Effects of river sediments on coral recruitment, algal abundance benthic community structure on Kenyan coral reefs

The effects of sediment concentration and season on coral recruitment algal abundance and benthic community structure were studied in Kenyan coral reef lagoons to determine their potential influence on coral recovery. Nutrient levels and recruit numbers were higher during the southeast monsoon (SEM) than during the northeast monsoon (NEM) season and in sediment-exposed compared to nonsediment exposed reefs. Mean algal biomass also exhibited the same seasonal trend (except at one site), but was higher in the non-sediment exposed reef compared to the other reefs. Corals in the sediment exposed reef exhibited morphological differences relative to the other reefs: fewer corymbose and plate-like but more branching, massive and solitary forms and increased colony and corallite sizes. However, sediments did not suppress coral recruitment rates. These morphological changes coupled with the interaction between biological and physico-chemical characteristics have important ecological and geological implications: by potentially modifying calcium carbonate production and ameliorating the adverse effects of climate induced stress events, this may minimize coral mortality and enhance reef recovery.

Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License 2014; McClanahan et al., 2012;Palumbi et al., 2014) and anthropogenic disturbances, especially increases in human activities and associated development (West and Woesik, 2001;Wolanski et al., 2009;Hughes et al., 2010).In many cases, these declines have been linked to human induced perturbations that influence gradients of water quality and levels of exploitation that lead to increased sea surface temperature (SST) and atmospheric CO 2 (Fabricius, 2005;Pandolfi et al., 2005;Fabricius and McCorry, 2006;Richmond et al., 2007;Doropoulos et al., 2012;Logan et al., 2014).Recently, local and regional risk assessments have ranked the degree of vulnerability of reefs to anthropogenic stressors based primarily on potential landbased sources of disturbances (Burke and Maidens, 2004;Nyström et al., 2000;Wolanski et al., 2009).Further, the 1997/1998 worldwide bleaching event was unprecedented in its global impact on coral reefs with reports of about 50 to 90% coral mortality in some areas (Goreau et al., 2000;McClanahan et al., 2005;McClanahan, 2014).The interaction of anthropogenic and natural disturbances is therefore of great concern to the future existence and the persistence of coral reefs worldwide.In order to develop management strategies that improve coral survival and recovery, a greater understanding of life history processes and terrestrial impacts on coral recruitment and settlement is needed (Fabricius, 2005;Nozawa and Harrison, 2008;Maina et al., 2011;Darling et al., 2012).
Reproduction, recruitment and settlement are important processes in the maintenance and replenishment of coral reef populations (Harii et al., 2002;Glassom et al., 2004;Gilmore et al., 2013), but water quality is considered a major factor influencing the above three population parameters (Nzali et al,. 1998;Miller et al., 2000;Belleveau and Paul, 2002;Harrington et al., 2004;Glassom et al., 2004).In particular, water quality through its sediment and nutrient components has been reported to affect early coral life history processes by (1) the influence of nutrients and sediments on coral disease dynamics (Kuta and Richardson, 2002;Bruno et al., 2003), (2) altering the structural strength and density of reef substrate by influencing rates of bioerosion (Tribollet and Golubic, 2011;Carreiro-Silva et al., 2012), (3) modifying predation and herbivory rates on coral reefs through changes in grazing populations (McClanahan et al., 2011;Baker et al., 2013), (4) modifying heterotrophic filter feeding populations and therefore, net carbonate production (Dunn et al., 2012;Godinot et al., 2012), (5) influence of sediments on coral settlement through changes in substrate composition (Fabricius et al., 2005;Birrell et al., 2005Birrell et al., , 2008;;Diaz-Pulido et al., 2010) and ( 6) the action of sediments in reducing the competitive success of corals for space against algae (Nugues and Roberts, 2003a;McClanahan et al., 2012).
Recent increases in human activities such as farming, coastal development and deforestation have elevated the amount of soil transported by rivers into the marine environment with subsequent increases in turbidity and sediment in coral reef areas (Rotmann and Thomas, 2012).Increased terrestrial sediments into marine and coastal habitats have been identified as a major cause of distribution, diversity, community structure changes and degradation in coral reefs (Rodgers, 1999;and Roberts, 2003a;Fabricius, 2005).At the individual level, sediments have been linked to modification of light, thus affecting coral photosynthesis, overpower the sediment cleansing mechanisms and generally interfere with coral physiology leading mortality (Rodgers, 1990;Nugues;Fabricius, 2005).Moreover, Rodgers (1990) put forward seven characteristics of coral reefs effected by sediments: (1) low species diversity and absence of some species, (2) reduced percent coral cover, (3) high abundance of forms and species resistant to smothering and low light intensity (4) generally smaller colonies or, (5) generally larger colonies depending on coral sediment-tolerance/adaptive capabilities, (6) upward shift in depth zone and (7) a greater abundance of branching forms.Despite many reports on these potential influences, studies on the effects of sediments on benthic algae (and consequently coral recruitment) have produced inconsistent results, with elevated as well as reduced algal abundance reported in areas with high sediment levels (Hughes and Tanner, 2000;Irving and Connell, 2000;Gorgula and Connell, 2004).Resolving these conflicting and inconsistent findings is a major priority for coral reef management.Consequently, we assessed the effect of increased sediment levels on the coral community structure and coral recruit-algal density relationship in four protected reef lagoons in Kenya.

Study sites
Field studies were carried out in four Kenyan coral reef lagoons: (1) Malindi near the Sabaki river mouth (2) Watamu close to the mangrove-fringed Mida creek, (3) Mombasa, and (4) the offshore Kisite (Shimoni) reef on the extreme southern end of Kenya's fringing reef (Figure. 1).All four reefs are marine protected areas (MPAs) with more than 15 years of protection from fishing, coral and shell collection, and all forms of harvesting.The study reefs have been described fully in previous reports (Katwijk et al., 1993;McClanahan et al., 1994;Watson and Ormond, 1994;Mangubhai et al., 2007).These four shallow water reefs are found at 0.3 to 2 m deep at low tide with Kenya having a 4 m tidal range.Study reefs were chosen due to their accessibility, the type of exploitation and the influence of river discharge.The Malindi reefs experience a strong pulse of land-derived sediments during the short rain (northeast monsoon, NEM) season from the Sabaki River as a result of the reversed monsoon driven currents (Kayanne et al., 2006;Fleitmann et al., 2007).Watamu reef is a fringing reef adjacent to the highly ramified, groundwater sustained Mida mangrove creek that is devoid of any external freshwater input.Mombasa reef is a stretch of lagoonal fringing reef about 6 km long and 1 km from the shore, situated about 5 km from the city of Mombasa.This reef is moderately influenced by local rainfall and is adjacent to two mangrove creeks, Tudor and Mtwapa.Shimoni is a remotely located reef in the Kisite MPA on the seaward side of Wasini Island offshore from Shimoni town.Minimal terrestrial anthropogenic influence was expected from this reef due to its remote location.Three study sites were selected in Malindi and Shimoni reefs, and two sites were used in Watamu and Mombasa reefs.Sampling was conducted during the northeast (NEM) and the southeast (SEM) seasons in order to capture the effects of changing oceanographic conditions (McClanahan, 1988).

Measurement of environmental parameters
Environmental sampling was conducted twice per month for three months during each of the NEM and SEM seasons.Temperature and salinity measurements were taken using an automatic probe and current velocity with a modified current drogue.Chlorophyll a, phosphate and nitrate concentrations were determined after filtration of 1 L water samples through 0.45 μM pore size glass filters and spectrophotometric analysis of the filtrate.Total suspended sediment (TSS) concentration was determined by filtering 1 L water samples through pre-weighed 0.45 μM pore size glass filters.The filters were then oven dried overnight at 60°C and weighed again for gravimetric determination of suspended sediment concentration.Sedimentation rates were measured using sediment traps consisting of a set of four plastic cups attached to PVC pipe driven firmly into the substrate at each site with cup mouths approximately 50 cm above the sea bottom.Accumulated sediments were collected biweekly and after collection, sediments were washed with freshwater, small animals and plants were removed and the sediment was oven dried at 60°C overnight.Further, at each reef, samples of bottom sediment were randomly collected monthly using a cylindrical corer (50 cm high and 2 cm in diameter).Sediment samples were dried at 60°C to constant weight after washing with distilled water to remove salts.Replicate 100 g samples were then sieved through 710, 500, 355, 180, 125 and 63 μM using an automatic sieve shaker.Differences in sediment quality were inferred from the average weight of coarse (> 180 μM) and fine (≤ 180 μM) sediment fractions.Number of samples varied between eight and sixteen cores per reef.Sediment organic matter (as an indicator of terrestrial influence) was determined as loss in weight after combusting samples in a furnace at 500°C for 4 h.

Coral spat settlement studies
Twenty to forty ceramic tiles (15 x 15 cm) with the unglazed surface facing up were randomly deployed in each site.Tiles were attached to the substrate with cable tiles and left in the field for four to six months during each of the two East African monsoon seasons over a three year period.After collection, tiles were thoroughly washed with sea and freshwater to remove sediments and salts, packed individually with old newspapers to avoid damage, sun dried at the laboratory and later bleached to reveal the coral recruits.The tiles were examined at low magnification, recruits were counted and their diameters were measured (to the nearest mm) under a microscope equipped with an ocular scale.

Algal settlement and substrate studies
Ten to fifteen tiles from each reef per season were preserved in formalin after collection and transported to the laboratory.Tiles were thoroughly washed with freshwater to remove formalin and marine sediments, and algae samples were scrapped from five areas of 2 x 2 cm from the 4 corners and the centre of each tile.Algae samples were dried overnight in an oven at 100°C for determination of dry algal biomass.At each of the sites, benthic cover was studied by completing nine to twelve 10 m benthic line transects randomly placed on the substrate for description of the benthic cover (McClanahan and Shafir, 1990).The length of benthic categories > 3 cm were and classified according to the following categories: hard and soft coral, coralline, calcareous and fleshy algae, sponge, seagrass and sand.Hard corals were identified to the genus level, their percent cover was estimated and colony diameters were measured.Hard corals were further classified into eight coral growth forms: massive, branching, foliose, plate, digitate, corymbose, solitary and encrusting.Substrate complexity (rugosity) was estimated using the procedure described in McClanahan and Shafir (1990).Diversity was assessed using species richness, Pielou's J evenness index and Shannon's H diversity (Shannon and Weaver, 1949).

Statistical analysis
A one way-analysis of variance (ANOVA) was employed to test for differences in substrate cover, coral genera and colony size between reefs, and a 2-way ANOVA was used to test for differences in algal biomass and environmental factors between reefs and seasons.A 3-way ANOVA test was used to test for differences in sediment quality (reef x season x gain size) and coral recruit density (reef x season x surface).Bartlett's and Shapiro-Wilk tests were used to test for homoscedasticity and normality of variance, respectively.Where differences were detected, comparisons were done using Tukey HSD test to determine which means were significantly different.Sediment and algal data were log (x+1) transformed in order to conform to the assumptions of ANOVA.Pearson's correlation coefficients were used to examine the degree of association between biological and oceanographic parameters.The statistical programs STATISTICA 6.0 for Windows and JMP 7.0 were used for the above analyses while PRIMER (Plymouth Routines in Multivariate Ecological Research) was used in the analysis of diversity indices.

Environmental parameters
Mean seasonal environmental variables showed that Malindi had higher nutrient and sedimentation concentrations compared to the other reefs with a general decline in nutrients and sediments toward Shimoni in the south (Figures 2 and 3).Current velocity did not differ between seasons but was high in Malindi relative to the other reefs (Figure 2a).Mean temperature and salinity values for all study sites were significantly lower during the SEM compared to the NEM period (except salinity in Malindi, Figure 2b and c).Water column chlorophyll-a concentration exhibited a unique pattern: being higher in Malindi and Watamu compared to Mombasa and Shimoni during the SEM but generally lower and more similar during the NEM (Figure 2d).Mean phosphate concentrations did not differ between reefs during the NEM season; however, during the SEM, phosphate concentrations were found to be significantly lower in Shimoni and higher in Watamu compared to the other study reefs (Figure 2e).No significant difference in nitrate concentration was detected between seasons, but nitrate concentration was higher in Malindi and declined to the south toward Shimoni reef (Figure 2f).
Total suspended sediment concentration (TSS) and TSS organic content exhibited little spatial variation but were higher in the SEM compared to the NEM for all sites, except Mombasa (Figures 3a and b).Sedimentation (sediment deposition) rates were found to be high during the SEM compared to the NEM for all studied Higher proportions of organic matter in trapped (deposited) and bottom sediments were measured in Malindi during the NEM compared to the SEM, but sediment organic content did not differ between seasons in the other study reefs (except bottom sediment organic content in Mombasa).Bottom sediment organic content was found to be higher in Mombasa and Shimoni compared to Malindi and Watamu (Figures 3d and e).A higher proportion of fine sediment particles (< 180 um) was observed in Watamu and Shimoni relative to Mombasa and Malindi during the SEM season.During the NEM season, a higher proportion of fine particles was found in the Watamu reef compared to all other reefs, however Malindi sediment-exposed reef exhibited a higher proportion relative to Shimoni and Mombasa reefs (Figure 3f, Table 1).

Coral recruitment studies
Overall, recruit density did not differ between seasons and all reefs showed similar seasonal patterns in number of coral recruits (Figures 4 and 5; Table 2).Average recruit density (mean of 2 seasons) in the sediment exposed reefs of Malindi (21.4 recruits per tile) was significantly higher compared to all other study reefs, with more recruits being counted during the NEM compared to SEM season in all sites.Further, significant high numbers of recruits were observed on the underside (lower surface) of tiles relative to the up-facing surface except for the Shimoni reef.Recruit size studies showed truncated size-frequency distributions with the 2.1-4.0 mm size-class constituting the majority of recruits in all reefs (Figure 5).

Algal settlement and substrate studies
Bare surface comprised 28.9, 52.0, 48.26 and 53.85% of     , did not differ between seasons.Algal biomass in all reefs was higher during the NEM compared to the SEM except in the sediment-exposed Malindi reef (Table 3 and Figure 6).Mean biomass for Malindi was significantly lower than Shimoni reef but not different from Mombasa and Watamu reefs.In all sites, colonization was mostly by lower canopy algae.Halimeda and turf algae were found to be the dominant algal categories in Malindi.Coral cover in the Shimoni reef was significantly higher than the other reefs with the sediment-exposed Malindi reef showing higher cover than Watamu reef but lower than Mombasa (Table 4).No significant difference in turf algae cover was detected between Shimoni and Malindi reefs, however, higher turf algae cover was measured in Mombasa and Watamu compared to the other two reefs.Malindi reef had higher soft coral, calcareous and coralline algae cover whereas high sponge cover was measured in Mombasa compared to all other reefs.Rugosity did not differ between reefs.
The relative abundance of plate-like and corymbose growth forms was higher in Shimoni compared to the other reefs (Figure 7).Solitary and branching forms dominated Malindi reef relative to all other reefs.Massive forms constituted the largest proportion of corals in all sites,   5).
The abundance of digitate forms did not differ between reefs.Malindi reef had higher cover of Acropora, Galaxea, Platygyra, and branching and massive Porites compared to all other reefs (Figure 8).Watamu reef was dominated by Montipora whereas Favia and massive Porites were dominant in Mombasa reef.
Branching Porites and Acropora were the dominant genera in the Shimoni reef.Size structure data revealed that Malindi reef had larger colonies of Echinopora, Galaxea, Goniastrea, Favites, Millepora and Pocillopora compared to the other reefs (Figure 9).Although not statistically tested, Shimon reef exhibited higher species diversity and richness indices compared to the sediment-exposed Malindi reef, but both were lower than values measured in the Watamu reef.All reefs exhibited similar evenness.

DISCUSSION
Recruitment and settlement are vital processes in the development of coral populations and are essential in community recovery after disturbance.In post-disturbance periods, the settlement of coral larvae is a key ecological process in the resilience and persistence of coral reefs   (Arnold et al., 2010;Venera-Ponton et al., 2011;Doropoulos et al., 2012).Likewise, optimal habitat characteristics are critical in the post-settlement survival of benthic larvae (Kendrick, 1991;Airoldi and Virgilio, 1998;Harrington et al., 2004), and benthic algae and sediments offer both inductive and inhibitive settlement  cues for planktonic larvae (Kuffner et al., 2006;Thomsen and McGlathery, 2006;Diaz-Pulido et al., 2010).Past studies have indicated that coral larvae often search for appropriate substrate that is devoid of fleshy algae, silt and shifting sediments for successful settlement (Thomsen and McGlathery, 2006;Doropoulos et al., 2012).Competition for space in reef communities means that organisms that colonize new and open substrates and factors that facilitate or inhibit coral settlement play critical roles in influencing the structure and resilience of reef communities (Casey et al., 2015).

Environmental parameters
This study reveals high sedimentation rates on Malindi reef during the SEM period despite the northerly direction of winds and local currents away from the Sabaki River (Brakel, 1984;Katwijk et al., 1993;Obura, 1995;McClanahan and Obura, 1997).The reason for the higher sedimentation rates during the SEM is due to resuspension of sediments by increased hydrodynamic energy (Brakel, 1984;Kayanne et al., 2006;Fleitmann et al., 2007), and this is pronounced in Malindi where large amounts of sediments from the River Sabaki are deposited during the NEM and are available for re-suspension over the SEM season.The calmer conditions in Malindi during NEM resulted in transportation and accumulation of fine sediments and sediments of higher silt fraction through suspended and saltation that may later be re-suspended during the rougher SEM sea conditions.The importance of sediment quality in coral ecology has previously been discussed (Rodgers, 1990;Fabricius, 2005).High hydrodynamic energy and runoff are prevalent during the SEM in East African coastal areas (McClanahan, 1988;Kitheka et al., 2003).Despite the southward plume and river discharge movement during the NEM, salinity and SST values from Malindi were not different from those measured from the other study reefs.This is likely due to complete mixing of marine and fresh waters and/or low discharge volume unable to cause any significant changes in salinity and SST (Kitheka et al., 2003).Spatial and temporal distribution of nutrient loads associated with high levels of human activities in both freshwater and coastal watersheds are controlled by seasonal hydrodynamic changes (Christiansen et al., 1996;Madsen et al., 2001).Elevated nutrient levels observed during the SEM relative to the NEM season and in Malindi compared to the other reefs were probably due to several reasons.Firstly, the proximity of Malindi to the Sabaki River compared to the other reefs and the high discharge volume (~400-650 m -3 s -1 ) during the SEM likely increase nutrient loads relative to the NEM period (low discharge volume, 60-80 m 3 s -1 ; McClanahan, 1988;  Kitheka et al., 2003).
Secondly, phosphates are continuously trapped by bottom sediments (Ohowa, 1996;Xie et al., 2003), and are eventually released into the water column through disturbances in the sea bottom during high hydrodynamic events, conditions that are prevalent during the SEM period.The high levels of chlorophyll a concentration observed during the SEM result from heavy rains, high river discharge and runoff (relative to the NEM), and affirm the effects of River Sabaki and the Mida mangrove creek in Malindi and Watamu, respectively.
Further, reported upwelling conditions off the southern Somali and northern Kenya coast during the SEM season (McClanahan, 1988) may also be a contributing factor to the observed (elevated nutrient levels) findings in the current study.

Coral recruitment patterns
Coral recruit density varied with reef, seasons and surface orientation in accordance with findings from other studies (Obura, 1995;McClanahan and Obura, 1997;Glassom et al., 2004;Obura et al., 2005;Karisa et al., 2008).Consistent with other studies, coral larvae showed high affinity for underside surfaces, which is considered a strategy to escape from algae competition (Kuffner et al., 2006;Arnold et al., 2010;Venera-Ponton et al., 2011), sedimentation and light stress on upper surfaces (Thomason et al., 2002;Birrell et al.,2005).We could not ascertain reasons for the no effect of surface orientation on recruit numbers in Shimoni, however un-interfered light penetration may be a likely candidate.High coral recruit densities found in the sediment-exposed Malindi reef may be attributed to several factors.The presence of swift-currents, moderate (wave action) tidal amplitude (~4 m in East Africa) and a deep-water channel likely enhanced flushing rates, thereby preventing the accumulation of sediment to chronic levels and thus increasing available space for coral settlement.However, increased hydrodynamics have been shown to sweep away larvae from substrates before attachment (Reidenbach et al., 2009) but this probably varies with organisms or substrate types, thus explaining the findings in the present study.Control of macroalgae by herbivorous fish populations may also be an important contributing factor in promoting increased coral settlement through increased substrate availability (Edmunds and Carpenter, 2001;Casey et al., 2015).In addition, the observed high coral recruit densities observed during the NEM period coincide with the period of warmest sea temperatures and the timing of gamete release by a number of coral species in the Western Indian Ocean region (Mangubhai et al., 2007).Therefore, the interplay and interaction of oceanographic and biological factors may play a significant role in the maintenance of coral populations in disturbed reefs such as the Malindi reef.
Recruit densities measured in the current study are similar to those measured in Sodwana Bay, South Africa (Glassom et al., 2006) but different from density patterns measured along the Australian coast (Harriot and Banks, 1995) though recruitment is generally known to decline with increasing latitude.This is attributed the relative position of Kenya and South Africa towards the extreme ends of two opposite flowing currents, the northerly flowing East African Coastal Current and the south flowing Mozambique Current that split from the South Equatorial Current near the Tanzania-Mozambique border (Mangubhai et al, 2007).This dispersal pattern differs from that observed along the eastern and western coasts of Australia (Harriott and Banks, 1995) that are both bathed by southerly flowing currents, the East Australian Current and Leeuwin Current, respectively (Harrison and Booth 2007).

Algal settlement and diversity studies
The high coral recruit density measured in the Malindi reef relative to the other reefs differed from the results of earlier studies on the settlement of sessile organisms where sediments were shown to negatively impact coral settlement (Kendrick, 1991;Bobcock and Smith, 2000;Fabricius and De'ath, 2001;Birrell et al., 2005;Thomsen and McGlathery, 2006).However, the interaction of high sedimentation and hydrodynamics, high herbivory and reduced algal biomass may have played a significant role in our findings.Previous studies have shown that benthic algae compete with corals through a variety of mechanisms, including the promotion of bacterial growth that cause high juvenile coral mortality (McCook et al., 2001, Nugues et al., 2004;Diaz-Pulido et al., 2010;Venera-Ponton et al., 2011;O'Leary et al., 2012).In addition, Connell and Karlson (2000) reported that sediments negatively affect algal abundance and biomass through reduction in growth and reproductive success, potentially causing enhanced coral settlement in sediment-exposed reefs.These findings suggest that although algae may have a substantial impact on coral recruitment, sedimentation may indirectly modify coral recruitment by influencing algal abundance and biomass.Consequently, high coral recruitment rates were found on Malindi reef despite the suspected detrimental effects of Sabaki river sediments flowing into the Malindi reef system, possibly due to the presence of an optimal breeding population of sediment tolerant coral species, in addition to the aforementioned fators.

Benthic and coral community composition
Larger coral colonies were found in the sedimentexposed reefs relative to other reefs, which is consistent with previous findings (Obura, 1995;McClanahan and Obura, 1997;Nugues and Roberts, 2003b;McClanahan, 2014).Further, in a separate study conducted in the same reefs, dominant coral genera in sediment exposed reefs showed larger corallite diameters compared to nonsediment impacted reefs (Mwachireya et al., 2015), suggesting this to be a strategy to survive exposure to high sediment levels.Similar findings of larger diameters in sediment influenced reefs have previously been reported by other workers (Todd, 2001;Todd et al., 2008).Increased corallite size has been associated with enhanced sediment rejection capability (Todd et al., 2008) and large colony sizes have been linked to reduced susceptibility of corals to high sediment levels as well as increasing reproductive success relative to smaller colonies (McClanahan and Obura, 1997).
Sediments have also been reported to cause partial mortality and reduced coral size (Nugues and Roberts, 2003b) and this may cause reduced coral recruitment and increased frequency of small-sized colonies in sediment intolerant species (McClanahan and Obura, 1997;Hughes and Tanner, 2000;Nugues and Roberts, 2003b).The effects of high sediment levels in Malindi corals is further reinforced by the presence of high relative abundance of sediment-tolerant branching, digitate and solitary but low relative abundance of sediment-intolerant (encrusting, corymbose and platelike) growth forms.
It can therefore be concluded that coral community characteristics observed in the sediment-exposed reef in this study are a reflection of (1) high recruitment rates and low macroalgal cover, (2) sediment rejection characteristics such as colony and corallite morphology (3) hydrodynamic influences on growth forms, sediment and larval dispersal, and (4) the effects of grazing activities of herbivorous fish on algae freeing space for coral settlement.Consistent with Rodgers (1990), we found that reef communities in the sediment exposed reef exhibited low species diversity, reduced coral cover, high abundance of sediment tolerant species, generally smaller or larger colonies depending on sediment tolerance and greater abundance of branching corals.The dominance of sediment-tolerant corals as well as high levels of coral recruitment may contribute to the thriving coral community in Malindi reef despite exposure to high levels of sediments from the Sabaki River.This has important ecological and geological implications by potentially influencing rates and patterns of calcium carbonate accretion in sediment disturbed coral reefs (Perry and Larcombe, 2003).

Biological and physicochemical parameters
Sedimentation rates were positively correlated with recruit density similar to current velocity but were negatively correlated with fleshy algal cover.Sediment organic content was negatively related to recruit density but not algal biomass.Interestingly, all nutrient types and chlorophyll a concentrations were positively correlated with recruit density but negatively correlated with algal biomass and cover, contrary to general perceived views (Belleveau and Paul, 2002;Fabricius, 2005;Venera-Ponton et al., 2011).Previous studies on the relationship between algae and nutrient concentrations have also provided conflicting results (Diaz-Pulido and Garzón-Ferreira, 2002;Diaz-Pulido and McCook, 2003;Fabricius, 2005), which may be due to the many interacting biological and physico-chemical factors known to modify the algae-nutrient interaction.The negative association between algae and sedimentation rate is similar to findings of Gorgula and Connell (2004) but different from those of Irving and Connell (2002) and Nugues and Roberts (2003b).Sediment quality, site hydrodynamics and microhabitat differences and their interactions may be the likely factors responsible for the observed differences between studies (Arnold et al., 2010;Nakamura et al., 2011).Further, our results suggest that the organic fraction of sediment has a greater negative influence on recruit density compared to the actual sedimentation rate values.
The positive relationship between sediments and coral recruitment is likely due to (a) increased sediment accumulation reducing the recruitment, attachment and survival of new algal recruits onto substrates (Umar et al., 1998;Birrell et al., 2005), (b) elevated sediment concentrations smothering short fronds and also preventing the vegetative regeneration of fleshy algae (Umar et al., 1998), (c) positive sediment effects on calcareous and coralline algae facilitating the recruitment and settlement of coral recruits (Kendrick, 1991;Airoldi and Virgilio, 1998;Harrington et al., 2004) and (d) protective management increasing the abundance of herbivorous fish that free substrate space through their grazing activities on algae (Venera-Ponton et al., 2011).The present results therefore suggest that high coral recruitment levels can occur in reefs with moderate sedimentation rates and that the effects of sedimentation on algae in coral reefs are much more complex than previously considered.It would seem that algae are merely opportunistic beneficiaries of sedimentassociated factors rather than being favoured by high sediment loads.Therefore, attention should also be directed toward the effects of other water quality characteristics such as nutrients, toxins (pesticides, heavy metals) and organic material on a wide range of life history parameters in reef organisms (Rodgers, 1990;Richmond et al., 2007) that can tip the balance toward algal proliferation and reduced coral recruitment, which may lead to the decline of coral reefs.
This study provides evidence that increased sediment influx from Sabaki River has a significant impact on Malindi coral communities based on the dominance of sediment-adapted coral morphologies and reduced coral cover despite high or comparable recruitment rates relative to reefs unaffected by sedimentation.The existence of river discharge around the Malindi area since the formation of the Rift Valley (~39 million years ago) and the presence of coral reefs at sufficient distance from the Sabaki estuary suggests that the combination of biological and oceanographic characteristics of sedimentexposed reefs probably help to modify the potential impacts of sediments to below critical levels necessary to cause severe coral reef decline.Moderate sedimentation rates and high current speeds may enhance coral recruitment by inhibiting substrate colonization and overgrowth by competing algae and this has important implication for the recovery of coral reefs in areas affected by terrestrial sediment influx.`

Figure 1 .
Figure 1.Map of the Kenyan coastline showing location of Malindi, Watamu, Mombasa and Shimoni protected reefs and study sites.

Figure 7 .
Figure 7. Relative abundance of coral growth forms from the four Kenya study reef, Malindi (MLD), Watamu (WTM), Mombasa (MSA) and Shimoni (SHM).Data are percentages of a minimum of 53 and maximum of 1363 colonies per reef.

Figure 8 .
Figure 8. Relative abundance of some common coral genera from the four Kenyan study reefs.Data are percentages of a minimum of 9 and a maximum of 12 line transects per reef.

Figure 9 .
Figure 9. Average colony sizes (mean ± SD) of common coral genera from three Kenyan protected reefs.Data are means from 9 > n < 12 lines transects per reef.

Table 1 .
Analyses of variance testing the effect of reef, season and grain size (coarse and fine) on bottom sediments of Kenyan coral reef closures (ns = not significant, * =significant at α = 0.05).

Table 2 .
Analyses of variance testing the effect of reef, season and tile surface on coral spat density on Kenyan protected coral reef lagoons (ns = not significant, * =significant at α = 0.05).

Table 3 .
Analyses of variance test for the effect of reef and season on algal biomass from tiles deployed in Kenyan protected coral reef lagoons (ns = not significant, * =significant at α = 0.05).

Table 4 .
Mean bottom cover (natural substrate) and standard deviations, rugosity and sphericity from four Kenyan protected reefs with their comparison tests (ns = not significant, * =significant at α = 0.05, nt = not tested).