2019) to test if rate of recovery differed between sites and clearing size. We used time (number of weeks since the experiment started; continuous fixed factor), site (fixed, four levels), and their interactions, on seagrass cover and clearing depth.

In all univariate analyses, data was first assessed visually to test if assumptions of normality and homoscedasticity were met, and transformed if necessary. Site was included as a random intercept factor in all models in Paper I-III, but not in Paper IV, where site was instead treated as a fixed factor (because only one MPA was included).

Multivariate analyses Protection effects on benthic community composition based on taxonomic groups (Paper I) and fish value group composition (Paper II), were analyzed using multivariate analyses PERMANOVA (Permutational Multivariate ANOVA) and DISTLM (Distance-based linear models), as implemented in PRIMER (v 6.1.15) (following Anderson et al. 2008). Multivariate analysis of variance (MANOVA) and covariance (MANCOVA) can be used for categorical values, but these tests assume normal distribution for variables (parametric) and that observation units are independent of one another, and are not robust when data departures these assumptions (Johnson and Field 1993). PERMANOVA, on the other hand, is more robust and flexible and make no explicit assumptions regarding the distribution of variables (non-parametric) but instead use ranking of dissimilarities (Anderson et al. 2008), which is preferable with ecological data that are usually skewed and/or contains zeros (McArdle 1990, Welsh et al. 1996), as the data in Paper I and II. Further-

25 more, the number of taxa in Paper I exceeded the number of quadrats (sampling units) making traditional statistical approaches, like MANOVA, problematic (Johnson and Field 1993). The analysis of similarities (ANOSIM) (Clarke 1993) is also a robust method to analyze multivariate data, but since PERMANOVA has the ability to analyze more complex experimental designs (Anderson et al. 2008) we chose to analyze our data with PERMANOVA. DISTLM, is also a robust method that can be used for continuous predictor variables, unlike PERMANOVA (Anderson et al. 2008), and was therefore used instead of e.g. MANCOVA, in Paper II when using time since protection (age of MPA) as a predictor.

Structural equation modeling In Paper III we assessed the direct vs. indirect effects of protection on seagrass species and trait composition across three habitats (shallow, mid-lagoon and reef) using multigroup piecewise SEM (structural equation modeling). SEM is a powerful, multivariate technique that can test and evaluate empirical data with complex networks of causal relationships in ecological systems (Grace 2006, Fan et al. 2016). SEM combines multiple predictor and response variables into a single causal network, usually represented by direc- tional arrows (pathways) (Bollen and Pearl 2013, Fan et al. 2016). One of the advantages with SEM, as opposed to other modeling approaches, is that each variable within the network may act as both predictor and response, and hence allowing for testing direct and indirect effects on pre-assumed causal relationships (Grace 2006, Grace et al. 2010, Fan et al. 2016). Piecewise SEM, in contrast to traditional SEM, can handle nested factor survey designs and random effects (Lefcheck 2016). Since the factor management was modelled as a composite variable and the sign of the coefficients cannot be interpreted (Grace 2006), we also run linear mixed-effects models to test the interactive effect of management and habitat zone on species composition. Models included site as a random factor nested within each level of management to account for random variability between sites. The advantages with SEM and piecewise SEM have enabled scientists to investigate relationships of usually complex ecological systems and these methods have therefore been increasingly used during the last decades (e.g. van der Heide et al. 2011, Duffy et al. 2016, El-Hacen et al. 2018, Gouezo et al. 2019).

Assessment of intra- vs. interspecific trait variability Finally, in Paper III we also assessed the relative influence of MPAs on seagrass species and trait composition using a combination of single- and multiple-trait statistical analyses (Lepš et al. 2011, Auger and Shipley 2013, Jung et al. 2013, Kichenin et al. 2013) in R v. 4.0.0. (R Core Team 2020). Following

26 Lepš et al. (2011) we calculated the trait values; ‘specific’ (based on average trait value per species per sampling station) and ‘fixed’ (average trait value per species across all sampling stations). The ‘specific’ trait value include both intra- and interspecific variability, while the ‘fixed’ value only includes the effect of interspecific variability. By using the decomposition of the total sum of squares from ANOVA models for specific, fixed and intraspecific trait values we were able to disentangle the effects of interspecific (species turnover) and intraspecific trait variability, and their covariation. Additionally, we estimated the relative importance of the predictors ‘management’ (3 levels) and ‘habitat zone’ (3 levels) by fitting ANOVA models to each source of trait variability.

27 Synthesis of results and discussion

Effects of MPAs on benthic communities in seagrass beds and on coral reefs By sampling benthic communities in twelve sites in coastal Kenya, we found support for the hypothesis that MPAs benefit seagrass and coral communities (Paper I). Protection appears to induce a shift in benthic community composition, from dominance of stress-tolerant and pioneering species to stress-sensitive and competitive seagrass and coral taxa (Paper I). Succession theory suggests that competition and colonization trade-offs can change both composition and diversity of communities over time (Grime 1977, Grime and Pierce 2012). Recently, division of corals (Darling et al. 2012) and seagrasses (Kilminster et al. 2015) into different life-history groups have helped to predict how communities develop over time when excluding disturbances from fishing. In line with these theories, results from Paper I suggest that, on coral reefs, the weedy and stress-tolerant genera Porites and Stylophora were re- placed by stress-sensitive and structurally complex Acropora. In seagrass beds, T. hemprichii and the small and colonizing Halophila ovalis were re- placed by the more structurally complex, slow-growing (Marbà and Duarte 1998) and advantageous competitor (Duarte et al. 1996) T. ciliatum. H. ovalis can form persistent seed reserves and cope with stress through rapid re-colonization (Inglis 2000), and this may explain why this, and other pioneering species, is more common in fished areas. Fishing-associated activities, e.g. walking and trawling, can be destructive for both corals (Jones 1992) and seagrasses (Walker et al. 1989; Daby 2003). The exclusion of fishing within the MPAs may therefore facilitate recovery of sensitive organisms, such a corals from the genera Acropora, whose complex but brittle structure (Marshall 2000) contribute to reef complexity. Both T. hemprichii and T. ciliatum are persistent climax species (Kilminster et al. 2015), but T. ciliatum is much more sensitive to grazing by sea urchins due to its exposed meristems (Alcoverro and Mariani 2002). Since protected areas in Kenya allow recovery of a keystone sea urchin predator (red-lined triggerfish; Balistapus undulatus) (McClanahan 2000), the lower abundance of sea urchins in protected areas (Alcoverro and Mariani 2004) may allow T. ciliatum to recover (Figure 8). Our results support similar shifts that have been demonstrated among coral reef communities (Darling et al. 2013). Space-for-time substitutions should

28 not replace long-term studies but are plausible and can yet contribute with important knowledge when assessing ecological effects of MPAs (Pickett 1989, Micheli et al. 2012, Fraschetti et al. 2013, Strain et al. 2019). To my knowledge Paper I is the first study to suggest that MPAs can alter benthic community composition from structurally simple to complex foundation species in soft-bottom systems, even though previous research indicates that T. ciliatum is more common in protected areas (McClanahan et al. 1994, Alcoverro and Mariani 2004).

Figure 8. The collector urchin (Tripneustes gratilla) is abundant in heavily fished areas in Kenya, here in a T. hemprichii dominated seagrass bed in Kanamai (fished site) (left). Photography of a T. ciliatum seagrass meadow in Watamu MPA, Kenya (right). (Photographer A. Chirico)

Habitat-forming organisms in general alter the local environment and offer shelter for many organisms (Jones et al. 1994, Stachowicz 2001). Studies from coral reefs demonstrate that structurally complex habitats harbor more fish than less structurally complex ones (e.g. Beukers and Jones 1998, Graham and Nash 2012). Even though we have limited knowledge about interactions between seagrass habitat complexity and their associated fish communities, Gullström et al. (2008) showed that the complex seagrass species T. ciliatum harbored more fish than less complex species. Additionally, increased habitat complexity of seagrass has been shown to increase invertebrate abundance and diversity in tropical seagrass beds (Heck and Wetstone 1977), which in turn can benefit fishes since many fishes feed on invertebrates. Given the high value of seagrass-associated fisheries in East Africa (Francis et al. 2002, de la Torre-Castro and Rönnbäck 2004, Okemwa 2017), these effects can indirectly benefit coastal societies. More structurally complex species can also have implications for carbon storage, since e.g. the large T. ciliatum can through their high below-ground biomass store more carbon than smaller seagrasses

29 (Duarte and Chiscano 1999, Gullström et al. 2018). To summarize, these effects can contribute to sustain important ecological functions, such as increased algae grazing by herbivorous fishes, and ecosystem services like fish production and increased carbon sequestration in MPAs.

We also found differences in community composition between the three habitat zones (Paper I). Taxa composition in the intertidal and mid-lagoon zones were significantly different from the reef zone. The reef zone harbored more corals (although cover was generally low, on average 2-5%) and macroalgae such as Sargassum spp., while the intertidal and mid subtidal zone had more seagrass. Low hard coral cover in Kenyan lagoons is a typical pattern in Kenya (e.g. Lilliesköld Sjöö et al. 2011), and also in the wider Indo-Pacific (Darling et al. 2019). Following the El Niño Southern Oscillation heat wave in 1998, hard coral cover within Kenyan MPAs are only slightly higher than fished areas (data from 1991-2011; Darling et al. 2013). Coral bleaching, together with overfishing and destruction of marine habitats (Obura 2001, McClanahan et al. 2008, McClanahan 2019) are likely explanations to the low coral cover.

Protection also increased seagrass cover and reef topographic complexity, but did not affect total coral cover (Paper I). Similarly, a large body of studies from coral reef research demonstrate that MPAs increase diversity and reef complexity (e.g. review Graham et al. 2011), whereas effects of MPAs on coral cover is variable (Strain et al. 2019). A review of coral reef studies sum- marizes that protection can increase cover of corals slightly and decrease macroalgae (Graham et al. 2011), but protection effects on coral cover based on life-history traits (e.g. level of stress-sensitivity), and not total coral cover, seem to be more strongly affected by protection (Darling et al. 2013). Most studies assessing ecological effects of protected areas have focused on hard- bottom systems, such as coral reefs and temperate macroalgal reefs (e.g. Wells et al. 2007, Duarte et al. 2008), and effects of protection on seagrasses have been much less studied. Previous studies from seagrass beds demonstrate higher seagrass shoot density and leaf production in response to protection (Marbà et al. 2003, Ferrari et al. 2008), which can explain our results of higher seagrass cover in MPAs (Paper I).

Our results also suggested that protection increased benthic diversity in both coral and seagrass areas (Paper I). Even though the community-managed MPAs are only recently established they still harbored very diverse communities, probably by harboring both early and late successional species (see succession theory above). Additionally, there was no interaction between management and habitat, suggesting that diversity increased at a similar rate in the habitat zones as a result of protection. It is known that protection can increase diversity on coral reefs (e.g. Halpern and Warner 2003, UNEP 2006), but to

30 my knowledge this is the first study demonstrating that MPAs increase diversity in seagrass habitats as well. High seagrass diversity can be linked to more effective retention of sediments, since different species act on different scales, and increased sediment accretion can stabilize coastal sediments and prevent erosion (Potouroglou et al. 2017). Even though protection effects were similar, benthic diversity was higher in the reef zone, compared to the mid and intertidal seagrass zone. Coral reefs have long been acknowledged for being hy- perdiverse ecosystems (e.g. Odum and Odum 1955), and high diversity can sustain more ecological functions and ecosystem services (e.g. Worm et al. 2006). However, even though diversity was lower in seagrass beds this does not necessarily mean that these habitats are less important. Since seagrass and coral communities harbor different types of species, they also generate different types of feeding, nursery and settlement grounds for associated organisms. Moreover, they generate different types of ecological functions and ecosystem services. For instance, seagrass beds have been shown to stabilize sediments and alleviate low pH stress (Semesi et al. 2009, Potouroglou et al. 2017), services that are not known to be generated by coral reefs.

Effects of MPAs on fish assemblages in seagrass beds and on coral reefs In paper II we found that MPAs appeared to benefit fish assemblages in both coral and seagrass habitats by increasing fish size, biomass and theoretical monetary value. For total fish abundance effects were less clear, fished and community-managed MPAs did not differ, whereas government-managed MPAs had higher abundances than community-managed protected areas, but not fished areas). Previous studies have also shown that MPAs is an effective tool in conserving fish stocks, primarily on coral reefs (Bohnsack 1998, Gell and Roberts 2002, Graham et al. 2011, McClanahan et al. 2016b), but also in seagrass beds (Valentine et al. 2008, Valle and Bayle-Sempere 2009, Unsworth et al. 2010, Fraschetti et al. 2013, Seytre and Francour 2014, Alonso Aller et al. 2017). Time since protection (years) is an important factor for recovery since fish stocks build up over time (Roberts and Hawkins 2000, Claudet et al. 2008, McClanahan et al. 2009). However, several studies (including ours; Paper II) show that only a few years of protection can be enough for fish stocks to recover (e.g. Polunin and Roberts 1993, Halpern and Warner 2002, Babcock et al. 2010). Protection effects on fish communities, like those found in Paper II, can have consequences for ecological functions in coral and seagrass ecosystems. For example, higher levels of fish biomass and diversity improve the health of coral reefs, by enhancing grazing rates on macroalgae that can outcompete corals (Mumby et al. 2006, Topor et al.

31 2019). Additionally, increased abundance of sea urchin predators, such as triggerfish (Balistidae), can reduce abundance of sea urchins, that may otherwise erode corals (Bak 1994, McClanahan 1995, Eakin 1996, Dumont et al. 2013) and overconsume seagrasses (Eklöf et al. 2008).

Recovered fish populations within MPAs can also benefit adjacent fisheries by larval dispersal or by fish migrating from protected areas to fished areas, so called ‘spillover’ (Bohnsack 1998). Evidence of spillover has been demonstrated globally (Roberts et al. 2001, Russ et al. 2004, Halpern et al. 2010, Harrison et al. 2012, da Silva et al. 2015), and there is also evidence of spillover from MPAs in Kenya (McClanahan and Mangi 2000). McClanahan and Mangi (2000) showed that after five years of protection Mombasa MPA increased the catch per fisher and catch per area with up to 50% outside the borders. Spillover was greatest for the most heavily fished species, e.g. herbivorous surgeonfish (Acanthuridae) and rabbitfish (Siganidae), but was at the time insufficient to compensate for reduced catches caused by loss of fishing grounds in connection to the MPA establishment. Consequently, the community- and government-managed MPAs in Paper II could in theory provide spillover of large and valuable fish to nearby fisheries, although this was not tested in our study. Whether there are spillover effects, and if the amount of spillover can compensate for the loss of fishing grounds remain to be explored.

Protection effects on fish potential monetary value were much stronger than the effects on fish density, body size and biomass (Paper II). Similar results have been found in in the Western Pacific (Unsworth et al. 2010) and in the Caribbean (Polunin and Roberts 1993). In our study this result was explained by larger fishes within the protected areas. Fish value per kg increases with fish body length because fishes can enter a more valuable market, e.g. be sold to hotels instead of households (McClanahan 2010, Thyresson et al. 2013), and therefore the size of fish has a stronger effect on value than biomass. Ad- ditionally, MPAs harbored higher densities of valuable fish such as rabbitfish (Siganidae) (Paper II).

Abundance, biomass and potential monetary value of fish per unit area were greater on coral reefs than on seagrass beds, whereas fishes were on average larger in seagrass beds than on coral reefs (Paper II). Moreover, fish density and biomass increased with coral cover, but decreased with seagrass cover, whereas fish size and value were unaffected by cover. These contrasting effects of coral and seagrass cover can be explained by various factors. For instance, many coral reef fishes are small, territorial and sedentary, e.g. damsel- fishes (Pomacentridae), and are dependent on the local habitat (Roberts and Ormond 1987). This can explain why we found a positive relationship between coral cover and fish biomass and abundance, and also smaller sized fish on the coral reef (Paper II). In contrast to other studies, e.g. Alonso Aller et

32 al. (2014) that demonstrate a positive relationship, we found a negative relationship between seagrass cover and fish density and biomass. This result was in our case most likely driven by the fact that the MPA that had been protected the longest time (Kisite Marine National Park; 33 years) had low seagrass cover, but very high fish biomass and abundance. Effects of cover of habitat- forming organisms on fish communities would ideally require more replicates per management type (fished and MPAs) and more replicates in each site that additionally should span a larger difference between cover (i.e. from 0% to 100% cover) (see e.g. Alonso Aller et al. 2014). Just like coral cover, seagrass cover has been shown to benefit fish communities in seagrass beds, as well as nearby coral reefs (Burke et al. 2001, Dorenbosch et al. 2005, Grober- Dunsmore et al. 2007, Barbier et al. 2011), and this is also most likely the case in Kenya.

Effects of MPAs on seagrasses species and trait composition To test whether protection can influence seagrass species and trait composition we measured five morphological traits (shoot density, leaf length and width, number of leaves per shoot, and above:below-ground biomass ratio) on multispecies seagrass assemblages within MPAs and fished areas in three habitat zones (shallow, mid-lagoon and reef) (Paper III). We found that protection influenced seagrass species composition and, mostly indirect, trait composition, and effects were strongest in the mid-lagoon. Since these analyses are based on the same data set as in Paper I, effects of MPAs on seagrass species composition were similar to those described in Paper I, in which T. ciliatum were more common in old government MPAs than community MPAs and fished areas (see above). Regarding seagrass trait composition government MPAs increased shoot density and decreased leaf length through species turnover, and in the case of shoot density, protection also seems to induce higher shoot density through plasticity (Paper III). Previous studies show that MPAs can support higher rates of seagrass herbivory by fish than unprotected areas in Kenya (Alcoverro and Mariani 2004), and in the Mediterranean Sea (Planes et al. 2011). Additionally, Planes et al. (2011) demonstrated that the high rates of herbivory within the MPAs reduced leaf length and induced higher shoot density of the seagrass Posidonia oceanica through plasticity, which is consistent with our study. The response of seagrasses to compensate in growth after biomass loss induced by herbivores have also been shown in other studies (Verges et al. 2008, Alonso Aller 2018). Seagrass trait composition, whether induced by species turnover or plasticity, can strongly influence ecological functions that seagrasses regulate and ecosystem structure of associated organisms (Jones et al. 1994, Hemminga and Duarte 2000). A previous

33 study from the Western Indian Ocean shows that seagrass meadows with high shoot density harbor more diverse and dense fish assemblages than meadows with low shoot density (Gullström et al. 2008). Shoot density is also suggested to be a better predictor of fish assemblages than e.g. canopy height (Gullström et al. 2008). High shoot density can also increase diversity and density of infaunal benthic communities, like crustaceans and mussels (Webster et al. 1998, Boström and Bonsdorff 2000). Therefore, more dense meadows within MPAs could in theory result in higher fish production that can indirectly benefit local fisheries (McClanahan and Mangi 2000) and ecological functions, like fish herbivory that control macroalgal growth (as shown in the Atlantic and Pacific Ocean; Topor et al. 2019).

Analyses of single traits show that phenotypic plasticity (intraspecific variability) explain most of the variability in the three traits shoot density, leaf length and above:below-ground biomass ratio (Paper III). Leaf width and number of leaves per shoot on the other hand were mostly determined by species composition (interspecific variation). All of the five traits showed some degree of plasticity which is consistent with previous studies demonstrating that seagrasses show high degree of plasticity in response to local conditions (Cabaço et al. 2009; Maxwell et al. 2013; McDonald et al. 2016; Soissons et al. 2017). For example, seagrasses can increase their leaf length to be able to maximize photosynthetic capacity in areas with poor water quality (Bulthuis 1983). Management and habitat explained very little of the overall trait variability (R2<0.20). The high levels of ‘unknown’ variability can be explained by both small-scale and large-scale factors. Since the core capture a small area (10 cm diameter), small-scale variations such as sediment conditions, may differ between samples. At a larger scale variation in seagrass traits may be driven by more distant disturbances such as land runoff of dissolved nutrients and sediment particles. Crowded beaches and growing human populations along the Kenyan coast has increased sewage pollution (Holden 2016), and in Malindi and Watamu (where two of the old government MPAs are situated) levels of runoff from land is extensive (Fleitmann et al. 2007). High levels of urban areas and agricultural land is a well-known driver of seagrass species composition and distribution in tropical and temperate areas (e.g. Short and Wyllie-Echeverria 1996; Quiros et al. 2017) and can reduce seagrass growth and survival (Terrados et al. 1998; Quiros et al. 2017).

Effects of management on species and trait composition clearly differed between habitat zones (Paper III). Protection had a strong effect on species composition in the mid-lagoon, weak effect in the reef zone and no effect in the shallow intertidal zone. The weaker effect in the reef zone is likely explained by that seagrasses occur relatively sparsely at the reef, and that increasing time of protection decrease seagrass cover (as shown in Paper I), presumably because seagrasses are outcompeted by corals that benefit from

34 protection (McClanahan et al. 1994). Tourism is a large industry along the Kenyan coast and very important for Kenya’s economy (UNEP et al. 2000, Ongoma and Onyango 2014). Even though extractive resource use (e.g. fishing and digging for shellfish) is prohibited in the MPAs, shallow intertidal zones in Kenya are still heavily utilized for tourism activities that could counteract protection effects. For example removal of seagrass beach cast (considered nuisance for tourists) is common outside hotels in Kenya. In Mombasa MPA, seagrass removal has increased wave action and resulted in beach erosion (Ochieng and Erftemeijer 1999). Additionally, further studies in the Western Indian Ocean (Mauritius) show that beach cast removal has increased turbidity in shallow areas and resulted in loss of infauna (Daby 2003). High rates of walking sun-bathing tourists and boating activities to and from sea safaris include high rates of trampling and anchored boat moorings – activities that are known to damage seagrasses (Zieman 1976, Walker et al. 1989, Short and Wyllie-Echeverria 1996, Eckrich and Holmquist 2000, Orth et al. 2006) and can take several years to recover (Furman et al. 2019). Together with high levels of land runoff in some of the MPAs (Malindi and Watamu; Fleitmann et al. 2007), these disturbances are likely to diminish protection effects and contributed to the lack of MPA effect in the shallow intertidal zone (Paper III). Moreover, the shallow intertidal zone in one of the community MPAs was not protected, because local communities regularly use it to extract resources like shellfish and bait, even though the mid-lagoon and the reef were protected. Due to limitations in the statistical methods we counted this unprotected zone as ‘protected’ which could potentially have weakened the effect of MPAs in the shallow zone.

In contrast to the lack of protection effect in the shallow intertidal zone (Paper III), the analyses in Paper I suggest that there are protection effects on seagrasses in the shallow intertidal zone. The contradictory results are most likely explained by differences in sampling method. First, the analyses in Pa- per I is based on seagrass % bottom cover estimates, while those in Paper III are based on biomass. Second, cover estimates for each station in Paper I is based on a larger area (mean values of three 0.5 m2 quadrats [total area 1.5 m2]), whereas biomass values in Paper III is based on core samples with 10 cm in diameter (see Methods for details). Both of these methods have pros and cons. On the one hand biomass samples are measured (g, dry weight) (Paper III) and therefore more exact than cover estimates. On the other hand a larger area (Paper I) would be preferable to more accurately represent habitat characteristics (especially if the habitat is heterogeneous). Taking both results into account, I would interpret the overall results to suggest that MPAs can protect shallow intertidal zones to a certain extent, but that effects are weak and therefore not detected in one of the analyses. Meanwhile, both of the analyses showed clear MPA effects on seagrasses within the mid-lagoon zone. To what

35 extent MPAs can protect shallow intertidal seagrass needs to be more carefully evaluated.

Effects of a MPA and size of disturbance on seagrass recovery and sediment erosion We cleared seagrasses from plots of two different sizes (0.25 and 1 m2) in intertidal seagrass beds dominated by T. hemprichii, to assess effects of protection (one community-managed MPA) on seagrass recovery and sediment erosion. Over two years of sampling we found that seagrass recovery in the protected area was faster (steeper recovery trajectory) than in the fished reference sites (Paper IV) (Figure 9). Seagrass recovery was primarily faster during the initial phase, and effects weakened after one year and was more or less absent after two years. At the end of the experiment the clearings still had lower seagrass cover than control plots, demonstrating that recovery was not yet complete. A recent study on Thalassia testudinum showed that seagrass vessel injuries take more than seven years to fully recover (Furman et al. 2019), and confirm the long recovery time for Thalassia spp. After and during the experiment all clearings were dominated by T. hemprichii. This demon- strates that the disturbances did not trigger a shift in community composition, as shown in a macroalgal community study where large areas (ca. 8 m2) did not recover and turned into a community state with different species composition (Petraitis and Dudgeon 2005). Theoretically, the largest experimental clearings (1 m2) in our study were too small for such shifts to occur, or local hydrodynamics are not strong enough to generate thresholds in seagrass clearing size at all.

Similarly, we found that the protected site experienced less sediment erosion (estimated as the relative depth of the clearings), and had lower depth than the other sites during the first year (Figure 9) (Paper IV). However, effects of protection were gone at the end of the experiment. Even though we did not test the actual mechanism, it is very likely that the clearing recovered faster and experienced less erosion in the MPA due to less disturbances from human activities. Fishing pressure is generally high in this region, and particularly in seagrass beds, and the use of destructive gear, such as net trawls, are common (Obura 2001, Evans et al. 2011, Okemwa 2017). Fishing gears, trampling and boat anchoring can be detrimental to seagrasses and sediment (Short and Wyllie-Echeverria 1996, Mangi and Roberts 2006, Orth et al. 2006). Exclud- ing these disturbances likely explain why we observed reduced levels of sediment erosion and facilitation of seagrass recovery by protecting the emerging seagrass shoots in the protected area (Paper IV). The MPA (Kuruwitu Clo- sure) is well-enforced, has few tourist visitors and is successful in terms of

36 protecting fish communities and diversity of habitat-forming organisms associated with seagrass beds and coral reefs (Paper I, II) (see also McClanahan et al. 2016a). Poaching is therefore unlikely and, hence, the MPA constitutes an excellent site to assess protection effects.

Even though MPAs can decrease resilience of certain communities by selecting for more sensitive species (Darling et al. 2013), protected areas can also increase resilience. For example, protection can increase temporal stability of seagrass fish communities (Alonso Aller et al. 2017) and reef fish (Babcock et al. 2010, Mellin et al. 2016). Additionally, protection can increase resilience among coral reef communities by corals experiencing a 30% lower impact and a 20% faster recovery in MPAs following disturbances from coral diseases, bleaching, storms, and crown-of-thorns sea star outbreaks (Mellin et al. 2016). In line with these hypothesis, our results suggest that the presence of a MPA speeds up the recovery of seagrass cover and reduce sediment erosion (Paper IV). Whitfield et al. (2002) showed that already disturbed and injured seagrasses are more vulnerable to further damage from additional disturbances such as storms, compared to undisturbed seagrasses that are more resilient. This suggests that even though protection effects were more or less absent after two years, the faster rate of recovery during the first year (in terms of seagrass cover and erosion) in the MPA (Paper IV) could at least in theory imply higher resistance towards additional disturbances, such as storms, during that particular time. However, seagrass resilience in terms of resistance to change and ability to recover from disturbances is yet to be better understood (O'Brien et al. 2018).

Large and small disturbances did not differ in terms of recovery of seagrass cover, but clearly differed in terms of sediment erosion (Paper IV) (Figure 9). The depth of small clearings in the protected site started to recover already in the beginning of the experiment, while small clearings in the fished reference sites did not start to recover until after ca. 45 weeks. Large clearings, on the other hand, eroded at high rates in both the MPA and fished areas until ca. one year after the experiment started, although erosion was less pronounced in the MPA compared to one of the fished reference sites. Large disturbances are expected to recover slower than small disturbances (scale-dependent recovery; Petraitis and Latham 1999). This hypothesis was supported by our study that demonstrated higher levels of sediment erosion and erosion also continued for a longer time in large clearings compared to small clearings.

Figure 9. Temporal changes in seagrass cover and clearing depth across treatments and three sites during 107 weeks (ca. 2 years). Depth of clearings are relative to the depth right after start. Number of observations = 678 and number of groups = 45 for seagrass cover, and 450 and 30, respectively for clearing depth.

In a similar seagrass recovery experiment, Eklof et al. (2011) demonstrated that bioturbators reduced the threshold disturbance size needed to facilitate seagrass recovery. Additionally, a disturbance experiment in Mauritania showed that seagrass recovery in experimental clearings was lower in clearings that also experienced additional stress from air exposure during low tides, particularly large clearings (El-Hacen et al. 2018). Based on these interactions between size and disturbance, we hypothesized that the presence of a MPA (by excluding local disturbances associated with fishing) could increase the disturbance size needed to alter recovery trajectories. Regarding the interaction between management and size of disturbance, we found effects, even though relatively weak. Generally, seagrass cover did not differ between small and large clearings within the same site. However, small and large clearings differed in seagrass cover and rate of change (in seagrass cover) between sites, primarily in that the MPA had higher seagrass cover and that seagrass cover increased faster. For example small clearings in the MPA had higher seagrass cover and seagrass cover increased faster during the two years of the experiment than large clearings in the two fished reference sites (Ref 1 and Ref 3) and small clearings in one of the fished reference sites (Ref1) (for details see Table S1 in Paper IV). Additionally, large clearings in the MPA recovered faster than large clearings in the fished site Ref3 and smaller clearings in the fished site Ref1. Despite some indications of interacting effects, these results suggest that the presence of a MPA did not change recovery rate in response

38 to size of disturbance, in contrast to our hypothesis. There can be several explanations to this result. Either, the hypothesis in not true, or more likely, effects could not be detected because disturbance sizes were too similar and/or because the study only included one MPA. Regarding clearing depth, we did, however, find an interaction between size of disturbance and site. Large clearings in the fished areas (Ref1 and 3) eroded more (increased more in depth) than small clearings in the MPA (Figure 9). Additionally, small clearings in the fished reference site Ref3, were deeper than the large clearing within the same site, in contradiction to the theory of scale-dependent recovery. If the protected area would have experienced less difference in erosion (or seagrass growth) between large and small clearing, compared to fished areas, the presence of a MPA could have increased temporal stability regarding sediment erosion, as found in fish communities (Babcock et al. 2010, Mellin et al. 2016, Alonso Aller et al. 2017). However, such effects were not found and it remains unknown whether MPAs can decrease fluctuations in seagrass recovery rate between different disturbance sizes.

Protection effects differ depending on metric This thesis illustrates that effects from MPAs differ between the metrics used. For example, total coral % cover was not affected by protection, while reef structural complexity and coral taxonomic composition were strongly influenced by protection from old MPAs (Paper I). Likewise, total fish density was less sensitive to protection effects, but protection clearly affected density of certain fish groups (e.g. of high-value rabbitfish), as well as the potential market value of fish communities (Paper II). This suggests that some metrics, e.g. total coral cover, are not sensitive enough for protection effects to be detected and the choice of metric should, therefore, be selected carefully. Previ- ous studies have also found similar results. For example, Edinger and Risk (2000) suggest that coral cover is a less suitable metric, whereas coral species richness and habitat complexity are more suitable when assessing effects of protection on coral reef status.

The choice of metric or approach should be based on the type of question(s), organism(s) or system(s) that are of interest. Fishing is an important source of generating income and sustain daily food requirements (Roberts and Hawkins 2000, UNEP 2006), particularly in East Africa (Jiddawi and Öhman 2002, de la Torre-Castro and Rönnbäck 2004, Okemwa 2017). Additionally, theoretical monetary value of protected fish assemblages better represent the income of a fisher than e.g. total fish density (Coulthard 2012). Therefore, the strong protection effect on fish with high market value (Paper II) can from the fisher´s point of view be more important in case of spillover of valuable fish (McClanahan and Mangi 2000, Roberts et al. 2001, Halpern and Warner 2003,

39 Russ et al. 2004, Harrison et al. 2012, da Silva et al. 2015). Moreover, high abundance of herbivorous fishes can improve coral reef health by higher rates of grazing on algae that may otherwise suppress coral growth (Abal et al. 1994, Hughes 1994, Mumby et al. 2006, Topor et al. 2019). Therefore, effects on certain fish functional groups, such as herbivores, can be of greater importance from an ecological perspective than e.g. total fish abundance. How- ever, total cover of benthic organisms and total fish abundance can still be valuable measurements, but if complemented with other metrics, such as taxonomic composition, structural complexity or fish size and theoretical monetary value measurements (as in Paper I-III), studies can more accurately capture protection effects and/or ecosystem health.

When time and money are constraints, identifying easy and cheap methods that capture variables that mirror the organisms and systems of interest are advantageous. For example, measuring reef structural complexity (standard rugosity estimation; McClanahan and Shafir 1990) is a fairly easy way to capture a variable that strongly predicts fish biomass, diversity and trophic structure, and is therefore promoted in coral reef monitoring (Darling et al. 2017). If the benthic study in Paper I only would have included coral cover, effects of MPAs on corals would have been neglected, and hence protection effects would have been underestimated. For seagrasses, above-ground plant biomass (Heck and Wetstone 1977), shoot biomass, shoot density and canopy height (Gullström et al. 2008) have been suggested to be measurements of seagrass complexity. Biomass variables are extractive and require more comprehensive analysis (drying, weighing etc., see Methods), whereas estimating shoot density is fairly easy, fast and cost-effective, and also a stronger predictor of fish assemblages than canopy height (Gullström et al. 2008). Shoot density could therefore be an equivalent measure as reef rugosity for coral reefs. Both reef structural complexity and seagrass shoot density seems to be sensitive metrics and were affected by protection (Paper I and Paper III, respectively).

Effects of protection on seagrass beds versus coral reefs In Paper I and II we used the same method (quadrat and point-counts) when investigating benthic and fish communities in seagrass and coral habitats. This approach enabled us to compare protection effects on seagrass beds versus coral reefs. Our results demonstrated that protection affected seagrass and coral communities in the same way, by promoting more diverse and structurally complex habitat-forming organisms, and benefit their associated fish communities. One metric that did differ in terms of protection was total cover. Total coral cover (all species pooled) was not affected by protection, whereas total seagrass cover increased with protection (Paper I). A possible explana- tion to why seagrass cover increased but not coral cover, could be that corals

40 have suffered much more from heat waves than seagrasses (even though warming can also reduce seagrass growth and increase seagrass mortality; e.g. Collier and Waycott 2014). The heat wave in 1998 resulted in extensive coral decline, and additionally loss was more comprehensive in Kenyan MPAs compared to fished areas, since protection selects for more stress-sensitive coral taxa (Darling et al. 2010, 2013). The more prevailing loss of corals in protected areas and the relatively low coral cover in general in Kenya (see discussion above) could have resulted in that protection effects were absent.

Community-managed MPAs as an alternative or complementary protection tool Establishment of government-managed MPAs that exclude local communities and reduce fishing grounds can have great socio-economic costs for the locals (Klein et al. 2008, Smith et al. 2010, Adams et al. 2011), and additionally, result in low acceptance and conflicts that jeopardize conservation goals (Christie 2004, Abunge et al. 2013, Giakoumi et al. 2018). Therefore the increasing amount of community-managed MPAs globally (e.g. White and Vogt 2000, Govan 2009, Pascal 2011), and also in Kenya, are promising due to their higher acceptance among local communities (Pollnac et al. 2001, Gell and Roberts 2002, Hicks et al. 2009, McClanahan et al. 2016a). The social aspects of MPAs are not covered in this thesis, but previous studies from Kenya indicate that community-managed MPAs are associated with high compliance, opportunities for generating tourism revenues that benefit the village and high maintenance of social values within the community (e.g. cultural values) (Hicks et al. 2009, McClanahan et al. 2016a). From an ecological point of view, this thesis show that the community-managed MPAs did not generate as strong protection effects as the much older (≤6 vs. ≥20 years) and larger (<0.5 vs. ≥6 km2) government-managed MPAs. For example, they did not change benthic community composition from structurally simple to complex species, and they did not increase coral complexity or seagrass cover over time (Paper I). However, despite that they are young and much smaller they did generate protection effects on benthic community diversity (Paper I), fish communities (Paper II) and seagrass recovery following experimental disturbance (Paper IV). Previous research demonstrate that protection strength- ens over time due to time lags in recovery that depend on e.g. re-colonization rates and reproduction cycles (e.g. Roberts and Hawkins 2000, Claudet et al. 2008). Also the size of a protected area can influence protection effects, for example by regulating population size of organisms (Claudet et al. 2008, McClanahan et al. 2009). Community-managed MPAs are usually small and established close to villages (Weeks et al. 2010, McClanahan et al. 2012), where they can more easily be managed (e.g. guarded and monitored). Since

41 management type (community vs. government MPA) in this study are strongly correlated with both the age and size of the protected areas, we can only spec- ulate whether protection effects from community-managed MPAs in the future will be as strong as the effects from government-managed MPAs. Given that they are much younger, and compliance to fishing regulations are usually high (McClanahan et al. 2016a), protection effects are likely to strengthen with time. Additionally, Paper I reveals that the community-managed MPAs harbored a mix of early and late successional species, indicating that with time, protection effects will likely resemble those in the older government- managed MPAs, in which more stress-sensitive climax species have taken over after pioneering taxa (Paper I). Small MPAs might not be able to generate the same protection effects as large MPAs, but can successfully protect sedentary species and sustain local fisheries through spillover of valuable fish to nearby areas (e.g. McClanahan and Mangi 2000, da Silva et al. 2015). Large MPAs can be more effective in protecting larger and more mobile organisms, e.g. sharks, and be able to contain different types of habitats (Hilborn et al. 2004, Nardi et al. 2004), and are therefore more likely to support some types of conservation goals, like sustaining high biodiversity. The demonstrated protection effects from community-managed MPAs, together with previous studies showing that they can also fulfill socio-economical goals (e.g. Hicks et al. 2009), suggests that community-managed MPAs can be a promising complement to government-managed MPAs.

Figure 10. Semi-conceptual table summarizing the main results comparing the effects of fished areas, and community- vs. government-managed MPAs in Paper I-IV.

43 Conclusions

MPAs are effective but not effective enough Using coastal Kenya as a case study, this thesis demonstrate both strengths and shortcomings of MPAs as a conservation tool (summarized in Figure 10). MPAs generated protection effects on habitat-forming organisms, diversity of benthic organisms (Paper I), their associated fish communities (Paper II), and speeded up recovery following disturbances (Paper IV). Additionally, protection induced a shift in community composition towards dominance by structurally complex climax coral and seagrass species (Paper I, III). These effects are most likely a result of protected areas reducing disturbances associated with resource extraction that damage habitat-forming organisms di- rectly and indirectly in fished areas. Therefore, place-based protection through MPAs that ban extractive resource use is an effective tool in marine conservation. However, from a seagrass perspective, this thesis also shows that MPAs did not seem to fully protect seagrasses within the shallow intertidal zone (Paper III), likely because these areas are easily accessible and heavily utilized for tourism purposes. Additionally, previous studies suggest that non- point disturbances, like pollution and sediment runoff, can reduce seagrass growth and survival (e.g. Short and Wyllie-Echeverria 1996, Terrados et al. 1998, Quiros et al. 2017). Altogether, the findings in this thesis and previous literature suggest that MPAs do not seem to be enough to protect the tropical seascape as a whole. Thus, urging for a more holistic approach with integrated coastal zone management to reduce impacts of both local disturbances, like tourism activities, and those deriving from more distant sources such as runoff from agriculture.

Reducing disturbances from human activities, such as fishing, can increase seagrass resilience (Orth et al. 2006). The faster recovery in experimental clearings in the community-managed MPA compared to the fished reference sites (Paper IV) indicated that MPAs can increase seagrass recovery and decrease erosion. However, effects were not very strong and diminished with time, and since we could only use one MPA (see Methods), conclusions remain tentative. Given that seagrasses, particularly in East Africa, are in decline and increasingly impacted by human disturbances (Waycott et al. 2009, Harcourt et al. 2018, Unsworth et al. 2018a), these ecosystems are in need for

44 more effective and targeted conservation (Green and Short 2003). Conse- quently, our findings that protection appears to speed up seagrass recovery and decrease sediment erosion following disturbance (Paper IV) are an important contribution to seagrass resilience research and promising from a conservation perspective. Together with previous studies suggesting that protection can actually decrease resilience, by selecting for more sensitive taxa (Darling et al. 2010, 2013), there is clearly a need to more thoroughly evaluate how protection affects the resistance to, and recovery from, disturbances on seagrasses and other habitat-forming organisms.

Community-managed MPAs as promising complements to state-run MPAs Even though protection effects were usually stronger in government-managed MPAs, this thesis suggests that also community-managed MPAs contribute to diverse seagrass and coral communities (Paper I), protect fish stocks (Paper II) and to some extent speed up seagrass recovery following disturbances (Pa- per IV). Given their usually smaller size, community-managed MPAs may never generate as strong protection effects on e.g. biodiversity or large and highly mobile organisms, as large MPAs (Weeks et al. 2010). But the fact that these bottom-up driven forms of marine natural resource management appear to generate positive effects on seagrass and coral communities even after just a few years of protection (Paper I, II and IV), suggests that they can be a successful complement to government-managed MPAs. Regardless of ecosystem type, conservation efforts are always working towards being as effective as possible, and excluding local communities have often proved problematic, by causing political conflicts that threaten ecological conservation goals (Hilborn 2007, Pomeroy and Douvere 2008, Agardy et al. 2011). This thesis has contributed to the increasing amount of literature that points towards that community involvement is an effective management strategy that can help sustain both ecological (e.g. White and Vogt 2000, Clements et al. 2012, da Silva et al. 2015, McClanahan et al. 2016a) as well as socio-economic values (e.g. Hicks et al. 2009).

Include seagrass beds in management plans This thesis shows that MPAs can benefit mid-lagoon seagrass beds, their associated communities and increase seagrass recovery (Paper I-IV), just as they effectively protect coral reefs. These findings suggest that place-based protection is an effective tool for reducing local disturbances in seagrass beds. Seagrasses are often ignored within marine management in the Western Indian

45 Ocean (Unsworth and Cullen 2010), and a recent study shows that there are very few seagrass restoration sites in the Indo-Pacific region (in Western Aus- tralia and Indonesia) (Duarte et al. 2020). This is despite that they often constitute the most important fishing ground and support communities with food and income (de la Torre-Castro and Rönnbäck 2004, Cinner et al. 2013). Given that different seagrasses generate different types of ecosystem services, e.g. large and complex seagrasses support more fish (Gullström et al. 2008) and store more carbon (Duarte and Chiscano 1999, Gullström et al. 2018) than smaller, seagrass beds that are particularly important should actively be incorporated in MPAs.

46 Future perspectives

In general we know much less about seagrass ecosystems compared to coral ecosystems, and much remains to be discovered (Orth et al. 2006, Wells et al. 2007, Duarte et al. 2008). This thesis (Paper I-IV) has demonstrated that MPAs can constitute an effective management tool to conserve seagrass ecosystems. However, if seagrasses should be included in protected areas, knowledge about seagrasses and their associated communities needs to be improved, so that conservation efforts can be as effective as possible. For example, T. ciliatum is known to be more heavily impacted by sea urchin grazing than e.g. T. hemprichii (Alcoverro and Mariani 2002). This is also supported by the results in Paper I and III, showing that T. ciliatum is much more common in old MPAs. Since T. ciliatum also seems to harbor more fish than smaller seagrasses (Gullström et al. 2008) and generate a wider range of ecosystem services (Nordlund et al. 2016), this suggests that T. ciliatum should (at least in theory) be of conservation priority. However, whether these effects are general is more or less unknown, since there are so few studies on seagrass ecosystems and their services, especially in East Africa. Therefore, I would suggest studies that assess if MPAs that favor more structurally complex seagrasses also generate more ecosystem services, e.g. harbor more diverse fish communities and store more carbon. In general, there is a need to assess more in detail what types of ecological functions and processes, and ecosystem services different seagrass species regulate and deliver. Likewise, the analyses of trait composition (Paper III) demonstrated that seagrasses within the old MPAs have seagrasses with shorter leaves and higher shoot density (the latter induced by plasticity) than seagrasses within fished areas and community MPAs. In theory this is likely to have consequences for associated organisms, like fish communities, and in turn the ecological functions that these fish communities regulate, like herbivory. These hypotheses could be explored in future studies.

Additionally, the results in Paper II show that fish protection had larger effect on potential monetary value of fish, compared to biomass and size. Therefore, from a fisheries perspective it is important to assess if the value of fish catches migrating out of MPAs (spillover) also increase more than biomass and size of fish over time, and if spillover is sufficient to compensate for reduced catches caused by loss of fishing grounds.

The lack of (or very weak) protection effects on shallow intertidal seagrass areas and the fact that much of the variation in seagrass species and trait composition were unexplained (Paper III), indicate that further studies of MPA effects on seagrasses are needed. Tourism and agriculture are widespread along the Kenyan coast, and in other tropical coastal areas worldwide, and can negatively impact coastal habitats like seagrasses (Orth et al. 2006, Waycott et al. 2009). Better understanding of how (and if) tourism activities and land- use (e.g. agriculture) affect water quality and coastal habitats in Kenya is needed to provide opportunities for adaptive and more effective management. A previous study from the Philippines suggests that land-use seems to be a better predictor of tropical seagrass condition than MPA presence (Quiros et al. 2017). Nutrient and sediment stresses originating from deforestation, agriculture and coastal development are not only detrimental to seagrasses, but also coral reef ecosystems (e.g. review by Bartley et al. 2014). Integrated coastal zone management have proven successful elsewhere; managing wa- tersheds in the USA have reduced runoff of nutrients (Lefcheck et al. 2018), and leaving undisturbed zones around rivers in cultivated landscapes have limited sediments to reach the ocean in North Europe (Borum et al. 2004). By identifying management practices that have minimal pollutant export rates and by having effective incentives for the adoption of these practices has proven to successfully restore coral reefs (Kroon et al. 2014). These results are promising, and studies that assess if land-based management can minimize disturbances deriving from outside borders of Kenyan MPAs are encouraging. I also suggest there is a need for future studies designed to test whether management of land-use together with protected areas can have synergistic effects on seagrass and coral communities.

There is also a need to continue to evaluate how and to what extent MPAs can increase resilience of seagrass communities. Paper IV is, to my knowledge, the first study that tests if MPAs can affect seagrass recovery from disturbance and it is also the first experimental test of how protection affect recovery in general. The results suggests that protection can speed up recovery of seagrass and decrease erosion (Paper IV). Since the study is only based on one protected area, it is not clear how general these results are and if they can also generate the same protection effects on other seagrass species than T. hemprichii. Based on the experiment in Paper IV, I have four recommenda- tions for future similar experiments: i) experiments should run much longer (up to seven years; see Furman et al. 2019), since some of the experimental plots had not yet recovered after two years, ii) use more replicates in terms of several MPAs, to be able to draw more general conclusions about protection effects, iii) test effects on seagrasses with different types of life-histories, e.g. pioneering, opportunistic and climax seagrass species, since their strategies differ in their ability to recover (Kilminster et al. 2015), and iiii) if the purpose

48 is to test effects of different size of disturbances and/or to evaluate disturbance size thresholds, clearing sizes should be more different from each other (e.g. 1 m2 and 4 m2) and/or use several different sizes. Additionally, previous studies from coral reefs suggest that protection can decrease resilience among coral communities, by selecting for more sensitive taxa (Darling et al. 2010, 2013). If T. ciliatum that benefits from protection (Paper I, III) is less re- sistant to disturbances, e.g. warming, this suggests that protection can decrease seagrass resilience, just as shown in coral reef studies. However, resilience – as in resistance to and recovery from disturbances – of seagrasses needs to be studied much more in detail.

49 Populärvetenskaplig sammanfattning (svenska)

Tropiska sjögräsängar och korallrev är några av de mest produktiva och artrika ekosystemen på jorden. Dessa ekosystem förser människor med ekosystemtjänster, såsom att skydda våra kuster från stormar, och förser också miljontals människor med fisk som är viktiga för deras livsuppehälle. Samtidigt hotas sjögräsängar och korallrev världen över av ökad mänsklig påverkan, främst överfiske, föroreningar och global uppvärmning. Att inrätta områdesskydd, t.ex. marina reservat eller fiskefredade områden, är en nyckelstrategi för att bekämpa många av dessa hot. Marina reservat har visat sig återställa den biologiska mångfalden och skydda korallrev och dess fiskar. Forskning på effekter av marina reservat har främst bedrivits på korallrev och det finns därför ett underskott av studier som undersöker hur marina reservat påverkar sjögräs och organismer som lever i sjögräsängarna. Dessutom ägs och förvaltas de flesta marina reservat av staten och kritiseras därför alltmer eftersom de utesluter och marginaliserar lokala samhällen som oftast är högst beroende av fisket som källa till inkomst och mat för dagen. Därför kan marina reservat som förvaltas och ägs av lokalbefolkningen utgöra ett lovande alternativ, bland annat eftersom de har mycket högre acceptans än statligt ägda reservat. Dock är studier av dessa typer av marina reservats effekter få och mer forskning behövs för att undersöka om och hur lokalt förvaltade reservat påverkar korallrev och sjögräsängar.

Syftet med denna avhandling är att undersöka ekologiska effekter av statligt och lokalt förvaltade marina reservat på sjögräs, koraller och de fisksamhällen som lever i grunda tropiska havslandskap. Därför undersöktes korallrev och sjögräsängar i Kenya i tre olika typer av områden; fiskade områden, nyligen etablerade lokalt förvaltade marina reservat (1-6 års skydd från fiske) och äldre statligt förvaltade marina reservat (20-44 års skydd). Resultaten från fältstudierna visade att endast några få år (≤6) av skydd i lokalt förvaltade marina reservat verkar öka den biologiska mångfalden i korall- och sjögrässamhällen (Studie I) och skydda värdefulla fiskbestånd (Studie II). Områdesskydd tycktes också påverka artsammansättningen, från dominans av snabbväxande, morfologiskt enkla och stresstoleranta arter i fiskade områden och nyligen etablerade reservat, till morfologiskt komplexa och stresskänsliga arter i gamla statligt förvaltade marina reservat (Studie I). Resultaten visade

50 också att sjögrässamhällen i marina reservat har högre skottäthet än sjögräs i fiskade områden och att denna effekt var främst indirekt och orsakad av förändringar i artsammansätting, men också direkt genom förändring av sjögräsarternas individuella morfologi (så kallad fenotypisk plasticitet). Effekterna av marina reservat på artsammansättning och sjögräsegenskaper var starkast i mitten av lagunen, svag på korallrevet och inte alls synbar i de grunda intertidala sjögräsängarna (Studie III). Troligen gynnas inte grunda sjögräsängar av det marina skyddet eftersom det samtidigt är mycket höga nivåer av turism-relaterade aktiviteter och störningar (t.ex. människor som trampar på sjögräsen och ankring av båtar). Resultaten från ett tvåårigt sjögräsexperiment tydde på att ett lokalt förvaltat marint reservat kan påskynda återhämtning av sjögräs och minska erosion av sediment efter experimentell rensning av sjögräs (Studie IV). Minskade nivåer av mänskliga störningar, främst från fiskeaktiviteter, är troligen orsaken till snabbare återhämtning av sjögräset i reservatet, liksom att marina reservat överlag hade mer artrika och strukturellt komplexa arter.

Baserat på dessa resultat drar jag tre slutsatser. För det första pekar resultaten på att marina reservat skyddar tropiska sjögräs på liknande sätt som koraller, vilket tyder på att områdesskydd inte bara är en effektiv strategi för att på lokal nivå skydda koraller, utan också sjögräs. Därför bör sjögräsängar, precis som korallrev, ingå i områden där marina reservat planeras. För det andra, även om de nyligen inrättade lokalt förvaltade reservaten inte var lika effektiva som de äldre statligt förvaltade reservaten gynnade de både sjögräs- och korallsamhällen (Studie I, II, IV). Med tanke på att tidigare studier visar att lokalt förvaltade reservat också kan uppfylla socio-ekonomiska värden, till exempel traditioner, bestämmanderätt och engagemang i förvaltning, är de ett lovande komplement till statligt förvaltade marina reservat. För det tredje, även om marina reservat är ett effektivt verktyg för att skydda sjögräs i mitten av lagunen och koraller på reven, verkar de inte kunna skydda de grunda sjögräsängarna (Studie III), vilket belyser behovet av förbättrade förvaltningsstrategier och strategier av mer holistisk karaktär, till exempel integrerad kustzonförvaltning, där reglering av både fiskeaktiviteter och turism förekommer.

51 Muhtasari (Kiswahili)

ATHARI ZA TENGEFU ZINAZOSIMAMIWA NA JAMII YA WENYEJI NA ZILE ZINAZOSIMAMIWA NA SERIKALI KWA JAMII ZA NYASI ZA BAHARINI NA MATUMBAWE KWENYE MAENEO YA JOTO

Nyasi za baharini kwenye vindimbwi na miamba ya matumbawe katika maneo ya joto ni kati ya mazingira yenye natija duniani. Mfumo huu wa mazingira hutoa huduma kama vile, uzalishashiji wa samaki, uhifadhi wa hewa ya kaboni ya kinga ya pwani ambayo inaweza kupunguza mabadiliko ya hali hewa na kuongeaza msaada kwa maisha ya mamilioni ya watu. Wakati huo huo mifumi hii ya mazingira imetishiwa (imehatarishwa) ulimwenguni kwa usumbufu unaosababishwa na binadamu kama vile uvuvi wa kukithiri, uchafuzi wa mazingira na ongezeko la joto dunian. Utekelezaji wa tengefu baharini (maeneo yalio tengwa/MPAs) ni zana nzuri/bora ya uhifadhi kupambana na vitisho hivi vingi. Tengefu (maeneo yaliyo tengwa/MPAs) ambazo zinakataza uchimbaji wa rasilimali imethibitishwa kurudisha bianuwai (diversity) na hifadhi ya samaki angalau kwenye miamba ya matumbawe. Walakini (ingawaje) tafiti zinazo tathmini (kisia) athari za kinga kwenye jamii ya nyasi za baharini ni adimu/kidogo hasa katika sehemu ama maeneo ya joto. Kwa kuongezea tengefu nyingi (maeneo yalio tengwa/MPAs) yanasimamiwa na serikali kwahivyo yanazidi kukosolewa ama kushutumiwa kwa kuzitenga jamii za wenyeji ambazo sana sana zatengemea uvuvi kwa chanzo cha chakula, mapato ya kiuchumi na maadili ya kitamaduni. Kwahivyo Tengefu (MPAs) ambazo zasimamiwa na jamii za wenyeji wenyewe zinaunda njia mmbadala ya kukidhi (kukimu) lakini bado hazijafanyiwa utafiti vizuri. Kusudi la dhamira hii (lengo hili) likuwa nikuchunguza athari za ikologia (ecological effects) za nyasi za baharini na matumbawe na jamii zao za senthic na samaki kwenya mazingira ya joto katika tengefu zinazosimamiwa na serikali na zile za jamii ya wenyeji. Tulikagua jamii za matumbawe na nyasi za baharini kweny maeneo yanayo vuliwa (fished areas) na maeneo ya tengefu (MPAs) zinazosimamiwa na jamii (miaka 1-6 ya ulizi) na vile vile kwenye tengefu (MPAs) za zamani zinzosimamiwa na serikali (miaka 20-44) katika pwani ya Kenya, Africa mashariki.

Matokeo yanaonyesha kwamba kwa uchache tu (miaka ≤ 6) ya ulinzi katika tengefu (MPAs) zinasimamiwa na jamii zinaweza kuongeza utofauti

52 (diversity) wa jamii za nyasi za baharini na matumbawe (karatasi I), na pia kulinda uchumi na ikolojia ya uhifadhi wa samaki wenye thamani (karatasi II). Ulinzi pia ama uhifadhi ulionekana kushawishi mabadiliko ya jamii za senthic kutoka kwa uwezaji/utawala upainia (pioneering), mkazo-uvumilivu (dhiki) na muundo rahisi ya nyasi za baharini na aina za matumbawe katika maeneo yanayo vuliwa hadi kwa kilele cha aina ngumu ya muundo katika tengefu (MPAs) za zamani zinazosimamiwa na serakali (karatasi I). Kwa kuongezea athari za ulinzi kwenye jamii za nyasi za baharini zinaonekana kati kati ya dimbwi (mid-lagoon) munamopendelewa na aina ya nyasi zenye miani mingi (shoot density). Athari za nyasi za baharini maeneo ya tengefu (MPAs) zenye sifa ya (msongamano wa wiani, (high shoot density), idadi ya majani kwa miani (shoot), biomasia, urefu na upana wa majani) vilikuwa visivyo moja kwa majo kusababishwa na mabadiliko katika muundo wa aina ya nyasi badala ya ujanibishaji wa phenotypic (mabadiliko ya mimea katika kukabiliana na mazingira). Wakati huo huo athari katika eneo lenye kina kifupi ama lenye maji machache na maeneo ya miamba zilikuwa dhaifu au hazikuwepo (karatasi III). Mwishoe, majaribio ya utafiti ya miaka miwili yanaonyesha kwamba (kufungwa kwa tengefu ya kuruwitu) (Kuruwitu closure) inayosimamiwa na jamii ya wenyeji, kunaharakisha kufufua nyasi za baharini na kupunguza mmomonyoko wa ardhi kufuatia usumbufu wa majaribio (kusafisha ambako nyasi ya baharini iliondolewa). Athari za kinga hii ina uwezikano mkubwa kwa sababu ya kupunguzwa zaidi usumbufu (k.v uvuvi na kutembea) kwenye mimea ya kupona na kwenye mchanga (karatasi IV).

Kulingana na matokeo haya mimi nafanya mahitimisho matatu. Kwanza tengefu (marine protected areas – MPAs) zinaonyesha kulinda nyasi za baharini kwa njia ile ile kama zinavyo linda matumbawe, ikionyesha kuwa tengefu (MPAs) inaweza kusaidia uhifadhi was nyasi za baharini. Vidimbwi (seagrass beds) vya nyasi za baharini kwhiyo vinapaswa kujumuishwa katika upangaji wa mazingira ya baharini. Pili, hata kama tengefu (MPAs) zinazosimamiwa na jamii zilianzishwa hivi karibuni hazikuwa nzuri kama zile tengefu (MPAs) za zamani zinazosimamiwa na serikali (labda kwa sababu ni tengefu changa na kupona kunachukua wakati) zinanufaisha nyasi za baharini na jamii za matumbawe na samaki ambao wanaishi maeneo haya (karatasi I, II, IV). Kwa kuzingatia kwamba tafiti zilizopita zinaonyesha kuwa tengefu zinaweza kutimiza viwango vya maadili ya jamii na uchumi (k.m kuhusika katika muundo na utekelezaji wa tenfefu za (MPAs), Matokeo yetu yanasisitiza uwezo wao kama pongezi kwa tengefu zinazo simamiwa na serikali. Tatu na mwisho, tengefu (MPAs) ni zana madhubuti ya kulinda nyasi za baharini na jamii ya matumbawe kutokana na usumbufu wa uchimbaji rasilimali wa eneo hilo haswa katika maeneo ya mwambao/ufuoni (beach) na maeneo au sehemu ya mwamba, lakini haionekani kulinda vidimbwi/vitanda vya nyasi za baharini (seagrass beds) kwenye maji kidogo (shallow intertidal)

53 (karatasi III). Sababu ya hali hii haikowazi (haieleweki), lakini inaweza kuwa ni kwa sababu ya kilimo na shughuli nyingi za utalii (K.v kutembe, maji taka kutoka kwa mahoteli/mikahawa) katika tengefu (MPAs) zinazosimamiwa na serikali. Hii inaonyesha umuhimu wa tathmini zaidi ya tengefu (MPAs), lakini pia hitaji la njia kamili za uhifadhi, kama usimamizi wa ukanda wa pwani uliojumuishwa.

54 Acknowledgements

First of all, thank you Johan Eklöf. I am grateful for your supervision and hours of discussions about science, experiments, fieldwork and writing. Thank you for sharing your experiences and wisdom, and for all the support during these years. I´ve learnt so much! I am truly grateful that you gave me the opportunity to do a PhD. Thank you for believing in me. Thanks also to my co- supervisor Nils Kautsky for your supervision, wealth of experience, scientific enthusiasm and all the floorball games!

My gratitude also goes to Tim McClanahan and Caroline Abunge at the Wild- life Conservation Society (Global Marine Programs, Mombasa). Thank you for sharing your expertise, technical and logistical support, scientific knowledge, inspiration, kind reception and for always making me feel wel- come in Kenya.

Masudi Juma Zamu, I am sincerely grateful for excellent field assistance and faithful, fun and exhausting sampling, swimming, walking, experimenting, hours of boat and car rides, and everything else associated with field work. I am happy to be your friend! Thank you also for writing the Swahili summary in this thesis. Nime shukuru.

I would also like to thank fishing communities and tengefu committees in Chamjale, Kanamai, Mayungu, Nyali, Kibuyuni, Wasini and Kuruwitu for kind reception and for allowing us to access and sample in the fishing grounds and closures. Furthermore, I extend my sincere gratitude to the fishers, con- servationists and colleagues James, Marsha, Katana, Ali, Dickson, Charo, John, Masudi, Shukuru, Hassan, Jimmy, Juma, Harun, Michael, Collins and Hussein. I am truly grateful for your time, cooperation, interesting discussions, and support in the field. I´ve learnt a lot from you and this work could not have been done without you. I really hope to see you again soon!

I express my deepest appreciation to colleagues at Kenya Marine and Fisheries Research Institute (KMFRI), and particularly Jacqueline Nduku Uku and Samuel Ndirangu for logistic and technical support during my field studies in Kenya. Additionally, to the KMFRI lab technicians – Ann, Eve, Purity, Ever- lyne, Kimanthi, Waithera, Winnie, Kilonzi, Anna, Kosieny, Christine, Philip,

55 Doreen, Lenjo, Muthama and Paul – for excellent lab work. I also thank the Kenya Wildlife Service for permission to access the marine parks, and the East African Wildlife Society for technical, logistical and fieldwork support.

I am sincerely grateful to the Department of Ecology, Environment and Plant Sciences (DEEP), at Stockholm University, particularly Ove Eriksson and Siw Hedin for supporting me during my leave and PhD studies. Thank you, I am DEEPly grateful. Thank you Charlotte Berkström and Kristoffer Hylander for careful revision and scientific input to my thesis. Thank you to former and current DEEP group members Serena D, Tiina S and Benjamin J, and particularly Elisa AA (for being an excellent coauthor and for always being available with clever scientific and statistical input), and Åsa A – I am really grateful to have you as a friend. To my friends and colleagues at DEEP – particularly Linda E, Sara F, Maria E, Sigi WH, Nadja S, Stina T and Josefine S – thank you for all great times at conferences and at DEEP. Thank you to the master students Amanda FG and Michael P. Thank you Amanda GB, Pil R and Lina MN in the DEEP communication group. It was an amazing time filled with enthusiasm and inspirational talks. I was always thriving after our meetings!

My gratitude also extends to Emily Darling. Thank you for supervising me during my MSc studies in Kenya, inspiring me to become a scientist and for introducing me to the tropical seascape (totally blew me away!). You are a role model in marine sciences.

To my friends Seth and Rodgers. Thank you for fun times and for making me feel at home in Kenya!

To the Chiricos’. Thank you for all your love and support during the years and especially Janne for guiding me within the research community and for supporting me during my MSc and PhD application.

To my excellent biology teacher in high school Lennart Risberg who gave me the opportunity to go to the Swedish West coast for my first marine course. Thank you for introducing me to the magical world of marine biology!

Finally, I would like to thank my family. Mum, thank you for playfully introducing me to biology and ecology when I grew up. I wouldn’t be here if it wasn’t for you. To my sisters Cecilia and Emilia, and my brother Jonathan. Thank you for making me laugh and feel loved, and for always being there for me and believing in me. It is an indulgence to share my life with you, in good and in bad times. Thank you to my grandmas Lina and Ninni, you are inspiring human beings and have meant more to me than you can ever imagine. To my extended family Erik, Emma, Jens, Lena N, Stina and Siri, and especially Molle, thank you for all the fun, support and love.

56 Financial support

This work was financially supported by Sida (the Swedish International De- velopment Cooperation Agency), the Swedish Research Council VR (Uforsk grants SW-2009-018 and SWE-2012-086 to Johan Eklöf) and C. F. Liljevalch J:ors scholarship.

57 References

Abal, E. G., N. Loneragan, P. Bowen, C. J. Perry, J. W. Udy, and W. C. Dennison. 1994. Physiological and morphological responses of the seagrass Zostera capricorni Aschers, to light intensity. Journal of Experimental Marine Biology and Ecology 178:113-129. Abunge, C., S. Coulthard, and T. M. Daw. 2013. Connecting Marine Ecosystem Services to Human Well-being: Insights from Participatory Well-being Assessment in Kenya. Ambio 42:1010- 1021. Aburto-Oropeza, O., B. Erisman, G. R. Galland, I. Mascarenas-Osorio, E. Sala, and E. Ezcurra. 2011. Large recovery of fish biomass in a no- take marine reserve. PLoS One 6:e23601. Adams, V. M., M. Mills, S. D. Jupiter, and R. L. Pressey. 2011. Improving social acceptability of marine protected area networks: A method for estimating opportunity costs to multiple gear types in both fished and currently unfished areas. Biological Conservation 144:350-361. Agardy, T., G. N. di Sciara, and P. Christie. 2011. Mind the gap: addressing the shortcomings of marine protected areas through large scale marine spatial planning. Marine Policy 35:226-232. Albert, C. H., W. Thuiller, N. G. Yoccoz, R. Douzet, S. Aubert, and S. Lavorel. 2010. A multi-trait approach reveals the structure and the relative importance of intra- vs. interspecific variability in plant traits. Functional Ecology 24:1192-1201. Alcoverro, T., and S. Mariani. 2002. Effects of sea urchin grazing on seagrass (Thalassodendron ciliatum) beds of a Kenyan lagoon. Mar Ecol Prog Ser 226:255-263. Alcoverro, T., and S. Mariani. 2004. Patterns of fish and sea urchin grazing on tropical Indo-Pacific seagrass beds. Ecography 27:361-365. Alongi, D. M. 2002. Present state and future of the world's mangrove forests. Environmental Conservation 29. Alonso Aller, E. 2018. Effects of Marine Protected Areas on Tropical Seagrass Ecosystems. Doctoral thesis, comprehensive summary. Department of Ecology, Environment and Plant Sciences, Stockholm University, Stockholm. Alonso Aller, E., J. S. Eklöf, M. Gullström, U. Kloiber, H. W. Linderholm, and L. M. Nordlund. 2019. Temporal variability of a protected

58 multispecific tropical seagrass meadow in response to environmental change. Environmental Monitoring and Assessment 191:774. Alonso Aller, E., M. Gullström, F. K. J. Eveleens Maarse, M. Gren, L. M. Nordlund, N. Jiddawi, and J. S. Eklöf. 2014. Single and joint effects of regional- and local-scale variables on tropical seagrass fish assemblages. Marine Biology 161:2395-2405. Alonso Aller, E., N. S. Jiddawi, and J. S. Eklöf. 2017. Marine protected areas increase temporal stability of community structure, but not density or diversity, of tropical seagrass fish communities. PLoS One 12:e0183999. Alvarez-Filip, L., N. K. Dulvy, J. A. Gill, I. M. Cote, and A. R. Watkinson. 2009. Flattening of Caribbean coral reefs: region-wide declines in architectural complexity. Proc Biol Sci 276:3019-3025. Anderson, M. J., R. N. Gorley, and K. R. Clarke. 2008. PERMANOVA+ for PRIMER: Guide to software and statistical methods. PRIMER-E: Plymouth, UK. Appolloni, L., R. Sandulli, G. Vetrano, and G. F. Russo. 2018. Assessing the effects of habitat patches ensuring propagule supply and different costs inclusion in marine spatial planning through multivariate analyses. J Environ Manage 214:45-55. Arkema, K. K., G. Guannel, G. Verutes, S. A. Wood, A. Guerry, M. Ruckelshaus, P. Kareiva, M. Lacayo, and J. M. Silver. 2013. Coastal habitats shield people and property from sea-level rise and storms. Nature Climate Change 3:913-918. Auger, S., and B. Shipley. 2013. Inter-specific and intra-specific trait variation along short environmental gradients in an old-growth temperate forest. Journal of Vegetation Science 24:419-428. Austin, T., A. Douglas, A. Edwards, N. Galal, J. P. Hawkins, P. Hogarth, C. Mees, R. Ormond, C. M. Roberts, T. van´t Hof, M. Watson, S. Wells, and A. T. White. 1997. The exploitation of coral reefs. Pages 571-571 in R. Ormond and A. Douglas, editors. Journal of the Marine Biological Association of the United Kingdom. Cambridge University Press, London. Babcock, R. C., N. T. Shears, A. C. Alcala, N. S. Barrett, G. J. Edgar, K. D. Lafferty, T. R. McClanahan, and G. R. Russ. 2010. Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proceedings of the National Academy of Sciences 107:18256-18261. Baden, S., A. Emanuelsson, L. Pihl, C. J. Svensson, and P. Åberg. 2012. Shift in seagrass food web structure over decades is linked to overfishing. Mar Ecol Prog Ser 451:61-73. Bak, R. P. M. 1994. Sea urchin bioerosion on coral reefs: place in the carbonate budget and relevant variables. Coral Reefs 13:99-103.

59 Barbier, B. E., D. S. Hacker, C. Kennedy, W. E. Koch, C. A. Stier, and R. B. Silliman. 2011. The value of estuarine and coastal ecosystem services. Ecological Monographs 81:169-193. Barnett, L. A. K., N. S. Jacobsen, J. T. Thorson, and J. M. Cope. 2019. Realizing the potential of trait-based approaches to advance fisheries science. Fish and Fisheries 20:1034-1050. Bartley, R., Z. T. Bainbridge, S. E. Lewis, F. J. Kroon, S. N. Wilkinson, J. E. Brodie, and D. M. Silburn. 2014. Relating sediment impacts on coral reefs to watershed sources, processes and management: A review. Science of The Total Environment 468-469:1138-1153. Bates, A. E., R. S. C. Cooke, M. I. Duncan, G. J. Edgar, J. F. Bruno, L. Benedetti-Cecchi, I. M. Côté, J. S. Lefcheck, M. J. Costello, N. Barrett, T. J. Bird, P. B. Fenberg, and R. D. Stuart-Smith. 2019. Climate resilience in marine protected areas and the ‘Protection Paradox’. Biological Conservation 236:305-314. Bell, J. D., G. J. S. Craik, D. A. Pollard, and B. C. Russell. 1985. Estimating length frequency distributions of large reef fish underwater. Coral Reefs 4:41-44. Beukers, J. S., and G. P. Jones. 1998. Habitat complexity modifies the impact of piscivores on a coral reef fish population. Oecologia 114:50-59. Bohnsack, A., James. 1998. Application of marine reserves to reef fisheries management. Australian Journal of Ecology 23:298-304. Bollen, K., and J. Pearl. 2013. Eight Myths About Causality and Structural Equation Models. Pages 301-328. Bortone, S. A., J. J. Kimmel, and C. M. Bundrick. 1989. A comparison of three methods for visually assessing reef fish communities: time and area compensated. Northeast Gulf Science 10:85-96. Borum, J., C. M. Duarte, D. Krause-Jensen, and T. M. Greve. 2004. European seagrasses: an introduction to monitoring and management. EU project Monitoring and Managing of European Seagrasses. Bos, A. R., T. J. Bouma, G. L. J. de Kort, and M. M. van Katwijk. 2007. Ecosystem engineering by annual intertidal seagrass beds: Sediment accretion and modification. Estuarine, Coastal and Shelf Science 74:344-348. Boström, C., and E. Bonsdorff. 2000. Zoobenthic community establishment and habitat complexity‹the importance of seagrass shoot-density, morphology and physical disturbance for faunal recruitment. Marine Ecology Progress Series 205:123-138. Bradshaw, A. D. 1965. Evolutionary Significance of Phenotypic Plasticity in Plants. Pages 115-155 in E. W. Caspari and J. M. Thoday, editors. Advances in Genetics. Academic Press. Brock, V. E. 1954. A Preliminary Report on a Method of Estimating Reef Fish Populations. The Journal of Wildlife Management 18:297-308. Bruno, J. F., K. E. Boyer, J. E. Duffy, S. C. Lee, and J. S. Kertesz. 2005. Effects of macroalgal species identity and richness on primary

60 production in benthic marine communities. Ecology Letters 8:1165- 1174. Bulthuis, D. A. 1983. Effects of in situ light reduction on density and growth of the seagrass Heterozostera tasmanica (Martens ex Aschers.) den Hartog in Western Port, Victoria, Australia. Journal of Experimental Marine Biology and Ecology 67:91-103. Burke, L., Y. Kura, K. Kassem, C. Revenga, M. Spalding, and D. McAllister. 2001. Pilot analysis of global ecosystems - Coastal ecosystems. World Resources Institute. Washington, DC. Cesar, H. 2011. Coral Reefs: Their Functions, Threats and Economic Value. Cheal, A., A. Macneil, E. Cripps, M. Emslie, M. Jonker, B. Schaffelke, and H. Sweatman. 2010. Coral-macroalgal phase shifts or reef resilience: Links with diversity and functional roles of herbivorous fishes on the Great Barrier Reef. Coral Reefs 29:1005-1015. Christie, P. 2004. MPAs as biological successes and social failures in Southeast Asia. Pages 155-164 in J. B. Shipley, editor. Aquatic protected areas as fisheries management tools: design, use, and evaluation of these fully protected areas. American Fisheries Society, Bethesda, MD, USA. Cinner, J., T. McClanahan, A. Wamukota, E. Darling, A. Humphries, C. Hicks, C. Huchery, N. Marshall, T. Hempson, N. Graham, Ö. Bodin, T. Daw, and E. Allison. 2013. Social-ecological vulnerability of coral reef fisheries to climatic shocks., FAO Fisheries and Aquaculture Rome, FAO. . Cinner, J. E., C. Folke, T. Daw, and C. C. Hicks. 2011. Responding to change: Using scenarios to understand how socioeconomic factors may influence amplifying or dampening exploitation feedbacks among Tanzanian fishers. Global Environmental Change 21:7-12. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18:117-143. Claudet, J., C. W. Osenberg, L. Benedetti-Cecchi, P. Domenici, J.-A. García- Charton, Á. Pérez-Ruzafa, F. Badalamenti, J. Bayle-Sempere, A. Brito, F. Bulleri, J.-M. Culioli, M. Dimech, J. M. Falcón, I. Guala, M. Milazzo, J. Sánchez-Meca, P. J. Somerfield, B. Stobart, F. Vandeperre, C. Valle, and S. Planes. 2008. Marine reserves: size and age do matter. Ecology Letters 11:481-489. Clements, C., V. Bonito, R. Grober-Dunsmore, and M. Sobey. 2012. Effects of small, Fijian community-based marine protected areas on exploited reef fishes. Marine Ecology Progress Series 449:233-243. Collier, C. J., and M. Waycott. 2014. Temperature extremes reduce seagrass growth and induce mortality. Marine Pollution Bulletin 83:483-490. Costanza, R., R. d'Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R. V. O'Neill, J. Paruelo, R. G. Raskin, P. Sutton, and M. van den Belt. 1997. The value of the world's ecosystem services and natural capital. Nature 387:253-260.

61 Costanza, R., R. de Groot, P. Sutton, S. van der Ploeg, S. J. Anderson, I. Kubiszewski, S. Farber, and R. K. Turner. 2014. Changes in the global value of ecosystem services. Global Environmental Change 26:152- 158. Coulthard, S. 2012. Can we be both resilient and well, and what choices do people have? Incorporating agency into the resilience debate from a fisheries perspective. Ecology and Society 17. Cullen-Unsworth, C. L., L. M. Nordlund, J. Paddock, S. Baker, J. L. McKenzie, and K. F. R. Unsworth. 2014. Seagrass meadows globally as a coupled social–ecological system: implications for human wellbeing. Marine Pollution Bulletin 83:387-397. da Silva, M. I., N. Hill, H. Shimadzu, M. V. M. A. Soares, and M. Dornelas. 2015. Spillover effects of a community-managed marine reserve. PLoS One 10:e0111774. Daby, D. 2003. Effects of seagrass bed removal for tourism purposes in a Mauritian bay. Environ Pollut 125:313-324. Daily, G. C., T. Söderqvist, S. Aniyar, K. Arrow, P. Dasgupta, P. R. Ehrlich, C. Folke, A. Jansson, B.-O. Jansson, N. Kautsky, S. Levin, J. Lubchenco, K.-G. Mäler, D. Simpson, D. Starrett, D. Tilman, and B. Walker. 2000. The Value of Nature and the Nature of Value. Science 289:395-396. Darling, E. S., L. Alvarez-Filip, T. A. Oliver, T. R. McClanahan, I. M. Cote, and D. Bellwood. 2012. Evaluating life-history strategies of reef corals from species traits. Ecol Lett 15:1378-1386. Darling, E. S., N. A. J. Graham, F. A. Januchowski-Hartley, K. L. Nash, M. S. Pratchett, and S. K. Wilson. 2017. Relationships between structural complexity, coral traits, and reef fish assemblages. Coral Reefs 36:561-575. Darling, E. S., T. R. McClanahan, and I. M. Côté. 2010. Combined effects of two stressors on Kenyan coral reefs are additive or antagonistic, not synergistic. Conserv Lett 3:122-130. Darling, E. S., T. R. McClanahan, and I. M. Côté. 2013. Life histories predict coral community disassembly under multiple stressors. Global Change Biology 19:1930-1940. Darling, E. S., T. R. McClanahan, J. Maina, G. G. Gurney, N. A. J. Graham, F. Januchowski-Hartley, J. E. Cinner, C. Mora, C. C. Hicks, E. Maire, M. Puotinen, W. J. Skirving, M. Adjeroud, G. Ahmadia, R. Arthur, A. G. Bauman, M. Beger, M. L. Berumen, L. Bigot, J. Bouwmeester, A. Brenier, T. C. L. Bridge, E. Brown, S. J. Campbell, S. Cannon, B. Cauvin, C. A. Chen, J. Claudet, V. Denis, S. Donner, Estradivari, N. Fadli, D. A. Feary, D. Fenner, H. Fox, E. C. Franklin, A. Friedlander, J. Gilmour, C. Goiran, J. Guest, J.-P. A. Hobbs, A. S. Hoey, P. Houk, S. Johnson, S. D. Jupiter, M. Kayal, C.-y. Kuo, J. Lamb, M. A. C. Lee, J. Low, N. Muthiga, E. Muttaqin, Y. Nand, K. L. Nash, O. Nedlic, J. M. Pandolfi, S. Pardede, V. Patankar, L. Penin, L. Ribas-Deulofeu, Z.

62 Richards, T. E. Roberts, K. u. S. Rodgers, C. D. M. Safuan, E. Sala, G. Shedrawi, T. M. Sin, P. Smallhorn-West, J. E. Smith, B. Sommer, P. D. Steinberg, M. Sutthacheep, C. H. J. Tan, G. J. Williams, S. Wilson, T. Yeemin, J. F. Bruno, M.-J. Fortin, M. Krkosek, and D. Mouillot. 2019. Social–environmental drivers inform strategic management of coral reefs in the Anthropocene. Nature Ecology & Evolution 3:1341-1350. de la Torre-Castro, M., and P. Rönnbäck. 2004. Links between humans and seagrasses—an example from tropical East Africa. Ocean & Coastal Management 47:361-387. Dearden, P., M. Theberge, and M. Yasué. 2010. Using underwater cameras to assess the effects of snorkeler and SCUBA diver presence on coral reef fish abundance, family richness, and species composition. Environmental Monitoring and Assessment 163:531-538. Diaz, S., J. G. Hodgson, K. Thompson, M. Cabido, J. H. C. Cornelissen, A. Jalili, G. Montserrat-Martí, J. P. Grime, F. Zarrinkamar, Y. Asri, S. R. Band, S. Basconcelo, P. Castro-Díez, G. Funes, B. Hamzehee, M. Khoshnevi, N. Pérez-Harguindeguy, M. C. Pérez-Rontomé, F. A. Shirvany, F. Vendramini, S. Yazdani, R. Abbas-Azimi, A. Bogaard, S. Boustani, M. Charles, M. Dehghan, L. de Torres-Espuny, V. Falczuk, J. Guerrero-Campo, A. Hynd, G. Jones, E. Kowsary, F. Kazemi-Saeed, M. Maestro-Martínez, A. Romo-Díez, S. Shaw, B. Siavash, P. Villar-Salvador, and M. R. Zak. 2004. The plant traits that drive ecosystems: Evidence from three continents. Journal of Vegetation Science 15:295-304. Díaz, S., S. Lavorel, F. de Bello, F. Quétier, K. Grigulis, and T. M. Robson. 2007. Incorporating plant functional diversity effects in ecosystem service assessments. Proceedings of the National Academy of Sciences 104:20684-20689. Dorenbosch, M., M. G. G. Grol, I. Nagelkerken, and G. van der Velde. 2005. Distribution of coral reef fishes along a coral reef-seagrass gradient: edge effects and habitat segregation. Marine Ecology Progress Series 299:277-288. Dorenbosch, M., M. G. G. Grol, I. Nagelkerken, and G. van der Velde. 2006. Different surrounding landscapes may result in different fish assemblages in East African seagrass beds. Hydrobiologia 563:45-60. Duarte, C. M., S. Agusti, E. Barbier, G. L. Britten, J. C. Castilla, J.-P. Gattuso, R. W. Fulweiler, T. P. Hughes, N. Knowlton, C. E. Lovelock, H. K. Lotze, M. Predragovic, E. Poloczanska, C. Roberts, and B. Worm. 2020. Rebuilding marine life. Nature 580:39-51. Duarte, C. M., and C. L. Chiscano. 1999. Seagrass biomass and production: a reassessment. Aquatic Botany 65:159-174. Duarte, C. M., W. C. Dennison, R. J. W. Orth, and T. J. B. Carruthers. 2008. The charisma of coastal ecosystems: addressing the imbalance. Estuaries Coasts 31:233-238.

63 Duarte, C. M., M. A. Hemminga, and N. Marbà. 1996. Growth and population dynamics of Thalassodendron ciliatum in a Kenyan back-reef lagoon. Aquatic Botany 55:1-11. Duarte, C. M., and H. Kirkman. 2001. Methods for the measurement of seagrass abundance and depth distribution. Pages 141–153 in F. T. Short and R. G. Coles, editors. Global seagrass research methods. Elsevier, Amsterdam. Duarte, C. M., N. Marbà, E. Gacia, J. W. Fourqurean, J. Beggins, C. Barrón, and E. T. Apostolaki. 2010. Seagrass community metabolism: Assessing the carbon sink capacity of seagrass meadows. Global Biogeochemical Cycles 24:n/a-n/a. Duarte, C. M., J. J. Middelburg, and N. Caraco. 2005. Major role of marine vegetation on the oceanic carbon cycle. Biogeosciences 2:1-8. Duffy, J. E., J. S. Lefcheck, R. D. Stuart-Smith, S. A. Navarrete, and G. J. Edgar. 2016. Biodiversity enhances reef fish biomass and resistance to climate change. Proc Natl Acad Sci U S A 113:6230-6235. Dumont, C., D. Lau, J. Astudillo, K. Fong, S. T. C. Chak, and J.-W. Qiu. 2013. Coral bioerosion by the sea urchin Diadema setosum in Hong Kong: Susceptibility of different coral species. Journal of Experimental Marine Biology and Ecology 441:71–79. Eakin, C. M. 1996. Where have all the carbonates gone? A model comparison of calcium carbonate budgets before and after the 1982–1983 El Nino at Uva Island in the eastern Pacific. Coral Reefs 15:109-119. Eckrich, C. E., and J. G. Holmquist. 2000. Trampling in a seagrass assemblage direct effects, response of associated fauna, and the role of substrate characteristics. Marine Ecology Progress Series 201:199-209. Edgar, G., R. Stuart‐Smith, T. Willis, S. Kininmonth, S. Baker, S. Banks, N. Barrett, M. Becerro, A. Bernard, J. Berkhout, C. Buxton, S. Campbell, A. Cooper, M. Davey, S. Edgar, G. Försterra, D. Galván, A. Irigoyen, D. Kushner, and R. Thomson. 2014. Global conservation outcomes depend on marine protected areas with five key features. Nature 506:216-220. Edinger, E., and M. Risk. 2000. Reef classification by coral morphology predicts coral reef conservation value. Biological Conservation 92:1- 13. Eklof, J. S., T. van der Heide, S. Donadi, E. M. van der Zee, R. O'Hara, and B. K. Eriksson. 2011. Habitat-mediated facilitation and counteracting ecosystem engineering interactively influence ecosystem responses to disturbance. PLoS One 6:e23229. Eklöf, J. S., M. de la Torre-Castro, M. Gullström, J. Uku, N. Muthiga, T. Lyimo, and S. O. Bandeira. 2008. Sea urchin overgrazing of seagrasses: A review of current knowledge on causes, consequences, and management. Estuarine Coastal Shelf Sci 79:569-580. Eklöf, J. S., S. Fröcklin, A. Lindvall, N. Stadlinger, A. Kimathi, J. N. Uku, and T. R. McClanahan. 2009. How effective are MPAs? Predation

64 control and ‘spill-in effects’ in seagrass–coral reef lagoons under contrasting fishery management. Mar Ecol Prog Ser 384:83-96. El-Hacen, E.-H. M., T. J. Bouma, G. S. Fivash, A. A. Sall, T. Piersma, H. Olff, and L. L. Govers. 2018. Evidence for ‘critical slowing down’ in seagrass: a stress gradient experiment at the southern limit of its range. Scientific Reports 8:17263. Emslie, M. J., M. Logan, D. H. Williamson, A. M. Ayling, M. A. MacNeil, D. Ceccarelli, A. J. Cheal, R. D. Evans, K. A. Johns, M. J. Jonker, I. R. Miller, K. Osborne, G. R. Russ, and H. P. Sweatman. 2015. Expectations and Outcomes of Reserve Network Performance following Re-zoning of the Great Barrier Reef Marine Park. Curr Biol 25:983-992. Evans, L. S., K. Brown, and E. H. Allison. 2011. Factors Influencing Adaptive Marine Governance in a Developing Country Context a Case Study of Southern Kenya. Ecology and Society 16. Fan, Y., J. Chen, G. Shirkey, R. John, S. R. Wu, H. Park, and C. Shao. 2016. Applications of structural equation modeling (SEM) in ecological studies: an updated review. Ecological Processes 5:19. Ferrari, B., N. Raventos, and S. Planes. 2008. Assessing effects of fishing prohibition on Posidonia oceanica seagrass meadows in the Marine Natural Reserve of Cerbère-Banyuls. Aquat Bot 88:295-302. Fleitmann, D., R. B. Dunbar, M. McCulloch, M. Mudelsee, M. Vuille, T. R. McClanahan, J. E. Cole, and S. Eggins. 2007. East African soil erosion recorded in a 300 year old coral colony from Kenya. Geophysical Research Letters 34. Fortunel, C., E. Garnier, R. Joffre, E. Kazakou, H. Quested, K. Grigulis, S. Lavorel, P. Ansquer, H. Castro, P. Cruz, J. DoleŽal, O. Eriksson, H. Freitas, C. Golodets, C. Jouany, J. Kigel, M. Kleyer, V. Lehsten, J. Lepš, T. Meier, R. Pakeman, M. Papadimitriou, V. P. Papanastasis, F. Quétier, M. Robson, M. Sternberg, J.-P. Theau, A. Thébault, and M. Zarovali. 2009. Leaf traits capture the effects of land use changes and climate on litter decomposability of grasslands across Europe. Ecology 90:598-611. Foster, N., S. Box, and P. Mumby. 2008. Competitive effects of macroalgae on fecundity of the reef-building coral Montastraea annularis. Marine Ecology Progress Series 367:143-152. Francis, J., A. Nilsson, and D. Waruinge. 2002. Marine Protected Areas in the Eastern African Region: How Successful Are They? AMBIO: A Journal of the Human Environment 31:503-511. Fraschetti, S., G. Guarnieri, S. Bevilacqua, A. Terlizzi, and F. Boero. 2013. Protection enhances community and habitat stability: Evidence from a Mediterranean marine protected area. PLoS One 8:e81838. Froese, R., and D. Pauly. 2013. World Wide Web Electronic Publication. Furman, B. T., M. Merello, C. P. Shea, W. J. Kenworthy, and M. O. Hall. 2019. Monitoring of physically restored seagrass meadows reveals a

65 slow rate of recovery for Thalassia testudinum. Restoration Ecology 27:421-430. Gagic, V., I. Bartomeus, T. Jonsson, A. Taylor, C. Winqvist, C. Fischer, E. M. Slade, I. Steffan-Dewenter, M. Emmerson, S. G. Potts, T. Tscharntke, W. Weisser, and R. Bommarco. 2015. Functional identity and diversity of animals predict ecosystem functioning better than species-based indices. Proceedings of the Royal Society B: Biological Sciences 282:20142620. Gell, F. R., and C. M. Roberts. 2002. The fishery effects of marine reserves and fishery closures. WWF, Washington (D.C.). Giakoumi, S., J. McGowan, M. Mills, M. Beger, R. H. Bustamante, A. Charles, P. Christie, M. Fox, P. Garcia-Borboroglu, S. Gelcich, P. Guidetti, P. Mackelworth, J. M. Maina, L. McCook, F. Micheli, L. E. Morgan, P. J. Mumby, L. M. Reyes, A. White, K. Grorud-Colvert, and H. P. Possingham. 2018. Revisiting “Success” and “Failure” of Marine Protected Areas: A Conservation Scientist Perspective. Frontiers in Marine Science 5. Glaesel, H. 2000. Community-Level Marine Resource Management and the Spirit Realm in Coastal Kenya. Women in Natural Resources. 21:35- 42. González-Correa, J., J. Bayle-Sempere, P. Sanchez-Jerez, and C. Valle. 2007. Posidonia oceanica meadows are not declining globally. Analysis of population dynamics in marine protected areas of the Mediterranean Sea. Marine Ecology Progress Series 336:111-119. Gouezo, M., Y. Golbuu, K. Fabricius, D. Olsudong, G. Mereb, V. Nestor, E. Wolanski, P. Harrison, and C. Doropoulos. 2019. Drivers of recovery and reassembly of coral reef communities. Proceedings of the Royal Society B-Biological Sciences 286:10. Govan, H. 2009. Status and potential of locally-managed marine areas in the Pacific Island Region: meeting nature conservation and sustainable livelihood targets through wide-spread implementation of LMMAs. University Library of Munich, Germany, MPRA Paper. Grace, J. 2006. Structural Equation Modeling and Natural Systems. Cambridge University Press, Cambridge, UK. Grace, J. B., T. M. Anderson, H. Olff, and S. M. Scheiner. 2010. On the specification of structural equation models for ecological systems. Ecological Monographs 80:67-87. Graham, M. H., B. P. Kinlan, L. D. Druehl, L. E. Garske, and S. Banks. 2007. Deep-water kelp refugia as potential hotspots of tropical marine diversity and productivity. Proceedings of the National Academy of Sciences 104:16576-16580. Graham, N. A., S. Jennings, M. A. MacNeil, D. Mouillot, and S. K. Wilson. 2015. Predicting climate-driven regime shifts versus rebound potential in coral reefs. Nature 518:94-97.

66 Graham, N. A. J., T. D. Ainsworth, A. H. Baird, N. C. Ban, L. K. Bay, J. E. Cinner, D. M. De Freitas, G. Diaz-Pulido, M. Dornelas, S. R. Dunn, P. I. J. Fidelman, S. Foret, T. C. Good, J. Kool, J. Mallela, L. Penin, M. S. Pratchett, and D. H. Williamson. 2011. From microbes to people: tractable benefits of no-take areas for coral reefs. Pages 105- 135 in R. N. Gibson, R. J. A. Atkinson, and J. D. M. Gordon, editors. Oceanography and Marine Biology: An Annual Review, Vol 49. Crc Press-Taylor & Francis Group. Graham, N. A. J., and K. L. Nash. 2012. The importance of structural complexity in coral reef ecosystems. Coral Reefs 32:315-326. Green, E. P., and F. T. Short. 2003. World Atlas of Seagrasses. First edition edition. University of California Press. Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am Nat 111:1169-1194. Grime, J. P., and S. Pierce. 2012. The evolutionary strategies that shape ecosystems. 7th edn. edition. Wiley-Blackwell, Oxford, UK. Grober-Dunsmore, R., T. K. Frazer, J. P. Beets, W. J. Lindberg, P. Zwick, and N. A. Funicelli. 2007. Influence of landscape structure on reef fish assemblages. Landscape Ecology 23:37-53. Gullström, M., M. Bodin, P. G. Nilsson, and M. C. Öhman. 2008. Seagrass structural complexity and landscape configuration as determinants of tropical fish assemblage composition. Mar Ecol Prog Ser 363:241- 255. Gullström, M., M. de la Torre-Castro, S. O. Bandeira, M. Bjork, M. Dahlberg, N. Kautsky, P. Ronnback, and M. C. Ohman. 2002. Seagrass ecosystems in the Western Indian Ocean. Ambio 31:588-596. Gullström, M., L. D. Lyimo, M. Dahl, G. S. Samuelsson, M. Eggertsen, E. Anderberg, L. M. Rasmusson, H. W. Linderholm, A. Knudby, S. Bandeira, L. M. Nordlund, and M. Björk. 2018. Blue Carbon Storage in Tropical Seagrass Meadows Relates to Carbonate Stock Dynamics, Plant–Sediment Processes, and Landscape Context: Insights from the Western Indian Ocean. Ecosystems 21:551-566. Halpern, B. S., S. E. Lester, and J. B. Kellner. 2010. Spillover from marine reserves and the replenishment of fished stocks. Environmental Conservation 36:268-276. Halpern, B. S., and R. R. Warner. 2002. Marine reserves have rapid and lasting effects. Ecology Letters 5:361-366. Halpern, B. S., and R. R. Warner. 2003. Matching marine reserve design to reserve objectives. Proc Biol Sci 270:1871-1878. Harcourt, W. D., R. A. Briers, and M. Huxham. 2018. The thin(ning) green line? Investigating changes in Kenya's seagrass coverage. Biology Letters 14. Harrison, H. B., D. H. Williamson, R. D. Evans, G. R. Almany, S. R. Thorrold, G. R. Russ, K. A. Feldheim, L. van Herwerden, S. Planes, M.

67 Srinivasan, M. L. Berumen, and G. P. Jones. 2012. Larval export from marine reserves and the recruitment benefit for fish and fisheries. Curr Biol 22:1023-1028. Heck, K. L., T. J. B. Carruthers, C. M. Duarte, A. R. Hughes, G. Kendrick, R. J. Orth, and S. W. Williams. 2008. Trophic Transfers from Seagrass Meadows Subsidize Diverse Marine and Terrestrial Consumers. Ecosystems 11:1198-1210. Heck, K. L., and G. S. Wetstone. 1977. Habitat Complexity and Invertebrate Species Richness and Abundance in Tropical Seagrass Meadows. Journal of Biogeography 4:135-142. Hemminga, M. A., and C. Duarte. 2000. Seagrass Ecology. Cambridge University Press, Cambridge. Hicks, C. C., T. R. McClanahan, J. Cinner, and J. M. Hills. 2009. Trade-offs in values assigned to ecological goods and services associated with different coral reef management strategies. Ecol Soc 14(1): 10. Hilborn, R. 2007. Moving to sustainability by learning from successful fisheries. Ambio 36:296-303. Hilborn, R., R. O. Amoroso, C. M. Anderson, J. K. Baum, T. A. Branch, C. Costello, C. L. de Moor, A. Faraj, D. Hively, O. P. Jensen, H. Kurota, L. R. Little, P. Mace, T. McClanahan, M. C. Melnychuk, C. Minto, G. C. Osio, A. M. Parma, M. Pons, S. Segurado, C. S. Szuwalski, J. R. Wilson, and Y. Ye. 2020. Effective fisheries management instrumental in improving fish stock status. Proceedings of the National Academy of Sciences 117:2218-2224. Hilborn, R., K. Stokes, J.-J. Maguire, T. Smith, L. W. Botsford, M. Mangel, J. Orensanz, A. Parma, J. Rice, J. Bell, K. L. Cochrane, S. Garcia, S. J. Hall, G. P. Kirkwood, K. Sainsbury, G. Stefansson, and C. Walters. 2004. When can marine reserves improve fisheries management? Ocean & Coastal Management 47:197-205. Hill, J., and C. Wilkinson. 2004. METHODS FOR ECOLOGICAL MONITORING OF CORAL REEFS. Australian Institute of Marine Science, Townsville, Australia. Hoegh-Guldberg, O. 1999. Climate change, coral bleaching and the future of the world's coral reefs. Marine and Freshwater Research 50. Hoegh-Guldberg, O., P. J. Mumby, A. J. Hooten, R. S. Steneck, P. Greenfield, E. Gomez, C. D. Harvell, P. F. Sale, A. J. Edwards, K. Caldeira, N. Knowlton, C. M. Eakin, R. Iglesias-Prieto, N. Muthiga, R. H. Bradbury, A. Dubi, and M. E. Hatziolos. 2007. Coral Reefs Under Rapid Climate Change and Ocean Acidification. Science 318:1737- 1742. Holden, A. 2016. Environment and Tourism. Taylor and Francis, London, UK. Hoorweg, J., B. Wangila, and A. Degen. 2009. Artisanal fishers on the Kenyan coast: Household livelihoods and Marine resource management. 14:1-160.

68 Hughes, T. P. 1994. Catastrophes, Phase Shifts, and Large-Scale Degradation of a Caribbean Coral Reef. Science 265:1547-1551. Inglis, G. 2000. Variation in the recruitment behaviour of seagrass seeds: Implications for population dynamics and resource management. Pacific Conservation Biology 5:256-259. Integration and Application Network. University of Mary-land Center for Environmental Science. Isbell, F., A. Gonzalez, M. Loreau, J. Cowles, S. Díaz, A. Hector, G. M. Mace, D. A. Wardle, M. I. O'Connor, J. E. Duffy, L. A. Turnbull, P. L. Thompson, and A. Larigauderie. 2017. Linking the influence and dependence of people on biodiversity across scales. Nature 546:65. Jackson, J. B., M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-637. Jiddawi, N. S., and M. C. Öhman. 2002. Marine fisheries in Tanzania. Ambio 31:518-527. Johnson, C. R., and C. A. Field. 1993. Using fixed‐effects model multivariate analysis of variance in marine biology and ecology. Oceanography and Marine Biology - An Annual Review 31:177-221. Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as Ecosystem Engineers. Oikos 69:373-386. Jones, G. P., M. I. McCormick, M. Srinivasan, and J. V. Eagle. 2004. Coral decline threatens fish biodiversity in marine reserves. Proc Natl Acad Sci U S A 101:8251-8253. Jones, J. 1992. Environmental impact of trawling on the seabed: A review. New Zealand Journal of Marine and Freshwater Research - N Z J MAR FRESHWATER RES 26:59-67. Jung, V., C. Albert, C. Violle, G. Kunstler, G. Loucougaray, and T. Spiegelberger. 2013. Intraspecific trait variability mediates the response of subalpine grassland communities to extreme drought events. Kaiser, M., F. Spence, and P. Hart. 2000. Fishing-Gear Restrictions and Conservation of Benthic Habitat Complexity. Conservation Biology - CONSERV BIOL 14:1512-1525. Kenworthy, W. J., S. Wyllie-Echeverria, R. G. Coles, G. Pergent, and C. Pergent-Martini. 2006. Seagrass conservation biology: an interdisciplinary science for protection of the seagrass biome. Pages 595-623 in A. W. D. Larkum, R. J. Orth, and C. M. Duarte, editors. Physiological and morphological responses of the seagrass Zostera capricorni Aschers, to light intensity. Springer, The Netherlands. Kichenin, E., D. A. Wardle, D. A. Peltzer, C. W. Morse, and G. T. Freschet. 2013. Contrasting effects of plant inter- and intraspecific variation on

69 community-level trait measures along an environmental gradient. Functional Ecology 27:1254-1261. Kilminster, K., K. McMahon, M. Waycott, G. A. Kendrick, P. Scanes, L. McKenzie, K. R. O'Brien, M. Lyons, A. Ferguson, P. Maxwell, T. Glasby, and J. Udy. 2015. Unravelling complexity in seagrass systems for management: Australia as a microcosm. Science of The Total Environment 534:97-109. Klein, C., C. Steinback, A. Scholz, and H. Possingham. 2008. Effectiveness of marine reserve networks in representing biodiversity and minimizing impact to fishermen: a comparison of two approaches used in California. Conservation Letters 1. Kroon, F. J., B. Schaffelke, and R. Bartley. 2014. Informing policy to protect coastal coral reefs: Insight from a global review of reducing agricultural pollution to coastal ecosystems. Marine Pollution Bulletin 85:33-41. Lamb, J. B., J. A. J. M. van de Water, D. G. Bourne, C. Altier, M. Y. Hein, E. A. Fiorenza, N. Abu, J. Jompa, and C. D. Harvell. 2017. Seagrass ecosystems reduce exposure to bacterial pathogens of humans, fishes, and invertebrates. Science 355:731-733. Lavorel, S., and E. Garnier. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16:545-556. Lee, K.-S., and K. H. Dunton. 1997. Effect of in situ light reduction on the maintenance, growth and partitioning of carbon resources in Thalassia testudinum banks ex König. Journal of Experimental Marine Biology and Ecology 210:53-73. Lefcheck, J. S. 2016. piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods in Ecology and Evolution 7:573-579. Lefcheck, J. S., R. J. Orth, W. C. Dennison, D. J. Wilcox, R. R. Murphy, J. Keisman, C. Gurbisz, M. Hannam, J. B. Landry, K. A. Moore, C. J. Patrick, J. Testa, D. E. Weller, and R. A. Batiuk. 2018. Long-term nutrient reductions lead to the unprecedented recovery of a temperate coastal region. Proceedings of the National Academy of Sciences 115:3658-3662. Lepš, J., F. de Bello, P. Šmilauer, and J. Doležal. 2011. Community trait response to environment: disentangling species turnover vs intraspecific trait variability effects. Ecography 34:856-863. Lester, S. E., and B. Halpern. 2008. Biological Responses in Marine No-Take Reserves versus Partially Protected Areas. Marine Ecology Progress Series 367:49-56. Lester, S. E., B. S. Halpern, K. Grorud-Colvert, J. Lubchenco, B. I. Ruttenberg, S. D. Gaines, S. Airamé, and R. R. Warner. 2009. Biological effects within no-take marine reserves: a global synthesis. Marine Ecology Progress Series 384:33-46.

70 Lieske, E., and R. Myers. 2002. Coral reef fishes: Indo-Pacific and Caribbean. Princeton University Press, Princeton, NJ. Lilliesköld Sjöö, G., E. Mörk, S. Andersson, and I. Melander. 2011. Differences in top-down and bottom-up regulation of macroalgal communities between a reef crest and back reef habitat in Zanzibar. Estuarine, Coastal and Shelf Science 91:511-518. Littler, M. M., and D. S. Littler. 1980. The Evolution of Thallus Form and Survival Strategies in Benthic Marine Macroalgae: Field and Laboratory Tests of a Functional Form Model. The American Naturalist 116:25-44. Lotze, H. K., H. S. Lenihan, B. J. Bourque, R. H. Bradbury, R. G. Cooke, M. C. Kay, S. M. Kidwell, M. X. Kirby, C. H. Peterson, and J. B. Jackson. 2006. Depletion, degradation, and recovery potential of estuaries and coastal seas. Science 312:1806-1809. Madin, J. S., M. O. Hoogenboom, S. R. Connolly, E. S. Darling, D. S. Falster, D. Huang, S. A. Keith, T. Mizerek, J. M. Pandolfi, H. M. Putnam, and A. H. Baird. 2016. A Trait-Based Approach to Advance Coral Reef Science. Trends in Ecology & Evolution 31:419-428. Mangi, S. C., and C. M. Roberts. 2006. Quantifying the environmental impacts of artisanal fishing gear on Kenya's coral reef ecosystems. Mar Pollut Bull 52:1646-1660. Marbà, N., and C. M. Duarte. 1998. Rhizome elongation and seagrass clonal growth. Marine Ecology Progress Series 174:269-280. Marbà, N., C. M. Duarte, M. Holmer, R. Martínez, G. Basterretxea, A. Orfila, A. Jordi, and J. Tintoré. 2003. Effectiveness of protection of seagrass (Posidonia oceanica) populations in Cabrera National Park (Spain). Environ Conserv 29:509-518. Marshall, P. A. 2000. Skeletal damage in reef corals: relating resistance to colony morphology. Marine Ecology Progress Series 200:177-189. Martínez, M. L., A. Intralawan, G. Vázquez, O. Pérez-Maqueo, P. Sutton, and R. Landgrave. 2007. The coasts of our world: ecological, economic and social importance. Ecol Econ 63:254-272. Maxwell, P. S., K. A. Pitt, D. D. Burfeind, A. D. Olds, R. C. Babcock, and R. M. Connolly. 2014. Phenotypic plasticity promotes persistence following severe events: physiological and morphological responses of seagrass to flooding. Journal of Ecology 102:54-64. McArdle, B. H. 1990. Detecting and displaying impacts of biological monitoring: spatial problems and partial solutions. Pages 249-255 Proceedings of Invited Papers, International Biometrics Conference, IBC, Budapest. McClanahan, T. 1995. A coral reef ecosystem-fisheries model: impacts of fishing intensity and catch selection on reef structure and processes. Ecological Modelling 80:1-19. McClanahan, T. 2000. Recovery of a coral reef keystone pred Balistapus undulatus. Biological Conservation 94:191-198.

71 McClanahan, T., J. Davies, and J. Maina. 2005a. Factors influencing resource users and managers' perceptions towards marine protected area management in Kenya. Environmental Conservation 32:42-49. McClanahan, T., and N. Graham. 2005. Recovery trajectories of coral reef fish assemblages within Kenyan marine protected areas. Marine Ecology Progress Series 294:241-248. McClanahan, T., and S. Mangi. 2000. Spillover of exploitable fishes from a marine park and its effect on the adjacent fishery. Ecological Applications 10:1792–1805. McClanahan, T., and N. Muthiga. 2016. Similar impacts of fishing and environmental stress on calcifying organisms in Indian Ocean coral reefs. Marine Ecology Progress Series 560. McClanahan, T., N. A. Muthiga, and C. A. Abunge. 2016a. Establishment of Community Managed Fisheries’ Closures in Kenya: Early Evolution of the Tengefu Movement. Coastal Management 44:1-20. McClanahan, T. R. 1999. Is there a future for coral reef parks in poor tropical countries? Coral Reefs 18:321-325. McClanahan, T. R. 2010. Effects of fisheries closures and gear restrictions on fishing income in a Kenyan coral reef. Conservation Biology 24:1519-1528. McClanahan, T. R. 2014. Recovery of functional groups and trophic relationships in tropical fisheries closures. Marine Ecology Progress Series 497:13-23. McClanahan, T. R. 2019. Coral reef fish communities, diversity, and their fisheries and biodiversity status in East Africa. Marine Ecology Progress Series 632:175-191. McClanahan, T. R., C. A. Abunge, and J. E. Cinner. 2012. Heterogeneity in fishers' and managers' preferences towards management restrictions and benefits in Kenya. Environmental Conservation 39:357-369. McClanahan, T. R., M. Ateweberhan, and J. Omukoto. 2007a. Long-term changes in coral colony size distributions on Kenyan reefs under different management regimes and across the 1998 bleaching event. Marine Biology 153:755-768. McClanahan, T. R., H. Glaesel, J. Rubens, and R. Kiambo. 1997. The effects of traditional fisheries management on fisheries yields and the coral- reef ecosystems of southern Kenya. Environmental Conservation 24:105-120. McClanahan, T. R., N. A. J. Graham, J. M. Calnan, and M. A. MacNeil. 2007b. Toward pristine biomass: reef fish recovery in coral reef marine protected areas in Kenya. Ecol Appl 17:1055-1067. McClanahan, T. R., N. A. J. Graham, S. K. Wilson, Y. Letourneur, and R. Fisher. 2009. Effects of fisheries closure size, age, and history of compliance on coral reef fish communities in the western Indian Ocean. Marine Ecology Progress Series 396:99-109.

72 McClanahan, T. R., C. C. Hicks, and E. S. Darling. 2008. Malthusian overfishing and efforts to overcome it on Kenyan coral reefs. Ecological Applications 18:1516-1529. McClanahan, T. R., and B. Kaunda-Arara. 1996. Fishery recovery in a coral- reef marine park and its effect on the adjacent fishery. Conservation Biology 10:1187-1199. McClanahan, T. R., J. M. Maina, N. A. J. Graham, and K. R. Jones. 2016b. Modeling reef fish biomass, recovery potential, and management priorities in the Western Indian Ocean. PLoS One 11:e0154585. McClanahan, T. R., M. J. Marnane, J. E. Cinner, and W. E. Kiene. 2006. A comparison of marine protected areas and alternative approaches to coral-reef management. Current Biology 16:1408-1413. McClanahan, T. R., S. Mwaguni, and N. A. Muthiga. 2005b. Management of the Kenyan coast. Ocean & Coastal Management 48:901-931. McClanahan, T. R., M. Nugues, and S. Mwachireya. 1994. Fish and sea urchin herbivory and competition in Kenyan coral reef lagoons: the role of reef management. J Exp Mar Biol Ecol 184:237-254. McClanahan, T. R., and S. H. Shafir. 1990. Causes and consequences of sea urchin abundance and diversity in Kenyan coral reef lagoons. Oecologia 83:362-370. McLeod, E., G. L. Chmura, S. Bouillon, R. Salm, M. Björk, C. M. Duarte, C. E. Lovelock, W. H. Schlesinger, and B. R. Silliman. 2011. A blueprint for blue carbon: toward an improved understanding of the role of vegetated coastal habitats in sequestering CO2. Frontiers in Ecology and the Environment 9:552-560. Mellin, C., M. Aaron MacNeil, A. J. Cheal, M. J. Emslie, and M. Julian Caley. 2016. Marine protected areas increase resilience among coral reef communities. Ecology Letters 19:629-637. Micheli, F., A. Saenz-Arroyo, A. Greenley, L. Vazquez, J. A. Espinoza Montes, M. Rossetto, and G. A. De Leo. 2012. Evidence that marine reserves enhance resilience to climatic impacts. PLoS One 7:e40832. Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well- being: Synthesis. Miner, B. G., S. E. Sultan, S. G. Morgan, D. K. Padilla, and R. A. Relyea. 2005. Ecological consequences of phenotypic plasticity. Trends in Ecology & Evolution 20:685-692. Mitra, A. 2020. Ecosystem Services of Mangroves: An Overview. Pages 1-32 Mangrove Forests in India: Exploring Ecosystem Services. Springer International Publishing, Cham. Moberg, F., and C. Folke. 1999. Ecological goods and services of coral reef ecosystems. Ecological Economics 29:215-233. Moberg, F., and P. Rönnbäck. 2003. Ecosystem services of the tropical seascape. Ocean & Coastal Management 46:27–46. Montefalcone, M., G. Albertelli, C. Morri, V. Parravicini, and C. N. Bianchi. 2009. Legal protection is not enough: Posidonia oceanica meadows in

73 marine protected areas are not healthier than those in unprotected areas of the northwest Mediterranean Sea. Marine Pollution Bulletin 58:515-519. Mumby, P. J., C. P. Dahlgren, A. R. Harborne, C. V. Kappel, F. Micheli, D. R. Brumbaugh, K. E. Holmes, J. M. Mendes, K. Broad, J. N. Sanchirico, K. Buch, S. Box, R. W. Stoffle, and A. B. Gill. 2006. Fishing, trophic cascades, and the process of grazing on coral reefs. Science 311:98-101. Muthiga, N. A. 2009. Evaluating the effectiveness of management of the Malindi–Watamu marine protected area complex in Kenya. Ocean & Coastal Management 52:417-423. Mörk, E., G. L. Sjöö, N. Kautsky, and T. R. McClanahan. 2009. Top–down and bottom–up regulation of macroalgal community structure on a Kenyan reef. Estuarine, Coastal and Shelf Science 84:331-336. Nakamura, Y., and M. Sano. 2004. Overlaps in habitat use of fishes between a seagrass bed and adjacent coral and sand areas at Amitori Bay, Iriomote Island, Japan: importance of the seagrass bed as juvenile habitat. Fisheries Science 70:788-803. Nardi, K., G. P. Jones, M. J. Moran, and Y. W. Cheng. 2004. Contrasting effects of marine protected areas on the abundance of two exploited fishes at the sub-tropical Houtman Abrolhos Islands, Western Australia. Environmental Conservation 31:160-168. National Research Council. 2001. Marine Protected Areas: Tools for Sustaining Ocean Ecosystems. The National Academies Press, Washington, DC. Nemeth, R., S. Herzlieb, and J. Blondeau. 2006. Comparison of two seasonal closures for protecting red hind spawning aggregations in the US Virgin Islands. Nordlund, L. M., E. W. Koch, E. B. Barbier, and J. C. Creed. 2016. Seagrass ecosystem services and their variability across genera and geographical regions. PLoS One 11:e0163091. Nordlund, L. M., R. K. F. Unsworth, M. Gullström, and L. C. Cullen- Unsworth. 2018. Global significance of seagrass fishery activity. Fish and Fisheries 19:399-412. Nyström, M., C. Folke, and F. Moberg. 2000. Coral reef disturbance and resilience in a human-dominated environment. Trends in Ecology & Evolution 15:413-417. O'Brien, K. R., M. Waycott, P. Maxwell, G. A. Kendrick, J. W. Udy, A. J. P. Ferguson, K. Kilminster, P. Scanes, L. J. McKenzie, K. McMahon, M. P. Adams, J. Samper-Villarreal, C. Collier, M. Lyons, P. J. Mumby, L. Radke, M. J. A. Christianen, and W. C. Dennison. 2018. Seagrass ecosystem trajectory depends on the relative timescales of resistance, recovery and disturbance. Marine Pollution Bulletin 134:166-176. Obura, D. O. 2001. Kenya. Marine Pollution Bulletin 42:1264-1278.

74 Ochieng, C., and P. Erftemeijer. 1993. The Seagrasses of Kenya and Tanzania. Pages 82-92. Ochieng, C. A., and P. L. A. Erftemeijer. 1999. Accumulation of seagrass beach cast along the Kenyan coast: a quantitative assessment. Aquatic Botany 65:221-238. Odum, H. T., and E. P. Odum. 1955. Trophic Structure and Productivity of a Windward Coral Reef Community on Eniwetok Atoll. Ecological Monographs 25:291-320. Ogden, J. C. 1988. The influence of adjacent systems on the structure and function of coral reefs. Pages 123-129 in J. H. Choat, D. Barnes, M. Borowitzka, J. C. Coll, P. J. Davies, P. Flood, B. G. Hatcher, D. Hopley, P. A. Hutchings, D. Kinsey, G. R. Orme, M. Pichon, P. F. Sale, P. Sammarco, C. C. Wallace, C. Wilkinson, E. Wolanski, and O. Bellwood, editors. Proceedings of the 6th International Coral Reef Symposium, Townsville, Australia. Ogden, J. C., and E. H. Gladfelter. 1983. Coral reefs, seagrass beds and mangroves: their interactions in the coastal zones of the Caribbean. UNESCO Reports of Marine Science. 23, 133. Ojeda-Martínez, C., J. Bayle-Sempere, P. Sanchez-Jerez, F. Salas, B. Stobart, R. Goñi, J. Falcón, M. Graziano, I. Guala, R. Higgins, F. Vandeperre, L. Ledireach, P. Martín-Sosa, and S. Vaselli. 2011. Review of the effects of protection in marine protected areas: Current knowledge and gaps. Animal Biodiversity and Conservation 34:191-203. Okemwa, G. M. 2017. Fisheries Catch Assessment Surveys (CAS) in coastal Kenya. Olds, A. D., K. A. Pitt, P. S. Maxwell, R. C. Babcock, D. Rissik, and R. M. Connolly. 2014. Marine reserves help coastal ecosystems cope with extreme weather. Global Change Biology 20:3050-3058. Olsen, E., A. R. Kleiven, H. R. Skjoldal, and C. H. von Quillfeldt. 2011. Place- based management at different spatial scales. Journal of Coastal Conservation 15:257-269. Ongoma, V., and O. Onyango. 2014. A Review of the Future of Tourism in Coastal Kenya: The Challenges and Opportunities Posed by Climate Change. Journal of Earth Science & Climatic Change 5. Orth, R. J., T. J. B. Carruthers, W. C. Dennison, C. M. Duarte, J. W. Fourqurean, K. L. Heck, A. R. Hughes, G. A. Kendrick, W. J. Kenworthy, S. Olyarnik, F. T. Short, M. Waycott, and S. L. Williams. 2006. A global crisis for seagrass ecosystems. BioScience 56:987- 996. Pakeman, R., J. Leps, M. Kleyer, S. Lavorel, E. Garnier, E. Kazakou, and C. Fortunel. 2009. Relative climatic, edaphic and management controls of plant functional trait signatures. Journal of Vegetation Science 20:148-159. Parker, J. D., J. Duffy, and R. Orth. 2001. Plant species diversity and composition: Experimental effects on marine epifaunal assemblages.

75 Marine Ecology-progress Series - MAR ECOL-PROGR SER 224:55- 67. Pascal, N. 2011. Cost-Benefit analysis of community-based marine protected areas: 5 case studies in Vanuatu, South Pacific. Petraitis, P. S., and S. R. Dudgeon. 2005. Divergent succession and implications for alternative states on rocky intertidal shores. Journal of Experimental Marine Biology and Ecology 326:14-26. Petraitis, P. S., and R. E. Latham. 1999. THE IMPORTANCE OF SCALE IN TESTING THE ORIGINS OF ALTERNATIVE COMMUNITY STATES. Ecology 80:429-442. Pickett, S. T. A. 1989. Space-for-Time Substitution as an Alternative to Long- Term Studies. Pages 110-135 in G. E. Likens, editor. Long-Term Studies in Ecology: Approaches and Alternatives. Springer New York, New York, NY. Piet, G. J. 1996. On the ecology of a tropical fish community. Piet, S.l. Pinheiro, J., D. Bates, S. DebRoy, D. Sarkar, and R. C. Team. 2018. {nlme}: Linear and Nonlinear Mixed Effects Models. R package version 3.1- 137. URL: https://CRAN.R-project.org/package=nlme. Planes, S., N. Raventos, B. Ferrari, and T. Alcoverro. 2011. Fish herbivory leads to shifts in seagrass Posidonia oceanica investments in sexual reproduction. Marine Ecology Progress Series 431:205-213. Pollnac, R. B., B. R. Crawford, and M. L. G. Gorospe. 2001. Discovering factors that influence the success of community-based marine protected areas in the Visayas, Philippines. Ocean Coastal Manage 44:683-710. Polunin, N. V. C., and C. M. Roberts. 1993. Greater biomass and value of target coral-reef fishes in two small Caribbean marine reserves. Marine Ecology Progress Series 100:167-176. Pomeroy, R., and F. Douvere. 2008. The engagement of stakeholders in the marine spatial planning process. Marine Policy 32:816-822. Potouroglou, M., J. C. Bull, K. W. Krauss, H. A. Kennedy, M. Fusi, D. Daffonchio, M. M. Mangora, M. N. Githaiga, K. Diele, and M. Huxham. 2017. Measuring the role of seagrasses in regulating sediment surface elevation. Scientific Reports 7:11917. Quiros, A. T. E., D. Croll, B. Tershy, M. D. Fortes, and P. Raimondi. 2017. Land use is a better predictor of tropical seagrass condition than marine protection. Biol Conserv 209:454-463. R Core Team. 2017. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. R Core Team. 2019. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.

76 R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. v. 3.4.3. http://www.r-project.org/. Rasher, D. B., and M. E. Hay. 2010. Chemically rich seaweeds poison corals when not controlled by herbivores. Proceedings of the National Academy of Sciences 107:9683-9688. Roberts, C. M., J. A. Bohnsack, F. Gell, J. P. Hawkins, and R. Goodridge. 2001. Effects of marine reserves on adjacent fisheries. Science 294:1920-1923. Roberts, C. M., C. J. McClean, J. E. N. Veron, J. P. Hawkins, G. R. Allen, D. E. McAllister, C. G. Mittermeier, F. W. Schueler, M. Spalding, F. Wells, C. Vynne, and T. B. Werner. 2002. Marine Biodiversity Hotspots and Conservation Priorities for Tropical Reefs. Science 295:1280-1284. Roberts, C. M., B. C. O’Leary, D. J. McCauley, P. M. Cury, C. M. Duarte, J. Lubchenco, D. Pauly, A. Sáenz-Arroyo, U. R. Sumaila, R. W. Wilson, B. Worm, and J. C. Castilla. 2017. Marine reserves can mitigate and promote adaptation to climate change. Proceedings of the National Academy of Sciences 114:6167-6175. Roberts, M. C., and P. J. Hawkins. 2000. Fully-protected marine reserves: A guide. WWF Endangered Seas Campaign, 1250 24th Street, NW, Washington, DC 20037, USA and Environment Department, University of York, York, YO10 5DD, UK. Roberts, M. C., and F. G. R. Ormond. 1987. Habitat complexity and coral reef fish diversity and abundance on Red Sea fringing reefs. Marine Ecology Progress Series 41:1-8. Rogers, A., J. L. Blanchard, and P. J. Mumby. 2014. Vulnerability of coral reef fisheries to a loss of structural complexity. Current Biology 24:1000-1005. Ruiz Fernandez, J. M., and J. Romero. 2004. Effects of disturbances caused by coastal constructions on spatial structure, growth dynamics and photosynthesis of the seagrass Posidonia oceanica. Marine Pollution Bulletin 46:1523-1533. Russ, G. R., and A. C. Alcala. 2004. Marine reserves: long-term protection is required for full recovery of predatory fish populations. Oecologia 138:622-627. Russ, G. R., A. C. Alcala, P. M. Aileen, H. P. Calumpong, and T. W. Alan. 2004. Marine Reserve Benefits Local Fisheries. Ecological Applications 14:597-606. Sale, P. F., and B. J. Sharp. 1983. Correction for bias in visual transect censuses of coral reef fishes. Coral Reefs 2:37-42. Samoilys, M. 1997. Manual for Assessing Fish Stocks on Pacific Coral Reefs.75. Samoilys, M., and G. Carlos. 1991. A survey of reef fish stocks in Western Samoa: application of underwater visual census methods for fisheries

77 personnel. A report prepared for the Forum Fisheries Agency, Honiara, Solomon Islands, and the Fisheries Division, Western Samoa, 26 pp. Samoilys, M. A., and G. Carlos. 2000. Determining methods of underwater visual census for estimating the abundance of coral reef fishes. Environmental Biology of Fishes 57:289-304. Samoilys, M. A., K. M. Martin-Smith, B. G. Giles, B. Cabrera, J. A. Anticamara, E. O. Brunio, and A. C. J. Vincent. 2007. Effectiveness of five small Philippines’ coral reef reserves for fish populations depends on site-specific factors, particularly enforcement history. Biological Conservation 136:584-601. Selig, E. R., D. G. Hole, E. H. Allison, K. K. Arkema, M. C. McKinnon, J. Chu, A. de Sherbinin, B. Fisher, L. Glew, M. B. Holland, J. C. Ingram, N. S. Rao, R. B. Russell, T. Srebotnjak, L. C. L. Teh, S. Troëng, W. R. Turner, and A. Zvoleff. 2019. Mapping global human dependence on marine ecosystems. Conservation Letters 12:e12617. Semesi, I. S., S. Beer, and M. Björk. 2009. Seagrass photosynthesis controls rates of calcification and photosynthesis of calcareous macroalgae in a tropical seagrass meadow. Mar Ecol Prog Ser 382:41-47. Seytre, C., and P. Francour. 2014. A long-term survey of Posidonia oceanica fish assemblages in a Mediterranean marine protected area: emphasis on stability and no-take area effectiveness. Marine and Freshwater Research 65:244-254. Short, F., T. Carruthers, W. Dennison, and M. Waycott. 2007. Global seagrass distribution and diversity: A bioregional model. Journal of Experimental Marine Biology and Ecology 350:3-20. Short, F. T., L. J. McKenzie, R. G. Coles, K. P. Vidler, and J. L. Gaeckle. 2006. SeagrassNet Manual for Scientific Monitoring of Seagrass Habitat, Worldwide edition. . University of New Hampshire Publication. Short, F. T., and S. Wyllie-Echeverria. 1996. Natural and human-induced disturbance of seagrasses. Environmental Conservation 23:17-27. Smith, M. D., J. Lynham, J. N. Sanchirico, and J. A. Wilson. 2010. Political economy of marine reserves: Understanding the role of opportunity costs. Proceedings of the National Academy of Sciences 107:18300- 18305. Spalding, M., L. Burke, S. A. Wood, J. Ashpole, J. Hutchison, and P. zu Ermgassen. 2017. Mapping the global value and distribution of coral reef tourism. Marine Policy 82:104-113. Spalding, M. D., I. Meliane, A. Milam, C. Fitzgerald, and L. Z. Hale. 2013. Coastal and Marine Spatial Planning. Protecting marine spaces: Global targets and changing approaches. Ocean Yearbook:213-248. Stachowicz, J. J. 2001. Mutualism, Facilitation, and the Structure of Ecological Communities: Positive interactions play a critical, but underappreciated, role in ecological communities by reducing

78 physical or biotic stresses in existing habitats and by creating new habitats on which many species depend. BioScience 51:235-246. Stelling-Wood, T. P., P. E. Gribben, and A. G. B. Poore. 2020. Habitat variability in an underwater forest: Using a trait-based approach to predict associated communities. Functional Ecology 34:888-898. Stjohn, J., G. R. Russ, and W. Gladstone. 1990. Accuracy and Bias of Visual Estimates of Numbers, Size Structure and Biomass of a Coral Reef Fish. Marine Ecology-progress Series - MAR ECOL-PROGR SER 64:253-262. Strain, E. M. A., G. J. Edgar, D. Ceccarelli, R. D. Stuart-Smith, G. R. Hosack, and R. J. Thomson. 2019. A global assessment of the direct and indirect benefits of marine protected areas for coral reef conservation. Diversity and Distributions 25:9-20. Sweet, M., and B. Brown. 2016. Coral Responses To Anthropogenic Stress In the Twenty-First Century: An Ecophysiological Perspective. Oceanography and Marine Biology: An Annual Review 54:271-314. Tanner, J. E. 1995. Competition between scleractinian corals and macroalgae: An experimental investigation of coral growth, survival and reproduction. Journal of Experimental Marine Biology and Ecology 190:151-168. Terrados, J., C. M. Duarte, M. D. Fortes, J. Borum, N. S. R. Agawin, S. Bach, U. Thampanya, L. Kamp-Nielsen, W. J. Kenworthy, O. Geertz- Hansen, and J. Vermaat. 1998. Changes in Community Structure and Biomass of Seagrass Communities along Gradients of Siltation in SE Asia. Estuarine, Coastal and Shelf Science 46:757-768. Thompson, A., and B. Mapstone. 1997. Observer Effects and Training in Underwater Visual Surveys of Reef Fishes. Marine Ecology Progress Series 154:53-63. Thresher, R. E., and J. S. Gunn. 1986. Comparative analysis of visual census techniques for highly mobile, reef-associated piscivores (Carangidae). Environmental Biology of Fishes 17:93-116. Thyresson, M., B. Crona, M. Nyström, M. de la Torre-Castro, and N. Jiddawi. 2013. Tracing value chains to understand effects of trade on coral reef fish in Zanzibar, Tanzania. Marine Policy 38:246-256. Topor, Z. M., D. B. Rasher, J. E. Duffy, and S. J. Brandl. 2019. Marine protected areas enhance coral reef functioning by promoting fish biodiversity. Conservation Letters 12:e12638. Udy, J. W., and W. C. Dennison. 1997. Growth and physiological responses of three seagrass species to elevated sediment nutrients in Moreton Bay, Australia. Journal of Experimental Marine Biology and Ecology 217:253-277. Underwood, A. J. 1993. The mechanics of spatially replicated sampling programmes to detect environmental impacts in a variable world. Australian Journal of Ecology 18:99-116.

79 UNEP. 2006. Marine and coastal ecosystems and human wellbeing: A synthesis report based on the findings of the Millennium Ecosystem Assessment. United Nations Environment Programme P.O. Box 30552, Nairobi, Kenya. UNEP, FAO, PAP, and CDA. 2000. Progress in integrated coastal management for sustainable development of Kenya’s coast: The case of Nyali-Bamburi-Shanzu Area. East African Regional Seas Technical Reports Series No. 6. , Split, Croatia. UNESCO. 1997. Sustainable Coastal Development: through integrated planning and management focused on mitigating the impacts of coastline instability. Nairobi, Kenya. Unsworth, K. F. R., J. J. Bell, and J. D. Smith. 2007. Tidal fish connectivity of reef and sea grass habitats in the Indo-Pacific. Journal of the Marine Biological Association of the United Kingdom 87:1287-1296. Unsworth, K. F. R., C. L. Cullen, N. J. Pretty, J. D. Smith, and J. J. Bell. 2010. Economic and subsistence values of the standing stocks of seagrass fisheries: potential benefits of no-fishing marine protected area management. Ocean & Coastal Management 53:218-224. Unsworth, R. K. F., R. Ambo-Rappe, B. L. Jones, Y. A. La Nafie, A. Irawan, U. E. Hernawan, A. M. Moore, and L. C. Cullen-Unsworth. 2018a. Indonesia's globally significant seagrass meadows are under widespread threat. Sci Total Environ 634:279-286. Unsworth, R. K. F., and L. C. Cullen. 2010. Recognising the necessity for Indo-Pacific seagrass conservation. Conservation Letters 3:63-73. Unsworth, R. K. F., L. J. McKenzie, C. J. Collier, L. C. Cullen-Unsworth, C. M. Duarte, J. S. Eklöf, J. C. Jarvis, B. L. Jones, and L. M. Nordlund. 2019. Global challenges for seagrass conservation. Ambio 48:801- 815. Unsworth, R. K. F., L. M. Nordlund, and L. C. Cullen-Unsworth. 2018b. Seagrass meadows support global fisheries production. Conserv Lett:1-8. Valentine, J. F., K. L. Heck, D. Blackmon, M. E. Goecker, J. Christian, R. M. Kroutil, B. J. Peterson, M. A. Vanderklift, K. D. Kirsch, and M. Beck. 2008. Exploited species impacts on trophic linkages along reef- seagrass interfaces in the Florida Keys. Ecological Applications 18:1501-1515. Walker, D. I., R. J. Lukatelich, G. Bastyan, and A. J. McComb. 1989. Effect of boat moorings on seagrass beds near Perth, Western Australia. Aquatic Botany 36:69-77. Valle, C., and J. T. Bayle-Sempere. 2009. Effects of a marine protected area on fish assemblage associated with Posidonia oceanica seagrass beds: temporal and depth variations. Journal of Applied Ichthyology 25:537-544. Walters, C., and C. Holling. 1990. Large-Scale Management Experiments and Learning by Doing. Ecology 71.

80 van der Heide, T., E. H. van Nes, M. M. van Katwijk, H. Olff, and A. J. Smolders. 2011. Positive feedbacks in seagrass ecosystems--evidence from large-scale empirical data. PLoS One 6:e16504. Waycott, M., C. M. Duarte, T. J. B. Carruthers, R. J. Orth, W. C. Dennison, S. Olyarnik, A. Calladine, J. W. Fourqurean, K. L. Heck, A. R. Hughes, G. A. Kendrick, W. J. Kenworthy, F. T. Short, and S. L. Williams. 2009. Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proc Natl Acad Sci U S A 106:12377- 12381. Webster, P. J., A. A. Rowden, and M. J. Attrill. 1998. Effect of Shoot Density on the Infaunal Macro-invertebrate Community within aZostera marinaSeagrass Bed. Estuarine, Coastal and Shelf Science 47:351- 357. Weeks, R., G. R. Russ, A. C. Alcala, and A. T. White. 2010. Effectiveness of marine protected areas in the Philippines for biodiversity conservation. Conservation Biology 24:531-540. Wells, S., N. Burgess, and A. Ngusaru. 2007. Towards the 2012 marine protected area targets in Eastern Africa. Ocean Coastal Manage 50:67-83. Welsh, A. H., R. B. Cunningham, C. F. Donnelly, and D. B. Lindenmayer. 1996. Modelling the abundance of rare species: statistical models for counts with extra zeros. Ecological Modelling 88:297-308. Verges, A., M. Perez, T. Alcoverro, and J. Romero. 2008. Compensation and resistance to herbivory in seagrasses: induced responses to simulated consumption by fish. Oecologia 155:751-760. White, A., and H. Vogt. 2000. Philippine Coral Reefs Under Threat: Lessons Learned After 25 Years of Community-Based Reef Conservation. Marine Pollution Bulletin 40:537-550. Whitfield, P., W. Kenworthy, K. Hammerstrom, and M. Fonseca. 2002. The role of a hurricane in the expansion of disturbances initiated by motor vessels on seagrass banks. J Coast Res 37:86-99. Villaseñor-Derbez, J. C., E. Aceves-Bueno, S. Fulton, A. Suarez, A. Hernández-Velasco, J. Torre, and F. Micheli. 2019. An interdisciplinary evaluation of community-based TURF-reserves. PLoS One 14:e0221660-e0221660. Williams, I., W. Walsh, B. Tissot, and L. Hallacher. 2006. Impact of observers' experience level on counts of fishes in underwater visual surveys. Marine Ecology Progress Series 310:185-191. Violle, C., M.-L. Navas, D. Vile, E. Kazakou, C. Fortunel, I. Hummel, and E. Garnier. 2007. Let the concept of trait be functional! Oikos 116:882- 892. Worm, B., E. B. Barbier, N. Beaumont, J. E. Duffy, C. Folke, B. S. Halpern, J. B. Jackson, H. K. Lotze, F. Micheli, S. R. Palumbi, E. Sala, K. A. Selkoe, J. J. Stachowicz, and R. Watson. 2006. Impacts of biodiversity loss on ocean ecosystem services. Science 314:787-790.

81 Zieman, J. C. 1976. The ecological effects of physical damage from motor boats on turtle grass beds in Southern Florida. Aquatic Botany 2:127- 139.