Western Indian Ocean Journal of Marine Science

EDITOR-IN-CHIEF: Prof. Michael Schleyer, Oceanographic Research Institute, Durban, South Africa EDITORIAL BOARD Dr Jared Bosire, Kenya Prof. Kassim Kulindwa, Tanzania Dr Blandina Lugendo, Tanzania Dr Francis Marsac, Reunion Dr Nyawira Muthiga, Kenya Prof. J. Paula, Portugal Prof. Chris Reason, South Africa

PUBLISHED BIANNUALLY Aims and scope: The Western Indian Ocean Journal of Marine Science provides an avenue for the wide dissemination of high quality research generated in the Western Indian Ocean (WIO) region, in particular on the sustainable use of coastal and marine resources. This is central to the goal of supporting and promoting sustainable coastal development in the region, as well as contributing to the global base of marine science. The journal publishes original research articles dealing with all aspects of marine science and coastal management. Topics include, but are not limited to: theoretical studies, recovery and restoration processes, legal and institutional frameworks, and interactions/relationships between humans and the coastal and marine environment. In addition, WIOJMS features state-of-the-art review articles, book reviews, short communications and opinions. The journal will, from time to time, consist of special issues on major events or important thematic issues. Submitted articles are peer-reviewed prior to publication. Editorial correspondence, including manuscript submissions should be sent by e-mail to the Editor- in-Chief ([email protected]). Details concerning the preparation and submission of articles can be found in each issue and at http://www.wiomsa.org Disclaimer: Statements in the Journal reflect the views of the authors, and not necessarily those of WIOMSA, the editors or publisher. Annual Subscription (2010). Institutional subscription: US$ 150, applicable to university libraries, laboratories, industrial and other multiple-reader institutions. Individual subscription: US$ 100. Members of WIOMSA are eligible for the discounted subscription rate of US $ 50. Prices are subject to amendment without notice Copyright © 2010 —Western Indian Ocean Marine Science Association (WIOMSA)) No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means without permission in writing from the copyright holder. ISSN 0856-860X

Volume 9 Final 20th Oct 2010.indd 2 10/27/2010 10:02:02 AM Western Indian Ocean Journal of Marine Sciences

Volume 9 Number 1 (2010)

Volume 9 Final 20th Oct 2010.indd 3 10/27/2010 10:02:02 AM PREFACE

The publication of this issue heralds change, as Alan Whittick is bowing out as Editor-in-Chief of the Western Indian Ocean Journal of Marine Science, having graciously and faithfully served in this capacity since inception of the Journal in 2002. The growth of the Journal is a tribute to his efforts, as well as to the scien- tists working in the region who have used it as a vehicle for publication of their work. I am now taking over the role of Editor-in-Chief and wish to build on the foundation laid by my predecessor.

Allan was ably assisted in his work by an Editorial Board and we are similarly grateful to its members for their services. The size and composition of the board is now changing as it is being enlarged to reflect the growing diversity of research undertaken in the region. Further changes are being investigated to improve and streamline publication of the Journal through procedures such as online submis- sion. This will have several advantages and will make accepted manuscripts available on the web prior to going to press.

There can be no doubt that the Western Indian Ocean Journal of Marine Science has grown in stature and the Editorial Board and I will work together to ensure its continued growth in importance, standard and distribution. We look forward to your continued support through the submission of manuscripts for publication in this Journal.

MH Schleyer September 2010

Volume 9 Final 20th Oct 2010.indd 4 10/27/2010 10:02:02 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 1-16, 2010 © 2010 WIOMSA

Monthly Variations in Sea Level at the Island of Zanzibar

S.B. Mahongo1 and J. Francis2 1Tanzania Fisheries Research Institute, P.O. Box 9750, Dar es Salaam, Tanzania; 2Department of Aquatic Sciences and Fisheries, University of Dar es Salaam, P.O. Box 65011, Dar es Salaam, Tanzania.

Keywords: Zanzibar, sea level variations, climate, spectral analysis, multiple regression.

Abstract—Meteorological and tide gauge data were used to analyze correlations between climatic parameters and variations in mean sea level at Zanzibar for the period 1985-2004. This involved spectral and multiple regression analysis of the monthly variables, as well as harmonic analysis of hourly sea level. Air pressure and rainfall remained relatively constant during the 20-year study period, but there were trends in sea level, northeast winds, southeast winds and air temperature. Monthly variations in mean sea level, composed predominantly of semi-annual, annual and 4-year oscillations, were represented by the steric effect proxies of rainfall and air temperature (45%), southeast and northeast monsoon winds (41%), and air pressure (5%). The trend in sea level (9%) appeared to be mainly correlated with northeast winds. The annual cycle in sea level (36%) was represented to a certain degree by rainfall (11%), air temperature (10%), southeast winds (8%) and northeast winds (7%). The semi-annual component (28%) was best represented by southeast winds (15%), with the remaining 13% of the variability being equally represented by rainfall, northeast winds and air pressure. The 4-year oscillations, which accounted for about 27% of the variation in sea level, were mainly represented by air temperature (12%), rainfall (8%) and southeast winds (6%). There is a strong likelihood that physical processes other than meteorology and tides influenced the observed variations in sea level, especially in the 4-year cycle.

Corresponding Author: SBM E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 5 10/27/2010 10:02:02 AM 2 S.B. MAHONGO & J. FRANCIS

INTRODUCTION level is investigated off Zanzibar, contributing towards climate Changes in monthly mean sea level monitoring and coastal management. reflect fluctuations in atmospheric The largest manifestation of pressure (the inverted barometer seasonal change at the Island of effect), wind regime, river runoff and Zanzibar (Fig.1) is the monsoon steric effects due to changes in the wind reversal and associated changes specific volume of the ocean associated in climate. The northeast monsoon with shifts in water temperature and extends from January to February salinity (Tsimplis and Woodworth, and is characterised by higher air 1994). The steric effect causes sea temperatures, lower wind speeds and level to rise when the water is warm, consequently calmer sea. The southeast or has low salinity, and vice versa. At monsoon begins in April, ends in low latitudes, variations in mean sea November and is usually strong. Wind level are controlled to a very large speeds are highest between June and extent by these steric effects and are October, and are lowest during the mainly thermal (Pattullo et al., 1955; inter-monsoon months of March and Lisitzin and Pattullo, 1961). December. The heavy (long) rains Variations in sea level off Tanzania are experienced from March to May, have received very little attention. while the light (short) rains fall during Pattullo et al., (1955) indicated that November and December. the sea level in Dar es Salaam (Fig. The prevailing current along the 1), which is about 80 km south of Tanzanian coast is the East African Zanzibar, is highest during March-May Coastal Current (EACC), which flows and lowest during July-September. northwards throughout the year. Apart The authors, however, did not outline from the EACC, winds and tides are factors that may be responsible for the the main forces that drive the oceanic observed seasonality. Using rainfall circulation in the Zanzibar Channel and sea level data from 1962 to1966, (Mayorga-Adame, 2007). The Ragoonaden (1998) observed that Channel, which is about 35-40 km variations in monthly mean sea level wide and 120 km long, is generally at Tanga (about 120 km north of shallow with a depth of between 30-40 Zanzibar town) were mainly due to m in the middle (Shaghude & Wannäs, elevation of the sea surface resulting 1998). According to Harvey (1977) from changes in the current velocity and Mohammed et al., (1993), the off the coast, ruling out the effect of tidal circulation inside the Channel is rainfall and river runoff as a significant complex. Flood streams enter and ebb factor. In this paper, the contribution streams exit the Channel at both its of meteorology and long-term trends north and south entrances. While the in tides to monthly variations in sea Wami and Ruvu rivers on Mainland

Volume 9 Final 20th Oct 2010.indd 6 10/27/2010 10:02:02 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 3

Fig. 1: Map showing the island of Zanzibar, the major direction of the Southeast (SE) and Northeast (NE) Monsoon winds, and the direction of the East African Coastal Current (EACC).

Tanzania discharge into the Zanzibar is protected by a number of offshore Channel directly opposite the island, islands, reefs and sandbanks that there are no permanent rivers on complicate current patterns in the Zanzibar. (Fig. 1). vicinity of the harbour area (Shaghude The Zanzibar tide gauge is located et al., 2002; Mayorga-Adame, 2007). on the seaward end of the main jetty The tide gauge was commissioned in Zanzibar Harbour, on the western in 1984, serving as one of the prime coast of Zanzibar town. The harbour Indian Ocean stations for observation

Volume 9 Final 20th Oct 2010.indd 7 10/27/2010 10:02:02 AM 4 S.B. MAHONGO & J. FRANCIS

of global sea level change. The station recorded at 1500 local time (GMT +3). has also been transmitting real-time sea Meteorological records were obtained level data for the Indian Ocean Tsunami from Chukwani station, about 5 km Warning System (IOTWS) through the from the harbour and considered Global Telecommunications System representative of the study area. The (GTS) since December 2006. MATLAB “wind-rose” script was used to determine the main axes of DATA AND METHODS the southeast and northeast monsoon OF ANALYSIS winds at Zanzibar. These were oriented at 135o and 29o respectively Two sets of data were used in the from True North. The data on winds investigation. The first included records were then decomposed into these two of hourly and monthly sea level data major wind directions, which also from the Zanzibar tide gauge (1985- conform to the general orientation of 2004). These were obtained from the the northern and southern entrances of University of Hawaii Sea Level Centre Zanzibar Channel (Fig. 1). (UHSLC), and can be accessed at www. The contribution of steric effects ilkai.soest.hawaii.edu. The hourly data to the observed sea level oscillations were subjected to harmonic analysis to was not directly computed due to obtain the amplitudes and phase lags paucity of data on salinity and sea of major tidal harmonics using the Sea surface temperature along the coast. Level Processing Software, SLPR2 Consequently, the steric effect proxies (Caldwell, 1998). of rainfall and air temperature were The UHSLC computes monthly mean used. The above data sets were cleaned sea levels through simple averaging of before analysis through removal of large the daily values. The daily series, on the spikes. These were outlying data points other hand, are derived using a two-step that were considered incorrect. Linear filtering operation. First, the dominant regressions of the monthly series were diurnal and semidiurnal tidal components plotted by the method of least squares are removed from the quality-controlled on SIGMAPLOT software to determine hourly values. Secondly, the remaining the presence of long term changes. higher frequency energy data are Spectral time series in the sea level removed using a 119-point convolution and meteorological data were extracted filter (Bloomfield, 1976) to prevent using Fast Fourier Transform (FFT) aliasing when the data are computed to in the STATISTICA software package daily values. which determined the significant The second group of data (1985- spectrum peaks (Shumway, 1988). Prior 2004) consisted of mean monthly to spectral analysis, the relatively few rainfall, air temperature, air pressure missing data were interpolated from and winds (speed and direction) linear trend regressions. The amplitude

Volume 9 Final 20th Oct 2010.indd 8 10/27/2010 10:02:02 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 5

and phase lags in significant spectral and discussed in detail in Makridakis & cycles were computed by Discrete Wheelwright (1978, 1989). The moving Fourier Transform (DFT) using a script average in the variable trend-cycle was downloaded from www.mathworks. determined using the Henderson curve com\matlabcentral\fileexchange. moving average, which is a weighted The Hamming window (Oppenheim moving average in which the magnitude & Schafer, 1989) was chosen to in the weighting follows a bell-shaped smooth the raw data for both the FFT curve. To avoid erroneous conclusions, and DFT analyses. The procedure residual analysis tests were carried involves weighted moving average out in STATISTICA to identify any transformation that assigns the greatest violations of the assumptions after weight to the observations in the centre fitting the multiple regression equations of the window, and increasingly smaller through case-wise plots of the residuals, weights to values that are further away case-wise plots of outliers and normal from the centre. This procedure was probability plots of the residuals. preferred to other lag windows as it provided better interpretation of the results. The amplitude, Table 1: Main tidal components at Zanzibar phase and nature of Harbour, 1 Jan 2003–31 Jan 2004. the meteorological SN Symbol Period (h/days) Amplitude (cm) Greenwich oscillations were related Phase (degrees) to fluctuations in sea level. 1 SA 365.3 days 2.0 84.5 STATISTICA was also 2 SSA 182.6 days 2.9 81.4 used to calculate multiple 3 MF 13.7 days 1.8 21.0 regressions of time series 4 Q1 26.9 h 2.5 31.9 of the meteorological 5 O1 25.8 h 11.2 42.7 parameters against 6 P1 24.1 h 5.6 39.3 monthly mean sea level 7 S1 24.0 h 2.7 0.2 (see e.g. Thompson, 1980; 8 K1 23.9 h 18.4 41.0 Abdelrahman, 1997; 9 2N2 12.9 h 2.1 44.0 Vilibić, 2006). Prior to 10 MU2 12.9 h 1.77 87.5 these calculations, the 11 N2 12.7 h 22.6 89.7 monthly sea level data 12 NU2 12.6 h 4.6 92.4 were isolated in terms of 13 M2 12.4 h 120.4 112.4 cyclical trends, seasonal 14 L2 12.2 h 4.3 120.2 effects, and remaining 15 T2 12.0 h 4.1 157.9 16 S2 12.0 h 60.9 154.6 (irregular) variability using 17 K2 12.0 h 16.7 151.6 the method known as 18 M4 6.2 h 1.56 354.8 seasonal decomposition, 19 2MN6 4.2 h 1.33 258.6 an algorithm described 20 M6 4.1 h 2.2 307.6

Volume 9 Final 20th Oct 2010.indd 9 10/27/2010 10:02:02 AM 6 S.B. MAHONGO & J. FRANCIS

RESULTS AND 2004 are listed in Table 2. Whereas the Ssa tides occur every year, the Sa tides DISCUSSION appear only once in every four years. Astronomic tides Monthly variations The results of harmonic analysis of tides Figure 2 illustrates monthly variations of amplitude >1 cm are listed in Table 1. in the meteorological and sea level data. The tides were strongly semidiurnal, with Atmospheric pressure, air temperature, a Form Factor F (Pugh, 2004) of 0.16. the southeast winds and rainfall clearly The corresponding mean spring tidal manifested annual variability with low range was 3.65 m, which compares well air temperatures corresponding to high with previous observations (Odido and air pressure and vice versa. The trends in Francis, 1999). Table 1 also shows that sea level and northeast winds displayed most of the constituent harmonics were inconsistent fluctuations throughout the diurnal and semi-diurnal. Results of the observation period. With the exception tidal harmonic analysis for long-period of air pressure and rainfall, all the other tides (Sa and Ssa) between 1985 and meteorological parameters exhibited Table 2: Amplitude and Greenwich Phases of long- increasing trends over period tidal components at Zanzibar. the 20-year study period Year Sa Ssa (1985-2004). The mean Amplitude (cm) Phase (deg) Amplitude (cm) Phase (deg) sea level has also been 1985 0 - 1.8 89.4 declining consistently as 1986 0 - 1.6 44.3 indicated by the linear 1987 0 - 5.1 87.2 regression and Mahongo 1988 3.6 128.1 0.9 332.5 (2009) has noted that 1989 0 - 3.7 74.0 this parameter has been 1990 0 - 3.4 70.7 declining at Zanzibar 1991 0 - 2.1 60.0 at the rate of 3.6 mm/yr 1992 3.0 109.0 1.0 70.5 during this period. 1993 0 - 1.5 107.9 Descriptive statistics 1994 0 - 2.2 157.4 of sea level and the 1995 0 - 1.2 25.1 meteorological parameters 1996 5.3 80.5 1.0 78.1 for the period 1985-2004 1997 0 - 2.8 99.5 are presented in Table 4. 1998 0 - 2.5 8.0 The range between the 1999 0 - 0.7 0.7 lowest and highest sea 2000 2.2 26.3 1.2 110.6 2001 0 - 2.7 106.2 levels (6.2 cm) is much 2002 0 - 3.1 53.0 smaller than that observed 2003 0 - 3.5 75.2 at higher latitude. For 2004 3.6 38.7 2.8 104.2 example, on the west

Volume 9 Final 20th Oct 2010.indd 10 10/27/2010 10:02:02 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 7

coast of India, the amplitude in annual were 3.5 cm and 2.3 cm in mean mean sea level is of the order of 20 cm amplitude at Zanzibar respectively. (Wijeratne et al., 2008) while, in the This conforms to an observation Red Sea, amplitudes of about 40 cm by Tsimplis & Woodworth (1994) have been recorded (Abdelrahman, that, generally, semi-annual tidal 1997). amplitudes are much smaller than The solar annual (Sa) and the solar the annual component. Differences in semi-annual (Ssa) tidal components atmospheric pressure at Zanzibar fell

Fig. 2: Monthly sea level and meteorological data for Zanzibar with linear regressions.

Volume 9 Final 20th Oct 2010.indd 11 10/27/2010 10:02:03 AM 8 S.B. MAHONGO & J. FRANCIS

within a range of 7.3 mb, this being small compared to higher latitudes such as the Arabian Sea, where differences >40 mb have been recorded (Sultan et al., 1995). The mean monthly air temperature was 26.5°C, ranging from a minimum of 24.7°C in July to a maximum of 28.2°C in February. Monthly means Monthly means in sea level and the m e t e o r o l o g i c a l and oceanographic parameters are presented in Figure 3. These were obtained by averaging monthly series, e.g. January, over all the years in the record. The Fig. 3: Monthly means of the sea level and associated highest sea levels meteorological parameters at Zanzibar (1985-2004). were observed at Tanga, which is closer to Zanzibar during March-May, while the lowest (Fig. 1), having the lowest levels in occurred during July-September. The February-May and the highest in July- highest and lowest mean sea levels September (Ragoonaden (1988). This were recorded in April and August, is rather surprising, since Tanga is respectively. sandwiched between Zanzibar and Dar The mean monthly variations in es Salaam in the south and Mombasa sea level at Zanzibar conformed to in the north. those of Dar es Salaam on mainland Mean variations representing the Tanzania and Kilindini harbour in sum of the Sa and Ssa tides were Mombasa, Kenya (Pattullo et al., highest in January and February, and 1955). However, the pattern differs lowest in October and November.

Volume 9 Final 20th Oct 2010.indd 12 10/27/2010 10:02:03 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 9

Table 3: Analysis of spectral components in sea level and meteorological and tidal data at Zanzibar. Amplitude Period (yrs) Sea level Air pressure Air temp Rainfall SE winds NE winds Sa tide Ssa tide N=240 (cm) (mb) (oC) (mm) (m/s) (m/s) (cm) (cm) (cm)

0.5 1.6 0.6 0.1 125.1 1.8 0.9 0.0 1.6 1 2.1 3.9 1.9 89.9 5.2 1.0 0.9 0.0 4 2.7 0.4 0.1 18.8 1.0 0.6 0.4 0.1

Phase (degrees) as of January Period (yrs) Sea level Air pressure Air temp Rainfall SE winds NE winds Sa tide Ssa tide N=240 0.5 149.5 5.5 196.3 145.8 159.2 315.5 31.3 342.6 1 290.9 164.3 333.4 296.7 169.6 66.5 310.0 66.7 4 163.1 254.0 144.3 269.0 175.0 139.8 145.3 41.4

The mean atmospheric pressures northeast winds, averaging 3.5 and 1.3 were highest during May-October, m/s, respectively (Table 4). whereas air pressures were lowest during November-April. As Spectral Analysis expected, the periods of high and The results of spectral analysis of low temperatures coincided with low the various parameters are presented and high pressure. The period with in Figure 4. In addition to annual the lowest mean air temperature (i.e. and semi-annual variations, the sea August) also corresponded with the level spectrum clearly displayed the lowest mean sea level. However, prominence of a long-period cycle higher air temperatures that occurred of four years. The annual and semi- in February did not correspond with annual oscillations were of the same higher sea levels. The total annual order of magnitude, while the 4-yearly rainfall over the study period averaged cycles were significantly larger. The 168 cm, with the heaviest rainfall in spectra of air pressure, air temperature April (42.3 cm) and the lowest in that influence the 4-year oscillations September (3.1 cm). The rainfall in sea level. Nevertheless, there was was highest during April and May, a cyclical trend in sea level which corresponding with the highest mean accounted for about 9% of the total sea level. Rainfall in the region is bi- variance in sea level, attributable and modal (falling twice a year), with a southeast winds largely manifested an second but smaller peak in November annual cycle. In addition to the annual and December. The southeast winds cycle, the southeast wind component were generally stronger than the also contained a small but significant

Volume 9 Final 20th Oct 2010.indd 13 10/27/2010 10:02:03 AM 10 S.B. MAHONGO & J. FRANCIS

Table 4: Descriptive statistics of monthly mean sea level and meteorological data at Zanzibar. N=240 Sea level Air pressure Air temp Rainfall SE winds NE winds (cm) (mb) oC (cm) m/s m/s Mean 204.1 1010.3 26.5 13.8 3.5 1.3 Min 201.1 1007.1 24.7 2.9 -2.3 0.1 Max 207.3 1014.4 28.2 40.7 6.9 3.5 Range 6.2 7.3 3.6 37.7 9.3 3.4 STD 4.5 2.8 1.3 14.9 4.0 3.0

semi-annual signal at 95% confidence level. The spectrum of the northeast winds was predominantly composed of annual and semi-annual signals, as well as numerous lower and higher frequency signals that were smaller. The rainfall spectrum contained three significant peaks (at 95% confidence level) at annual, semi-annual and 4-monthly periods with the semi-annual signal being more pronounced than the other two signals. Background noise of low energy oscillations in the sea level and rainfall spectra were of less significance. Amplitude and phase lags Table 3 lists the amplitude of and phase lags in the Fig. 4: Spectral analysis of the sea level and meteorological and sea level meteorological data (1985-2004). Horizontal variables corresponding to lines show the upper limit of the 95% confidence the spectral frequencies in sea level.

Volume 9 Final 20th Oct 2010.indd 14 10/27/2010 10:02:03 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 11

level. In the semi-annual cycle, changes were in phase) were comparatively in sea level were nearly in phase with small. Hence, apart from meteorology rainfall and southeast winds, but were and tides, it is very likely that other about 1.6 months later in phase than air physical processes significantly temperature. Air pressure and northeast influence the 4-year oscillations in sea winds were out of phase with changes level at Zanzibar. Seismic activity and in sea level by 4.8 and 5.5 months, oceanographic effects such as the El respectively. Rainfall had the highest Niño Southern Oscillation can also amplitude in the semi-annual cycle, yet contribute to variations in mean sea the corresponding amplitude in sea level level (Aung et al., 1998). was smaller than in the annual cycle. This suggests that other physical processes Multiple regression analysis are more important than rainfall as a Regression coefficients, normalized cause or proxy of the semi-annual cycle regression coefficients, semi-partial in sea level. The solar semi-annual tide correlations and percent variance in Ssa, with an amplitude of 1.6 cm (equal the meteorological data are listed in to that of sea level), was opposite in Table 5. The advantage of normalizing phase with sea level, preceding it by 6.4 the regression coefficients was to months. compare the relative contribution of In the annual cycle, changes in each meteorological variable in the sea level were almost in phase with prediction of sea level, regardless rainfall and the solar annual tide, Sa. of the units used. The semi-partial Again, changes in air temperature correlation provided a measure of the occurred 1.4 months earlier than correlation with each meteorological those of sea level. Southeast winds, variable after eliminating the effects air pressure and northeast winds were of the other variables. out of phase with changes in sea level In Table 5a-d, the variance in the by 4, 4.2 and 7.5 months respectively. meteorological variables fitted to In the 4-year cycle, southeast winds, that of sea level totalled 100%. The air temperature, northeast winds and annual rainfall had a regression rate the annual astronomic tides (Sa) were of 0.51 cm/10 cm. This implies that, almost in phase with changes in sea at the recorded maximum rainfall of level. Changes in sea level relative 42.3 cm, the sea level could increase to air pressure and rainfall were by 2.2 cm. In the Adriatic, Vilibić out of phase by 3 and 3.5 months, (2006) obtained a corresponding respectively. Changes in sea level were rate of 0.53 cm/10 cm. The seasonal of much higher amplitude in the 4-year cycle of sea level, represented by cycle than in the annual and semi- the sum of semi-annual and annual annual cycles, yet the amplitude of the oscillations, accounted for about 64% meteorological and tidal factors (that of the sea level variance at Zanzibar.

Volume 9 Final 20th Oct 2010.indd 15 10/27/2010 10:02:03 AM 12 S.B. MAHONGO & J. FRANCIS

Table 5: Results of multiple regressions of meteorological and sea level data (*not significant at 95% confidence level). (a) In the semi-annual cycle N=240, R = 0.64 Regression Normalized Semi-partial Percent variance Total variance = 28.2% coefficients regression correlations coefficients SE winds 0.203 ± 0.027 0.618 ± 0.081 0.39 14.9 Rainfall 0.023 ± 0.005 0.265 ± 0.062 0.22 4.6 NE winds -0.111 ± 0.027 -0.241 ± 0.058 -0.21 4.4 Air pressure -0.242 ± 0.059 -0.513 ± 0.124 -0.21 4.3 Air temperature* -0.039 ± 0.115 -0.039 ± 0.114 -0.02 0.0 (b) In the annual cycle N=240, R = 0.81 Regression Normalized Semi-partial Percent variance Total variance = 28.2% coefficients regression correlations coefficients Rainfall 0.051 ± 0.006 0.407 ± 0.047 0.33 11.0 Air temperature 1.022 ± 0.124 0.713 ± 0.087 0.31 9.9 SE winds 0.216± 0.029 0.460± 0.061 0.29 8.2 NE winds -0.192 ± 0.029 -0.292 ± 0.044 -0.25 6.5 Air pressure* -0.021 ± 0.064 -0.032 ± 0.094 -0.01 0.0 (c) In the 4-year cycle N=240, R = 0.62 Regression Normalized Semi-partial Percent variance Total variance = 28.2% coefficients regression correlations coefficients Air temperature 1.662 ± 0.245 0.789 ± 0.116 0.35 12.0 Rainfall 0.062 ± 0.012 0.336 ± 0.063 0.27 7.5 SE winds 0.275 ± 0.057 0.398± 0.082 0.25 6.1 Air pressure* 0.221 ± 0.126 0.223 ± 0.126 0.09 0.8 NE winds* -0.084 ± 0.057 -0.086 ± 0.059 -0.08 0.6 (d) In the trend cycle N=241, R = 0.33 Regression Normalized Semi-partial Percent variance Total variance = 28.2% coefficients regression correlations coefficients NE winds -0.378 ± 0.082 -0.327 ± 0.071 -0.3 8.1 Air pressure* -0.151 ± 0.181 -0.127 ± 0.152 -0.1 0.3 Air temperature* -0.278 ± 0.353 -0.110 ± 0.140 0.0 0.2 Rainfall* -0.012 ± 0.017 -0.052 ± 0.076 0.0 0.2 SE winds* -0.024 ± 0.082 -0.030 ± 0.099 0.0 0.0

Volume 9 Final 20th Oct 2010.indd 16 10/27/2010 10:02:03 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 13

In other areas such as the Arabian a subsequent drop in sea level. The Gulf, the corresponding variance is stronger winds observed over the 20- about 80% (Sultan et al., 1995). In the year study period have not had similar spectral analysis, the 4-year cycle was effects at Zanzibar; both the northeast significantly larger than the annual and southeast winds have intensified cycle (at the 95% confidence level), at Zanzibar, but the southeast winds in yet the fitted variance was smaller. the annual, semi-annual and the 4-year This again suggests the presence of cycles were associated with an increase factors other than meteorology largely in sea level. The declining trend in to northeast winds. sea level at Zanzibar is therefore best Air temperature, rainfall and represented by strengthening of the southeast winds appeared to increase northeast winds, which may have led the sea level as indicated by the sign to a decrease in the flow of the EACC of the regression and normalized and, consequently, a reduction in the regression coefficients (Table 5a- trend of sea level. According to Han et d). In contrast, the northeast winds al., (2010), the decrease in sea level at and air pressure appeared to depress Zanzibar and the south tropical Indian the sea level. According to Harvey Ocean is driven by changing surface (1977), the EACC flows faster during winds associated with a combined the SE monsoon (speed 1.5 - 2 m/s), invigoration of the Indian Ocean being accelerated by the SE winds. Hadley and Walker cells, which is During the NE monsoon, the current partly attributable to rising levels is retarded by the northeast winds of atmospheric greenhouse gases. to 0.5 m/s. The southeast monsoon Mid-latitude westerly winds have winds thus increase the speed strengthened in both hemispheres and consequently the geostrophic since the 1960s as a result of climate flow of the EACC, which flows change (Trenberth et al., 2007). permanently northwards parallel to the coast of Tanzania. As a result of CONCLUSIONS the Coriolis Effect, water is pumped In conclusion, analyses of the monthly into the Zanzibar Channel through the means of sea level and meteorological southern entrance, thereby increasing parameters during 1985-2004 at the sea level. In contrast, the northeast Zanzibar revealed a large proportion winds reduce the geostrophic flow of (64%) of the spectral signal to be the EACC, consequently reducing the seasonal, essentially comprising the sea level in Zanzibar Channel. sum of the annual and semi-annual In the Maldives, Mőrner et al., components. Seasonality was largely (2004) linked intensification of evidenced by changes in winds and the the northeast winds over a 30-year steric effect proxies of rainfall and air period to increased evaporation and temperature. A 4-year cycle accounted

Volume 9 Final 20th Oct 2010.indd 17 10/27/2010 10:02:04 AM 14 S.B. MAHONGO & J. FRANCIS

for about 27% of the variance in sea on sea level available. Prof. D.T. level and was best represented by air Pugh of the University of Liverpool temperature, rainfall, southeast winds (UK) provided useful comments and as well as the annual astronomic suggestions which have substantially tides. Long-term changes in sea level improved the manuscript. Lastly, constituted about 9% of the total we appreciate the support of Online variance, apparently due to changes in Access to Research in the Environment the northeast winds. This corroborated (OARE), and Access to Global Online an observation by Pattullo et al., Research in Agriculture (AGORA) (1955) that, in low latitudes, variations for assistance in acquiring pertinent in mean sea level are regulated to a references. very large extent by steric effects. Results of multiple regression REFERENCES analysis of the data generally Abdelrahman, S.M. (1997). Seasonal conformed to the results of the spectral fluctuations of mean sea level at Gizan, analyses. Discrepancies, especially in Red Sea. J. Coast. Res. 13: 1166-1178. the 4-yearly cycle, were presumably Aung, T. H., Kaluwin, C. & Lennon, G.W. attributable to the effects of other (1998). Climate Change and Sea Level, physical processes superimposed on Part 1: Physical Science, National Tidal the measured variables. The mean sea Facility, The Flinders University of level, winds (NE and SE monsoons) South Australia, Adelaide. 106 pp. and air temperature have been Bloomfield, P. (1976). Fourier analysis of changing at Zanzibar steadily over time series: An Introduction. John Wiley & Sons, New York. 288 pp. the 20-year study period, probably as a consequence of the changing global Caldwell, P. (1998) Sea Level Data Processing On IBM PC Compatible Computers climate. Version 3.0 (Year 2000 Compliant). Acknowledgments–We wish to JIMAR Contribution No. 98-319. National Oceanographic Data Center express our sincere gratitude to the and University of Hawaii Sea Level Western Indian Ocean Marine Science Center. 72pp. Association (WIOMSA) for providing Han, W., Meehl, G.A., Rajagopalan, B., a fellowship grant to present this Fasullo, J.T., Hu, A., Lin, J., Large, paper at the Fifth WIOMSA Scientific W.G., Wang, J., Quan, X., Trenary, L.L., Symposium in South Africa during Wallcraft, A., Shinoda, T. & Yeager, S. October 2007. We also thank the (2010) Patterns of Indian Ocean Sea- level Change in a Warming Climate. Tanzania Meteorological Agency for Nature Geosci. 3: 546 - 550. the provision of atmospheric data, and the University of Hawaii Sea Level Harvey, J. (1977) Some Aspects of The Hydrography of the Water off the Centre (UHSLC) for making the data coast of Tanzania: A Contribution to CINCWIO. Univ. Sci. J. (University of Dar es Salaam) 3: 53-92.

Volume 9 Final 20th Oct 2010.indd 18 10/27/2010 10:02:04 AM MONTHLY VARIATIONS IN SEA LEVEL AT THE ISLAND OF ZANZIBAR 15

Lisitzin, E. & Pattullo, J.G. (1961) The Ragoonaden, S. (1998) Mean Monthly Sea Principal Factors Influencing the Level Variation and its Relation to Seasonal Oscillation of Sea Level. J. Large-scale Ocean Circulation in the Geophy. Res. 66: 845-852. Southwest Indian Ocean. In: Sherman, K. Okemwa, E.N. & Ntiba, M.J. (eds.) Mahongo, S.B. (2009) The changing Global Large Marine Ecosystems of the Indian Climate and its Implication on Sea Level Ocean: Assessment, Sustainability and Trends in Tanzania and the Western Management. Blackwell Science, Inc. Indian Ocean region. Western Indian Malden, MA. pp. 193-214. Ocean J. Mar. Sci. 8: 147-159. Shaghude, Y. & Wannäs, K. (1998) Makridakis, S.G. & Wheelwright, S.C. (1978) Morphology and Sediment Distribution Interactive Forecasting: Univariate and of the Zanzibar Channel. Ambio 27: Multivariate Methods (2nd ed.). Holden- 729-733. Day, San Francisco, CA. 650 pp. Shaghude, Y., Wannäs, K. & Mahongo, Makridakis, S.G. & Wheelwright, S.C. (1989) S. (2002) Biogenic Assemblage and Forecasting Methods for Management Hydrodynamic Settings of the Tidally (5th ed.). John Wiley, New York. 241 pp. Dominated Reef Platform Sediments of Mayorga-Adame, C.G. (2007) Ocean the Zanzibar Channel. Western Indian Circulation of the Zanzibar Channel: A Ocean J. Mar. Sci. 1: 107-116. Modelling Approach. Technical Report Shumway, R.H. (1988) Applied Statistical for Theiss Research and IMS (Zanzibar). Time Series Analysis. Prentice Hall. La Jolla, CA. 8 pp. Englewood Cliffs, NJ. 379 pp. Mohammed, S., Ngusaru, A. & Mwaipopo, Sultan, S.A.R., Ahmad, F., Elghribi N.M. & O. (1993) Determination of the Effects Al-Subhi, A.M. (1995) An Analysis of of Pollutants on Coral Reefs around Arabian Gulf Monthly Mean Sea Level. Zanzibar Town. A Technical Report to Cont. Shelf Res. 15 (11/12): 1471-1482. NEMC. IMS, Zanzibar. 34 pp. Thompson, K.R. (1980) An Analysis of Mörner, N., Tooley, M & Possnert, G. (2004) British Monthly Mean Sea Level. New Perspectives for the Future of the Geophys. J. R. Astron. Soc. 63: 57-13. Maldives. Global Planet. Change 40: 177–182. Trenberth, K.E., Jones, P.D., Ambenje, P., Bojariu, R., Easterling, D., Klein Odido, M. & Francis, J. (1999) Sea Level Tank, A., Parker, D., Rahimzadeh, F., Measurement and Analysis in the Renwick, J.A., Rusticucci, M., Soden, B. Western Indian Ocean. Regional Report & Zhai, P. (2007) Observations: Surface for IOC/UNESCO. 59 pp. and Atmospheric Climate Change. In: Oppenheim, A.V. & Schafer, R.W. (1989) Solomon, S., Qin, D., Manning, M., Discrete-Time Signal Processing, Chen, Z., Marquis, M., Averyt, K.B., Prentice-Hall. 870 pp. Tignor, M. & Miller, H.L. (eds.) Climate Pattullo, J.G., W. Munk, R. Ravelle & Strong, Change 2007: The Physical Science E. (1955) The Seasonal Oscillations in Basis. Contribution of Working Group I Sea level. J. Mar. Res. 14: 88-156. to the IPCC AR4. Cambridge University Press. Cambridge, UK and New York, Pugh, D.T. (2004) Changing Sea Levels: NY, USA. pp. 235-336. Effects of Tides, Weather, and Climate. Cambridge University Press, 265 pp.

Volume 9 Final 20th Oct 2010.indd 19 10/27/2010 10:02:04 AM 16 S.B. MAHONGO & J. FRANCIS

Tsimplis, M.N. & Woodworth, P.L. (1994) The Global Distribution of the seasonal Sea Level Cycle Calculated from Coastal Tide Gauge Data. J. Geophys. Res. 99 (c8): 16031-16039. Vilibić, I. (2006) Seasonal Sea Level Variations in the Adriatic. Acta Adriat. 47: 141-158. Wijeratne, E.M.S., Woodworth, P.L. & Stepanov, V.N. (2008) The seasonal Cycle of Sea Level in Sri Lanka and Southern India. Western Indian Ocean J. Mar. Sci. 7(1): 29-43.

Volume 9 Final 20th Oct 2010.indd 20 10/27/2010 10:02:04 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 17 -30, 2010 © 2010 WIOMSA

Coastal Marine Pollution in Dar es Salaam (Tanzania) relative to Recommended Environmental Quality Targets for the Western Indian Ocean

J.F. Machiwa Department of Aquatic Sciences and Fisheries, University of Dar es Salaam, PO Box 35064, Dar es Salaam, Tanzania.

Keywords: Dar es Salaam, environmental quality targets, heavy metals, marine pollution, microbial pollution, persistent organic pollutants.

Abstract—Pollution surveys were undertaken during 2007 and 2008 in the coastal marine environment of Dar es Salaam and the remote Ras Dege Creek. The objective was to determine the levels of microbial contamination, heavy metals and persistent organic pollutants and compare these with the recommended environmental quality targets (EQTs) for the West Indian Ocean (WIO). Levels of microbial pollution in urban coastal waters off Dar es Salaam were excessive, indicating that water within the port channel was not safe for contact recreation. Seafood from areas adjacent to Msimbazi Creek and the Ocean Road sewer outfall was unfit for human consumption. Conversely, the water quality of Ras Dege Creek was excellent for contact recreation as well as for the collection of seafood. Concentrations of heavy metals, even in the coastal marine environment off Dar es Salaam, were not significantly high compared with the recommended EQTs. Although some persistent organic pollutants exceeded the recommended EQTs in sediment and oysters along the coast of Dar es Salaam, this was not the case at Ras Dege. The lack of sufficient wastewater treatment facilities is the main cause of current levels of some pollution in the coastal marine environment off Dar es Salaam. The implementation of industrial and municipal wastewater management would greatly improve this situation. The results show that the proposed EQTs would constitute appropriate standards for coastal marine water quality in Tanzania.

E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 21 10/27/2010 10:02:04 AM 18 J.F. MACHIWA

INTRODUCTION battery industries; motor vehicle servicing; glass product manufacture; The WIO region is composed of ginnery and tobacco processing; five mainland states (Somalia, and mining. In addition, tourism and Kenya, Tanzania, Mozambique and the hospitality industry are centred South Africa) and four island states around the city’s beaches. The marine (Comoros, Madagascar, Mauritius and environment off Dar es Salaam Seychelles) with a combined coastline receives pollutants from these sources of more than 15,000 km. The region through streams, sewer and storm has a wide variety of marine habitats, water outfalls. Some of the streams such as coral reefs, estuaries, lagoons, also drain the densely-populated areas sandy beaches, mangrove wetlands and of the city. Even though the level of seagrass meadows. For generations industrialization is still low, untreated the coastal area has sustained coastal municipal and industrial wastewater communities and fisheries have discharges cause significant localised played a major role in their social pollution (Machiwa, 1992, Machiwa, and economic development. In recent years, coastal tourism and mariculture have become important economic activities in the WIO region. The sustainability of these activities and the coastal marine habitats and resources they depend on relies, to a large extent, on a healthy environment. In Tanzania, about 80% of its industries are located in Dar es Salaam. These include: Food, beverage and animal feed processing; chemical and cosmetic production; metal product manufacture; paper product, printing and publishing industries; wood product Fig. 1. Map of Dar es Salaam showing the sampling manufacture and construction stations in the coastal marine areas (Station 1- Mtoni Creek; Station 2 – Port entrance channel; Station 3 materials production; – Ocean Road; Station 4 – Msimbazi Creek; Station electrical appliance and 5 – Ras Dege Creek).

Volume 9 Final 20th Oct 2010.indd 22 10/27/2010 10:02:04 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 19

2000, DeWolf et al., 2001, Mwevura annual growth rate of about 8%. et al., 2002, Mremi & Machiwa, About 70% of the urban population 2003; Mtanga & Machiwa, 2007). live in unplanned settlements (NBS, The coastline of Dar es Salaam 2004, 2009). Ras Dege Creek (RDC), from Msimbazi Creek to Mzinga situated about 100 km from the city Creek is considered to be a pollution of Dar es Salaam, was sampled as an hotspot (Mohammed et al., 2006). unpopulated reference location remote The Msimbazi River drains a large from the city. proportion of the city’s industrial and residential area, while the Sample collection and analysis Mzinga and Msimbazi rivers receive Samples were collected in 2007 and domestic waste and effluents from 2008 at four sampling stations adjacent industries located along Nyerere, to wastewater or river discharges Morogoro and Mandela Roads. Apart in Dar es Salaam and one station in from these land-based sources of RDC (Fig. 1). Station 1 was located pollution, pollutants are also directly at Mtoni (MTC) near the confluence introduced to the coastal waters from of the Mzinga and Kizinga Creeks, shipping activities. The purpose of Station 2 was located within the port this paper was to present current data entrance channel (PEC), Station 3 on pollution in the coastal marine was located in the subtidal zone at environment of Dar es Salaam relative Ocean Road (ORA), and Station 4 to the recommended environmental was located in the subtidal zone near quality targets (EQTs) for the Western Msimbazi Creek (MSC). Station 5 Indian Ocean (UNEP & CSIR, 2009). was the reference site located within RDC. MATERIALS AND METHODS Water samples for microbial analysis were taken from the surface Study site using 250 ml sterile serum bottles The coastline of Dar es Salaam is rinsed with sea water from at the located between latitudes 6º27´and sampling sites. The bottles were 7º15´S and longitudes 39º and immersed below the water surface and 39º33´E, extending about 100 km turned upright to collect the samples. from the Mpiji River in the north to Water samples were immediately the Mzinga River in the south (Fig. 1). transported to the laboratory in a cool The Dar es Salaam Region includes box for analysis of total coliforms, the Ilala, Kinondoni and Temeke faecal coliforms, E. coli, enterococci municipal districts, with a total land (faecal streptococci), Salmonella and area of about 1,350 km2. Dar es Salaam Staphylococcus aureas. The analysis has an estimated population of about of faecal bacteria concentrations was three million people with an average based on most probable numbers

Volume 9 Final 20th Oct 2010.indd 23 10/27/2010 10:02:04 AM 20 J.F. MACHIWA

Fig. 2. Mean (± SD) concentration of (a) copper, lead and nickel (b) chromium and zinc, and (c) cadmium and total mercury in sediment samples with their Recommended Environmental Quality Targets (EQTs).

Volume 9 Final 20th Oct 2010.indd 24 10/27/2010 10:02:04 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 21

(MPNs) according to the standard soft tissues were removed and oven- methods (APHA, 2005). dried at 50ºC. The dried oyster tissue Subtidal sediment samples were was ground into fine powder using a collected using a van Veen grab mortar and pestle. and placed in clean glass bottles for About 2 g sub-samples of finely organic contaminants analysis and in powdered sediment samples and 1 g polyethylene bags for metals analysis. of well-homogenized oyster tissue The samples were kept frozen were accurately weighed and placed in until they were analysed. Sediment separate 100 ml flasks. The sediment samples were oven-dried at 60º C, and oyster samples were digested

sieved through 1 mm nylon sieve using aqua regia (HNO3 : HCl, 1:3 and thoroughly homogenized before v/v) and analyzed for copper (Cu), analysis. cadmium (Cd), lead (Pb), chromium Oyster samples (Saccostrea (Cr), zinc (Zn) and nickel (Ni) using cucullata) were collected from Mtoni ICP-OES (model Vista MPXTM ) in Creek and RDC mangrove stands. accordance with Thomson and Wash Samples for heavy metal analysis (2003). The concentrations of these were stored in polyethylene bags metals were recorded per unit dry and those for persistent organic weight. pollutant analysis were kept in pre- Total mercury (Hg) was determined cleaned aluminium foil. Samples in the sediment and undried oyster were transported to the laboratory in flesh following the procedure of Akagi a cool box where they were washed and Nishimura (1991). Mercury was with distilled water, weighed and analysed using a semi-automatic kept frozen at –20ºC until processed. mercury analyser (Sanso Seisakusho The frozen oyster samples were Co. Ltd model HG 201). thawed before they were opened, the Table 1. Microbial counts (MPN/100 ml) in coastal waters at various locations in Dar es Salaam and Ras Dege. The 95th percentile for Enterococcus/100 ml is in parentheses.

Site Total Faecal Escherichia Enterococcus coliforms coliforms coli

MTC 79 – 2 400 33 – 1 600 4 – 17 133 – 1 340 (1034)

PEC 49 – 240 23 – 49 4 – 9 17 – 33 (31)

ORA 3 500 – 24 000 1 800 – 3 500 7 – 70 110 – 920 (680)

MSC 9 200 – 24 000 2 400 – 9 200 9 – 17 33 – 220 (220)

RDC 13 – 30 6 – 14 0 – 7 5 – 10 (10)

Volume 9 Final 20th Oct 2010.indd 25 10/27/2010 10:02:05 AM 22 J.F. MACHIWA

Table 2. Recommended risk-based Environmental Quality Targets (EQTs) for application in coastal waters used for contact recreation in the WIO region based on estimated risk per exposure to gastrointestinal illness (GI) and acute febrile respiratory illness (AFRI).

Risk Category 95th Percentile of Estimated risk enterococci/100ml per exposure A (very good) <40 <1% GI risk; <0.3% AFRI risk B (good) 40 – 200 1 – 5% GI risk 0.3 – 1.9% AFRI risk C (fair) 201 – 500 5 – 10% GI risk 1.9 – 3.9% AFRI risk D (poor) >500 >10% GI risk >3.9% AFRI risk Quality control for metal analysis 417 (sediment). Recoveries ranged included procedural blanks, and between 60 % and 110 %. measurement of the Certified Persistent organic pollutants were Reference Materials (CRMs): extracted from sediment and oyster DORM-2 (Dogfish muscle) DOLT- tissue samples in dichloromethane 2 (Dogfish liver) PACS–2 (Marine using Soxhlet apparatus. Samples for sediment) from the National Research petroleum hydrocarbon determination Council, Canada, and IAEA–436 were cleaned using pre-heated silica from the International Atomic Energy gel, alumina and activated copper. Agency. The recovery values for the (Fig. 2b), the former value being analysed metals were above 80%. above the recommended EQT (52.3 Certified reference materials used for mg kg-1). This may pose a threat to organic pollutants were IAEA–432 marine life in the area. Indeed, the (mussel homogenate) and IAEA–

Table 3. Mean concentration (± SD) of total mercury (mg kg-1 ww) and other metals (mg kg-1 dw) in oysters from Mtoni and Ras Dege compared with FAO/WHO (2002) Codex alimentarius guideline levels for contaminants and toxins in food.

Mtoni Ras Dege FAO/WHO Cadmium 0.79 ± 0.21 0.04 ± 0.01 1.0 Chromium 24.3 ± 4.6 0.3 ± 0.1 - Copper 38.8 ± 9.1 0.5 ± 0.2 - Lead 1.2 ± 0.8 0.3 ± 0.2 1.0 Zinc 78.4 ± 16.3 21.9 ± 4.5 - Mercury 0.536 ± 0.036 0.055 ± 0.026 0.5

Volume 9 Final 20th Oct 2010.indd 26 10/27/2010 10:02:05 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 23

concentration of Cr in oyster samples Apart from copper and lead, these from Mtoni was an order of magnitude results suggest that the current levels higher than samples from RDC (Table of heavy metals in the marine coastal 3), suggesting that oysters at Mtoni waters off Dar es Salaam are not accumulate Cr from the environment, excessive relative to the recommended affecting their quality as seafood. The EQTs. It is unfortunate that land-based concentration of Zn in sediment was sources of heavy metal pollution cannot highest (71.2 mg kg-1) in samples be pinpointed because of lack of data from Mtoni (Fig. 2b) but was below on the quality of industrial effluents. the recommended EQT (124 mg kg-1). However, the nearby refuse dumpsite Oyster samples from Mtoni similarly is a likely source of some of these had higher concentrations (78.4 ± 16.3 heavy metal pollutants. Inappropriate mg kg-1) of Zn, about four times higher industrial wastewater handling is the than samples from RDC (Table 3). other major cause of the elevated The nearby solid waste disposal site levels of some heavy metals. The is a potential source of this element. establishment and implementation of Concentrations of Cd in sediment an industrial wastewater management samples from Mtoni, the port entrance framework would improve the quality channel, Ocean Road and RDC were of industrial wastewater, and hence the below the detection limit (<0.01 mg quality of the Dar es Salaam coastal kg-1 dw), and detectable amounts of marine environment. Cd in samples from Msimbazi were well below the recommended EQT Persistent organic pollutants in (Fig. 2c). Cadmium in oyster samples sediment and oysters from Mtoni was 0.79 ± 0.21 mg kg-1 Organochlorine pesticide residues dw, and 0.04 ± 0.01 mg kg-1 in samples were generally higher in sediment from RDC. The relatively high and oyster samples from Dar es concentration of Cd in oyster samples Salaam, compared with RDC from Mtoni suggests contamination of (Table 4). The concentration of total the water. Concentrations of Hg were dichloro-diphenyl trichloroethane generally below the recommended (DDT), however, was lower in EQT for sediment (0.13 mg kg- sediment and oysters compared with 1) except at Mtoni where the Hg its metabolites DDD and DDE, for concentration (0.093 ± 0.033 mg kg- which concentrations were above 1) approached the recommended EQT the recommended EQTs in sediment. (Fig. 2c). The concentration of Hg in Apart from its previous use in the oysters from Mtoni also approached control of malaria, the insecticide the upper limit recommended in FAO/ DDT has also been used in vegetable WHO guidelines (Table 3). gardens in the city (Mwevura et

Volume 9 Final 20th Oct 2010.indd 27 10/27/2010 10:02:05 AM 24 J.F. MACHIWA

al., 2002). Other organochlorine water quality where shellfish are compounds like the polychlorinated collected for direct human consumption biphenyls (PCBs) Samples for (UNEP & CSIR, 2009). The median polycyclic aromatic hydrocarbons concentration of faecal coliforms (PCBs) and organochlorine pesticides should not exceed 14 MPN/100ml were analysed according to the in a 5 tube, 3-dilutions method. In methods described by Brownawell coastal waters that are used for contact & Farrington (1986) and Åkerblom recreation, risk-based EQTs have been (1995). PCBs, pesticides and petroleum recommended (UNEP & CSIR, 2009; hydrocarbons were analysed by high WHO, 2003; Kay et al., 2004). The resolution dual capillary column gas risk levels and a range of target values chromatography (Hewlett-Packard are presented in Table 2. Enterococci 5840) using ECD/FID and confirmed (faecal streptococci) are proposed as by GC/MS. the microbial indicator for recreational waters, and are considered to be the RESULTS AND DISCUSSION most suitable indicator for marine waters (although in the case where Microbial contamination in seawater faecal pollution originates from a The highest counts of faecal coliforms, waste stabilisation pond, E. coli may enterococci and total coliforms were be the more appropriate indicator). found in water samples from Mtoni, Based on the recommended risk- Ocean Road and Msimbazi (Table 1). based EQTs for microbial water quality Total and faecal coliform counts were assessment, it is clear that coastal highest in samples from Msimbazi waters off the city of Dar es Salaam, followed by samples from the Ocean from Msimbazi Creek southward Road area. It was clear that the cause to Mtoni, are not safe for contact of high microbial pollution at the two recreation or the collection of seafood locations was untreated wastewater for direct human consumption. RDC, discharge; there is a sewer discharge an area which is remote from the city extending about 150 m into the of Dar es Salaam, has low levels of water at Ocean Road beach and the microbial contamination. Based on outfall does not extend far enough the microbial indicator results, the for offshore transport to prevent water quality of RDC is good for pollution of the inshore water. High contact recreation and the collection levels of contamination in Msimbazi of shellfish for human consumption. Creek are also caused by a significant Generally, the marine areas where wastewater discharge. microorganism densities were high are The proposed EQTs recommend located close to point pollution sources. that faecal coliforms be the preferred Proper management of municipal and microbial indicator to assess coastal industrial wastewater in urban Dar es

Volume 9 Final 20th Oct 2010.indd 28 10/27/2010 10:02:05 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 25

Salaam would improve the quality The concentration of Pb was of its wastewater, thus reducing the similarly higher in sediment samples level of the microbial pollutants in the from Mtoni and the port entrance receiving coastal waters. channel compared to the rest of the sites (Fig. 2a). A few stations in the Heavy metals in sediment and oysters port entrance channel had Pb levels The concentrations of metals measured higher than the EQT (30.2 mg kg- in sediment samples are presented 1) for this element. Concentrations in Figure 2a-c. The concentration of Pb in oyster samples were 1.2 ± of Cu in sediment samples from 0.8 mg kg-1 and 0.3 ± 0.2 mg kg-1 in Mtoni (MTC) (28.8 ± 5.5 mg kg-1 samples from Mtoni and the RDC, dw) and the port entrance channel respectively. The Pb level in oysters (PEC) (25.7 ± 10.2 mg kg-1 dw) was from Mtoni was slightly above the about twofold the concentration in guideline recommended by the FAO/ samples from RDC. This is excessive WHO (2002). The potential sources relative to the recommended EQT for of Pb in the coastal waters off Dar es Cu (18.7 mg kg-1 dw) (Fig. 2a). The Salaam include contaminated storm concentration of Cu was low in the water runoff from the city roads and sediment samples from Ocean Road garages as a result of previous usage and Msimbazi Creek, despite the high of leaded petrol. Further contributions input of wastewater at these sites. The possibly come from runoff from paint concentration of Cu in oyster samples shops and a nearby dumpsite. from Mtoni was higher compared to Nickel concentrations in sediments oyster samples for the RDC (Table 3), were low relative to the recommended suggesting the accumulation of this EQT (Fig. 2a). This is not element in these filter-feeding marine uncommon since Ni is infrequently organisms. used in industrial manufacturing. Previous studies also reported high Concentrations of Cr in sediment Cu concentrations in sediment in the samples from Mtoni and the port entrance channel were 57.6 ± 8.2 and mangrove wetlands close to Dar es -1 Salaam compared with mangrove 35.6 ± 6.1 mg kg respectively (Fig. 2b), the former value being above the stands in areas remote from the city -1 (DeWolf et al., 2001, Mremi and recommended EQT (52.3 mg kg ). Machiwa, 2003). The potential sources This may pose a threat to marine life of Cu to the coastal waters include in the area. Indeed, the concentration the various industries in the city but of Cr in oyster samples from Mtoni it is unfortunate that the available was an order of magnitude higher data does not implicate any specific than samples from RDC (Table 3), industry in view of the high levels of suggesting that oysters at Mtoni Cu in the sediment. accumulate Cr from the environment,

Volume 9 Final 20th Oct 2010.indd 29 10/27/2010 10:02:05 AM 26 J.F. MACHIWA

affecting their quality as seafood. The sources of heavy metal pollution cannot concentration of Zn in sediment was be pinpointed because of lack of data highest (71.2 mg kg-1) in samples on the quality of industrial effluents. from Mtoni (Fig. 2b) but was below However, the nearby refuse dumpsite the recommended EQT (124 mg kg-1). is a likely source of some of these Oyster samples from Mtoni similarly heavy metal pollutants. Inappropriate had higher concentrations (78.4 ± 16.3 industrial wastewater handling is the mg kg-1) of Zn, about four times higher other major cause of the elevated than samples from RDC (Table 3). levels of some heavy metals. The The nearby solid waste disposal site establishment and implementation of is a potential source of this element. an industrial wastewater management Concentrations of Cd in sediment framework would improve the quality samples from Mtoni, the port entrance of industrial wastewater, and hence the channel, Ocean Road and RDC were quality of the Dar es Salaam coastal below the detection limit (<0.01 mg marine environment. kg-1 dw), and detectable amounts of Cd in samples from Msimbazi were Persistent organic pollutants in well below the recommended EQT sediment and oysters (Fig. 2c). Cadmium in oyster samples Organochlorine pesticide residues from Mtoni was 0.79 ± 0.21 mg kg-1 were generally higher in sediment dw, and 0.04 ± 0.01 mg kg-1 in samples and oyster samples from Dar es from RDC. The relatively high Salaam, compared with RDC concentration of Cd in oyster samples (Table 4). The concentration of total from Mtoni suggests contamination of dichloro-diphenyl trichloroethane the water. Concentrations of Hg were (DDT), however, was lower in generally below the recommended sediment and oysters compared with EQT for sediment (0.13 mg kg- its metabolites DDD and DDE, for 1) except at Mtoni where the Hg which concentrations were above the concentration (0.093 ± 0.033 mg kg- recommended EQTs in sediment. Apart 1) approached the recommended EQT from its previous use in the control of (Fig. 2c). The concentration of Hg in malaria, the insecticide DDT has also oysters from Mtoni also approached been used in vegetable gardens in the the upper limit recommended in FAO/ city (Mwevura et al., 2002). Other WHO guidelines (Table 3). organochlorine compounds like the Apart from copper and lead, these polychlorinated biphenyls (PCBs) results suggest that the current levels were low in concentration, below the of heavy metals in the marine coastal recommended EQTs for sediment waters off Dar es Salaam are not (Table 4). Sediment and oyster samples excessive relative to the recommended from RDC were not contaminated EQTs. It is unfortunate that land-based with any of these pesticides.

Volume 9 Final 20th Oct 2010.indd 30 10/27/2010 10:02:05 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 27

Table 4. Concentration (µg kg-1 ± SD dw ) of organic pollutants in sediment and oyster from Dar es Salaam and Ras Dege with the Recommended Environmental Quality Targets (EQTs) for sediment. Values that exceed the EQT are in bold. Organic Dar es Salaam Ras Dege EQTs Pollutant Sediment Oyster Sediment Oyster (sediment) Organochlorine pesticides Total DDTs <1.0 2.1±1.1 <1.0 <1.0 3.89 DDD 6.5±3.4 9.2±2.7 <1.0 <1.0 - DDE 7.1±1.6 31.3±11.4 <1.0 <1.0 2.2 Dieldrin 1.0±1.0 3.0 <1.0 <1.0 0.72 Polychlorinated biphenyls Total PCBs 11.2±2.1 - - - 21.6 Polycyclic Aromatic 27.4±8.5 45.1±11.9 10.7±3.6 12.6 34.6 Hydrocarbons (PAHs) Naphthalene Acenaphthylene <1.0 <1.0 <1.0 <1.0 -

Acenaphthene 6.5±0.5 10.5±3.5 3.0±2.0 2.6 6.71 Fluorene 21.0±3.8 13.5±4.5 10.0±6.4 3.0 21.2 Phenanthrene 55.0±46.0 11.0±2.8 46.5±33.2 20.3 86.7 Anthracene 39.2±28.7 1.0 30.0±14.1 1.5 46.9 Chrysene 43.0±15.4 8.5±0.7 6.1±2.8 1.0 108 Fluoranthene 59.2±47.4 4.5±2.1 42.5±19.5 6.4 113 Pyrene 38.2±14.9 20.0±4.2 26.0±13.2 16.6 153 Benzo[a]anthracene 53.6±17.2 2.5±0.7 5.0±2.8 1.0 74.8 Benzo[b]fluoranthene 61.3±31.3 4.0 2.5±2.1 1.1 - Benzo[a]pyrene 24.2±9.7 <1.0 2.0±1.4 <1.0 88.8 Indeno[c,d]pyrene 38.7±13.3 1.0 4.5±2.5 <1.0 - Dibenzo[a,h]anthracene 3.5±2.1 1.0 <1.0 <1.0 6.22 Benzo[g,h,i]perylene 4.3±1.9 1.5±0.7 2.5±1.3 <1.0 -

Polyaromatic hydrocarbons were below the recommended EQT (PAHs), resulting mainly from (Table 4). Oyster samples did not petroleum spillage, were detected in contain detectable concentrations sediment and oyster samples from Dar of this compound. Concentrations es Salaam and Ras Dege. Benzo[a] of phenanthrene, anthracene and pyrene, a potent carcinogen and an fluoranthene were higher in the indicator of petroleum hydrocarbon subtidal sediments off Dar es Salaam pollution, was detected in sediment than in samples from RDC but below samples from both Dar es Salaam the recommended EQTs. and RDC. However, the levels

Volume 9 Final 20th Oct 2010.indd 31 10/27/2010 10:02:05 AM 28 J.F. MACHIWA

The presence of petroleum exceeded the recommended EQTs, hydrocarbons in the marine concentrations of persistent organic environment is presumed to result pollutants in sediments and oyster from shipping activities including were below the recommended EQTs accidental spillage, runoff containing in the rural area (Ras Dege). oil from service stations and garages It is obvious that insufficient and municipal wastewater. Even wastewater treatment facilities are though levels of most petroleum the main cause of the current elevated hydrocarbons were below the levels of some pollutants in Dar recommended EQTs, efforts should be es Salaam. The establishment and made to control accidental spillages implementation of industrial and as well as reduce the levels of these municipal wastewater management compounds in wastewater and runoff would greatly enhance the quality from the city. of the coastal marine environment of the urban areas of the city. CONCLUSION Furthermore, this study showed that the recommended EQTs for the WIO Based on a comparison with proposed region, even though developed at a EQTs for the WIO region, results more generic level, are applicable to from this study suggest that the level the coastal marine environment of of microbial pollution in the coastal Tanzania and would provide standards marine environment adjacent to urban for wastewater disposal. areas of Dar es Salaam is excessive. In fact, the current status indicates that Acknowledgments:The UNEP- the waters within the port channel were GEF WIO-LaB Project addressing not safe for contact recreation. Seafood land-based activities and sources of from areas adjacent to Msimbazi Creek pollution provided funds for logistics, and the Ocean Road sewer outfall was sample collection and purchase unfit for human consumption due to of the necessary chemicals and high microbiological counts. The standards. The Project Management water quality in RDC, situated in a Unit, headed by Dr. Peter Scheren, rural area, was excellent for contact is thanked for this timely support. recreation as well as the collection Ms. Lydia Gaspare and other of seafood for human consumption. technical staff in the former Faculty Concentrations of heavy metals at of Aquatic Sciences and Technology, all sites were not significantly high University of Dar es Salaam, assisted relative to the recommended EQTs. with the sample collection and Although some persistent organic preparation. Analysis of pollutants pollutants in sediment and oysters was undertaken at the University of along the coast of Dar es Salaam Dar es Salaam and at the Centre for

Volume 9 Final 20th Oct 2010.indd 32 10/27/2010 10:02:05 AM COASTAL MARINE POLLUTION RELATIVE TO RECOMMENDED ENVIRONMENTAL QUALITY TARGETS 29

Environmental Research(Zentrum für Kay, D., Bartram, J., Prüss, A., Ashbolt, N., Umweltforschung; UFZ) in Germany. Wyer, M.D., Fleisher, J.M., Fewtrell, L., Anonymous reviewers are thanked for Rogers, A. & Rees, G. (2004) Derivation of Numerical Values for the WHO their useful comments. Guidelines for Recreational Waters. Water Res. 38: 1296-304. REFERENCES Machiwa, J.F. (1992) Anthropogenic Pollution in the Dar es Salaam harbour Akagi, H. & Nishimura, B. (1991) Speciation area, Tanzania. Mar. Pollut. Bull. 24: 562- of Mercury in the Environment In: Suzuki, 567. T., Imura, N. & Clarkson, T.W. (eds.) Machiwa, J.F. (2000) Heavy Metals and Advances in Mercury Toxicology. Plenum Organic Pollutants in Sediments of Dar es Press, New York. pp 53-76. Salaam harbour Prior to Dredging in 1999. Åkerblom, M. (1995) Environmental Tanzania J. Sci. 26: 29-45. Monitoring of Pesticide Residues: Mohammed, S.M., Machiwa, J., Njau, K.N. & Guideline for SADC Region: Monitoring Mato, R.R.A.M. (2006) Tanzania National Technique Series, vol. 3. Southern Africa Status Report on Priority Land-Based Development Community, Environment Sources of Pollution and Pollutant Levels and Land Management Sector, Maseru, in Water and Sediment. Report submitted 141 pp. to WIO-LaB Project Management Unit, APHA/AWWA/WEF (2005) Standard UNEP, Nairobi, Kenya. 111 pp. Methods for the Examination of Water & Mremi, S.D. & Machiwa, J.F. (2003) Heavy st and Wastewater, 21 edition, American Metal Contamination Of Mangrove Public Health Association, American Sediments and the Associated Biota in Dar Water Works Association and Water es Salaam, Tanzania. Tanzania J. Sci. 29: Environment Federation. Washington, 61-75. DC. 1368 pp. Mtanga, A. & Machiwa, J. (2007) Assessment Brownawell, B.J. & Farrington, J.W. (1986) of Heavy Metal Pollution in Sediment and Biogeochemistry of PCBs in Interstitial Polychaete Worms from the Mzinga Creek Waters of Coastal Marine Sediment. and Ras Dege Mangrove Ecosystems, Geochim. et Cosmochim. Acta 50: 157- Dar es Salaam, Tanzania. Western Indian 169. Ocean J. Mar. Sci. 6: 125-135. DeWolf, H., Ulomi, S.A., Backeljau, T., Mwevura, H., Othman, O.C. & Mhehe, G.L. Pratap, H.B. & Blust, R. (2001) Heavy (2002) Organochlorine Pesticides Residue Metal Levels in Sediments of Four in Edible Biota from the coastal area of Mangroves: Accumulation in and Effect Dar es Salaam city. Western Indian Ocean on the Morphology of the Periwinkle, J. Mar. Sci. 1: 91-96. Littolaria scabra (Mollusca: Gastropoda). NBS (2004) The 2002 Population and Housing Environ. Int. 26: 243-246. Census. National Bureau of Statistics, FAO/WHO 2002. General Standards for Ministry of Finance and Economic Affairs, Contaminants and Toxins in Food and Dar es Salaam, Tanzania. www.nbs.go.tz. Feed. Codex Alimentarius Commission. NBS (2009) Tanzania in Figures 2008. Joint Food and Agriculture Organization National Bureau of Statistics, Ministry of the United Nations and World Health of Finance and Economic Affairs, Dar es Organisation, Food Standards Programme, Salaam, Tanzania. www.nbs.go.tz. Rome. 44 pp.

Volume 9 Final 20th Oct 2010.indd 33 10/27/2010 10:02:05 AM 30 J.F. MACHIWA

Thompson, M. & Walsh, J.N. (2003) The Handbook of Inductively Coupled Plasma spectrometry. Viridian Publishing, Surrey. 316 pp. UNEP & CSIR (2009) Guidelines for the Establishment of Environmental Quality Objectives and Targets in the Coastal Zone of the Western Indian Ocean (WIO) Region. WIO-LaB Technical Report Series No. 01/2009, United Nations Environment Programme (Nairobi Convention Secretariat), Kenya, Nairobi. 146 pp. WHO (2003) Guidelines for Safe Recreational Water Environments, Volume 1: Coastal and Freshwater, Geneva. ISBN 9241545801. www.who.int/water-sanitation-health/ bathing/srwel/en 33pp.

Volume 9 Final 20th Oct 2010.indd 34 10/27/2010 10:02:05 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 31 - 42, 2010 © 2010 WIOMSA

Bioavailability of Sediment-bound Heavy Metals on the East African Coast

E.O. Okuku1,2, V.K. Mubiana2, K.G. Hagos3,2, H.K. Peter4,2 and R. Blust2 1Kenya Marine and Fisheries Research Institute, P.O. Box 81651, Mombasa, Kenya; 2Department of Biology, University of Antwerp, Groenenborgerlaan 171, B-2020, Antwerp, Belgium; 3Ministry of Fisheries, P.O. Box 932, Asmara, Eritrea; 4Tanzania Fisheries Research Institute, P.O. Box 9750, Dar es Salaam, Tanzania.

Keywords: Metals, sediments, bioavailability, aqua regia extraction, sequential extraction.

Abstract—Currently, environmental risk assessment of metals is based on comparisons of toxicity thresholds against environmental exposure levels measured as total metals, despite the fact that not all heavy metals in the sediments are in bioavailable form to the biota. Analysis was undertaken for nine elements (Al, Cd, Co, Cr, Cu, Fe, Mn, Ni, and Zn) in sediment samples collected at eight sites along the Eastern African Coast following aqua regia extraction (to determine the total quantity of metals) and three-step BCR (Community Bureau of Reference) sequential extraction (to obtain the metal fractionation patterns in the sediments in this region). The results revealed that heavy metal levels at some locations were higher than at others. Cd, Mn and Co were more concentrated in labile fractions compared to the other elements. These metals are easily liberated into overlying water, making them available for biological uptake. More than 62% of the total concentrations of Fe and Zn were present in the residual fraction, being immobile and less bioavailable.

Corresponding Author: EOK E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 35 10/27/2010 10:02:05 AM 32 E.O. OKUKU ET AL

INTRODUCTION In order to accurately estimate the effects and potential risks associated Human activities have been increasing with heavy metal contamination, the more rapidly and widely in the recent environmentally available fraction past than ever before in human needs to be determined. Knowledge history. In developing countries, of this fraction would be more attempts to achieve long-term growth informative in setting environmental in the economy have resulted in quality criteria and standards due to its consumption of enormous amounts of direct relation to metal toxicity. natural resources and the subsequent Estimates of the available fraction production of large amounts of waste. of metals in sediments can be obtained Contamination of aquatic ecosystems through tiered analytical methods, by heavy metals has been identified especially sequential extraction as one of the major concerns since techniques (Tessier and Campbell, the industrial revolution (Förstner & 1987). Sediments are usually Wittmann, 1983; Vinogradov, 1953). subjected to sequential extraction Most developing countries have, using chemicals of decreasing pH in the past, concentrated pollution and increasing oxidizing strength to research on inorganic contaminants remove various operationally-defined other than heavy metals. Of the few geochemical phases (Ngiam and Lim, studies carried out on heavy metals, 2001). The use of sequential extraction a number have been focused on the is effective in providing detailed total concentration of heavy metals information on the origin, species, in sediments (Kamau, 2001, 2002), biological and physiochemical despite the fact that this approach availability, mobilization and does not provide information on the transport of trace metals (Perin mobility, bioavailability and toxicity et al., 1997; Tessier et al., 1979). of metals and, as such, is not of This study aimed at comparing the direct biological significance. The environmental availability of heavy total metal concentration approach metals in sediments samples collected has recently been criticized as an from selected locations along the inaccurate measure of potential East African coast using three step- risk (Harmsen et al., 2005). This BCR sequential extraction. This can be attributed to the fact that the technique was used to compare total proportion of contaminants that is metal concentrations with individual bioavailable and liable to cause adverse metal concentrations in the residual biological effects is much smaller in fractions, viz. the F1-Exchangeable magnitude due to sorption processes F2-Reducible F3-Oxidizable, and F4- in sediments (Alexander, 2000). Residual fraction.

Volume 9 Final 20th Oct 2010.indd 36 10/27/2010 10:02:06 AM BIOAVAILABILITY OF SEDIMENT-BOUND HEAVY METALS USING BCR SEQUENTIAL EXTRACTION 33

MATERIALS AND METHODS Mbegani (T1) in Tanzania. The sites at Gazi and Chale were located within Study area Gazi Bay, a shallow, tropical coastal Sediments samples were collected at system located on the southern coast eight sampling sites along the East of Kenya, approximately 50 km from African coast (Figure 1), selected Mombasa City. Mikindani and Mtwapa on the basis of their proximity to sites were both located in the Tudor urban areas or industrials activities. Creeks. The site at Msimbazi was at According to this criterion, sites the mouth of the Msimbazi River, an close to urban or industrialized areas urban river that runs through the centre included Mikindani (K3) and Mtwapa of the Dar es Salaam. Mbegani was in (K4) in Kenya, the port of Massawa the town of Bagamoyo, 70 kilometres (E2) in Eritrea and Msimbazi (T2) north of Dar es Salaam. Massawa in Tanzania. Sites relatively distant is the main harbour of Eritrea and from sources of contamination were is located on the western side of the Chale (K1) and Gazi (K2) in Kenya, Red Sea. Gurgussum is located in the Gurgussum (E1) in Eritrea and northern part of Massawa.

Gurgussum

Massawa Red Sea Gulf Mtwapa Eritrea

Mikindani Kenya Ethiopia

Kenya Somalia

Chale Tanzania

Gazi

Tanzania Mbegani

Msimbazi

Figure 1. Map of the sampling sites at which coastal sediments were sampled for heavy metal analysis.

Volume 9 Final 20th Oct 2010.indd 37 10/27/2010 10:02:06 AM 34 E.O. OKUKU ET AL

Sample collection and of the sediments were performed preparation in aqua regia. The three replicate subsamples (approximately 0.5g Sampling was conducted in August each) were carefully weighed and and September 2006. Three random placed in digestion vessels to which sediment samples were collected to a 1.5 ml analytical grade HNO3 (69%) depth of 10 cm at each site using plastic and 4.5 ml HCl (37%) were added hand corers (Ø 8 cm), and extraneous and digested in a digestion microwave material (such as debris and stones) oven (model ETHOS 900, Milestone, removed. The samples were pooled Shelton, CT, USA). The samples at each location, mixed by hand, air- were digested at 90, 200, 350 and dried and placed in clean polythene 500 watts for 5, 3, 5 and 5 minutes bags for transport to the laboratory respectively. Three replicate standard in Belgium. In the laboratory, about and procedural blank samples were 1 kg of each sediment sample was included with each batch of samples freeze-dried. The samples were that were digested in the microwave further mixed, ground in a mortar and oven. Products of digestion were sieved through a 500 µm sieve. Three transferred to polypropylene vials and replicate subsamples of approximately diluted to 50 ml with Milli-Q water 1 g each were weighed for three-step and stored at -20°C for metal analysis. sequential extraction, as well as three Certified sediment reference material additional subsamples (approximately (CRM 141R) was used for quality 0.5 g each) for aqua regia extraction. control. Sample digestion Three-step sequential extraction Total metal extraction (Aqua regia The three-step sequential extraction extraction) procedure used in this study was based Extractions of total metal content on the protocol recommended by The

Table 1: Analytical procedure for European Community Bureau of Reference (BCR) three-step sequential extraction of metals from sediments. Phase definition Chemical reagents and experimental protocol Exchangeable and carbonate Step 1: 1g subsample with 40 ml of 0.11M CH3COOH, shaken for 16 hours at a speed of 400 rpm and 20°C.

Reducible Step 2: 40 ml 0.5M NH2OH.HCl (pH 2 adjusted with HNO3) added to residue, shaken for 16 h (at 20°C and speed of 3000 rpm).

Oxidizable Step 3: 10 ml of 8.8M H2O2 (pH 2-3) added to residue, di- gested for 1 h at 20°C, heated to 85°C, digested for 1 h, 10 ml

H2O2 added, digested at 85°C for 1 h, 50 ml 1M CH3COONH4 (pH 2) added and shaken for 16 h.

Volume 9 Final 20th Oct 2010.indd 38 10/27/2010 10:02:06 AM BIOAVAILABILITY OF SEDIMENT-BOUND HEAVY METALS USING BCR SEQUENTIAL EXTRACTION 35

European Institute for Reference The mean heavy metal Material and Measurements (Geel, concentrations in the sediments Belgium, http://www.irmm.jrc.be), samples are presented in Figure 2. Fe also described by (Ure et al., 1993). occurred at the highest concentrations Analyses were performed on three and Cd at the lowest concentrations replicate sediment samples from at all the sampling sites. The each site, following the procedure order of abundance of the various summarized in Table 1. Concentrations metals at all the sampling sites was of metals in the residual fractions Fe>Al>Mn>Zn>Cr>Ni>Cu>Co>Cd. were calculated as the difference The Kenyan site 3 (Mikindani) and between the total metal concentration Eritrean sites 1 and 2 (Gurgussum and in the sediments and the sum of the port area) were found to be the most concentrations of specific metals in polluted in terms of total heavy metals. the extracted fractions. Samples of On the other hand, Kenyan sites K1 and Freshwater Sediments Reference K2 (Chale and Gazi) were found to be Material (BCR 701) were included the least polluted (Figure 2). There were with each batch of samples (through significant differences between the the three steps) for verification of the urbanized and non-urbanized sites in measurements. Kenya (p<0.01, t=7.452) and Tanzania (p<0.05, t=2.415). This confirms the Metal determination results of a previous study (e.g. Biney et Metal concentrations in the solutions al., 1994) in which it was demonstrated were analysed with an Inductively that, in most parts of Africa, the effects Coupled Plasma Mass Spectrometer of pollution were mostly evident (ICP-MS, Varian, Australia). Yttrium around urban areas; more remote areas was used as an internal standard to manifested relatively low background correct for signal interference due levels of metal contamination. A lack to differences between calibration of significant differences (p>0.05, solutions and the samples. t=0.2039) between Eritrean sites 1 and 2 may be due to the fact that they are RESULTS AND DISCUSSION subjected to the same anthropogenic influences or have similar elemental Total metal concentrations composition in the surrounding Recovery of metals in the reference bedrock. materials used to check the accuracy of the analytical procedures were between Three-step sequential extractions: 95-107% of the certified values in the Exchangeable and associated aqua regia extractions and between 97- carbonate phases (F1) 105%, 81-92% and 83-97% in the F1, The most abundant metals in F2 and F3 fractions respectively. Fraction 1 were Cd, Mn and Co,

Volume 9 Final 20th Oct 2010.indd 39 10/27/2010 10:02:06 AM 36 E.O. OKUKU ET AL

Figure 2. Total heavy metal concentrations (mean ± S.D. in µg g-1) measured in East African coastal sediments following microwave assisted acid digestion.

comprising 19.3-49.7%, 8.1-28.2% sampling sites. The higher abundance and 2.1-22.4% of the total metals of metals such as Cd, Mn, Ni and Co respectively at most of the sites in this fraction may be attributable to (Figure 3). The order of abundance the ability of these metals to adsorb on of the metals in this fraction was: sediments or to their presence in clays, Cd>Mn>Co=Zn>Ni>Al>Fe>Cu>Cr. Fe and Mn hydrated oxides. Changes Low concentrations of most metals in in water ionic strength and pH could the exchangeable fraction indicated result in release of these elements that anthropogenic sources contributed into the overlying water (Tessier et less to the sedimentary metal pool at all al., 1979). This makes metals in this

Volume 9 Final 20th Oct 2010.indd 40 10/27/2010 10:02:06 AM BIOAVAILABILITY OF SEDIMENT-BOUND HEAVY METALS USING BCR SEQUENTIAL EXTRACTION 37

fraction the most mobile (Table 2) and biota with changes in environmental readily-available metal for biological conditions. uptake in the environment. Oxidizable fraction or bound to Reducible fraction or fraction organic matter (F3) associated with Fe and Mn oxides (F2) Al, Co, Cr and Ni were the most Cu, Zn, Fe were the most abundant abundant metals in this fraction, metals in Fraction 2, comprising comprising 2.0-27.8%, 4.7-33.4%, between 3.6-32.3%, 4.1-29.1% and 9.0- 2.3-51.5% and 3.1-32.4% of the 17.4% of the total metals respectively total metals respectively (Figure 3). (Figure 3). The order in abundance of The order in abundance of metals metals in this fraction at most sites was in this fraction at most sites was Cu>Zn>Fe>Cd>Co>Al>Mn>Cr>Ni. Cr>Co>Ni>Al>Cd>Mn>Cu>Fe> Cu, Zn and Fe were found to be Zn. The oxidizable fraction of some important in the reducible fraction metals was relatively important at compared to the other labile fractions. some stations, especially Al (K1, K4, This could be due to the fact that Fe T1 and T2), Cr (K2, K4, Ti and T2), and Mn oxides constitute a significant Co (K2, K1), Ni (K1, K2, K3, K4 and sink for heavy metals in aquatic T2) and Cd (K2) as shown in Figure 3. system under oxidizing conditions as This is in agreement with the findings explained by Gibbs et al., (1997). The of Perin et al., (1997) which indicated higher concentration of elements such that organic matter is important in as Cu, Zn, and Fe associated with this regulating the amount of free or fraction are adsorbed to the Fe-Mn bioavailable metals in sediment. colloids (Jenne, 1968), illustrating Metals in this fraction are not the ability of Fe-Mn oxides to considered mobile or freely available scavenge trace metals from solution (Table 2) as they are thought to be through processes such as adsorption associated with stable, high molecular and co-precipitation (Lim and Kiu, weight humic substances that slowly 1995). Site-to-site variations in the release only small amounts of metals. concentration of metals in this fraction may be attributable to the fact that Residual fraction (F4) the mechanisms of their adsorption Most metals were more abundant in and co-precipitation are sensitive to the residual fraction than in the labile changes in redox potential, rendering fraction at a number of the sites (Figure them moderately mobile and affecting 3). Fe was the most abundant element their relative concentration (Table 2). in the residual fraction in terms of This fraction thus carries a significant relative abundance (in comparison to burden of metals, making substantial the other fractions), constituting 62.0- quantities potentially available to 91.2% of the total metal concentration

Volume 9 Final 20th Oct 2010.indd 41 10/27/2010 10:02:06 AM 38 E.O. OKUKU ET AL

Table 2: Relative mobility and availability of trace metals in sediments (modified from Salomons, 1995).

Metals and association Mobility Exchangeable and associated with carbonated phases High; changes in major cationic composition (e.g. in estuarine environments) may cause their release. Metals associated with Fe-Mn oxides Medium; changes in redox conditions may cause mobilization. Metals associated with organic matter Medium/low; follows trends in decomposition/ oxidation of organic matter. Metals fixed in crystalline phase (residual fraction) Low; only available with weathering.

in the sediments. The relative Cd>Mn>Co>Cr>Cu>Zn>Ni>Al>Fe. abundance of the other elements in Despite the fact that the total level of the residual fraction varied between metals at Gazi was low (compared to 62.4-95.1% for Zn to 18.8-68.4% for the other sites), a substantial amount Cd (Figure 3). The order in abundance (41.9-65% of the total) of most metals of metals in the residual fraction was: (Al, Cd, Co and Mn) at this site were Fe>Zn>Al>Ni>Cu>Cr>Co>Mn>Cd. in the mobile fraction. A relatively The metals in the residual fraction high bioavailability of extractible are usually retained within the crystal metals was notable at Gazi and may lattice of minerals and in well- be attributable to naturally-occurring crystallized oxides, and are thus processes such as weathering of considered to be immobile (Tessier the parent rock, the grain size of et al., 1979). As reported elsewhere the sediment, or the prevailing (Gibbs, 1977), absolute concentrations environmental conditions which of metals in this fraction are not could be enhancing dissolution of the affected by anthropogenic inputs. The metals. Sites K2 (Co, Cr), T1 (Cd, Co, results of this work therefore indicate Cr, Ni) and T2 (Cu) had substantial that sediment-bound metals along quantities (>40%) of these metals the East African coast are mostly of in the labile fraction. As stated by natural geochemical origin. Fernandes (1997), the occurrence of metals in more easily-leached phases Labile fraction would characterize samples collected When considering the labile fraction at polluted sites. This could be true (the sum of F1, F2 and F3), the most of T1 and T2, since the likelihood of mobile elements at most sites were these sites being polluted is evident found to be Cd, Mn and Co. The order from the amount of metals extracted of abundance of labile metals was: using aqua regia.

Volume 9 Final 20th Oct 2010.indd 42 10/27/2010 10:02:06 AM BIOAVAILABILITY OF SEDIMENT-BOUND HEAVY METALS USING BCR SEQUENTIAL EXTRACTION 39

Figure 3. Exchangeable (F1), reducible (F2), oxidizable (F3) and residual (F4) metal fractions in East African coastal sediments.

Volume 9 Final 20th Oct 2010.indd 43 10/27/2010 10:02:07 AM 40 E.O. OKUKU ET AL

Metals of interest CONCLUSIONS Heavy metals manifested differing In this study, the extraction of metals concentrations in the various from sediments using aqua regia sediment fractions. Fe and Cu in the yielded differences between the sedimentary fractions followed the concentration of metal at sites in the order, F4>F2>F3>F1; Mn and Cd, proximity of urban areas and those F4>F1>F2>F3; Cr, Ni, Al and Co, less influenced by anthropogenic F4>F3>F2>F1; and Zn, F4>F2>F1>F3. activities. However, the results of A high proportion of most of the metals sequential extraction revealed that was found in the residual fraction most of the metals were present in (F4), making them relatively less the residual fraction. The results of bioavailable. This has been reported sequential extraction thus indicate elsewhere by Usero et al., (1998), that the metals in East African coastal Chester et al., (1988) and Belzunce- sediments are natural in origin. Segarra et al., (1997). Substantial This therefore has implications in amounts of Mn (16.4-31.9%), Cd Environmental risk Assessments (31.3-81.2%), Cr (3.84-51.5%), Co (ERAs), as the consideration of (18.2-65%) were also found in the non- total metal concentrations will residual fraction, as variously reported obviously lead to an overestimation by Ngiam and Lim (2001), Usero et of environmental pollution. Future al., (1998), Lin and Chen (1996), exposure assessments (especially if Tessier et al., (1979) and Olazabal the results are to be used for ERAs) et al., (1997). This suggests that should focus on bioavailable metals. considerable quantitites of Mn and Cd Some sites, in particular Gazi, had may be released into the environment if relatively high levels of bioavailable the environmental conditions change, metals, despite the fact that they were favouring their desorption and making relatively unpolluted in terms of total them potentially bioavailable for metal concentrations. Future studies uptake by organisms. should strive toward determining local Concentration of metals such factors that affect the bioavailability as Co (K1, T1, K2 and K4), Al of metals in such unpolluted systems. (Gazi), Zn (Msimbazi) and Cr (T1) were exceptionally high in the non- Acknowledgments: This work was part residual fraction at the sites indicated, of East African metal contamination suggesting that these metals could be pilot study organized by Antwerp anthropogenic in origin and easily University with partial funding from mobilized into the water compartment, Vlamse Interuniversitaire Raad (VLIR)- making them more bioavailable Belgium. We appreciate the assistance depending on the environment accorded to this study by these two conditions. Institutions. We are also indebted to

Volume 9 Final 20th Oct 2010.indd 44 10/27/2010 10:02:07 AM BIOAVAILABILITY OF SEDIMENT-BOUND HEAVY METALS USING BCR SEQUENTIAL EXTRACTION 41

all the technical staff in the Tanzania Gibbs, P.E., Bebianno, M.J. & Coelho M.R. Fisheries Research Institute, Ministry of (1997) Evidence of the Differential Fisheries Eritrea and Kenya Marine and Sensitivity of Neogastropods to Tributyltin Tbt Pollution, With Notes on Fisheries Research Institute that helped a Species Columbella Rustica Lacking with sampling in the field. the Imposex Response. Environ. Technol. 18: 1219-1224. REFERENCES Gibbs, R.J. (1977) Transport Phases of Alexander, M. (2000) Aging, Bioavailability, Transition Metals in the Amazon and and Overestimation of Risk from Yukon rivers. Geol Soc Am Bull. 88: Environmental Pollutants. Environ Sci. 829-943. Technol. 34: 4259-4265. Harmsen, J., Rulkens, W. & Eijsackers, H. Belzunce-Segarra, M.J., Bacon, J.R., Prego, (2005) Bioavailability; Concept for R. & Wilson, M.J. (1997) Chemical Understanding or Tool for Predicting. Forms of Heavy Metals in Surface Land Contam. Reclamation 13: 161- Sediments of the San Simon inlet, Ria 171. de Vigo, Galicia. J. Environ. Sci. Health Jenne, E.A. (1968) Controls on Mn, Fe, Co, A32: 1271-1292. Ni, Cu and Zn Concentrations in Soils Biney, C., Amuza, A.T., Calamari, D., Kaba, and Water: The Significant Role of N., Mbone, II., Naeva, II., Ochumba, Hydrous Mn and Fe Oxides. In: Gould, P.B.O., Osibanjo, O., Radegonde, V. R.F. (ed.) American Chemistry Society, & Saad, M.AM. (1994) Review of Washington, DC. pp 337-387. Heavy Metals in the African Aquatic Kamau, J.N. (2001) Heavy Metals Environment. Ecotoxicoology and Distribution in Sediments along the Environ. Safety 28: 134-159. Kilindini and Makupa creeks, Kenya. Chester, R., Thomas, A., Lin, F.J., Basaham, Hydrobiol. 458: 235-240. A.S. & Jacinto, G. (1988) The Solid Kamau, J.N. (2002) Heavy Metal Distribution State Speciation of Copper in Surface and Enrichment at Port-Reitz Creek, Water Particles and Oceanic Sediments. Western Indian Ocean J. Mar. Sci. 1: Mar. Chem. 24: 261-292. 64-70. Fernandes, H.M. (1997) Heavy Metal Lim, P. & Kiu M. (1995) Determination Distribution in Sediment and Ecological and Speciation Of Heavy Metals in Risk Assessment: The Role of Sediments of the Juru river, Penang, Diagenetic Processes Reducing Metal Malaysia. Environ. Monit. Assess. 35: Toxicity in Bottom Sediment. Environ. 85-95. Pollut. 97: 317-325. Lin, S. & Chen, C.M. (1996) Spatial Variations Förstner, U. & Wittmann, G.T.W. (1983) of Heavy Metals in the east China Sea Metal Pollution in the Aquatic Continental Shelf Surface Sediments. Environment. Springer Verlag, Berlin, Chem. Ecol. 13: 77- 91. New York. 486 pp. Luther, G.W.I., Wilk Z., Ryans R.A. & Meyerson L. (1986) On the Speciation of Metals in the Water Column of a Polluted Estuary. Mar. Poll Bulletin. 17: 535-542.

Volume 9 Final 20th Oct 2010.indd 45 10/27/2010 10:02:07 AM 42 E.O. OKUKU ET AL

Morse, J.W. (1995) Dynamics of Trace Metal Shuman, L.M. (1977) Adsorption of Zn by Fe Interactions with Authigenic Sulfide and Al Oxides as Influenced by Aging Minerals in Anoxic Sediments. In: and pH. J. Soil Sci. Soc. Amer. 41: 703- Allen, H.E. (ed.) Metal Contaminated 706. Aquatic Sediments. Ann Arbor, MI. Ann Tessier, A. & Campbell, P.G.C. (1987) Arbor Press. pp 187-199. Partitioning of trace Metals in Ngiam L.S. & Lim P.E. (2001) Speciation Sediments: Relationship With Patterns of Heavy Metals in Tropical Bioavailability. Hydrobiol. 149: 43-52. Estuarine Anoxic and Oxidized Tessier, A., Campbell, P.G.C. & Bisson, M. Sediments by Different Sequential (1979) Sequential Extraction Procedure Extraction Schemes. Sci. Tot. Environ. for The Speciation of Trace Metals. 275: 53-61. Anal.Chem. 51: 844-851. Olazabal, M.A., Nikolaidis, N.P., Suib, S.A. Ure, A.M., Quevauviller, P.H., Muntau, H. & Madaraiga, J.M. (1997) Precipitation & Griepink, B. (1993) Speciation of Equilibria of the Chromium (vi)/ Heavy Metals in Soils and Sediments. Iron (iii) System and Spectroscopic An Account of the Improvement and Characterization of the Precipitates. Harmonization of Extraction Techniques Environ. Sci. Technol. 31: 2898-2902. Undertaken Under the Auspices of the Perin G., Fabris R., Manente S., Rebello BCR of the Commission of European Wagener A., Hamacher C. & Scotto S. Communities. Int. J. Environ. Anal. (1997) A Five-Year Study on the Heavy Chem. 51: 135-151. Metal Pollution of Guanabara Bay Usero, J., Gamero, M., Morillo, J. & Gracia, Sediments (Rio de Janeiro, Brazil) and I. (1998) Comparative Study of Three Evaluation of the Metal Bioavailability Sequential Extraction Procedures by Means of Geochemical Speciation. for Metals in Marine Sediments. Wat. Res. 31: 3017-3028. Environment International 24: 478-496. Prohic, E. & Kniewald, G. (1987) Heavy Vinogradov, A.P. (1953) The Elementary Metal Distribution in Recent Sediments Chemical Composition of Marine of the Krka River Estuary: An Example Organisms. New Haven, CT, USA: of Sequential Extraction Analysis. Mar. Sears Found. 647 pp. Chem. 22: 279-297. Salomons, W. (1995) Environmental Impact of Metals Derived from Mining Activities: Processes, Predictions and Prevention. J. Geochem. Expl. 52: 5-23.

Volume 9 Final 20th Oct 2010.indd 46 10/27/2010 10:02:07 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 43 - 55, 2010 © 2010 WIOMSA

Development of a Regional Habitat Classification Scheme for the Amirante Islands, Seychelles

Sarah Hamylton, Annelise Hagan and Tom Spencer Cambridge Coastal Research Unit, Department of Geography, University of Cambridge, Downing Place, Cambridge, CB2 3EN, UK.

Keywords: Remote sensing, Geographical Information Systems, coastal management, habitat.

Abstract—A collaborative expedition between Khaled bin Sultan Living Oceans Foundation, Cambridge Coastal Research Unit and Seychelles Centre for Marine Research and Technology – Marine Parks Authority (SCMRT-MPA) was conducted to the southern Seychelles, western Indian Ocean, in January 2005. This resulted in a series of habitat maps of the reefs and reef islands of the Amirantes Archipelago, derived from remotely-sensed Compact Airborne Spectrographic Imager (CASI) data. The procedures used in map development, image processing techniques and field survey methods are outlined. Habitat classification, and regional-scale comparisons of relative habitat composition are described. The study demonstrates the use of remote sensing data to construct digital habitat maps for the comparison of regional habitat coverage, a key function for coastal management.

INTRODUCTION diversity and ecological status, allow comparison of the status within and Surveys of coastal environments between ecoregions, and facilitate provide information on the the detection of changes in coastal distribution and abundance of shallow ecosystems relative to established water benthic communities, as well baselines (English et al., 1997). In the as local topography and bathymetry. shallow water environments of the Standardised survey protocols permit tropics and sub-tropics, standardised assessment of regional biological surveys can be applied to ecosystems

Corresponding author: SH E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 47 10/27/2010 10:02:07 AM 44 S. HAMYLTON ET AL.

such as coral reefs, seagrasses and temporally dynamic benthic surfaces mangroves, which collectively provide as discrete units. This requires valuable ecological functions and some form of classification scheme. services in the marine environment. Habitat classification schemes play Remote sensing instruments an important role in standardising provide a synoptic portrait of the the thematic content of regional Earth’s surface by recording numerical benthic maps, facilitating their use information on the radiance measured as a common baseline against which in a series of picture elements (pixels) regional subsets can be interpreted. In across a number of spectral bands a spatial context, classes are assigned (Green et al., 2000). When mounted to homogenous patches of surface on an airborne platform, they sample that differ in appearance from their habitat cover over extensive spatial surroundings but may vary widely in ranges. The ability to survey detailed size, shape, type, relative heterogeneity information from relatively large areas and boundary characteristics (Forman is particularly valuable at the coast and Godron, 1986). Ecological because changing weather, locations units can be divided hierarchically which are inaccessible or difficult to to accommodate user requirements. access, and the logistical challenges Top-level descriptions often identify of carrying out underwater fieldwork geomorphological context, while all compromise the ability to survey lower tiers describe the relative cover a representative sample of these of benthic organisms discriminated shallow water environments directly. from ground-verified data (e.g. Remote sensing data benefit the Mumby and Harborne, 1999). A mapping process by enabling accurate hierarchical classification scheme that extrapolation of information to broader subsumes both geomorphological and scales, providing information on areas ecological characteristics is systematic not accessible in the field. In this way, in structure with tier-level descriptors cost savings can be achieved. Mumby that cannot be used interchangeably. et al., (1999) found the combined use Geomorphological classes (e.g. fore- of remote sensing and field surveys to reef slope coral spur) can, therefore, be a cost-effective means of acquiring be coupled with ecological ones (e.g. meaningful survey data on the Caicos high cover of calcareous algae) within Bank when compared to a solely field- the same tier level. based approach. Many coastal mapping strategies employ coarse mapping at the Hierarchical classification schemes regional scale, augmented by finer for coastal zone management resolution maps at locations of To map coastal habitats, it is necessary specific interest (e.g. Borstadt et to represent spatially continuous and al., 1997). Despite the potential use

Volume 9 Final 20th Oct 2010.indd 48 10/27/2010 10:02:07 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 45

of habitat classification schemes in discrete data structure on communities regional comparisons, most coastal that often present themselves as a habitat mapping has been conducted continuum of changing densities, on an ad hoc basis and displays such data formats lend themselves little consistency in terminology. well to analysis within Geographical This limits the interpretation of map Information Systems (GIS) (Burrough products, particularly where regional and McDonnell, 1998). Such an comparisons would be of value. The approach to data storage provides a use of digital habitat maps derived computationally efficient means by from remotely-sensed imagery is also which to carry out statistical analyses limited by these difficulties as they on landscape-scale datasets, e.g. range considerably in resolution. extracting coverage values for the Habitat maps are commonly different habitats. compiled from vector data, which partition an area in such a way that STUDY AREA each location falls into a polygon The Amirantes Archipelago, which assigned a value that is assumed to be lies SW of the extensive, shallow homogenous for all locations within water Seychelles Bank in the western that polygon. Although this imposes a Indian Ocean, comprises a group of (i.)

Seychelles

Figure 1. (i.) Location of the Seychelles Islands, western Indian Ocean.

Volume 9 Final 20th Oct 2010.indd 49 10/27/2010 10:02:07 AM 46 S. HAMYLTON ET AL.

(ii.)

53º00 E

African Banks

Remire

D’Arros

St. Joseph

Sand Cay

Desroches Poivre Etoile

06º00 S Boudeuse Marie Louise Desnoeufs

Alphonse 07º00 S Bijoutiere and St. Francois

Figure 1. (ii.) The Islands of the Amirantes Bank, Seychelles.

carbonate islands and islets extending at sea level with varying degrees of over a distance of ~152 km, from subaerial sand cay and coral island o 4°52’S (African Banks) to 6 14’S development. They have evolved (Desnoeufs) (Fig. 1). The majority over the last 6,000 years since the of the islands are coral reef platforms post-glacial sea level approached its

Volume 9 Final 20th Oct 2010.indd 50 10/27/2010 10:02:07 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 47

cyclonic gyre to the south (between Table 1. Aerial coverage of CASI data. 40–15°S) and reversing monsoon Number of islands surveyed 13 gyres north of 10°S. The northern Area flown 268 sq kms boundary of the subtropical gyre Flight lines flown 110 lines is formed by the South Equatorial Current (SEC). During the northern Data volume (raw) 65 Gbytes hemisphere summer the SEC is Data volume processed (estimated) 150 Gbytes displaced northwards as far as present level on the Amirantes Bank 6°S, and thus into the region of the (Stoddart, 1984). The Amirantes Bank Amirantes Archipelago, with typical -1 is an elongate structure, measuring current speeds of 0.25 m s and an approximately 180 km by 35 km, estimated transport rate of 50 Sv 6 3 -1 deepest in its centre (up to ~70 m) with (where 1 sverdrup (Sv) = 10 m s ). a marginal rim 11-27 m deep. With These climatic patterns influence the exception of the islands of Etoile both the structural form and surface and Boudeuse, the reef platforms lie communities of the reefs of the towards the eastern margin of the Amirantes, which display notable Bank. The atolls of Alphonse and leeward-windward contrasts in reef Bijoutier/St François which form the platform development (Spencer et Alphonse Group are approximately 95 al., 2009). km further south. Of the 14 islands, 13 were mapped, the exception being METHODS Desroches, a shallow submerged atoll, The primary aim of this collaborative 19-21 km in diameter, lying 16 km to expedition was to use a Compact the east of the Amirantes Bank. Airborne Spectrographic Imager The climate of the western Indian (CASI) remote sensor onboard a Ocean is humid and tropical, with seaplane to conduct large-scale mean monthly temperatures always >20°C and an annual rainfall >700 mapping of the reefs and islands of the mm. Seasonal and inter-annual Amirantes Bank. The sites comprised climatic variability is determined laterally extensive shallow water by i) the SE Asian Monsoon and the landforms, which were ideally suited associated seasonal reversal of winds; to airborne mapping. Concurrent field ii) monsoon-related movements of surveys were conducted alongside the Inter-Tropical Convergence Zone (ITCZ); iii) changes in the position the airborne surveys. Data were and intensity of the South Indian collected on the terrestrial and marine Ocean subtropical high pressure; and habitats. Results from the CASI image iv) variations in ocean circulation processing provided the first detailed and sea surface temperature. maps of the distribution of shallow The surface ocean circulation marine habitats for each of these of the western Indian Ocean is characterised by a subtropical, anti- locations.

Volume 9 Final 20th Oct 2010.indd 51 10/27/2010 10:02:08 AM 48 S. HAMYLTON ET AL.

Airborne remote sensing surveys between 0.005 and 10 km, in spatial Airborne remote sensing data were extent was employed in the analyses, acquired over 13 islands in the encompassing both intra-reef and reef Amirantes (Table 1). Reflectance data system landforms (Perry et al., 2008). were recorded between 430-850 nm Ground-referencing using a CASI sensor. The sensor was calibrated to measure radiance in 19 Ground-referencing was conducted spectral bands at a pixel size of 1 m2, in the terrestrial and marine yielding continuous data layers for the environments at eight of the islands. area. Synoptic coverage was acquired Over 1,500 ground-reference points by following a predetermined flight were recorded; locations were marked pattern over each island at an altitude with hand-held GPS units (horizontal of 1,000 m. This generated a series of accuracy of ±10 m). In shallow water, parallel flight lines, each 512 m wide, ground-referencing was conducted which could be geo-corrected and from the surface using a glass- processed. A landscape scale, ranging bottomed bucket lowered over the side

Table 2. Two-tier classification scheme for the marine habitats of the Amirantes Islands. First tier Second tier 1. Terrestrial vegetation: Trees and shrubs 1.1 Coconut woodland 1.2 Other trees and shrubs 2. Herbs and grasses 3. Saline ponds 4. Cleared/bare ground 5. Littoral hedge 6. Mangrove woodland 7. Coarse beach material & rocks 7.1 Coral sandstone/raised reef 7.2 Coral boulders 7.3 Beachrock 8. Beach sand 9. Rock pavement 10. Reef-flat sand 11. Seagrass 11.1 Low density seagrass/macroalgae 11.2 Medium density seagrass 11.3 High density seagrass 12. Lagoon patch reef 13. Lagoon sand 14. Fore-reef slope (not sand) 14.1 Coral rubble with coralline algae 14.2 Fore-reef slope coral spurs with coralline algae 14.3 Rocky fore-reef slope 14.4 Fore-reef slope (rubble and sand) 14.5 Fore-reef slope with coral 15. Fore-reef slope sand

Volume 9 Final 20th Oct 2010.indd 52 10/27/2010 10:02:08 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 49

iv

i ii iii

vii vi viii xi

v x ix

Figure 2. Habitat maps of the Amirantes Islands: i. African Banks, ii. Alphonse, iii. Marie-Louise, iv. Boudeuse, v. Bijoutier & St François, vi. Desnoeufs, vii. D’Arros & St Joseph, viii. Etoile, ix. Remire, x. Sand Cay, xi. Poivre. See Figure 1(ii.) for island locations.

of a small boat. The boat was driven Correction Now (ACORN) algorithm at a constant speed along a consistent to retrieve radiance values at the line of bearing perpendicular to the water surface (ACORN, 2001). beach from a depth of ca. 20 m to Raw data were geo-corrected using the shallow (ca. 3 m), stopping at ground control points and a first order one minute intervals to record the polynomial model was applied to substratum type and the GPS position. correct for linear offset, with nearest Image pre-processing neighbour resampling (Erdas inc., 1997). Strips were mosaiced and a Prior to processing of the remote band-wise linear colour balancing sensing data, the effects of scattering model was applied to minimise and absorption in the atmosphere across-track variance, with histogram were corrected using the Atmospheric matching to adjust for radiance offset (Rees, 1990).

Volume 9 Final 20th Oct 2010.indd 53 10/27/2010 10:02:08 AM 50 S. HAMYLTON ET AL.

100% 90%

80% 70%

60%

50%

40%

30%

20%

10%

0% African Remire D’Arros St. Joseph Sandy Cay Poivre Etoile Boudeuse Marie Desnoeufs Alphonse Bijoutier Banks Louise and St. Francois Terrestrial vegetation; trees and shrubs Herbs and grasses Buildings and other structures Littoral hedge Beach sand Rock pavement High density seagrass Lagoon patch reef Fore-reef slope sand Cleared/bare ground Saline pond Coarse beach material and rocks Mangrove woodland Seagrass Reef-flat sand Fore-reef slope material Lagoon sand Figure 3. Tier 1 breakdown of habitat coverage in the Amirantes Islands, as mapped by the Compact Airborne Spectrographic Imager (CASI).

Water column correction was where the radiance just above the

performed using the method outlined water surface Rrs (z=a) was provided by Purkis and Pasterkamp (2004), by atmospherically-corrected in which the formulae derived by imagery, the water column reflectance

Bierwirth et al., (1993) are adjusted of optically deep water (Rw) was to account for the refractive influence estimated from image statistics, of the water surface at the air/water and the depth (z) of each pixel was boundary and rearranged to solve for extracted from a bathymetric map

substratum reflectance, bR : generated using the Stupmf et al., band ratio model (2003). For each band, R = 1/0.54 x R –(1 – e-2kjz) R Equation 1 b z w the diffuse attenuation coefficient e-2kjz (k) was estimated through regression

Volume 9 Final 20th Oct 2010.indd 54 10/27/2010 10:02:08 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 51

53º0053º00 E E 1 Terrestial vegetation:trees and shrubs Littoral hedge African Banks Rock pavement 05º00 S Beach sand Remire Herbs and grasses High density seagrass 10 Reef flat sand 5 Lagoon sand D’Arros Coarse beachmaterial and rocks 11 St. Joseph Seagrass Fore-reef slope material Sand Cay Fore-reef slope sand 7 Desroches Poivre Image classification Etoile 9 and development of 06º00 S the Amirantes Islands 4 Boudeuse classification scheme Marie Louise Desnoeufs 8 A maximum likelihood classification was used to

6 assign each pixel of the image to the most likely thematic class on the basis of statistical probability 2 (Mather, 2004). This supervised classification Alphonse 07º00 S required the user to define Bijoutiere and St. 3 substratum and vegetation Francois type in a number of pixels where the content had Figure 4. Tier 1 breakdown of habitat coverage in the Amirantes Islands. Pie chart key: 1. African been verified in the field. Banks, 2. Alphonse, 3. Bijoutier & St François, 4. Training signatures were Boudeuse, 5. D’Arros & St Joseph, 6. Desnoeufs, 7. thus established to analyse Etoile, 8. Marie-Louise, 9. Poivre, 10. Remire, 11. reflectance over each strip Sand Cay. of imagery across a number of spectral subclasses. In of digital data taken from a series of this way, a statistical population of white targets deployed at known and reflectance values was built up for varying depths. The gradient of the each island habitat. For each individual regression line of log-transformed flight strip, as much heterogeneity was data plotted against depth yielded an captured in the images as possible. estimation of k. A total of 2,018 signatures were

Volume 9 Final 20th Oct 2010.indd 55 10/27/2010 10:02:08 AM 52 S. HAMYLTON ET AL.

collected and evaluated before being loaded into a hand-held GPS. These merged into a subset of 720 signatures. locations were visited in the field This merging was undertaken on the and the habitat type at each location basis of spectral similarity, using was recorded in terms of the overall the Erdas Imagine signature editor classification scheme. Validation data (Spencer et al., 2009). As a parametric were compared against the habitat class classifier, the maximum likelihood assigned by the mapping process and function assumes that each thematic scored as being correct or incorrect. The sample category can be represented overall accuracy was expressed as the by a Gaussian probability density proportion of patches that were correct function, derived from the varying (Congleton, 1991). This approach image radiance across each spectral encompassed both the locational band (Thomas et al., 1987). Given the and thematic aspects of accuracy. user-specified training signatures, it Where islands could not be visited, a is possible to compute the statistical similar approach was undertaken by probability of a pixel belonging to comparing the maps with oblique aerial each thematic spectral class. photographs of the islands. Habitats At the regional scale, individual were identified in these photographs island habitat map keys were by three independent validators with combined to produce an overall expert knowledge of western Indian scheme for the Amirantes Islands. A Ocean shallow water geomorphology habitat classification scheme with a and ecology. hierarchical structure was developed Using the conversion tools in to accommodate user requirements, ArcMap (ArcGIS 9), a raster to vector field data availability and the spatial conversion was carried out on the and spectral resolution of the CASI resultant habitat maps and the cover sensor. The classification process of the different habitat types was yielded 15 classes at Tier 1; four of computed with respect to both the first these Tier 1 classes were subsequently and second tiers of the classification sub-divided into a series of Tier 2 scheme. classes, totalling 24 habitat classes (Table 2). RESULTS Map Validation Habitat classification scheme Classification accuracy was assessed Image classification provided a clear with field data collected in situ on the and accurate representation of the eight islands that were visited. Large, heterogeneity apparent in the raw heterogeneous patches of the island images (Fig. 2). The overall accuracy habitat maps were randomly selected of the maps when ground-truthed and their central coordinates were ranged from 67-77%.

Volume 9 Final 20th Oct 2010.indd 56 10/27/2010 10:02:09 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 53

Seagrasses constituted the most structure and benthic characteristics widely represented shallow marine as class descriptors. In doing so, it habitat class (13-84% cover), drew on both of these aspects of the encompassing low, medium and remotely-sensed image. Langrebe high density communities. The (1998) describes three domains in most abundant seagrass species which spectral datasets can be viewed: were Thalassodendron ciliatum and image space, spectral space and feature Thalassia hemprichii. Fore-reef slope space. Geomorphological zones have material, reef-flat sand and lagoon more distinct contextual boundaries sand were also abundant (Fig. 3). than benthic assemblages and, as Habitat types particularly in such, the position of a pixel in typical subaerial conditions (Fig. 4) on the reef zones (Done, 1983), determined islands on the western margin of the by viewing the image space, can Amirantes Bank were characterised facilitated contextual editing. The by a restricted range of terrestrial classification algorithm on the other and littoral habitats, whereas those hand, operated in feature space to on the eastern side of the Bank were fully exploit the spectral information more diverse in habitat, particularly in of the image. subaerial conditions. The inclusion of a substantial set of training signatures facilitated DISCUSSION AND the application of image statistics CONCLUSIONS in the development of the habitat classification scheme. A top-down The map classification scheme approach in resolving the different incorporated the variable habitats from the CASI data was geomorphological nature of the islands found to be an appropriate technique within a consistent, standardised in devising the habitat classification approach. This was largely made scheme for the Amirantes Islands. possible by the use of the CASI Seagrasses were the most widely sensor. The image-based classification represented class in the maps as the employed data from an extensive set of extensive reef-flats provided the habitat training signatures, which represented needs of the genera Thalassodendron a combination of user and computer and Thalassia. Their requirements input into the classification process. include an adequate rooting The hierarchical structure of the substratum, sufficient immersion in classification scheme reflected the seawater and adequate illumination capability of the CASI sensor to resolve to maintain growth (Hemminga the various tiers at a comparably and Duarte, 2000). Seagrasses have high spectral resolution. The scheme broad, splayed leaves, a growth form employed both geomorphological that maximises light capture. These

Volume 9 Final 20th Oct 2010.indd 57 10/27/2010 10:02:09 AM 54 S. HAMYLTON ET AL.

appeared extensive when viewed from such comparisons can be extended above and remote sensing thus lends to the temporal domain, adding to its itself well to mapping their spatial usefulness in coastal management. coverage. They were, therefore, Acknowledgments: The following well represented in habitat coverage organisations are thanked for funding statistics, alongside the equally and logistical support in the field: extensive cover of reef-flat and Khaled bin Sultan Living Oceans lagoonal sands (Fig. 3). Conversely, Foundation, Seychelles Centre for some habitats, such as beach sand and Marine Research and Technology the fore-reef categories, appeared to - MPA, Island Development be limited in their spatial extent being Company, Seychelles and Great characterised by relatively steeply Plains Seychelles. Dr Chris Banks is sloping surfaces. thanked for assistance with image data There was a clear distinction in the processing. habitat maps between stable, vegetated islands located on a rock or reef REFERENCES platform (e.g. Alphonse) and islands ACORNTM (2001) Atmospheric Correction characterised by ephemeral sand cays Now: Analytical Imaging and surrounded by extensive seagrass Geophysics LLC, Version 3.12 beds (e.g. Etoile). These differences Bierwirth, P.N., Lee, T.J. & Burne, R.V. relate to the presence or absence (1993) Shallow Sea-Floor Reflectance of suitable antecendent topography and Water Depth Derived by Unmixing for postglacial coral growth and the Multispectral Imagery. Photogramm. variation in environmental parameters Eng. Remote Sens. 59: 331–338. such as wave and current fields across Borstad, G., Brown, L., Cross, W., Nallee, the Amirantes Bank. M. & Wainwright, P. (1997) Towards a Management Plan for a Tropical Comparisons of habitat coverage Reef-lagoon System Using Airborne between the 13 islands of the Multispectral Imaging and GIS. Fourth Amirantes revealed biological and International Conference on Remote geomorphological trends that would Sensing for Marine and Coastal not have been apparent without a Environments, Florida, March, 1997. standardised classification scheme Burrough, P.A. & McDonnell, R.A. (1998) applied at the regional scale. This study Principles of Geographical Information demonstrates how remotely-sensed Systems. Oxford University Press, Oxford. 327pp. imagery can be used for resource assessment and regional comparison Congleton, R. (1991) A Review of Assessing the Accuracy of Classifications of of habitat coverage. Such an approach Remotely Sensed Data. Remote Sens. would not have been possible in a Env. 37: 35-46. cost-efficient manner through field surveys alone. With repeat surveys,

Volume 9 Final 20th Oct 2010.indd 58 10/27/2010 10:02:09 AM DEVELOPMENT OF A REGIONAL HABITAT CLASSIFICATION SCHEME 55

Done, T.J. (1983) Coral Zonation: Its Nature Perry, C.T., Spencer, T. & Kench, P.S. (2008) and Significance. In: D.J. Barnes Carbonate Budgets and Reef Production (ed.) Perspectives on Coral Reefs. States: A Geomorphic Perspective on Australian Institute of Marine Science, the Ecological Phase-Shift Concept. Townesville. pp. 107-147. Coral Reefs 27: 853–866. English, S., Wilkinson, C.R. & Baker, V. Purkis, S.J. & Pasterkamp, R. (2004) (1997) Survey Manual for Tropical Integrating in situ Reef-top Reflectance Marine Resources. Australian Institute Spectra with Landsat Tm Imagery to of Marine Sciences, Townesville. Aid Shallow-Tropical Benthic Habitat 390pp. Mapping. Coral Reefs 23: 5–20. Erdas Inc. (1997) Erdas Field Guide, Fourth Rees, W.G. (1990) Physical Principles of edition, Revised and Expanded. Atlanta, Remote Sensing. Cambridge University Georgia. 394pp. Press, Cambridge. 247pp. Forman, R.T. & Godron, M. (1986) Landscape Spencer, T., Hagan, A.B., Hamylton, S.M. Ecology. John Wiley & Sons, New York. & Renaud, P. (2009) Atlas of the 379pp. Amirantes. Cambridge Coastal Research Green, E.P, Mumby, P.J., Edwards, A.J. & Unit, University of Cambridge, UK. vi Clark, C.D. (2000) Remote Sensing + 66pp. Handbook for Tropical Coastal Management. UNESCO Publishing, Stoddart, D.R. (1984) Coral Reefs of the Paris. 316pp. Seychelles and Adjacent Regions. In: Stoddart, D.R. (ed.) Biogeography Hemminga, M.A. & Duarte, C.M. (2000) and Ecology of the Seychelles Islands. Seagrass Ecology. Cambridge W.Junk, The Hague. pp. 63-81. University Press, Cambridge. 292pp. Langrebe, D. (1999) Information Extraction Stumpf, R.P., Holdereid, K & Sinclair, M Principles and Methods for Multispectral (2003) Determination of Water Depth and Hyperspectral Image Data. In: Chen, with High-Resolution Satellite Imagery L. (ed.) (1999) Information Processing over Variable Bottom Types. Limn & for Remote Sensing. World Scientific Oceanogr. 48: 547–556. Publishing Company, Singapore. 30pp. Thomas, I.L., Benning, V.M. & Ching, N.P. Mather, P.M. (2004) Computer Processing (1987) Classification of Remotely of Remotely Sensed Images: An Sensed Images. A. Hilger Publishing, introduction. Second edition, John Bristol, UK. 268pp. Wiley & Sons, New York. 324pp. Mumby, P.J. & Harborne A.R. (1999) Development of a Systematic Classification Scheme of Marine Habitats to Facilitate Regional Management and Mapping of Caribbean Coral Reefs. Biol. Conserv. 88: 155-163. Mumby, P.J., Green, E.P., Edwards, A.J., & Clark, C.D. (1999) The Cost- effectiveness of Remote Sensing for Tropical Coastal Resources Assessment and Management. J. Env. Man. 55: 157- 166.

Volume 9 Final 20th Oct 2010.indd 59 10/27/2010 10:02:09 AM Volume 9 Final 20th Oct 2010.indd 60 10/27/2010 10:02:09 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 57 - 73, 2010 © 2010 WIOMSA

Effects of Seismic Exploration on Mangrove Habitat in Tanzania

Isobel Pring1 and Nicholas V. C. Polunin2 1 eco2 Diving, Marine Research and Education Centre, PO Box 784, Mtwara, Tanzania; 2 School of Marine Science and Technology, Newcastle University, NE1 7RU, UK.

Keywords: Mangrove, recovery, oil exploration, seismic survey, Mnazi Bay Ruvuma Estuary Marine Park, Tanzania, East Africa.

Abstract—Global demand for oil and gas has resulted in increased seismic exploration for new resources in environmentally sensitive areas. Reports of damage to ecosystem function have been reported, but the environmental effects of seismic exploration are largely undocumented. Key impacts in mangroves include tree removal and trampling. This paper reports on their effects in southeast Tanzania, through assessments of tree density and species distribution, incidence of local harvesting and changes in environmental conditions that might influence the biota. Seismic survey-related gaps in the canopy have not resulted in increases in mangrove recruitment, or affected microhabitat temperatures or salinity. However, seismic lines may have become access routes, leading to increased mangrove harvesting. There were few signs of recovery in the immediate vicinity of seismic lines, which appeared to be related to trampling effects on soil stability and changes in hydrology attributable to the loss of trees. Future research should target seedling and sapling abundance and growth rates, and soil structure, composition and nutrient levels. Recommended mitigation measures would involve the promotion of mangrove regeneration and the prevention of secondary impacts such as the use of lines as access routes, with monitoring of forest recovery.

Corresponding author: IP E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 61 10/27/2010 10:02:09 AM 58 I. PRING AND N.V.C. POLUNIN

INTRODUCTION turbidity, reducing photosynthetic activity (Zabbey, 2004), but very little Increasing global demand and rising scientific evidence has been presented prices for oil and gas have made it illustrating ecosystem change. Two economically viable to prospect for studies examined the impacts of 4D new resources in areas that might seismic exploration in the Niger previously have been thought politically Delta (Osuji et al., 2006; Osuji et or environmentally sensitive. Where al., 2007); both focused on chemical coastal regions become the focus for pollution and acknowledged that the such exploration, prospecting may results were affected by ongoing oil take place in environments such as extraction, historical oil spills and the wetlands (Browning et al., 1996), salt presence of base camps. marshes and mangroves (Osuji et al., This paper reports on the effects 2006; Osuji et al., 2007). of seismic exploration on mangroves Seismic surveying is one of the in the Mtwara region of southeast first stages in oil and gas exploration Tanzania. and is used to investigate an area’s Field research took place in the geological potential for resource Mnazi Bay - Ruvuma Estuary Marine discovery. The procedure involves the Park (MBREMP) (Fig. 1), where generation of low frequency sound gas reserves were first discovered in waves and measuring their reflection the 1980s. In 2005, a programme of from subsurface geological structures. 2D seismic exploration began and These reflections are captured by seismic surveys were undertaken in geophone receivers and analysed to March 2005 and July 2007. In this assess potential oil or gas resources. In study, the effects of seismic surveying mangroves, this involves the creation on the mangroves were examined by of seismic survey lines in which the comparing plots on seismic survey vegetation is cleared from strips 1.5- lines with adjacent plots located 20 2 m wide and holes are drilled for m away, assessing: (1) the density dynamite charges at ~50 m intervals. and species distribution of trees in Documented impacts of seismic all age-classes, (2) the incidence of surveys in mangroves include land local mangrove harvesting, and (3) clearance, drilling, explosions, noise, changes in environmental conditions an influx of people, the creation of that might have influenced the biota. camps, and increased traffic (IUCN, In addition, the comparison of plots 1993), but the environmental effects on the 2005 and 2007 seismic survey are largely undocumented. It has been lines enabled an assessment of suggested that surveys destabilise mangrove recovery rates. sedimentary material and increase

Volume 9 Final 20th Oct 2010.indd 62 10/27/2010 10:02:09 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 59

BACKGROUND AND Field research was undertaken in METHODS two survey regions, one including a seismic line cleared in March 2005 The study area (three years and two months earlier), A survey of coastal mangroves in and the other a line cleared in July Tanzania completed in 2000 identified 2007 (nine months earlier). The 94.58 km2 of forest in the Mtwara seismic lines studied run parallel to region, describing it as relatively each other ~500 m apart and are ~1.5 unexploited and in better condition km (2005) and ~2 km (2007) from the than other areas along the Tanzanian village of Chui (Fig. 2). Both regions coast (Wang et al., 2003). The present support mixed stands of Rhizophora study area fell in this region and is mucronata and Ceriops tagal and are located in the south of the MBREMP inundated at high tides for between near the village of Chui. It lies ~7 km 15-30 days a month. Avicennia inland, in the upper inter-tidal area, and marina and Sonneratia alba occur includes an extensive channel and creek seaward in the area, but were not system but no major freshwater inputs. recorded in the survey transects. The It forms part of the northern extension 2005 survey region was accessed of the mangrove system in the Ruvuma on foot, while the 2007 region was delta 10 km to the south on the border reached in dugout canoes. with Mozambique (Fig. 1).

Fig. 1. Map of the study region with mangrove areas indicated by dark shading.

Volume 9 Final 20th Oct 2010.indd 63 10/27/2010 10:02:09 AM 60 I. PRING AND N.V.C. POLUNIN

Fig. 2. Map of survey region showing the first (S) and last (F) plots on the 2005 and 2007 seismic survey lines. Seismic survey procedures firing, removal crews (10) cleared Seismic survey procedures vary the area of equipment (Gwyther, A., depending on the location and the Artumas Geophysical Operations scale of the operation. In sensitive Manager, pers. comm.). environments such as mangroves, We can thus assume that the seismic work is generally carried out on foot lines were ‘walked’ a minimum of 84 using a minimum of equipment. In times (out and back) and, where lines the study area, each seismic line was were long and otherwise inaccessible, as traversed by at least six separate work is probable in larger areas of mangrove, crews (minimum crew numbers are crews may have again traversed ground given in brackets): surveyors and line they had already worked on in order to cutters (8) defined the line and cleared progress. This would have resulted in vegetation; drilling crews (6-7) took sections of line being walked 168 or a compressed air hose down the line 252 times if two or three passes were and drilled holes for seismic charges; made. In addition, at 50 m intervals, loading crews (4) laid charges; drilling crews drilled shot holes, which recording crews (10) laid geophones; then became the focus of subsequent shooting crews (4-5) laid detonators, work crews (Gwyther, A., Artumas possibly travelling the lines more than Geophysical Operations Manager, pers. once to resolve problems; and, after comm.), resulting in further trampling.

Volume 9 Final 20th Oct 2010.indd 64 10/27/2010 10:02:09 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 61

Field sampling and analysis General procedures Site selection Surveying began at the start of each Two 300 m transects were surveyed in transect, progressing to a new plot each survey region, one with the seismic each day to ensure that plots were line running down the centre, and an not disturbed by the team prior to adjacent transect 20 m away. Twelve surveying. Quadrats and replicates 10 m x 10 m plots were located on were laid using coloured rope, each seismic line transect, six over shot sediment and water samples were holes and six between shot holes. From collected, mud temperatures taken and these, three plots of each type were mangrove surveys completed. selected for surveying using a random Mangrove surveys number generator. Plots on the adjacent transects mirrored the position of the Mangrove surveys focused on species seismic line plots. These 12 plots (six on diversity, density of seedlings, saplings each seismic line transect with their six and mature trees, and an assessment adjoining plots) formed the basis of the of their status and damage. Trees were mangrove and sediment surveys. counted within each plot, identified Within each plot, three 1 m x 1 m and categorised in the following age replicate quadrats were selected by classes: seedlings, <1 m tall; saplings, dividing the plot into nine equal sections >1 m tall but lower than the main and randomly selecting one on each canopy and without fully formed bark; side of the seismic line (quadrats 1 and and mature trees with fully formed 3), and one on the seismic line (quadrat bark near or within the main canopy. 2). A similar pattern of replicates was Trees were also categorised as living, sampled in the adjacent plots. The damaged or dead, and the cause of any replicates were used as loci for the impacts was noted: seismic cutting, for measurement of mud temperature and cut trees within 0.5 m of the seismic pore water salinity and pH. line; other cutting, for cut trees >0.5 m from the seismic line; and unknown, Survey times for fallen, damaged or dead trees with Surveys were conducted over three no evidence of cutting. four-day periods in 2008: 11-14 and Environmental variables 27-30 May, and 8-11 June. Seismic line transects were surveyed between Environmental conditions were 10:00 and 12:00 and adjacent transects assessed by sampling sediment in between 12:00 and 14:00. All surveys replicate 2 in the mangrove plots. were conducted shortly before or up to Mud temperatures were measured five hours after high tide. at the surface and at 10 cm in all the replicates, as were the salinity and pH of the pore water.

Volume 9 Final 20th Oct 2010.indd 65 10/27/2010 10:02:09 AM 62 I. PRING AND N.V.C. POLUNIN

Sediment samples were collected were LOG10(x+1)-transformed and to a depth of 10 cm using a corer sediment data were arcsine transformed. 10.2 cm in diameter, bagged on site In a small number of instances, data and sieved through 1 mm mesh under were used that passed Levene’s Test running water later the same day. but, in some data combinations, did not Material retained in the sieve was pass the Anderson-Darling Test. This placed in a 175 mm diameter dish route was taken as statistical analysis and the percentage coverage of the of variance is robust when dealing with total amount and the proportions of non-normal distributions (Underwood, medium-coarse sand (>1 mm grain 1997). In such cases, Anderson-Darling size), living root material, small results are given in the text. living/dead root fragments, and plant debris was recorded. Parametric tests Mud temperatures were measured Analysis of variance was determined with a non-mercury thermometer using Minitab Statistical Software to an accuracy of 0.5°C. Pore water (Minitab, 2007) for One-way and samples were collected by digging a GLM-nested ANOVAs, the latter shallow hole and allowing sufficient involving comparisons using water to collect to fill sample bottles. Bonferroni’s Pairwise Comparisons. Salinity was measured with a hand Mangrove data were analysed held refractometer to an accuracy according to tree density, species, age, of 0.5. Indicator solution and octet damage status and damage impact, comparators were used to assess pH to and retained sediment materials were an accuracy of 0.5, adjusted for salinity analysed for quantity and relative according to tables provided with the composition. pH measurement kits (LaMotte Inc, GLM-nested ANOVAs employed MD, USA). the following nesting designs: [1] All samples were analysed in the Region, Transect (Region); [2] Region, evening on the day of collection, Transect (Region), Plot (Region, except samples from plot 5 of the 2007 survey, which were analysed the Transect); and [3] Region, Transect following day. (Region), Plot (Region, Transect), Replicate (Region, Transect, Plot). Statistical analysis Regional differences Data preparation To ensure that differences determined All data were tested using the between data sets were not a result Anderson-Darling Normality Test of underlying differences between and Levene’s Test for equal variance. the two survey regions, additional Mangrove and environmental data analyses that excluded the effects of that were not normally distributed seismic clearance were conducted.

Volume 9 Final 20th Oct 2010.indd 66 10/27/2010 10:02:09 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 63

Tree populations were analysed Due to the missing data, line and by region with One-way ANOVA, adjacent plots 5 were omitted from and seismic clearance effects were both survey regions in the parametric excluded by combining living and analyses but incorporated in the dead trees in the analysis. The multivariate analyses. Data from the remaining data were also analysed by mangrove surveys, sediment samples region with One-way ANOVA, but and environmental variables are data gathered on seismic line transects summarised in Tables 1-3. were excluded. Differences between the two survey Non-parametric tests regions Data that manifested non-normal Total tree densities (including living distributions and did not pass tests and dead trees) differed between for equal variance were analysed the two survey regions (p = 0.018); using non-parametric Kruskal-Wallis Ceriops tagal was more abundant tests in Minitab statistical software in the 2005 region (p = 0.048), but (Minitab, 2007). Rhizophora mucronata showed no significant difference. Within age Multivariate analysis classes, only mature trees manifested Parametric correlation analysis was not a difference between regions (p = possible due to lack of homogeneity 0.004), but a breakdown by species and non-normal distributions in the showed all age classes of C. tagal were environmental data. Instead, principal more abundant in the 2005 region component analysis (PCA) was (mature trees p = 0.045, saplings p = carried out using PRIMER software 0.008, and seedlings p = 0.030). (Clarke and Gorley, 2006). Sediment Sediment analysis revealed that the data were square root-transformed but total amount of material retained by temperature, salinity and pH variables the 1 mm sieve did not differ between were not manipulated. All variables the two regions but the 2005 transects were then normalised. A PRIMER had significantly more coarse sand Draftsman plot (Clarke and Gorley, than the 2007 transects (p <0.001). 2006) suggested that none of the Mud surface temperatures were variables were significantly correlated, indistinguishable between the regions. so all data were used for the PCA. However, mud temperatures at 10 cm differed (p <0.001), with the 2007 RESULTS region having higher temperatures. Kruskal-Wallis tests revealed While the survey plan included salinity differences between the two twenty-four plots, tidal flooding survey regions (p <0.001), with the prevented surveying of plot 5 on the 2005 transects exhibiting higher adjacent transect in the 2007 region.

Volume 9 Final 20th Oct 2010.indd 67 10/27/2010 10:02:10 AM 64 I. PRING AND N.V.C. POLUNIN

Table 1. Condition and number of mangrove trees in the survey transects. Tree condition Living Damaged Dead Total Impact None Seismic Local Unknown Seismic Local Unknown 2005 seismic line 653 91 7 7 77 26 13 874 2005 adjacent plots 737 0 8 5 0 11 12 773 2007 seismic line 406 27 1 7 34 11 6 492 2007 adjacent plots 426 0 0 0 0 0 5 431

Table 2. Environmental variables recorded in replicate 2 in the survey plots. 2005 Seismic transect Mud surface temp. (°C) Mud temp. at 10 cm (°C) Salinity pH Plot 1 27.0 26.0 35.0 6.73 Plot 2 26.5 25.0 30.0 6.74 Plot 3 28.0 24.5 29.5 7.24 Plot 4 25.0 25.0 33.0 6.73 Plot 5 23.5 24.0 33.0 6.73 Plot 6 23.0 23.0 33.0 6.73 2005 Adjacent plots Plot 1 26.5 25.0 32.0 7.23 Plot 2 27.0 25.0 31.0 6.24 Plot 3 25.0 24.0 31.5 6.73 Plot 4 28.0 24.0 34.0 6.73 Plot 5 25.0 24.0 33.5 6.73 Plot 6 22.0 22.0 35.0 6.73 2007 Seismic transect Plot 1 30.0 26.5 28.0 7.24 Plot 2 30.0 27.0 30.0 7.24 Plot 3 27.0 26.5 30.0 7.24 Plot 4 27.0 26.0 28.5 6.74 Plot 5 26.5 25.0 31.0 6.74 Plot 6 25.5 24.5 32.5 6.73 2007 Adjacent plots Plot 1 30.0 29.0 29.0 6.74 Plot 2 28.0 27.0 29.0 6.74 Plot 3 27.5 26.5 29.0 7.24 Plot 4 26.0 26.0 30.5 7.24 Plot 5 - - - - Plot 6 24.0 24.0 33.0 6.73

Volume 9 Final 20th Oct 2010.indd 68 10/27/2010 10:02:10 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 65

Table 3. Sediment composition recorded in replicate 2 in the survey plots.

2005 Seismic transect Coarse sand Living/dead Living root Plant debris root fragments material

Plot 1 0.33 0.22 0.11 0.33 Plot 2 0.43 0.14 0.43 0.00 Plot 3 0.43 0.29 0.14 0.14 Plot 4 0.33 0.22 0.33 0.11 Plot 5 0.29 0.14 0.29 0.29 Plot 6 0.20 0.40 0.00 0.40 2005 Adjacent plots Plot 1 0.33 0.17 0.50 0.00 Plot 2 0.43 0.14 0.29 0.14 Plot 3 0.29 0.43 0.14 0.14 Plot 4 0.17 0.50 0.17 0.17 Plot 5 0.22 0.22 0.33 0.22 Plot 6 0.20 0.30 0.30 0.20 2007 Seismic transect Plot 1 0.00 0.20 0.00 0.80 Plot 2 0.00 0.50 0.00 0.50 Plot 3 0.00 0.60 0.00 0.40 Plot 4 0.00 0.25 0.25 0.50 Plot 5 0.14 0.43 0.14 0.29 Plot 6 0.20 0.20 0.00 0.60 2007 Adjacent plots Plot 1 0.00 0.75 0.00 0.25 Plot 2 0.00 0.50 0.00 0.50 Plot 3 0.00 0.33 0.33 0.33 Plot 4 0.00 0.40 0.20 0.40 Plot 5 - - - - Plot 6 0.11 0.33 0.33 0.22

salinity levels, having a median of 33 compared to 30 in the 2007 region. Tree density and distribution The pH also differed between regions The number of living trees in the (Kruskal-Wallis p = 0.012), with the seismic line transects was not 2007 region having a higher median pH. significantly different from the associated adjacent transects but there Differences between seismic lines were more dead trees in the seismic and adjacent areas line transects in both survey regions

Volume 9 Final 20th Oct 2010.indd 69 10/27/2010 10:02:10 AM 66 I. PRING AND N.V.C. POLUNIN

(nested ANOVA [1], 2005 p = 0.030, Underlying environmental conditions 2007 p = 0.008). In the 2005 region, Sediment analysis yielded no 13% of the trees in the seismic line significant difference in the amount plots were dead compared with only of material retained by a 1 mm 3% in the adjacent transect. Figures sieve between seismic line plots and for the 2007 region were 10% and 1% adjacent plots. The composition of respectively. retained material was also similar; a As expected, the number of trees regional difference in the amount of manifesting damage from seismic plant debris (nested ANOVA [1]; p = clearing was greater in line transects 0.045) was explained by a difference than in the adjacent controls for both between the 2007 line transect the regions (Kruskal-Wallis p = 0.005 in 2005 adjacent transect, these having each case). Seismic surveying resulted the most and least of this material, in a loss of 8.8% of trees in the 2005 respectively (nested ANOVA [1]; p = line plots and 6.9% in the 2007 line 0.006). plots. In addition, 17% of the living Mud temperatures at the surface trees in the 2005 plots, and 10% in the and at 10 cm did not differ significantly 2007 plots had been damaged through between the corresponding seismic the removal of branches or prop-roots and adjacent transects. However, during seismic surveying. some data failed the Anderson- Living trees manifested no Darling test (mud surface temperature differences in density in terms of in the 2005 adjacent transect, p = species, age class and combined 0.010; temperature at 10cm in the species and age class variables in the 2005 adjacent transect and 2007 line corresponding seismic and adjacent transect, p = 0.032 and p = 0.042). transects. Corresponding plots manifested no Effects of local mangrove harvesting significant differences in mud surface temperature, but the mud temperature Local cutting accounted for 3% of the at 10 cm for plot 1 in the 2007 region dead trees along the 2005 line transect, was lower in the seismic plot than in the 2.2% along the 2007 line transect, and adjacent plot (nested ANOVA [1]; p < 1.4% in the 2005 adjacent transect. No 0.001). Analysis of the seismic survey other cutting was evident in the 2007 transects comparing the different adjacent plots. There were significantly positions on the line (nested ANOVA more dead trees from other cutting in [3]) yielded no significant differences the 2005 line transect than in the 2007 in either temperature (but replicate 3 adjacent transect (nested ANOVA [1], in the 2005 adjacent transect failed the p = 0.0353). Anderson-Darling test, p = 0.039).

Volume 9 Final 20th Oct 2010.indd 70 10/27/2010 10:02:10 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 67

Table 4. PCA results for the first three components of environmental variables. Variable PC1 PC2 PC3

Mud surface temperature 0.391 -0.368 0.134 Mud temperature at 10cm 0.404 -0.257 0.183 Salinity -0.381 0.265 0.327 pH 0.343 -0.227 -0.266 Coarse sand in sediment -0.407 -0.152 0.201 Root fragments in sediment 0.318 0.178 0.696 Living roots in sediment -0.293 -0.523 -0.223 Plant debris in sediment 0.261 0.591 -0.442 % variation explained 57.7 17.9 9.4 Cumulative % variation 57.7 75.5 85.0

Kruskal-Wallis tests on salinity and in the sediment were the drivers of pH yielded no significant differences variation in the 2005 region, and mud between the seismic and adjacent temperature, pH, and plant debris transects. in the sediment were the drivers of PCA ordination of the environmental variation in the 2007 region. variables revealed relatively clear regional clustering, although the Differences in mangrove post- seismic and adjacent transects seismic survey recovery overlapped (Fig. 3b). The first three Apart from differences already components listed in accounted for presented in terms of regional 85% of the environmental variation. variation, there were no significant The PCA ordination plot indicated differences between the seismic that the sites are split by region at a survey line transects in the two survey Euclidian distance of 4, with all the regions suggestive of differences in 2005 sites grouping together (Fig. 3b). mangrove recovery. The association between the seismic DISCUSSION line plots and adjacent plots was less clear. However, the majority of Were there seismic-related seismic plots fell in the area associated differences in tree density and with plant debris in the sediment, and distribution? the majority of adjacent plots in the Given the nature of seismic area associated with living roots in the exploration in mangroves with its sediment (Fig. 3a). The eigenvectors associated removal of all trees in lines (Fig 3a) indicated that salinity, ~2m wide, the greater number of dead coarse sand and living root material trees in the seismic line transects was

Volume 9 Final 20th Oct 2010.indd 71 10/27/2010 10:02:10 AM 68 I. PRING AND N.V.C. POLUNIN

a) b)

Fig. 3. PCA of environmental variables from the 23 mangrove plots. The depicted components account for 75.5% of the variation in environmental factors between plots. a) Eigenvectors indicate the direction and level that variables contribute to variation. b) Ordination plot clustered at various Euclidian distance similarity levels. (▲) 2005 seismic survey line, (∆) 2005 adjacent plots, (■) 2007 seismic survey line, (□) 2007 adjacent plots.

expected. However, during surveying abundance and growth rates (Clarke it was apparent that both areas had & Kerrigan 2000, Sherman et al., also been subjected to small-scale 2000). However, despite the fact that harvesting, and trees in this category vegetation clearance along the seismic were recorded separately. Although survey lines had the effect of creating differences in the number of dead small canopy gaps, there were no trees in the line transects were not significant differences in the number significant, the higher percentage of of seedlings and saplings in seismic trees affected by seismic impacts in line and adjacent plots in either region. the 2005 region may be an indication What was the effect of local that environmental guidelines were mangrove harvesting? less stringent at the time this line was created. Seismic survey clearance was not the It has been suggested that canopy only cause of mangrove damage. Cut gaps in mangroves might play an stumps were recorded in line plots important part in recovery from well outside the area of seismic survey disturbance by improving survival clearance and in the surrounding rates and promoting increased areas of both survey regions. Cutting

Volume 9 Final 20th Oct 2010.indd 72 10/27/2010 10:02:10 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 69

was focused particularly on the 2005 dugout canoe, yet had been subjected line transect, which was significantly to more cutting than the 2005 adjacent different from the 2007 adjacent line, suggesting that these factors transect, indicating that the combined were not the only influence. Instead, impact of seismic surveying and local the lines appear to be targeted as cutting had resulted in a significant they provide greater ease of timber reduction in tree densities on the 2005 removal afforded by the clearance. It line. is also possible that more remote areas The main target of cutting was are targeted when harvesting is illegal. Ceriops tagal, with Rhizophora With no data on harvesting in the mucronata being subjected to minimal region prior to seismic surveying, damage. Both R. mucronata and C. it is not possible to establish if the tagal are target species for building creation of the seismic lines has materials, charcoal, boatbuilding, fish resulted in increased harvesting or traps and firewood, and C. tagal is simply facilitated access to areas that particularly favoured due to its termite- could not be reached before. However, resistant properties (Muhando et al., local community members confirmed 1999). Within the Mtwara district, that clearance lines were perceived 90% of the houses are built with as paths, and evidence of firewood mangrove poles and, in 1999, it was collection was observed on the 2005 estimated that ~1600 mangrove trees line transect during surveying. were harvested annually, most without the proper authorisation (Muhando Was there a change in the underlying et al., 1999); mangroves in Tanzania environmental conditions? are protected under the National The PCA ordination (Fig. 3a) indicated Mangrove Management Plan (1994), that, apart from mud temperature, which allows licensed harvesting by sediment content was the next most local residents for personal use, but important variable between regions not for resale. and transects (Table 4), the main The pattern of cutting suggests that driver being the amount of plant seismic survey lines may have become debris in the sediment. This suggests access routes to the mangroves, that the seismic surveying may have facilitating harvesting from previously resulted in compositional changes in inaccessible areas. Cutting was highest the sediment. in the 2005 region and the passage of Although mud temperatures did time since its creation and its closer not differ between transects or within proximity to the village of Chui may replicates, mean temperatures for have facilitated the greater damage. replicate 2 of the line transects (the However, the 2007 line, created more replicate on the cleared line) were recently, can only be reached by consistently slightly higher, both

Volume 9 Final 20th Oct 2010.indd 73 10/27/2010 10:02:10 AM 70 I. PRING AND N.V.C. POLUNIN

at the surface and at 10 cm. Raised increased levels of (toxic) sulphide, temperatures in mangrove mud have which inhibits seedling and mangrove been shown to cause increases in growth (Hogarth, 2007). salinity and nutrient concentrations In the PCA analysis, the position due to increased evaporation (Kaly of individual plots relative to their et al., 1997; Alongi & de Carvalho, sediment content showed that the 2008). However, salinity data did not 2007 seismic line plots were typified reflect such a change. The gaps in the by dead, organic material that might canopy created by the seismic lines have been generated during clearing, may have been too small to cause a while the control plot sediments temperature rise, although this may contained living root material. Many vary with season; surveying took 2005 seismic line plots had neither. place at the end of the Tanzanian This suggests that, while some level rainy season when temperatures were of recovery may have occurred in the falling. three years since seismic surveying, the surface root layer of the mangroves Was there evidence of recovery remained affected. Ongoing trampling over time? and cutting would further hamper During surveying, it was noticeable recovery by preventing seedling that there were few seedlings and establishment. saplings in the immediate area of Trampling also disrupts the surface the seismic lines, despite the time topography and soil stability of difference in their clearance. Reasons mangrove mud and, combined with for this are probably the trampling the loss of trees, can lead to increased effects when the lines were originally tidal flushing, which slows down cleared, ongoing trampling, and mangrove recruitment; seedlings are changes in the soil hydrology due to washed away before they become the loss of trees. established if the soil lacks stability Trampling has been shown to break (Kaly et al., 1997; Kairo et al., 2001). down the mangrove surface root layer Flushing can also result in reduced and alter the structure of the mangrove levels of nutrients (Kaly et al., 1997; sediments, and recovery from such Alongi & de Carvalho, 2008), and impacts probably takes several research on the effects of nutrients on years due to the slow growth rate of Rhizophora mangle seedlings shows mangrove root systems (Dye, 2006). that phosphorous is a limiting factor in This loss of root material can also lead seedling development and, even when to a reduction in anaerobic conditions, present, water-logged and anoxic resulting in reduced microbial activity, soils render it ineffective (Koch & especially that of sulphate reducers Snedaker, 1997). This suggests it may (Alongi & de Carvalho, 2008), and be necessary to artificially maintain

Volume 9 Final 20th Oct 2010.indd 74 10/27/2010 10:02:11 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 71

that the more dispersed nature of local harvesting meant that the whole area was benefiting from increased light, but a lack of seedlings and saplings on the cleared lines suggested that this was not the case. The present study was short in duration and based on a relatively small number of samples. Although some results were inconclusive, the seismic lines surveyed did not show signs of recovery (Fig. 4). Future research should target seedling and sapling abundance and growth in the lines, and soil structure, organic content and nutrient levels. This should elucidate any persistent effects of mangrove clearing and provide the information needed to develop appropriate mitigation measures to Fig. 4. The 2005 seismic survey line at facilitate recovery. An assessment of plot 2 showing the path cleared between local use of the mangroves might also the mangroves. establish the degree to which seismic lines are being used as access routes after their creation. soil profiles and increase nitrogen and Future seismic surveys in phosphorus levels in damaged systems mangroves need to incorporate to aid their recovery (Kaly et al., 1997; monitoring of forest recovery, Alongi & de Carvalho, 2008). activities to promote regeneration, and CONCLUSIONS the prevention of secondary impacts. Current guidelines specify that the Areas cleared along the seismic area to be cleared should be minimal, survey lines in the study region had mature trees should not be cut (the been subjected to additional local path should go around them) and mangrove cutting. The combined branches should not be trimmed above effect of seismic clearing and local the line of sight in an effort to retain cutting was significant, but it was not the canopy (Artumas, 2008). They possible to establish a link between also specify the ‘blocking’ of access canopy gaps and changes in seedling routes, but this appears to be directed and sapling densities. It is possible at preventing vehicle access.

Volume 9 Final 20th Oct 2010.indd 75 10/27/2010 10:02:11 AM 72 I. PRING AND N.V.C. POLUNIN

At present, vegetation cut during Browning, G., Dillane, T., van Baaren, P., clearance is left at the side of the lines Geco-Prakla & Dietz Unocal, D. (1996) since its removal would exacerbate Environmental Considerations for 3D Seismic in Louisiana Wetlands. In: the environmental impact. It also acts SPE Third International Conference on as a barrier, discouraging access to the Health, Safety and Environment in Oil forest. This material could be replaced and Gas Exploration and Production, on and across the lines after surveying 9-12 June 1996, New Orleans, Louisiana. to create a more effective barrier to the Society of Petroleum Engineers Inc. pp. 213-226. area. This would have to be done over some distance to prevent its removal Clarke, P.J. & Kerrigan, R.A. (2000) Do Forest Gaps Influence the Population and make passage difficult, thus Structure and Species Composition of protecting mature trees in the centre Mangrove Stands in northern Australia? of the forest. Biotropica. 32: 642-652. Clarke, K.R. & Gorley, R.N. (2006) PRIMER Acknowledgements: This research was v6: User Manual/Tutorial. PRIMER-E, funded by Artumas Group Inc. I thank Plymouth, UK. http://www.primer-e. the following for their assistance: Dr com/ M. Guard, Environmental Consultant Dye, A.H. (2006) Persistent Effects of to Artumas; H. Mkomba, for his Physical Disturbance on Meiobenthos invaluable knowledge of the seismic in Mangrove Sediments. Mar Environ Res. 62: 341-355 lines in the study region; and D. Hogarth, P.J. (2007) The Biology of Reynolds and staff of the Mnazi-Bay Mangroves and Seagrasses. Oxford Ruvuma Estuary Marine Park. University Press, Oxford, UK. 273 pp. IUCN (1993) Oil and Gas Exploration and REFERENCES Production in Mangrove Areas. IUCN, Gland, Switzerland and Cambridge, Alongi, D.M. & de Carvalho, N.A. (2008) UK, with E&P Forum, London, UK. 58 The Effect of Small-scale Logging pp. on Stand Characteristics and Soil Kairo, J.G., Dahdouh-Guebas, F., Bosire, Biogeochemistry in Mangrove Forests J.O., & Koedam, N. (2001) Restoration of Timor Leste. Forest Ecology and and Management of Mangrove Systems Management. 255: 1359-1366 - A Lesson for and from the East African Artumas (2008) Environmental Impact region. S. Afr. J. Bot. 67: 383-389. Statement for the Proposed Mtwara Kaly, U.L., Eugelink, G. & Robertson, A.I. Energy Project. Artumas, Dar es Salaam, (1997) Soil Conditions in Damaged Tanzania. 255 pp. North Queensland Mangroves. Estuaries. 20: 291-300. Koch, M.S. & Snedaker, S.C. (1997) Factors Influencing Rhizophora mangle L. Seedling Development in Everglades Carbonate Soils. Aquat Bot. 59: 87-98.

Volume 9 Final 20th Oct 2010.indd 76 10/27/2010 10:02:11 AM EFFECTS OF SEISMIC EXPLORATION ON MANGROVE HABITAT IN TANZANIA 73

Minitab Statistical Software. (2007) Minitab Sherman, R.E., Fahey, T.J. & Battles, J.J. Inc., PA, USA. http://www.minitab. (2000) Small-scale Disturbance and com/ Regeneration Dynamics in a Neotropical Muhando, C., Mndeme, Y.E.S. & Kankuru, Mangrove Forest. J. Ecol. 88: 165-178. A.T. (1999) Mnazi Bay Marine Park: Underwood, A.J. (1997) Experiments in Environmental and Social Impact Ecology, their Logical Design and Assessment, Dar es Salaam, Tanzania. Interpretation Using Analysis of 35 pp. Variance. Cambridge University Press, Osuji, L.C., Ayolagha, G., Obute, G.C. & Cambridge, UK. 504 pp. Ohabuike, H.C. (2007) Chemical Wang, Y., Bonynge, G., Nugranad, J. & and Biogeophysical Impact of Four- Traber, M. (2003) Remote Sensing of dimensional (4D) seismic Exploration in Mangrove Change along the Tanzania sub-Saharan Africa, and Restoration of coast. Mar. Geod. 26: 35-48. Dysfunctionalized Mangrove Forests in Zabbey, N. (2004) Impacts of Extractive the Prospect Areas. Chem. Biodiversity. Industries on the Biodiversity of the 4: 2149-2165. Niger Delta region. Nigeria National Osuji, L.C., Ndukwu, B.C., Obute, G.C. & Workshop on Coastal and Marine Agbagwa, I. (2006) Impact of Four- Biodiversity Management, Cross-River dimensional Seismic and Production State, Nigeria. 11 pp. Activities on The Mangrove Systems of the Niger Delta, Nigeria. Chem. Ecol. 22: 415-424.

Volume 9 Final 20th Oct 2010.indd 77 10/27/2010 10:02:11 AM Volume 9 Final 20th Oct 2010.indd 78 10/27/2010 10:02:11 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 75 - 89, 2010 © 2010 WIOMSA

Lunar Cycles, Catchability of Penaeid Shrimps and Implications for the Management of the Shrimp Fishery on Sofala Bank in Mozambique

A. Brito a, b a Instituto de Investigação Pesqueira (IIP), P.O. Box 4603, Maputo, Mozambique; b Environmental Sciences Institute, Florida A&M University, Tallahassee, FL 32307, USA.

Keywords: Penaeidae, fishing effort, lunar phases, profitability, spatial closures.

Abstract—This study investigates the relationship between lunar cycles and catch rates of penaeid prawns on the Sofala Bank, where the fishery occurs for 6.5 to 9 months a year (starting in February or March), and assesses the potential for effort reduction and economic benefits from short-term closures during periods of the lunar cycle with predictably low catch-per- unit-effort (CPUE). A comparative analysis of the day and night CPUE for the daylight-active “banana” shrimps, Fenneropenaeus (Penaeus) indicus and Metapenaeus monoceros and the night CPUE for the nocturnally active Melicertus latisulcatus and Marsupenaeus japonicus was undertaken for each lunar phase and month using two-way ANOVA. Significant monthly variations in CPUE were found in both day and night samples of banana shrimps, for which the CPUE declined throughout the fishing season regardless of time of the day. M. latisulcatus manifested significant variations in night-time CPUE during the lunar cycle, with full moons yielding the lowest catches. This is thought to be caused by the burrowing behaviour of this nocturnal shrimp which decreases its catchability. The benefits of fishery closure during full moons and from June to the end of the fishing season were tested assuming two fishery scenarios of differing duration: a 6.5-month fishery (15 March - 30 August) and a 9- month fishery (15 March – 15 December). These closures resulted in an effort reduction of 5% and 8%, respectively, with corresponding catch reductions of 4% and 6%. However, profitability of the fishing companies would improve by 3% and 7%, respectively, as a result of more efficient trawling.

E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 79 10/27/2010 10:02:11 AM 76 A. BRITO

INTRODUCTION shrimp species: Fenneropenaeus (Penaeus) indicus, Metapenaeus Lunar cycles have been found to monoceros, Penaeus monodon, be particularly important in the P. semisulcatus, Marsupenaeus reproduction patterns, catchability japonicus and Melicertus latisulcatus, (Garcia, 1988) and abundance of in depths varying from 5-70 m. Most shallow water penaeid shrimps (Slack- of the catch is taken during daylight Smith, 1969), especially in strongly hours in depths between 5-24 m. nocturnal species such as Melicertus The latter three species are mainly latisulcatus (Penn, 1976). caught while night fishing, a practice In a Western Australian stock of this which developed when the summer species, it was observed that, during closed-season was first introduced in full moon, catch rates fell to minimum 1991. These nocturnal species now levels (Penn, 1976). This information represent about 10-20 % of the total led to the introduction of short-term shrimp landings of about 8 600 t p.a. closures during full moon in a number and contribute significantly to the of Western Australian Melicertus profitability of the industry in the latisulcatus fisheries, to provide a second half of the season (Palha de cost-effective effort reduction that Sousa et al., 2006; Dr. L. Palha de would improve biomass levels during Sousa, IIP, unpublished data). They the remaining fishing periods (Sporer are mainly taken in depths greater than & Kangas, 2005). 25 m (Palha de Sousa et al., 2006). The Sofala Bank, located off The expansion of the fishery to fishing central Mozambique (Fig. 1), supports at night, together with increased day a multi-species shrimp trawl fishery, fishing on the major target species including a number of species common (F. indicus and M. monoceros) with the Western Australian fisheries. in the late 1990’s, has raised the The fishery is fully exploited by a potential for growth and, ultimately, fleet of about 70 industrial vessels, 30 recruitment overfishing (Palha de m in average length, which operate Sousa et al., 2006). This has led to for about 20 fishing hours/day/boat, recommendations for a reduction of during a 6.5 to 9-month fishing season 30% in fishing effort to achieve bio- beginning in February or March economic sustainability of the fishery. each year (Dr L. Palha de Sousa, IIP, The purpose of this study was to unpublished data). The remaining investigate the relationship between months are currently subject to a lunar cycle and catch rates (CPUE) complete fishery closure, to protect the for all species, but particularly recruitment of juvenile shrimps from the nocturnal species, Melicertus estuarine waters (Palha de Sousa et latisulcatus and Marsupenaeus al., 2006). The fleet targets six penaeid japonicus, which are the focus of

Volume 9 Final 20th Oct 2010.indd 80 10/27/2010 10:02:11 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 77

Fig.1. Location of Sofala Bank in Mozambique. night fishing in the Sofala Bank Maputo, Mozambique. A more fishery. The second objective of the detailed data-set was selected from study was to assess the potential for these records, for one of the major economic benefits from short-term companies for the 2001 fishing season spatial closures during the periods of (March 15th to December 15th), the lunar cycle with predictably low this being a representative year for CPUE. the study. The ‘experimental fleet’ consisted of 14 industrial trawlers, of DATA AND METHODS similar power and employing similar fishing methods, with a gross tonnage Data source and processing of 186-272t (Palha de Sousa et al., Shrimp catch and effort data are 2006). This fleet comprehensively obtained annually from log-book covered the Sofala Bank fishing records completed by all industrial grounds during the 2001 season. vessels operating in the Sofala Bank Information on the species fishery, stored in a database at the composition of the catch, the period Instituto de Investigação Pesqueira of the day fished, the geographic (Fisheries Research Institute) in coordinates, and the depth of the

Volume 9 Final 20th Oct 2010.indd 81 10/27/2010 10:02:11 AM 78 A. BRITO

fishing location were recorded for each The relationship between CPUE, trawl. The predominantly daylight- lunar phase, and month were examined active species, Fenneropenaeus using two-way ANOVA for both the indicus and Metapenaeus monoceros, day-active and nocturnal species. were processed onboard as a single One-way ANOVA was undertaken to category (banana shrimp) for assess the relationship between effort marketing reasons, and they therefore and lunar phase (Zar, 1999). could not be distinguished from one another in the catch records. Economic analysis For the lunar phase analysis, the A profitability analysis of the fishing catch and effort of trawls at night fleet was undertaken by comparing (6 pm–6 am) were grouped in two estimated revenue from the catch and depth intervals, 5-24 m and 25-70 m, operating costs per hour of trawling. while daytime trawls (6 am–6 pm) The average landed price per kilogram between 5-24 m were combined with of catch at the time of the study was the insignificant catches >24 m to US$8.7 based on price information provide average day/night catch rates supplied to the Mozambique Ministry at all depths. Trawls coinciding with of Fisheries by the fishing industry. each of the four different lunar phases The average operational cost per were separated by month, using vessel-hour during the same period astronomical tables (Inahina, 2001). was estimated at US$164.1, which was To maximize contrast in the data, only derived from the Fishing Company’s the data for the day before, after and monthly average cost of operation for of each phase of the moon were used. each vessel and included all fixed and After this data sorting, 3 229 and 603 variable costs. night trawls and 3 739 day-time trawls The net profit of the experimental were available respectively for each of fleet during each lunar phase was the above depth and diurnal intervals obtained by subtracting the total cost for analysis (Table 1). of operation from the total revenue of the fleet. Table 1.Summary of trawl data analysed in the Sofala Bank shrimp fishery, Mozambique Period Day-time Night-time Depth intervals (m) 5 – 24 5 - 24 25 - 70 Main active Fenneropenaeus indicus Fenneropenaeus indicus Melicercus latisulcatus Species Metapenaeus monoceros Metapenaeus monoceros Marsupenaeus japonicus Penaeus monodon Penaeus monodon Number of trawls 3739 3229 603

Volume 9 Final 20th Oct 2010.indd 82 10/27/2010 10:02:11 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 79

Due to the fact that the catch and In contrast, the nocturnal species effort data used were gathered (2001) Melicertus latisulcatus manifested when the fishery operated for a considerable variation in the mean 9-month fishing season, the economic night-time CPUE during each of analyses were re-run assuming a the four lunar phases (Fig. 3). This 6.5-month fishing season (15th March relationship between CPUE (night) to 30th August) to take into account and lunar phase was significant (p = the current fishing scenario (2008- 0.0055; Table 2), with catches during 2009) on the Sofala Bank. the new moon and last quarter generally being higher; the relationship between RESULTS CPUE and month was very close to significance (p = 0.0548), with catch CPUE by moon phase rates peaking in June (Fig. 3). Note that the analysis for this species was Results of ANOVA analysis of the only undertaken for night catches at CPUE (kg/h) for the diurnally active depths between 25 - 70 m where this banana shrimps showed that there was species occurs in fishable quantities. no significant relationship between The ANOVA assessment of lunar phase and CPUE during either monthly variations in the CPUE of day (p = 0.3832) or night (p = 0.5569; Marsupenaeus japonicus (Fig. 4) Table 2) trawling. There were, indicated that there were no significant however, significant variations in differences in CPUE between the CPUE according to month (p <0.0001) lunar phases of a particular month (p for both day and night trawling, and = 0.4469), nor between months for a the highest CPUE at the start of the particular lunar phase (p = 0.8434); 2001 season was associated with the catch rates of this species were lower first quarter of the new moon (Fig. 2a and 2b).

Table 2. Results of two-way ANOVA of variability in shrimp catch rates during different lunar phases and months in the Sofala Bank fishery, Mozambique. ANOVA Species Marsupenaeus Melicertus ‘Banana’ daylight ‘Banana’ night japonicus latisulcatus Source of Lunar phases Months Lunar phases Months Lunar phases Months Lunar phases Months Variation Df 3 9 3 9 3 9 3 9 SS 197 7323 57 3806 25 43 606 762 MS 66 814 19 423 8 5 202 85 F 1.05 13.10 0.71 15.59 0.91 0.53 5.25 2.00 P-value 0.3832 <0.0001 0.5569 <0.0001 0.4469 0.8434 0.0055 0.0548

Volume 9 Final 20th Oct 2010.indd 83 10/27/2010 10:02:12 AM 80 A. BRITO

than those of the "banana" shrimp night other shrimps. The catches of 80

Penaeus semisulcatus 60 were negligible New moon 40 First quarter and therefore not kg/h Full moon analyzed in this study. 20 Last quarter Fishing effort 0 MAR MAY JUL SEP NOV distribution by Months depth and lunar "banana" shrimp day

phase 120 An analysis of the 100 distribution of day and 80 New moon 60 First quarter

night fishing effort at Kg/h Full moon 40 various depths and Last quarter lunar phases over 20 the fishing season is 0 MAR MAY JUL SEP NOV presented in Figure Months 5 (a, b, c). These data indicated a small Fig. 2. Lunar variation in CPUE of ‘banana’ shrimp (Fenneropenaeus indicus + Metapenaeus monoceros) in shift in night-time a) daytime catches and b) at night in 5-24 m on Sofala effort from shallower Bank, Mozambique. waters (5-24 m) in the first three months of the fishing season Melicertus latisulcatus (Fig. 5b) to deeper 40 waters (25-70 m) 35 30 New moon in the latter part of 25 First quarter 20

the fishing season Kg/h Full moon 15 (Fig. 5c), revealing Last quarter 10 the increasing 5 importance of 0 nocturnal species MAR MAY JUL SEP NOV Months in the second part Fig. 3. Lunar variation in CPUE of Melicertus latisulcatus of the season (see harvested at night in 25-70 m on Sofala Bank, Mozambique. slope in trend lines in both). The total (Fig 5a). ANOVA did not reveal any daytime effort for the four lunar phases significant relationship between lunar increased slightly over the season phase and effort in the shallower

Volume 9 Final 20th Oct 2010.indd 84 10/27/2010 10:02:12 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 81

fishing zone (5- Marsupenaeus japonicus 24 m) in which banana shrimp were 15 predominant, at 10 New moon either day (F3,35 = First quarter 1.0156; p = 0.3974) Kg/h Full moon 5 or night (F3,35 = Last quarter 0.7752; p = 0.5157). 0 In contrast to MAR MAY JUL SEP NOV the absence of Months a relationship Fig 4. Lunar variation in CPUE of Marsupenaeus japonicus between effort and harvested at night in 25-70 m on Sofala Bank, Mozambique. lunar phase in the a) 5-24 m Effort Day shallower, banana 1600 shrimp zone, fishing 1400 effort in night trawls 1200 New moon 1000 First quarter differed significantly 800 y = 7,499x + 1240 600 ANOVA: F3,35 = 1.0156; p = 0.3974 Full moon

according to lunar Effort (hours) 400 Last quarter 200 Total 0 phase at 25-70 m Linear (Total) MAR MAY JUL SEP NOV on the Sofala Bank Months (F3,33 = 4.6525; p = 0.0080; Fig. 5). 5-24 m Effort night b) 1600 Profitability 1400 New moon 1200 First quarter 1000 y = -11.305x + 1273.7 Full moon Vessel profitability 800 ANOVA: F3,35 = 0.7752; p = 0.5157 was assessed in terms 600 Last quarter Effort (hours) 400 Total of the monthly mean 200 Linear (Total) 0 CPUE at different MAR MAY JUL SEP NOV lunar phases. The Months minimum CPUE 25-70 m Effort Night

required to attain c) 350 300 y = 16.395x + 121.23 profitability was ANOVA: F3,33 = 4.6525; p = 0.0080 250 New moon estimated to be 18.9 200 First quarter 150 Full moon kg/h, which was 100 Effort (hours) Last quarter 50 obtained by dividing Total 0 Linear (Total) the US$164.1 MAR MAY JUL SEP NOV operational cost per Months hour by the value Fig. 5. Fishing effort during each lunar phase and total of 1 kg of shrimp, effort per month expended in a) trawling at day in 5 – averaged at US$8.7. 24 m on Sofala Bank, Mozambique, b) trawling at night in 5-24 m, and c) trawling at night in 25-70 m.

Volume 9 Final 20th Oct 2010.indd 85 10/27/2010 10:02:14 AM 82 A. BRITO

First quarter 5-24 m First quarter 25-70 m 100 100 Day 80 80 Night Night 60 Day+Night 60

Kg/h 40 Kg/h 40

20 20

0 0 MAR MAY JUL SEP NOV MAR MAY JUL SEP NOV Month Month

Full moon 5-24 m Full moon 25-70 m 80 40 Day Night 60 Night Day+Night 40 20 Kg/h Kg/h

20

0 0 MAR MAY JUL SEP NOV MAR MAY JUL SEP NOV Month Month

New moon 5-24 m New moon 25-70 m 60 60

Day Night Night 40 40 Day+Night Kg/h Kg/h 20 20

0 0 MAR MAY JUL SEP NOV MAR MAY JUL SEP NOV Month Month

Last quarter 5-24 m Last quarter 25-70 m 60 60 Night Day Night 40 40 Day+Night Kg/h Kg/h 20 20

0 0 MAR MAY JUL SEP NOV MAR MAY JUL SEP NOV Month Month

Fig. 6. CPUE of shrimp trawlers during different lunar phases, times of day and at 5-24 m (left) and 25-70 m (right) on the Sofala Bank, Mozambique. The horizontal line reflects the break-even profitability at 18.9 kg/h. Note: y-axes are at different scales to afford greater clarity.

Volume 9 Final 20th Oct 2010.indd 86 10/27/2010 10:02:16 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 83

Figure 6 shows that the profitability point in the last quarter then increasing of the fishing season in shallow water again with the new moon (Fig. 6). declined from March to December, The right hand side of Figure 6 regardless of time of trawl; in deeper shows the relationship between income water, this trend was not apparent of the fleet and the operating costs at except during the first lunar quarter. 25-70 m, these depths being subject Overall, March-May was more to significant fishing only at night. It profitable than June-December within is evident that profits were attained in each lunar phase, at both day and the new and waxing and waning lunar night. In the predominantly banana phases in both periods of the fishing shrimp area (5-24 m), the daytime season, with March-May being more profit margins were 37-54% higher in profitable (up to 48% more) than June- March-May than in June-December. December. However, the full moon Similarly, at night, profit margins was typically unprofitable. The profit were 29-50% higher in the first fishing margins respectively covered only period than the second (Fig. 6). Over 94% and 96% of the operating costs the entire fishing season, day profit during each fishing season. margins were 7-30 % higher than the A comparison of the economic night-time margins. performance of night fishing in the Additionally, the profit margins two depth ranges, 5-24 m and 25- varied with the course of the lunar 70 m, indicated that the deeper zone cycle, particularly during the first was more profitable during the last fishing period, being highest during quarter lunar phase and new moon the first quarter, decreasing to a low in both March-May (respectively

Fig. 7. Monthly shrimp landings (t) by the experimental fleet in 2001 on Sofala Bank, Mozambique, showing the proportion of the catch harvested at new and full moon.

Volume 9 Final 20th Oct 2010.indd 87 10/27/2010 10:02:17 AM 84 A. BRITO

by 7% and 4%) and June-December day and night. In each case, the option season (respectively by 8% and assessed was for the complete closure 27%). Conversely, the shallow zone of the fishery i.e. all depths to 70 m, yielded higher profit margins than since a closure in the deeper zone the deeper zone during the first lunar alone (25-70 m) would be unlikely quarter and full moon in March-May to achieve an effort/cost reduction, (respectively by 1% and 84%) and in as a shift in effort to the shallow June-December (respectively by 1% (open) zones could be expected. The and 20%). impact assessment was undertaken The preceding relationships by quantifying the expected effort between profitability, depth and lunar reduction, the reduction in catch, and phase provide a basis to assess the the net effect on profit. potential economic benefits of short- Figure 7 presents the 2001 monthly term closures in the Sofala Bank catches, showing the contribution of shrimp fishery. the catch taken during the full moon (three nights and days) relative to the Impact of spatial full moon closures total catch by the experimental fleet of The preceding results suggest that 2 100 t during the fishing season (15 there is some potential to improve March-15 December). Shrimp caught the economic performance of the during the three full moon nights industrial fleet. The impact of a series between June and December totalled of fishery closures over the three days 56 t, which represents only 2.7% of around each full moon were assessed the total landings by the experimental for the period from June to the end fleet during this period. The catch of the fishing season in December taken during both the day and night (9–month fishing season) and for over these full moon periods was 125 June-August (in a 6.5-month fishing t or 5.9% of the total catch harvested season scenario). Two closure options by the experimental fleet. were considered: (1) closures only Assuming a 6.5 month trawling at night and (2) closures during both season (15 March–30 August), the catch Table 4. Profits (US$) from shrimp fishing by the experimental fleet on the Sofala Bank in 2001over different periods and the relative profit improvement from fish- ery closure over full moon. Profits Baseline: no closure % improvement from full moon closures, (in Million US$) June to end of fishing season Fishing season scenario night day + night nights* day+ night** March-December 0.9 2.3 + 5.2 % + 6.7 % March-August 1.0 2.3 + 1.8 % + 3.2 %

* closure option 1, ** closure option 2

Volume 9 Final 20th Oct 2010.indd 88 10/27/2010 10:02:17 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 85

of shrimp would have totalled 1700 t. DISCUSSION The closure of full moon nights from June to August would have resulted in The results of this study have revealed a loss in catch of 32 t, while a closure significant differences in the CPUE of both day and night harvesting would of the nocturnally active shrimp, have resulted in the loss in harvest Melicertus latisulcatus, relative to increasing to 65 t (4%; Fig. 7). lunar phase, with full moons yielding The fishing effort assessment the lowest CPUE. This finding (Table 3) revealed that 2 515 trawl corroborates the results of Penn (1976) hours (3.8% of the annual effort) for an Australian shrimp fishery for were spent harvesting during the M. latisulcatus. The varying CPUE three full moon nights between June is a consequence of changes in the and December and 5 058 trawl hours species’ behaviour in response to a (7.7%) were spent during both the range of environmental factors. M. day and night full moon periods. In latisulcatus is thought to burrow considering the shorter fishing season, in sand as a predator-avoidance the full moon periods from June to mechanism (Tanner & Deakin, 2001). August contributed 1 101 trawl hours This behaviour has also been observed (2.6%) to the harvest at night and 2 in Marsupenaeus japonicus (Egusa 093 trawls hours (5.0%) during both &Yamamoto, 1961) and many other day and night (Table 3). penaeid shrimps during periods of A profit margin assessment was high light intensity (Barnes, 1985; carried out, based on the above catch Wassenberg & Hill, 1994; Griffiths, and effort information and on the fleet 1999). A consequence of burrowing profitability. It is estimated that profit is that they avoid capture by the would have improved by 5.2% if there fishing gear. An important point to had been a three night closure during note is that, on the Sofala Bank, the the full moon periods from June to nocturnally active M. latisulcatus and December (option 1), while closing M. japonicus occur more offshore the fishery for three days and nights than the diurnally active ‘banana’ over full moon (option 2) would have shrimp which are found closer inshore resulted in a 6.7% improvement in (Brinca et al., 1983; Palha de Sousa profit (Table 4). These calculations et al., 2006). Therefore, declines in were re-run assuming a shorter fishing catchability of M. latisulcatus during season of 6.5 months, with closures full moon have a negative impact on from June to August, resulting in the profitability of the fleet in the 25- profit improvements of 1.8 % and 3.2 70 m zone, as shown in the profitability % using fishery closure options 1 and assessment. 2 respectively (Table 4).

Volume 9 Final 20th Oct 2010.indd 89 10/27/2010 10:02:17 AM 86 A. BRITO

As demonstrated by these results, Closed seasons are important and earlier studies (Dr L. Palha de for the biological and economic Sousa, IIP, unpubl. data.), Sofala management of tropical shrimps that Bank trawlers fish mostly at depths are fast-growing and short-lived, to shallower than 25 m, where they obtain control fishing effort, prevent growth the highest catches of banana shrimps. overfishing and maximize catch value In the first two to three months of the (Watson et al., 1993). Ye (1998) fishing season, the intensity of effort showed that the benefits of seasonal is similar in both day and night, but closures were greater at higher levels then shifts to deeper waters (25- rather than at lower levels of fishing 70 m) and increases at night in the effort, with more benefit in the value latter half of the season as banana per recruit than the yield per recruit. species become less abundant (Palha Palha de Sousa et al., 2006 provide de Sousa et al., 2006). The fact that a detailed description of the closed relatively low effort was recorded in seasons on Sofala Bank, which were this study at 25-70 m during full moon first introduced in 1991 following indicates that crews are avoiding these sharp declines in the shrimp catches depths to some extent because they and growth overfishing at high levels are unprofitable. Anecdotal evidence of fishing effort. The main objective indicates that vessel captains know of the closed seasons was to protect that Melicertus latisulcatus catches the recruitment of juveniles of the are low at full moon. main species, F. indicus, to the fishery Despite the general belief that and to make larger and more valuable banana shrimps are more diurnally shrimps available when the fishing active, a laboratory experiment season opens. While these objectives revealed that Metapenaeus monoceros were achieved, in recent years (2008 - juveniles burrow during daylight 2009), fishery closure has also become hours in the presence of predatory fish a tool to test ways of decreasing fishing (Macia et al., 2003). It is therefore effort and operational costs in the possible that M. monoceros may fishery by expanding the season from burrow, as do congenerics in Western 3.5 to 6 months (pers. obs.) due to Australia (Ruello, 1973). Differences the financial difficulties in the fishing in behavioural characteristics of the industry (Dr. L. Palha de Sousa, IIP, target species may affect their temporal unpublished data). Thus, closed and spatial distribution, patterns seasons are potentially an important in abundance and, ultimately, their management tool to maximize the catchability by the fleet. However, value of the Sofala Bank shrimp because F. indicus and M. monoceros fishery at its high level of exploitation were not separated onboard, it was not (Palha de Sousa et al., 2006). possible to assess this.

Volume 9 Final 20th Oct 2010.indd 90 10/27/2010 10:02:17 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 87

Aspects of shrimp distribution in profitability of the fishery. The and behaviour, and fleet operations, improvement would be 3.2% if the profitability and fishing effort were closure was implemented in June– considered in a semi-quantitative August, assuming the shorter fishing evaluation of two proposed closure season scenario. These improvements options (night closure and day + would probably result from night closure) in the June-December considerable savings in operational season. This evaluation was also re- costs, as fuel would not be used for run assuming a shorter fishing season unprofitable trawling over the full of 6.5 months ending in August to take moon periods. Vessel captains and into account the current (2008-2009) company managers have reported on fishing on Sofala Bank. Assuming many occasions that fuel represents that the ‘experimental’ fleet activity is 50% or more of the harvesting cost representative of the remainder of the and that as much as 85-90% of their industrial shrimp fishing fleet on the fuel is spent in trawling operations, the Sofala Bank, a closure of the fishery for remainder being used for sailing. Other three days and nights over full moon foreseeable economic benefits would (option 2) from June to December emanate from the closure, especially would result in approximately 8% in the current (2008 - 2009) situation. reduction in effort versus a 5% The Sofala Bank shrimp fishery is reduction in June–August, with a experiencing financial problems due respective catch reduction of about to low shrimp prices from competition 6% versus 4% in the two scenarios. with aquaculture products and high It is assumed that most of the loss in fuel costs (Dr L. Palha de Sousa, catch during the three closed days IIP, unpublished data). Preliminary and nights would still be available indications are that the operating cost to harvest during the ensuing open of US$214.6 per vessel hour in 2007- periods. The day + night closure option 2008 had risen to US$264 at the start is preferable as a closure at night alone of 2009. These figures are 24% and would result in an increase in the 38% higher than the 2001 estimate daylight effort at 5-24 m, which would used for the experimental fleet in this reduce profitability as the profits from study, the increases being attributable daylight catches are below average at to rising fuel prices. In addition, the this time, particularly at full moon. average shrimp price at this time fell Although profits gained from below that of the 2001 level of US$8.7, the closure could not be accurately so the 2007-2009 profitability would quantified, it was estimated that have been even further reduced. This application of the day + night closure is borne out by the large number of (option 2) over June to December trawlers which did not operate in the would generate a 6.7% improvement second half of 2008 (pers. obs.).

Volume 9 Final 20th Oct 2010.indd 91 10/27/2010 10:02:17 AM 88 A. BRITO

The profitability analysis has REFERENCES indicated that it would be economically Barnes, R.D. (1985) Zoologia de los beneficial to implement the proposed Invertebrados. Tomo II. 4a.ed. closure during full moons. Although Revolucionaria.Havana.1080 pp. the benefits would be lower in the Brinca, L., Budnitchenko, V.A., da Silva, A.J. current scenario of a 6.5–month & Silva C. (1983) Report on a Survey with the R/V “Ernst Haeckel” in July- fishery, the closure would further August 1980. Rev. Investig. Pesq. benefit the fishery when it eventually Maputo, 6: 1-90. returns to the 9-month fishing season. Egusa, S. & Yamamoto, T. (1961) Studies of Should such a closure be the Respiration of the Kuruma Prawn implemented by the fishery Penaeus japonicus Bate. 1. Burrowing management authority, a number of Behavior with Special Reference to its Relation to Environmental Oxygen changes could be introduced in fleet Concentrations. Bull. Jap. Soc. Scient. planning to operate more efficiently Fish. 27: 22-26. over the lunar cycle. For instance, Garcia, S. (1988) Tropical Penaeid Prawns. trawlers could return to their home In: J.A. Gulland (ed.) Fish Population port for maintenance over the Dynamics: The Implications for three-day closure every 27 days as Management. Chichester, John Wiley and Sons Ltd. pp. 219-249. opposed to the current average of three days in every 45. Alternatively, Griffiths, S. P. (1999) Effects of Lunar Periodicity on Catches of Penaeus anchoring at sea could be considered plebejus (Hess) in an Australian Coastal to allow onboard maintenance and Lagoon. Fish. Res. 42: 195-199. the crew to rest. However, for closure INAHINA (2001) Tabela de marés do to be successfully implemented, Instituto Nacional de Hidrografia e Navegação de Moçambique.INAHINA, enforcement capacity would need . to be improved and the current Maputo 101 pp management measures supplemented Macia, A., Abrantes, K.G.S. & Paula J. (2003) Thorn Fish Terapon jarbua (Forskal) with associated technology such as a Predation on Juvenile White Shrimp Vessel Monitoring System (VMS). Penaeus indicus H.Milne Edwards and Acknowledgments: I thank Dr Jim Brown Shrimp Metapenaeus monoceros (Fabricius): The Effect of Turbidity, Penn (Australia) and Dr Ross Shotton Prey Density, Substrate Type and (FAO) for their substantial comments Pneumatophore Density. J. Exp. Mar. on earlier versions of the manuscript Biol. Ecol. 4142: 1-28 and for providing important literature. Palha de Sousa, L., Brito, A., Abdula, S. & I appreciate the support and words Caputi, N. (2006) Research Assessment of encouragement by Dr L. Palha de for the Management of the Industrial Shallow-Water Multi-Species Shrimp Sousa (IIP), Dr David Die (University Fishery in Sofala Bank in Mozambique. of Miami) and Larry Robinson Fish. Res. 77: 207-219. (Florida A. & M. University). I also appreciate helpful comments from two anonymous reviewers.

Volume 9 Final 20th Oct 2010.indd 92 10/27/2010 10:02:17 AM LUNAR CYCLES, CATCHABILITY OF PENAEID SHRIMPS AND IMPLICATIONS FOR THE MANAGEMENT 89

Penn, J.W. (1976) Tagging Experiments Wassenberg, T.J. & Hill, B.J. (1994) with the Western King Prawn P. Laboratory Study of the Effect of Light latisulcatus Keshinouye. II. Estimation on the Emergence Behavior of Eight of Population Parameters. Aust. J. Mar. Species of Commercially Important Freshwater Res. 27: 239-250. Adult Penaeid Prawns. Aust. J. Mar. Ruello, N.V. (1973) Burrowing, Feeding, Freshwater Res. 45: 43-50. and Spatial Distribution of the School Watson, R., Die, D. & Restrepo, V. (1993) Prawn Metapenaeus macleay (Haswell) Closed Seasons and Tropical Penaeid in the Hunter river region, Australia. J. Fisheries: A Simulation including Fleet Exp. Mar. Biol. Ecol. 13: 189-206. Dynamics and Uncertainty. N. Am. J. Sporer, E. & Kangas, M. (2005) Shark Bay Fish. Manage. 13: 326-336. Prawn Managed Fishery Status Report. Ye, Y. (1998) Assessing Effects of Closed In: Penn, J.W. Fletcher W.J. & Head Seasons in Tropical and Subtropical F. (eds.) State of the Fisheries Report Penaeid Shrimp Fisheries Using Length- 2004/2005. Department of Fisheries, Based Yield-Per-Recruit Model. ICES J. Western Australia. pp. 86–91. Mar. Sci. 55: 1112-1124. Slack-Smith, R.J. (1969) The Prawn of Shark Zar, J. H. (1999) Biostatistical Analysis. 4th Bay, Western Australia. FAO Fish. Rep. ed. New Jersey: Prentice Hall. 661p. 57: 717-734. Tanner, J.E. & Deakin, S. (2001) Active Habitat Selection for Sand by Juvenile Western King Prawns, Melicertus latisulcatus (Kishinouye). J. Exp. Mar. Biol. Ecol. 261: 199-209.

Volume 9 Final 20th Oct 2010.indd 93 10/27/2010 10:02:17 AM Volume 9 Final 20th Oct 2010.indd 94 10/27/2010 10:02:17 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 91 - 102, 2010 © 2010 WIOMSA

Catch Composition, Abundance and Length- Weight Relationships of Groupers (Pisces: Serranidae) from Inshore Waters of Kenya

Simon Agembe1, 3, Chrisestom M. Mlewa2 and Boaz Kaunda-Arara3 1Kenya Marine & Fisheries Research Institute, P.O. Box 81651, Mombasa, Kenya; 2School of Pure and Applied Sciences, Pwani University College, P.O. Box 195, Kilifi, Kenya;3 Department of Fisheries and Aquatic Sciences, Moi University, P.O. Box 1125, Eldoret, Kenya.

Keywords: Groupers, Serranidae, catch composition, abundance, length-weight relationship.

Abstract—Groupers (family Serranidae) support important artisanal fisheries in most of the Western Indian Ocean (WIO) region. However, despite their economic and ecological importance, they are poorly studied in the region. This study describes, for the first time, the species composition, abundance and length-weight relationship of groupers from Kenya’s inshore artisanal fisheries. Data were obtained from landings by artisanal fishers on the south coast of Kenya from February to July 2007. A total of 37 species belonging to six genera, viz. Anyperodon, Cephalopholis, Dermatolepis, Epinephelus, Plectropomus and Variola, were landed by fishers. The genus Epinephelus was the most speciose in the landings, with 20 species. Significantly higher numbers of groupers were landed during the southeast monsoon (n = 616) compared to the northeast monsoon (n = 184) season (χ2 = 125.812, df = 1, p <0.001). Of the three sites studied, more species were recorded at Shimoni (n = 36) compared to the Msambweni and Vanga sites (14 and 8 species, respectively). Length-weight relationships derived for 15 species indicated that most groupers exhibited isometric growth. This work provides baseline data on the composition, distribution and abundance of grouper species in Kenya useful for comparison with the rest of the WIO region.

Corresponding Author: SA E-mail: [email protected], [email protected]

Volume 9 Final 20th Oct 2010.indd 95 10/27/2010 10:02:17 AM 92 S. AGEMBE ET AL.

INTRODUCTION Management of groupers must be based on scientific information if The family Serranidae contains about populations are to be maintained at 500 species in more than 60 genera, commercially viable levels (Rhodes including sea basses and groupers of the & Sadovy, 2002). However, bio- subfamily Epinephelinae (Smith, 1965; ecological data on these important Heemstra & Randall, 1993). Groupers fishes is scarce in the WIO region. In are the target of valuable fisheries in the Kenya, data on exploited groupers are Western Indian Ocean (WIO) region restricted to descriptive information (Nzioka, 1979; Kaunda-Arara, 1997; on a few species (Nzioka, 1979), Jiddawi & Stanley, 1999). In addition movement studies on the greasy to their importance to local economies, grouper, Epinephelus tauvina groupers are apex predators thought (Kaunda-Arara & Rose, 2004), to play important roles in ecosystem and preliminary investigations function (Huntsman et al., 1999; Dulvy on their spawning aggregations et al., 2004; Campbell & Perdede, (Samoilys et al., 2006). Data on 2006). Loss or reduction of these species composition and abundance species from coral reefs can therefore are important in evaluating spatial adversely affect local biodiversity effects of fishing effort (Jennings and ecosystem stability (Dulvy et al., & Polunin, 1996), determining 2004). Despite their ecological and recruitment variability (Caley et al., economic importance, there is concern 1996) and for trend analysis, while that many groupers are already being length-weight parameters are useful overfished (Sadovy, 1994; Luckhurst, inputs in length-structured models for 1996) and some grouper species are stock assessment (Pauly, 1984). The listed as threatened or endangered by objective of this study was therefore to IUCN (http://www.redlist.org). Many build on the limited grouper database grouper species are long-lived, slow by providing information on species to mature, sedentary (Sadovy, 1996), composition, abundance and length- and form spawning aggregations weight relationships of groupers from that are predictable in time and space Kenya’s artisanal fisheries. (Domeier & Colin, 1997). These factors make them highly vulnerable to MATERIAL AND METHODS overfishing. In Kenya, grouper catches Study sites have declined by about 80% in the last decades (Kaunda-Arara et al., 2003). The study was undertaken at three The extent to which these declines are fisheries landing beaches (Msambweni, attributable to species-specific fishing Shimoni and Vanga) on the southern vulnerability and/or environmental coast of Kenya (Fig. 1). These landing factors remains unknown. beaches were chosen because they

Volume 9 Final 20th Oct 2010.indd 96 10/27/2010 10:02:17 AM CATCH COMPOSITION, ABUNDANCE AND LENGTH-WEIGHT RELATIONSHIPS OF GROUPERS 93

Figure 1: Map of south coast of Kenya showing the fish landing sites (Msambweni, Shimoni and Vanga) monitored in the study. are some of the most active, with Data collection high artisanal fisheries landings for Fisher catches were sampled at the three the Kenyan coast. Landings from the beaches for ~10 days/beach/month study sites were therefore considered during the northeast monsoon (February, more representative and likely to March, April) and the southeast capture the variability in grouper monsoon (May, June, July) months in composition on the Kenyan coast. 2007. Groupers were separated from the Fishing activities off these landing catches and identified to species level beaches are concentrated within using keys from Heemstra & Randall nearshore reef lagoons as fishermen 1993 and Smith & Heemstra, 1986. infrequently venture beyond the outer Total length (to the nearest 0.1 cm) reef because of the low power of their and weight (to the nearest gram) were fishing craft.

Volume 9 Final 20th Oct 2010.indd 97 10/27/2010 10:02:19 AM 94 S. AGEMBE ET AL.

measured for each fish, using a top- and Vanga sites yielded 14 and 8 loading Salter digital balance for small species, respectively. At Shimoni, fish (<2.0 kg) or on a hanging Salter grouper species landed belonged to 6 scale (100 kg in 100 g increments) for genera; Anyperodon, Cephalopholis, bigger individuals. Dermatolepis, Epinephelus, Plectropomus and Variola (Table 1). Data analysis The genus Epinephelus was the most Percentage abundance was used to speciose in the samples, with 20 species, determine the numerical dominance of while the genera Cephalopholis, species at each of the sites. The non- Plectropomus, Dermatolepis and parametric Chi-square test was used Variola were represented by 7, 4, 2 and to test for differences in numerical 2 species, respectively, and Anyperodon abundance of species between the by only one species (Table 1). The northeast monsoon (NEM) and number of individuals landed at sites southeast (SEM) seasons. Length- varied considerably in some species, weight relationships were determined for example, it ranged from 1 to 124 in using the equation: Log10 W = log10 a Epinephelus caeruleopunctatus. Only + b log10 TL, where W is body weight, three species, Cephalopholis boenak, TL is total length, a is the intercept, E. caeruleopunctatus and E. fasciatus, and b is the slope of the regression had more than 100 individuals landed line (Wootton, 1990). This relationship at all the sites (Table 1). was estimated for fifteen species for In Shimoni, the white- which there were data for 15 or more spotted grouper, Epinephelus specimens. The slope of the length- caeruleopunctatus, was numerically weight relationship for each species was the most abundant (15.6%) in the catch tested for significant difference from the followed by Cephalopholis boenak isometric growth value of 3.0 (Ricker, (13%) and E. fasciatus (12.9%) (Fig. 1975) using a one tailed t-test. Seasonal 2a). Some species, such as E. areolatus, abundance of groupers was analysed E. spilotoceps, Plectropomus punctatus, for Shimoni, which had highest and P. laevis, E. hexagonatus, P. maculatus, most diverse landings of groupers. All E. tukula, Dermatolepis striolata, D. statistical tests followed Zar (1996). dermatolepis, E. melanostigma, and C. taeniops comprised <1% of the catches. RESULTS In Vanga, the commonly landed Species catch composition species were the white-spotted grouper, Epinephelus caeruleopunctatus, the Thirty seven species of groupers were malabar grouper, E. malabaricus, landed at the three beaches. Shimoni and the chocolate hind, C. boenak, landings were the most diverse each of which comprised over 20% of with 36 species, while Msambweni the grouper landings (Fig. 2b). Four

Volume 9 Final 20th Oct 2010.indd 98 10/27/2010 10:02:19 AM CATCH COMPOSITION, ABUNDANCE AND LENGTH-WEIGHT RELATIONSHIPS OF GROUPERS 95

Table 1: A comparison of numerical abundance of grouper species landed during the northeast (NEM) and southeast (SEM) monsoon seasons at Shimoni site. N = numbers landed, χ2 is chi-square statistic, p is probability of significance of the dif- ference at α = 0.05. Species SEM NEM N χ2 p-value Anyperodon leucogrammicus 5 3 8 0.254 0.614 Cephalopholis argus 13 15 28 0.072 0.789 Cephalopholis boenak 100 4 104 56.30 0.000 Cephalopholis leopardus 3 3 6 0.000 1.000 Cephalopholis miniata 10 5 15 0.883 0.347 Cephalopholis sonnerati 18 0 18 12.00 0.001 Cephalopholis taeniops 1 2 3 0.194 0.659 Cephalopholis urodeta 59 0 59 39.44 0.000 Dermatolepis dermatolepis 1 0 1 0.750 - Dermatolepis striolata 2 1 3 0.194 0.659 Epinephelus acanthistus 0 1 1 0.750 - Epinephelus areolatus 7 0 7 4.773 0.029 Epinephelus bentoides 2 0 2 1.333 - Epinephelus caeruleopunctatus 95 29 124 18.90 0.000 Epinephelus coioides 55 19 74 9.307 0.002 Epinephelus fasciatus 92 12 102 36.11 0.000 Epinephelus flavocaeruleus 1 0 1 0.750 - Epinephelus fuscoguttatus 23 1 24 12.77 0.000 Epinephelus hexagonatus 0 3 3 2.100 - Epinephelus lanceolatus 1 0 1 0.750 - Epinephelus longispinis 2 0 2 1.333 - Epinephelus macrospilos 2 4 6 0.343 0.558 Epinephelus malabaricus 45 22 67 4.096 0.043 Epinephelus melanostigma 1 0 1 0.750 - Epinephelus merra 14 10 24 0.336 0.562 Epinephelus multinotatus 25 18 43 0.567 0.451 Epinephelus poecilonatus 0 1 1 0.750 - Epinephelus socialis 1 0 1 0.750 - Epinephelus spilotoceps 6 0 6 4.000 0.046 Epinephelus tauvina 3 16 19 5.132 0.023 Epinephelus tukula 4 11 15 1.777 0.183 Plectropomus laevis 3 1 4 0.533 0.465 Plectropomus maculatus 2 0 2 1.333 - Plectropomus pessuliferus 1 0 1 0.750 - Plectropomus punctatus 3 1 4 0.533 0.465 Variola albimarginata 1 0 1 0.750 - Variola louti 17 2 19 7.127 0.008 Total 616 184 800

species, E. multinotatus, E. tauvina, In Msambweni, only two E. tukula and E. fasciatus, were low in species (Epinephelus fasciatus and relative abundance (1-5%). Cephalopholis urodeta) had relative

Volume 9 Final 20th Oct 2010.indd 99 10/27/2010 10:02:19 AM 96 S. AGEMBE ET AL.

Table 2: Length-weight relationships for fifteen species of groupers (Serranidae) landed on the south coast of Kenya. N = sample size; a = constant; b = length exponent. Species TL (cm) L-W parameters t-test statistics N min Max a b t^ tab r2 P Cephalopholis argus 28 19.0 37.5 -1.90 3.08 I -0.07 2.05 97.3 <0.001 Cephalopholis boenak 103 10.0 22.1 -1.67 2.89 I 0.36 1.98 85.9 <0.001 Cephalopholis miniata 15 19.0 30.8 -2.12 3.22 I -0.18 2.18 96.8 <0.001 Cephalopholis sonnerati 18 16.4 54.3 -1.75 2.98 I 0.01 2.12 98.5 <0.001 Cephalopholis urodeta 59 13.5 25.0 -1.74 2.94 I 0.16 2.00 95.6 <0.001 Epinephelus caeruleopunctatus 124 16.5 52.0 -1.75 2.92 I 0.12 1.98 95.1 <0.001 Epinephelus coioides 72 20.5 104.5 -1.93 3.04 I -0.02 1.99 99.3 <0.001 Epinephelus fasciatus 102 12.0 28.1 -1.77 2.93 I 0.24 1.98 93.0 <0.001 Epinephelus fuscoguttatus 24 24.1 90.1 -1.93 3.09 I -0.02 2.07 97.9 <0.001 Epinephelus malabaricus 67 23.5 108.0 -1.71 2.90 I 0.04 1.99 97.7 <0.001 Epinephelus merra 24 16.0 34.5 -1.84 2.99 I 0.01 2.07 97.6 <0.001 Epinephelus multinotatus 43 14.8 65.0 -1.75 2.93 I 0.04 2.02 89.8 <0.001 Epinephelus tauvina 19 23.0 50.7 -1.95 3.05 I -0.03 2.11 97.8 <0.001 Epinephelus tukula 15 27.5 77.5 -2.08 3.17 I -0.04 2.16 98.4 <0.001 Variola louti 19 19.0 37.5 -1.64 2.77 I 0.06 2.11 97.1 <0.001

abundances >16% (Fig. 2c). Five Seasonal variations in catches species were landed in relatively The total number of groupers landed abundance (2-5%) and seven at an at the sites differed seasonally (Table abundance of 1% (Fig. 2c). The black- 2). At Shimoni, higher numbers tip grouper, E. fasciatus, was the most were caught during the SEM (n = commonly (19%) landed species at 616) compared to the NEM (n = Msambweni, followed by C. urodeta 184) season (χ2 = 125.812, df = 1, which comprised 17% of the landings. p <0.001) (Table1). The dominant Catch composition by weight species caught during the NEM season was the white spotted grouper, Three species (Epinephelus coioides, Epinephelus caeruleopunctatus, E. malabaricus and E. fuscoguttatus) whereas the chocolate hind, formed the bulk of the biomass of Cephalopholis boenak, was dominant groupers landed at the three landing during the SEM season (Table 1). At beaches. The white-spotted grouper, Vanga, significantly higher numbers E. caeruleopunctatus, although of groupers were landed during the dominant in numerical abundance, SEM (n = 75) compared to the NEM ranked only fourth in terms of the total (n = 22) (χ2 = 13.033, df = 1, p weight landed. Seven species totalled <0.05). Similarly, significantly higher ≥20 kg in landed weight, while 27 number of groupers were landed at species weighed in at ≤15 kg (Fig. 3). Msambweni during the SEM (n = 42)

Volume 9 Final 20th Oct 2010.indd 100 10/27/2010 10:02:19 AM CATCH COMPOSITION, ABUNDANCE AND LENGTH-WEIGHT RELATIONSHIPS OF GROUPERS 97

16 14 (a) n = 800 12 10 8 6 4 2 0 V. louti V. P. laevis P. C. argusC. merra E. E. tukula E. E. tukula E. E. tauvina E. C. boenakC. C. miniataC. C. urodetaC. E. coioides E. D. striolata D. taeniopsC. E. fasciatus E. E. areolatus E. E. bentoides E. C. sonneratiC. P. punctatus P. P. maculatus P. C. leopardusC. E. longispinis E. E. spilotoceps E. E. lanceolatus E. pessuliferus P. E. acanthistus E. E. macrospilos E. E. malabaricus E. E. multinotatus E. E. hexagonatus E. D. dermatolepis D. E. multinotatus E. E. fuscoguttatus E. E. melanostigma E. V. albimarginata V. A.leucogrammicus E.caeruleopunctatus 35 (b) n = 45 30 25 20 15 10 5

Relative abundance (%) 0 C. argusC. E. tukula E. E. tauvina E. E. C. urodetaC. E. coioides E. C. taeniopsC. E. fasciatus E. Variolalouti E. diacanthus E. E. longispinis E. E. hexagonatus E. E. chlorostigma E. E. flavocaeruleus E. caeruleopunctatus

(c) n = 97 30 25 20 15 10 5 0 E. tukula E. E. tauvina E. E. boenakC. E. coioides E. E. fasciatus E. E. malabaricus E. E. multinotatus E. caeruleopunctatus Figure 2: Percentage abundance of grouper species landed at (a) Shimoni, (b) Msambweni, and (c) Vanga landing sites on the south coast of Kenya from February to July 2007.

Volume 9 Final 20th Oct 2010.indd 101 10/27/2010 10:02:20 AM 98 S. AGEMBE ET AL.

16 14 12 (a) 10 8 6 4 2 0 V. louti V. C. argusC. merra E. laevis P. E. tukula E. E. tukula E. E. tauvina E. C. boenakC. C. miniataC. C. urodetaC. E. coioides E. D. striolata D. taeniopsC. E. fasciatus E. E. areolatus E. E. bentoides E. C. sonneratiC. P. punctatus P. P. maculatus P. C. leopardusC. E. longispinis E. E. spilotoceps E. E. lanceolatus E. pessuliferus P. E. acanthistus E. E. macrospilos E. E. malabaricus E. E. multinotatus E. E. hexagonatus E. D. dermatolepis D. E. multinotatus E. E. fuscoguttatus E. E. melanostigma E. V. albimarginata V. A.leucogrammicus E.caeruleopunctatus Figure 3: Total landings (kg) of grouper species from Shimoni, Vanga and Msambweni on the south coast of Kenya (n = 942) from February to July 2007. compared to the NEM (n = 3) (χ2 = with Shimoni having the highest 18.861, df = 1, p <0.05). diversity of grouper landings (36 species) compared to Vanga and Length-weight relationships Msambweni where 8 and 14 species The length-weight regressions for 15 were landed, respectively. Higher grouper species were highly significant relative abundances of Epinephelus (p <0.001) with r2 values ranging from caeruleopunctatus, Cephalopholis 85.9-99.3% (Table 2). A one tailed t-test boenak and E. fasciatus were recorded result showed that the length exponents at Shimoni, while at Vanga, E. (b values) were not significantly caeruleopunctatus, E. malabaricus, different from 3 and were indicative of and C. boenak were the most isometric growth (Table 2). abundant. E. fasciatus, C. urodeta, E. diacanthus were the dominant catch at DISCUSSION Msambweni. This variation at the sites is probably due to spatial differences in A total of 37 grouper species were fishing pressure, habitat characteristics landed by fishers at the three south or variability in recruitment. Groupers Kenyan coast fish landing beaches, are sedentary fishes with high site-

Volume 9 Final 20th Oct 2010.indd 102 10/27/2010 10:02:20 AM CATCH COMPOSITION, ABUNDANCE AND LENGTH-WEIGHT RELATIONSHIPS OF GROUPERS 99

fidelity, often around coral heads fishers are known to concentrate their (Kaunda-Arara & Rose 2004a), a trait efforts on the inner reef during the making their distribution dependent SEM due to the rough sea conditions on habitat complexity. Additionally, (McClanahan, 1988). While the higher site fidelity and spawning of some effort probably contributes to enhanced species in aggregations make them catches in this season, the factors vulnerable to fishing mortality due to responsible for the higher diversity and predictability in their abundance and numerical landings of groupers during distribution (Domeier & Colin, 1997). the rougher SEM season are not clear. Recruitment data on Serranidae are However, it may be that they take refuge lacking for the WIO region; however, in the calmer inshore waters during the recruitment variability is known to SEM season, thereby increasing their influence the spatial structure of fish vulnerability to fishing, and become populations in Kenya (Kaunda-Arara more widely dispersed during the NEM, et al., 2009) and elsewhere (Caley et resulting in lower fishing mortality. al., 1996), and could contribute to the F i v e s p e c i e s , E p i n e p h e l u s spatial variations observed in species caeruleopunctatus, E. coioides, E. diversity and sample sizes between tauvina, E. tukula, and E. fasciatus, the three sites. The higher diversity at were common to all the landing sites. Shimoni could, among other factors, This may be attributed to the overall be related to the influence of the similarity in the geospatial distribution nearby Kisite-Mpunguti Marine Park, of the fringing reef from Vanga to probably through larval re-seeding and Msambweni, providing a uniformity spill-over effects of the park on the in macro-habitat that favours the adjacent fisheries (Kaunda-Arara & common species. However, E. Rose, 2004a, b; Bostford et al., 2007). diacanthus and E. chlorostigma were The results indicate that only landed in Msambweni. This Cephalopholis boenak, C. urodeta, restricted distribution is perhaps Epinephelus caeruleopunctatus, attributable to recruitment variability, E. coioides and E. fasciatus were differences in micro-habitat preference numerically higher in catches during and differential effects of fishing the southeast monsoon season, while amongst other factors (Hixon & Beets, the potato grouper, E. tukula, and the 1993, Kaunda-Arara & Rose, 2004). greasy grouper, E. tauvina, were more Although the Vanga and Shimoni abundant during the northeast monsoon. sites share a similar reef structure More species were caught during the (Samoilys, 1988), the low species rougher SEM conditions compared to diversity at Vanga may be attributable the calmer NEM season. These seasonal to dynamiting by fishers from Pemba differences may be a reflection of Island (Tanzania) which destroyed the seasonal variations in fishing effort as reefs at this site (Samoilys, 1988).

Volume 9 Final 20th Oct 2010.indd 103 10/27/2010 10:02:20 AM 100 S. AGEMBE ET AL.

All the length exponents in the Acknowledgments: We wish to length-weight relationships for the thank the director of Kenya Marine fifteen grouper species were between and Fisheries Research Institute for 2.77 and 3.22 and were indicative of financial support. We are grateful to isometric growth (Bagenal & Tesch, the technical staff, Messrs. B. Orembo, 1978). Length–weight relationships D. Odongo, J. Gonda, R. Anam, P. are important in fisheries science Nyalele, D. Ocharo, K. Omondi and (Ricker, 1975) and for modelling J. Omweri, for assisting with data stocks (Pitcher, 1995 ); the data collection. We thank the anonymous generated in this study will be therefore reviewers for constructive criticisms be useful in future stock assessment and useful suggestions. and modelling studies on these species in the WIO region. REFERENCES This study has documented a higher diversity of groupers on the Bagenal, T.B., Tesch, F.W. (1978) Age and Kenyan south coast at Shimoni Growth. In: T. B. Bagenal (ed.). Methods for Assessment of Fish Production in (37 species) than Msambweni (14 Fresh Waters. Blackwell Scientific, species) and Vanga (8 species). Spatial Oxford. pp101-136. differences in diversity and catch Bostford, L.W., Micheli, F. & Parma, A. between the sites may be attributable M. (2007) Biological and Ecological to differences in fishing effort, habitat Considerations in the Design, structure and recruitment variability. Implementation and Success of MPAs. Fishers land more groupers during In: Report and Documentation of the Expert Workshop on Marine Protected the rough SEM season compared to Areas and Fisheries Management: A the calm NEM season. The study Review of Issues and Considerations. has provided isometric length-weight Rome, 12-14 June 2006. FAO Fish. Tech. relationships for 13 species useful Rep. No.825. Rome, FAO. 332 pp. for modelling the grouper stocks. Caley, M.J., Carr, M.H. Hixon, M.A., Hughes, The species checklist generated in T.P. Jones, G.P. Menge B.A. (1996) Recruitment and the Local Dynamics this work will form useful future of Open of Marine Populations. Annual reference material for studies on the Reviews of Ecology and Systematics 27: distribution and ecology of individual 477-500. grouper species in Kenya and the WIO Campbell, S.J. & Perdede, S.T. (2006) Reef in general. It is recommended that Fish Structure and Cascading Effects the Kenyan government maintains in Response to Artisanal Effects in Response to Artisanal Fishing Pressure. landing statistics at the species level Fish Res. 79: 75-83. for purposes of monitoring changes Domeier, M.L. & Colin, P.L. (1997) Tropical in abundance and diversity of the fish Reef Fish Spawning Aggregations stocks over time. Defined and Reviewed. Bull. Mar. Sci. 60: 698-726.

Volume 9 Final 20th Oct 2010.indd 104 10/27/2010 10:02:20 AM CATCH COMPOSITION, ABUNDANCE AND LENGTH-WEIGHT RELATIONSHIPS OF GROUPERS 101

Dulvy, N.K., Freckleton, R.P.& Polunin, Kaunda-Arara, B. & Rose, G.A. (2004a) N.V.C. (2004) Coral Reef Cascades and Effects of Marine Reef National Parks the Indirect Effects of Predator Removal on Fishery CPUE in coastal Kenya. by Exploitation. Ecology Letters 7: 410- Biol. Cons. 118: 1-13 416. Kaunda-Arara, B. & Rose, G.A. (2004b) Heemstra, P.L. & Randall, J.E. (1993) FAO Outmigration of Tagged Fishes from Species Catalogue. Vol.19. Groupers of Marine Reef National Parks to Fisheries the World. Family Serranidae, Subfamily in coastal Kenya. Environ. Biol. Fish. Epinephelinae. An Annotated and 70: 363-372 Illustrated Catalogue of the Groupers Kaunda-Arara, B., Rose, G.A., Muchiri, M.S. and Lyretail Species Known to Date. & Kaka, R. (2003) Long-term Trends in FAO Fish. Syn., vol. 125, no.16, FAO, coral Reef Fish Yields and Exploitation Rome. 382pp. Rates of Commercial Species from Hixon, M.A & Beets, J.P. (1993) Predation, coastal Kenya. Western India Ocean J. Prey Refuges, and the Structure of Coral Mar. Sci. 2: 105-116. Reef Fish Assemblages. Ecol. Monogr. Kaunda-Arara, B. (1997) Analysis of Catch 63: 77-101. Data from 1985 to 1994 in the Kenyan Huntsman, G.R., Potts, J., Mays, R.W., Inshore Marine Waters. Afr. J. Trop. Vaughan, D. (1999) Groupers Hydrobiol. 7: 1-6. (Serranidae, Epinephelinae): Luckhurst, B.E. (1996) Trends in Commercial Endangered Apex Predators of Reef Fishery Landings of Groupers and Communities. Am. Fish. Soc. Symp. 23: Snappers in Bermuda from 1975 to 1992 217-231. and Associated Fishery Management Jennings, S. & Polunin, N.V.C. (1996) Impacts Issues. In: Pauly D (ed.) Biology, of Fishing on Tropical Ecosystems. Fisheries and Culture of Tropical Ambio 25: 44-49. Groupers and Snappers. ICLARM Conf. Jiddawi, N.S. & Stanley R.D. (1999) Fisheries Proc. 48: 277-288 Stock Assessment in the Traditional McClanahan, T.R. (1988) Seasonality in East Fisheries Sector: The Information Africa’s Coastal Waters. Mar. Ecol. Needs. In: N.S. Jiddawi & R.D. Stanley Prog. Ser. 44: 191-199. (eds). Proceedings of the National Nzioka, R.M. (1979) Observation on the Workshop on the Artisanal Fisheries Spawning Seasons of East African Reef Sector, Zanzibar, September 22–24, Fishes. J. Fish Biol. 14: 329-342. 1997. 50-70 pp. Pauly, D. (1984) Length-converted Catch Kaunda-Arara, B., Mwaluma, J.M., Gamoe, Curves: A Powerful Tool for Fisheries A.L. Oresland, V. & Osore, M.K. Research in the Tropics. Part II. (2009) Temporal Variability in Fish ICLARM Fishbyte 1: 17-19. Larval Supply to Malindi Marine Park, Pitcher, T.J. (1995) Thinking the Unthinkable: coastal Kenya. Aquatic Conservation: a Candidate Model for Predicting Marine and Freshwater Ecosystems 19: Sustainable Yields of Introduced Fish S10-S18. Species in African Lakes. In: Pitcher, Kaunda-Arara, B. & Rose, G.A. (2004) T.J. and Hart P.J.B. (eds.) Impact of Homing and Site Fidelity in Greasy Species Changes in the African Lakes. Grouper Epinephelus tauvina Chapman and Hall, London. pp 495- (Serranidae) within a Marine Protected 526. Area in coastal Kenya. Mar. Ecol. Prog. Ser. 277: 245-251.

Volume 9 Final 20th Oct 2010.indd 105 10/27/2010 10:02:20 AM 102 S. AGEMBE ET AL.

Rhodes, K.L. & Sadovy, Y. (2002) Temporal Samoilys, M.A., Church, J., Kaunda-Arara, and Spatial Trends in Spawning B., Kamukuru, A. & Jiddawi, N. (2006) Aggregations of Camouflage Grouper, Preliminary Findings on Spawning Epinephelus polyphekaidon, in Pohnpei, Aggregations of Reef Fishes in East Micronesia. Env. Biol. Fish. 63: 27-39. Africa. Proc. 10th Int. Coral Reef Symp. Ricker, W.E. (1975) Computation and Okinawa, July 2004: 1335-1346. Interpretation of Biological Statistics Smith, M.M. & Heemstra, P.C. (1986) Smith’s of Fish Populations. Bull. Fish. Res. Sea Fishes. Macmillan South Africa, Board Canada 191: 1-32. Johannesburg. 1047 pp. Sadovy, Y.J. (1996) Reproduction of Reef Smith, C.L. (1965) The Patterns of Sexuality Fishery Species. In: Polunin, N.V.C. and the Classification of Serranid & Roberts, C.M. (eds.) Reef fisheries. Fishes. Am. Mus. Novitates, 2207: 1-20. Chapman & Hall, London. pp 15-59. Wootton, R.J. (1990) Ecology of Teleost Sadovy, Y.J. (1994) Grouper Stocks of the Fishes. Fish and Fisheries Series. Western Central Atlantic: The need for Chapman and Hall, New York. 404 pp. Management and Management needs. Zar, J.H. (1996) Biostatistics Analysis. Proc. Gulf Carib. Fish. Inst. 43: 43-65. Prentice Hall, New Jersey. 662 pp. Samoilys, M.A. (1988) Abundance and Species Richness of Coral Reef Fish on the Kenyan Coast: The effects of Protective Management and Fishing. Proc. Int. Coral Reef Symp. 2: 261-266.

Volume 9 Final 20th Oct 2010.indd 106 10/27/2010 10:02:20 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 103 - 114, 2010 © 2010 WIOMSA

Calibration of Community-based Coral Reef Monitoring Protocols: Tanzanian Case Study

C.A. Muhando Institute of Marine Sciences, P.O. Box 668, Zanzibar, TANZANIA.

Keywords: coral reef monitoring, community-based, calibration

Abstract—Coral reef monitoring (CRM) has been recognised as an important management tool and has consequently been incorporated in Integrated Coastal Area Management (ICAM) programmes in the Western Indian Ocean (WIO). Community-based coral reef monitoring (CB-CRM), which uses simplified procedures suitable for local conditions, was introduced in Tanzania in 1996. Despite its widespread use, the method has not been calibrated and the validity of merging CB-CRM results with those gained using other techniques has not been determined. In this study, CB-CRM protocols adopted by the Tanga Coastal Zone Conservation and Development Programme (TCZCDP) were tested against SCUBA-based coral reef monitoring (SB-CRM) as practiced by the Institute of Marine Sciences, University of Dar es Salaam. Calibration showed no significant differences in measuring percent cover of live hard corals, sponges, dead corals and substrata (non-biotic cover). However, CB-CRM monitors recorded higher soft coral and lower fleshy algal cover. Larger differences were observed in deeper (>6 m) transects. Counts of sea cucumbers, clams, gastropods and bivalves categories were not significantly different. However, CB-CRM underestimated the abundance of sea urchins, starfish and younger macro-invertebrates in crevices or under overhangs. There were no differences in the identification of reef fish categories but CB- CRM recorded slightly higher reef fish densities than SB-CRM. If properly trained, CB-CRM monitors can generate results that are comparable to those obtained from SB-CRM on shallow reefs. Although a powerful tool which engenders community involvement and a sense of ownership in the sustainable use of coastal resources, CB-CRM has limitations of which managers need to be aware.

E-mail:[email protected]

Volume 9 Final 20th Oct 2010.indd 107 10/27/2010 10:02:21 AM 104 C.A. MUHANDO

INTRODUCTION (SB-CRM) yields detailed results, the method is relatively expensive Concern over the worldwide (Wilkinson et al., 2003; Muhando, 2009) degradation of coral reefs due to and was found inadequate for ICM in various natural and anthropogenic developing countries (Makoloweka factors (Jackson et al., 2001; Hughes and Shurcliff, 1997; Horrill et al., et al., 2003; McClanahan et al., 2007; 2001). Instead, community-based coral Hoegh-Guldberg et al., 2007) has reef monitoring (CB-CRM), based on highlighted the need for effective English et al., (1994) and Reef Check coral reef management programmes (Hodgson and Liebeler, 2002) protocols, in Tanzania (Makoloweka et al., 1997; was introduced in the Tanga Coastal Verheij et al., 2004; Muhando, 2008). Zone Conservation and Development The declaration of marine protected Programme (TCZCDP) in 1996 (Horrill areas (parks, reserves, conservation et al., 2001; Verheij et al., 2004) and later areas) and, more recently, collaborative adopted by the Kinondoni Integrated management areas (Christie et al., Coastal Area Project (KICAMP) 2002; Verheij et al., 2006; Wells et al., (Wagner, 2004). CB-CRM is executed 2007), among others, have constituted by trained local fishers and fisheries attempts to protect and conserve the officers, with scientists as supervisors, Tanzanian coral reefs from human as practiced in the Philippines damage (UNEP, 1989; Johnstone et (Uychiaoco et al., 2005). CB-CRM is al., 1998a; Johnstone et al., 1998b; more popular than SB-CRM, mainly Muhando and Francis, 2000; Horrill because it is cost effective and enhances et al., 2001; Verheij et al., 2006; the education of local communities, Samoilys et al., 2007). their environmental awareness and their Effective implementation of stewardship of natural resources (Hill integrated coastal management and Wilkinson, 2004; Subade et al., (ICM) programmes is dependent on 2008). More Tanzania coastal district information gained from ecological and ICM programmes, e.g., Bagamoyo, socio-economic monitoring and research Kilwa, and Mafia, have introduced the (McManus et al., 1988; Wilkinson et CB-CRM protocols. al., 2003). Ecological monitoring of Despite the wide use of this coral reefs using protocols described recognised technique, its reliability in English et al., (1994) was first and comparability have not been tested instituted in Tanzania by the Institute of against data gained from methods Marine Sciences (IMS) in 1994 using such as SB-CRM. Cross-calibration self-contained underwater breathing was considered necessary to raise apparatus (SCUBA) (Mohammed et al., the confidence of information users, 2000, 2002; Muhando, 2008). Although especially ICM managers, in CB-CRM SCUBA-based coral reef monitoring data. In this study, cross-calibration

Volume 9 Final 20th Oct 2010.indd 108 10/27/2010 10:02:21 AM CALIBRATION OF COMMUNITY-BASED CORAL REEF MONITORING PROTOCOLS 105

was conducted in a joint venture by the calibration exercise was conducted on Tanga Coastal Zone Conservation and Taa and Makome reefs, Tanga (Fig. 1). Development Programme (TCZCDP) and the Institute of Marine Sciences with Similarities and differences the objectives of: i) establishing whether between the CB-CRM and results obtained by CB-CRM and SB- SB-CRM protocols CRM are comparable, elucidating The CB-CRM team recorded all the live sources of error and ii) recommending hard corals in one category, ‘Matumbawe modifications for improvement in the hai’, while the SB-CRM team used the 13 CB-CRM protocols. coral growth forms listed in Table 1. The five algal groups in the SB-CRM records MATERIALS AND METHODS were grouped as one category, ‘Mwani’, The CB-CRM protocols are fully in the CB-CRM. Categories such as described by Horrill et al., (2001), soft corals, sponges, seagrass, sand, Verheij et al., (2006), and Samoilys and rock were common to both groups. et al., (2007), while the SB-CRM Three further categories in the SB-CRM protocols are described by Mohammed method, dead coral, dead coral with algae et al., (2000) and Muhando (2008). and rubble, were lumped in one category, The coral reef categories used in the ‘Matumbawe yaliyokufa’ (dead corals), calibration are listed in Tables 1-3. The in the CB-CRM. Categories such as silt

Fig. 1: Map of the Tanga coastline and location of the study sites.

Volume 9 Final 20th Oct 2010.indd 109 10/27/2010 10:02:21 AM 106 C.A. MUHANDO

Table 1. Coral reef benthic categories used in the SB-CRM, CB-CRM and calibration process. SB-CRM CB-CRM CALIBRATION GENERAL DESIGNATION BENTHIC CATEGORIES Acropora, branching (ACB) Acropora, encrusting (ACE) Acropora, submassive (ACS) Acropora, digitate (ACD) Live hard Acropora, tabulate (ACT) Matumbawe hai (MH) Live coral corals (HC) Coral, branching (CB) Coral, encrusting (CE) Coral, foliose (CF) Coral, massive (CM) Coral, submassive (CS) Coral, mushroom (CMR) Coral, Millepora (CME) Coral, Heliopora (CHL) Matumbawe yaliyokufa Partly dead coral kidogo (MKK) Matumbawe hai maeupe (MHM) Bleached coral Soft corals (SC) Soft coral (SC) Matumbawe laini (ML) Soft coral Sponges (SP) Sponges (SP) Spongi (SP) Sponge Coralline algae (CA) Algae (AL) Algal assemblage (AA) Mwani (MN) Algae Algae, Halimeda (HA) Macroalgae (MA) Turf algae (TA) Seagrass (SG) Majani (MJ) Seagrass Zoanthids (ZO) Others (OT) Clam (CLAM) Corallimorpharian (RH) Others (OT) Other organisms Others (OT) Sand (S) Mchanga (MC) Sand Substratum (SU) Silt (SI) Rock (RCK) Mwamba (MW) Rock Rubble (R) Dead coral (DC) Matumbawe yaliyokufa (MK) Dead coral Dead coral with algae (DCA)

(which represented sediment stress), meupe’) in the CB-CRM had no zoanthids, clams and corallimorpharians equivalent categories in the SB-CRM (mostly Rhodactis) were not recorded in monitoring system. the CB-CRM programme. Partially dead Important macro-invertebrate coral (‘Matumbawe yaliyokufa kidogo’) categories such as lobsters, clams, and bleached corals (‘Matumbawe gastropods, sea cucumbers, starfish,

Volume 9 Final 20th Oct 2010.indd 110 10/27/2010 10:02:21 AM CALIBRATION OF COMMUNITY-BASED CORAL REEF MONITORING PROTOCOLS 107

Table 2. Coral reef macro-invertebrate categories used in the SB-CRM, CB-CRM and calibration process. PHYLUM SB-CRM CB-CRM CALIBRATION Crustacea Lobsters Kamba koche (Lobsters) Lobsters Clams Nyera (e.g. Tridacna) Clams Mollusca Gastropods Nyale (e.g. Lambis) Gastropods Bivalves Makome (Shells) Bivalves Pweza (Octopus) Octopus Crown-of-thorns starfish (COTS) Matokambe (COTS) COTS Sea urchins Ufuma macho Sea Ufuma mawe Urchins Ufuma moto Echinodermata Ufuma bondo Starfish Kiti cha pweza or Starfish Tawangwe (starfish) Sea cucumbers Jongoo bahari Sea cucumbers

Table 3. Fish recording template for CB-CRM. Category designation Description Chafi Family Siganidae Chewa Family Serranidae Changu Family Lethrinidae and some Lutjanidae Chazanda Lutjanus argentimaculatus Kangu wadogo Selected smaller members of Scaridae and Labridae Kangu wakubwa Selected larger members of Scaridae and Labridae Kangaja Family Acanthuridae: Species of the genera Ctenochaetus and Acanthurus, except A. triostegus, Kolekole Family Carangidae Kitamba Plectorhinchus sordidus, P. playfairi, P. flavomaculatus. Kidui Family Balistidae Kipepeo Family Chaetodontidae Mlea Plectorhinchus gaterinus, and P. orientalis Mwasoya Family Pomacanthidae: Only species of the genera Pomacanthus and Pygoplites Mkundaji Family Mullidae Haraki Lutjanus bohar Tembo Lutjanus fulviflamma, L.lutjanus, L. ehrenbergii Mbono Family Caesionidae

Volume 9 Final 20th Oct 2010.indd 111 10/27/2010 10:02:21 AM 108 C.A. MUHANDO

sea urchins and crown-of-thorns cucumbers, sea stars and crown- starfish were included in both the CRM of-thorns starfish (Table 2) were programmes. However, the community counted in 2 x 20 m belt transects monitors divided sea urchins (‘Ufuma’) after recording the reef cover. Unlike into four categories: ‘Ufuma macho’ the SB-CRM monitors, the CB-CRM (Diadema setosum), ‘Ufuma mawe’ monitors dived up and down to identify (Echinometra mathaei), ‘Ufuma moto’ and record the benthos and count (Diadema savignyi) and ‘Ufuma bondo’ macroinvertebrates. This procedure (Echinothrix diadema). Molluscs were was repeated on eight transects: four subdivided into clams, gastropods, at Mwamba Taa (at 1 m and at 6 m) bivalves and octopus (Table 2). Octopus and four at Mwamba Makome (at 4 m were only recorded by the community and at 9 m). The CB-CRM monitors monitors. SB-CRM monitors counted recorded data using Kiswahili names, macro-invertebrates in 2 x 20 m long while the life-form categories of belt transects, while this was done in English et al., (1994) were retained in wider but shorter 5 x 10 m plots in the the SB-CRM . CB-CRM. The number of fish categories Reef fish: Reef fish were counted in recorded by the SB-CRM group was four 5 x 50 m belt transects, set between far too detailed (Mohammed et al., two parallel 50 m ropes set 5 m apart 2002) for use in the CB-CRM (Table at the lower end of the reef slope. Fish 3), the latter also being biased towards counts were undertaken 10 minutes commercial reef fish species. or more after setting the transects Routine CRM monitoring in and between counts to allow fish to both protocols involved the use of resume normal behaviour. The reef randomly-set, line-intercept transects fish counted included commercially (LITs) within permanent marked plots. and ecologically important families Reef benthos and Macro- or groups and were categorised in 17 invertebrates: A 20 m measuring tape groups, conveniently adopted from was laid over the reef and attached the CB-CRM protocols (Table 3). The with iron stakes to ensure that all eight CB-CRM monitors, two at a monitors followed the same transect time, counted fish with the SB-CRM line. Live coral cover, coralline monitors following about 1-2 meters algae, soft corals, sponges, fleshy below and to their rear. algae and non-biotic cover (Table 1) The CB-CRM and SB-CRM reef were assessed using the line-intercept benthos and macro-invertebrate data transect (LIT) method (English et al., were tested for differences using 1994) by eight CB-CRM monitors and the Student’s two-tailed t-test. Reef three SB-CRM monitors. Similarly, fish densities were compared by macro-invertebrates such as lobsters, calculating the percentage difference clams, gastropods, sea urchins, sea in each category. The performance

Volume 9 Final 20th Oct 2010.indd 112 10/27/2010 10:02:21 AM CALIBRATION OF COMMUNITY-BASED CORAL REEF MONITORING PROTOCOLS 109

Table 4. Comparison (Student’s two-tailed t test) of benthic reef cover recorded by the SB-CRM and CB-CRM monitors. Benthic category T df p Difference between SB-CRM and CB-CRM Hard coral 0.70 12 0.4963 Not significant Bleached coral * * * Significant - not observed by SB-CRM team Coralline algae * * * Significant - not observed by CB-CRM team Algae 1.41 12 0.1855 Not significant Soft coral 7.37 12 < 0.0001 Significant (CB-CRM > SB-CRM) Sponge 0.44 12 0.665 Not significant Other Organisms 3.03 12 0.0105 Significant (SB-CRM > CB-CRM) Dead coral 0.98 12 0.3112 Not significant Substratum 1.36 12 0.2003 Not significant

of the CB-CRM team in identifying deeper water; correspondingly greater categories was evaluated by estimating differences were observed between the the number of categories they recorded groups in transects on deeper reefs. relative to those by the SB-CRM Comparison of results revealed monitors. Performance in counting that CB-CRM overestimated the reef fish was evaluated by estimating abundance of soft corals due to their the Pearson correlation coefficient misidentification of corallimorphs, and similarity of counts (by category) zoanthids, sea anemones and turf between the monitoring groups. algae. Whitish/pinkish coralline algae were incorrectly recorded as bleached RESULTS coral by CB-CRM monitors. Calibration of CB-CRM Macro-invertebrate densities Reef benthic cover While no lobsters, octopus, and crown-of-thorns starfish were There was no significant difference observed, sea urchins (‘Ufuma’) were between the CB-CRM and SB-CRM the dominant macro-invertebrate records of hard coral cover, sponges, recorded in the transects (Table 5). dead coral or reef substrata (non-living The density of sea cucumbers, clams, components) (Table 4). However, gastropods and bivalves recorded by CB-CRM monitors recorded a higher the CB-CRM and SB-CRM monitors cover of soft corals and lower cover of was not significantly different (Table ‘other organisms’. The SB-CRM team 5). However, CB-CRM monitors recorded no bleached coral and the counted fewer sea urchins and starfish CB-CRM group reported no coralline than the SB-CRM monitors (Table 5), algae. CB-CRM monitors were more at especially on the deeper transects. ease in water <6 m deep and less so in

Volume 9 Final 20th Oct 2010.indd 113 10/27/2010 10:02:21 AM 110 C.A. MUHANDO

Table 5. Comparison (Student’s two-tailed t test) of macro-invertebrate counts recorded by the SB-CRM and CB-CRM monitors. Category t df p Difference between SB-CRM and CB-CRM Sea urchins 2.21 8 0.045 Significant (SB-CRM > CB-CRM) Sea cucumbers 1.66 8 0.162 Not significant Starfish 2.74 8 0.031 Significant (SB-CRM > CB-CRM) Crown-of-thorn-starfish - - - Not observed Gastropods * * * Not significant Bivalves * * * Not significant Clams * * * Not significant Octopus - - - Not observed Lobsters - - - Not observed

Reef fish densities ‘Chewa’ (Serranidae) (-30%), ‘Mlea’ The total fish count showed that CB- (-25%) and ‘Kangu s’ (small Scaridae CRM recorded higher fish densities and Labridae) (-19%). CB-CRM (42.1 fish per 250 2 m ) than the SB- and SB-CRM counts of ‘Mwasoya’ CRM monitors (34.6 fish per 250 2m ). (Pomacanthidae) were identical. Neither Identification of the reef fish categories group of monitors recorded ‘Chazanda’ was similar, with the CB-CRM group (Lutjanus argentimaculatus), identifying 12 and the SB-CRM ‘Kolekole’ (Carangidae), ‘Kitamba’ group 11 of the seventeen pre-selected (Plectorhinchus sordidus, P. playfairi, fish categories (Table 6). ‘Changu’ P. flavomaculatus), ‘Harak’ (Lutjanus (Lethrinidae and some members of bohar) or ‘Tembo’ (Lutjanus Lutjanidae) were recorded only by fulviflamma, L. lutjanus, L. ehrenbergii) the CB-CRM monitors. Differences (Table 6). in counting were notable amongst the Reef fish identification by individual ‘Chafi’ (Siganidae), with CB-CRM CB-CRM monitors was generally good. monitors recording densities of these Over 75% of the monitors were able to fish 1450% higher than SB-CRM identify >80% of the fish categories monitors (Table 6). Other categories recorded by the SB-CRM group (which recorded in higher densities by the CB- comprised the control) (Table 7); two CRM monitors included: ‘Mbono’ monitors were 100% accurate in their (Caesionidae) (325%), ‘Kangu l’ fish identification. However, most (large Scaridae and Labridae) (205%), reported one or more categories in ‘Kipepeo’ (Chaetodontidae) (76.2%), addition to those targeted observation is ‘Mkundaji’ (Mulidae) (56.3%) and always more useful as it generates data ‘Kidui’ (Balistidae) (50%). On the other specific to the management of fisheries hand, they recorded lower densities of problems (Labrosse et al., 2002).

Volume 9 Final 20th Oct 2010.indd 114 10/27/2010 10:02:21 AM CALIBRATION OF COMMUNITY-BASED CORAL REEF MONITORING PROTOCOLS 111

Table 6. Density of reef fish recorded per 250 m-2 by the SB-CRM, CB-CRM moni- tors and the mean difference (%) for each fish category counted. Name CB-CRM SB-CRM % difference Chafi Siganidae 1.107 0.071 1450 Chewa Serranidae 0.25 0.357 -30 Changu Lethrinidae and some Lutjanidae 1 0 - Chazanda Lutjanus argentimaculatus 0 0 0 Kangu l Large Scaridae and Labridae 2.179 0.714 205 Kangu s Small Scaridae and Labridae 11.18 13.79 -18.9 Kangaja Acanthuridae (Ctenochaetus spp. and 13.79 13 6.0 Acanthurus spp. except A. triostegus) Kolekole Carangidae 0 0 0 Kitamba Plectorhinchus sordidus, P. playfairi, 0 0 0 P. flavomaculatus Kidui Balistidae 0.107 0.071 50 Kipepeo Chaetodontidae 7.679 4.357 76.2 Mlea Plectorhinchus gaterinus, and P. orientalis 0.214 0.286 -25 Mwasoya Pomacanthidae (Pomacanthus spp.and 1.321 0.786 2 Plygoplites spp.) Mkundaji Mulidae 0.893 0.571 56.25 Haraki Lutjanus bohar 0 0 0 Tembo Lutjanus fulviflamma, L.lutjanus, L. ehrenbergii 0 0 0 Mbono Caesionidae 2.429 0.571 325 Total 42.14 34.57 21.9 Total fish categories 17 12 11 The reef fish calibration undertaken overestimate the fish stocks and in this study revealed that the may wrongly encourage managers identification of reef fish categories to allow more fishing. Analysis of did not pose any problems. However, the performance of the CB-CRM on average, CB-CRM monitors monitors in fish identification and recorded higher densities than their counting graded six out of the eight as SB-CRM counterparts (by about 22%; good to very good. Such calibrations Table 6a). A similar calibration study are important to maintain the quality in the Philippines yielded greater of the CB-CRM datasets (Gaudian variation and higher fish abundance in et al., 1995). In this study, it was CB-CRM than SB-CRM (Uychiaoco recommended that the quality of the et al., 2005). Such biases may be data would be improved by excluding attributable to the greater area view data from the two poor fish monitors. covered in CB-CRM as the monitors are positioned above the SCUBA CONCLUSIONS divers. Taking this into consideration, This study has shown that CB-CRM the 22% error (Table 6) probably is useful in monitoring coral reef falls within tolerable levels (Carr benthic cover. Modifying or removing et al., 2002). An error >50% would confusing categories, the use of

Volume 9 Final 20th Oct 2010.indd 115 10/27/2010 10:02:22 AM 112 C.A. MUHANDO

illustrative underwater guides and Carreiro-Silva, M. & McClanahan, T.R. frequent calibration should further (2001) Echinoid Bioerosion and improve the method. CB-CRM is Herbivory on Kenyan Coral Reefs: The Role of Protection from Fishing. J. Exp. depth-dependant and is most effective Mar. Biol. Ecol. 262: 133-153. in shallow water; hence additional Chabanet, P., Adjeroud, M., Andréfouët, S., strategies are needed in deeper or Bozec, Y., Ferraris, J., Garcìa-Charton, more complex coral reef habitats. The J. & Schrimm, M. (2005) Human- method of counting coral reef fish induced Physical Disturbances and adopted in the CB-CRM was simple their Indicators on Coral Reef Habitats: A multi-scale approach. Aquat. Living and convenient and should be effective Resour. 18: 215-230. in marine protected area (MPA) Christie, P., White, A. & Deguit, E. (2002) management. A CB-CRM manual Starting point or solution? Community- describing indicator categories would based Marine Protected Areas in the be a useful reference for monitors, and Philippines. Journal of Environmental would provide reef managers a tool to Management 66: 441- 454. assist in the interpretation of reef data. English, S., Wilkinson, C. & Baker, V. (1994) Survey Manual for Tropical Marine Resources. Australian Institute of Acknowledgements—I wish to Marine Science, Townsville, 358 pp. acknowledge Sida SAREC and Gaudian, G., Medley, P.A. & Ormond, R.F.G. the Institute of Marine Sciences, (1995) Estimation of the Size of a Coral Zanzibar, for financing this study; Reef Fish Population. Marine Ecology Solomon Mwakaloweka, Eric Verheij Progress Series 122:107-113. and Hassan Kolombo for providing Hill, J. & Wilkinson, C. (2004) Methods for logistical support and organizing Ecological Monitoring of Coral Reefs: A Resource for Managers. Australian fishers and fisheries officers as Institute of Marine Science. Townsville. community-based coral reef monitors; 117 pp. and the Institute of Marine Sciences Hoegh-Guldberg, O., Mumby, P.J., Hooten, Coral Reef Monitoring Team, A.J., Steneck, R.S., Greenfield, P., Mohammed Suleiman, Nsajigwa Gomez, E., Harvell, C. D., Sale, P. F., Mbije, Haji Machano, Evans Edward Edwards, A. J., Caldeira, K., Knowlton, and Mohammed Nur, for assistance N., Eakin, C. M., Iglesias-Prieto, R., Muthiga, N., Bradbury, R. H., Dubi, A. with this study. & Hatziolos, M. E. (2007) Coral Reefs Under Rapid Climate Change and Ocean REFERENCES Acidification.Science 318: 1737-1742. Horrill, J.C., Kalombo H & Makoloweka Carr, M.H., Anderson T.W. & Hixon, M.A. S., (2001). Collaborative Reef and (2002) Biodiversity, Population Reef Fisheries Management in Tanga, Regulation, and the Stability of Coral- Tanzania. IUCN EA-Program, Nairobi. Reef Fish Communities. Ecology 99: 37 pp. 11241–11245.

Volume 9 Final 20th Oct 2010.indd 116 10/27/2010 10:02:22 AM CALIBRATION OF COMMUNITY-BASED CORAL REEF MONITORING PROTOCOLS 113

Hodgson, G. & Liebeler, J. (2002) The Global McManus J.W., Ferrer E.M. and Campos Coral Reef Crisis: Trends and solutions. W.L. (1988) A Village-Level Approach Reef Check Foundation. 79 pp. to Coastal Adaptive Management and Hughes, T.P., Baird, A.H., Bellwood, D.R., Resource Assessment (CAMRA). Card, M., Connolly, S.R., Folke, C., Proceedings of the 6th International Grosberg, R. & Hoegh-Guldberg, Coral Reef Symposium 2: 381-386. O. (2003) Climate Change, Human McClanahan, T.R. (1994). Kenyan Coral Impacts and the Resilience of Coral Reef Lagoon Fish: Effects of Fishing, Reefs. Science 301: 929-933. Substrate Complexity, and Sea Urchins. Jackson, J.B.C., Kirby, M.X., Berger, W.H., Coral Reefs 13: 231-241. Bjorndal, K.A., Botsford, L.W., McClanahan, T.R. & Shafir S. H. (1990) Bourque, B.J., Bradbury, R.H., & Cooke Causes and Consequences of Sea R. (2001) Historical Overfishing and Urchin Abundance and Diversity in Recent Collapse of Coastal Ecosystems. Kenyan Coral Reef Lagoons. Oecologia Science 293: 629-638. 83: 362-370. Johnstone, R., Muhando, C. & Francis, J. McClanahan, T., Ateweberhan, M., Muhando, (1998a) The Status of Coral Reef of C., Maina, J. & Mohammed, M.S. Zanzibar: One Example of a Regional (2007) Effects of Climate and Seawater Predicament. Ambio 27: 700-707. Temperature Variation on Coral Johnstone, R.W., Francis, J. & Muhando, Bleaching and Mortality. Ecological C.A. (eds.) (1998b) Coral Reefs: Values, Monographs 7: 503–525 Threats and Solutions. Proceedings Mohammed, M.S., Muhando, C.A. & of the National Conference on Coral Machano, H. (2002) Coral Reef Reefs, Zanzibar, Tanzania, December Degradation in Tanzania: Results of 1997. Institute of Marine Sciences. Monitoring 1999-2002. In: O. Linden, Zanzibar, 124 pp. D. Souter, D. Wilhelmsson, & D. Obura Labrosse, P., Kulbicki, M & Ferraris J (2002) (Eds.) Coral Reef Degradation in the Underwater Visual Fish Census Surveys: Indian Ocean. Status Report 2002. Proper Use and Implementation. CORDIO. Kalmar Sweden, pp 21-30. Secretariat of the Pacific Community Mohammed, M.S., Muhando, C.A. & (SPC). Noumea, New Caledonia, 54 pp. Machano, H. (2000) Assessment of Malleret-King, D., Glass, A., Bunce, L. & Coral Reef Degradation in Tanzania: Pomeroy, B. (2006) Socioeconomic Results of Coral Monitoring 1999. In: Monitoring Guidelines for Coastal Souter, D., Obura, D., and Linden, O. Managers of the Western Indian Ocean. (eds.) 2000. Coral Reef Degradation in Socmon WIO. CORDIO East Africa the Indian Ocean: Status Report 2000. Publication (version 1), 108 pp. CORDIO Stockholm, pp 35-42. Makoloweka, S. & Shurcliff, K. (1997) Muhando, C.A. (1999) Assessment of the Coastal Management in Tanga, Extent of Coral Damage, Socio- Tanzania: a Decentralized Community- Economics Effects Mitigation and Based Approach. Ocean and Coastal Recovery of Coral Reefs in Tanzania. Management 37: 349-357. In: Linden, O & Sporrong N (eds). Coral Reef Degradation in the Indian Ocean: Status Report and Project Presentation. CORDIO, Stockholm, pp 43-47.

Volume 9 Final 20th Oct 2010.indd 117 10/27/2010 10:02:22 AM 114 C.A. MUHANDO

Muhando, C.A. (2008) Approaches to Coral Verheij, E., Makoloweka, S. & Kalombo, Reef Monitoring in Tanzania. In: H. (2004) Collaborative Coastal Obura, D, Tamelander J & Linden O. Management Improves Coral Reefs and (eds.) Coastal Oceans Research and Fisheries in Tanga, Tanzania. Ocean Development in the Indian Ocean: and Coastal Management 47: 309–320. Status Report 2008. CORDIO, pp 129- Verheij, E., Samoilys, M.A & Kalombo, 138. H. (2006) Assessing the Impact of a Muhando, C.A. & Francis, J. 2000. The Status Community-Based Network of Marine of Coral Reefs in the Dar-es-salaam Protected Areas Through Long Term Marine Reserves System and the State Monitoring of Coral Reef Resources. of Reefs in Other Marine Protected Proceedings of 10th International Coral Areas of Tanzania. IMS/UNEP/ICLAN Reef Symposium 2: 1396-1404. Report, 32 pp. Wagner, G.M. (2004) Coral Reefs and Their Muhando, C.A (2009) Coral Reef Monitoring Management in Tanzania. Western in Tanzania: An Analysis of the Last 20 Indian Ocean J. Mar. Sci. 3: 227–243. years. Western Indian Ocean J. Mar. Wells, S., Horrill, C., Kalombo, H., Kabamba, Sci. 8: 203-214. J. & Verheij, E. (2007) Collaborative Samoilys, M., Horrill, C., Kalombo, H., Management Area Planning. In: Wells Kabamba, J. & Wells, S. (2007) Coral S., Makoloweka S and Samoilys M. (ed.) reefs and Mangroves – Maintaining (2007): Putting Adaptive Management Ecosystem Health. In: Wells, S., into Practice: Collaborative Coastal Makoloweka S & Samoilys, M. (eds.) Management in Tanga, northern (2007) Putting Adaptive Management Tanzania, pp 22-45. into Practice: Collaborative Coastal White, AT and Vogt, HP (2000) Philippine Management in Tanga, Northern Coral Reefs Under Threat: Lessons Tanzania, pp 77-102. Learned After 25 Years of Community- Subade, A.L.A., Subade, R.F. & Catalan, Based Reef Conservation. Marine Z.B. (2008) Towards Local Fishers Pollution Bulletin 40: 537-550. Participation in Coral Reef Monitoring: Wilkinson, C., Green, A., Almany, J. & A Case in Tingloy, Batangas, Dionne, S. (2003) Monitoring Coral Philippines. Proceedings of the 11th Reef In Marine Protected Areas: A International Coral Reef Symposium, Practical Guide on How Monitoring Session 21: 983-987 Can Support Effective Management of UNEP (1989) Coastal and Marine MPAs. Australian Institute of Marine Environmental Problems of the United Science and the IUCN Marine Program, Republic of Tanzania. UNEP Regional 68 pp. Seas Reports and Studies No. 106. 33 pp. + annex. Uychiaoco, A J., Arceo, H.O., Green S.J., De La Cruz, M.T, Gaite, P.A. & Alino, P.M. (2005) Monitoring and Evaluation Of Reef Protected Areas by local Fishers in the Philippines: Tightening The Adaptive Management Cycle. Biodiversity and Conservation 14: 2775–2794.

Volume 9 Final 20th Oct 2010.indd 118 10/27/2010 10:02:22 AM Western Indian Ocean J. Mar. Sci. Vol. 9, No. 1, pp. 115 - 134, 2010 © 2010 WIOMSA SHORT COMMUNICATION

Rapid Visual Assessment of Fish Communities on Selected Reefs in the Bazaruto Archipelago

Jade Q. Maggs1, Camilla Floros1, Marcos A.M. Pereira2 and Michael H. Schleyer1 1Oceanographic Research Institute, P.O. Box 10712, Marine Parade, 4056 South Africa; 2Associação para Investigação Costeira e Marinha (AICM), P.O. Box 2046, Maputo, Mozambique.

Keywords: Bazaruto Archipelago, Mozambique, ichthyofauna, fish surveys, underwater visual census, coral reefs.

Abstract— Rapid visual censuses were conducted of fish on eight coral reefs in the Bazaruto Archipelago, Mozambique, in 2007. SCUBA and snorkelling were used for the censuses in depths between 1-20 m, yielding an inventory of 249 fish species belonging to 50 families. This is intended to serve as a baseline for more detailed studies and monitoring programmes in the future. Although fewer species were recorded relative to other studies conducted in the Western Indian Ocean, the trophic structure on Bazaruto’s reefs proved typical for the region, indicating a relative measure of reef health. However, other regional studies were not directly comparable, differing in habitat, duration of sampling effort and methodology. This highlighted the need for a long-term monitoring programme specifically adapted for the Bazaruto reef types to provide a basis for their sound management and conservation.

INTRODUCTION Mozambique in the Western Indian Ocean (WIO). The seas around the The Bazaruto Archipelago consists archipelago are rich in marine life of five islands and is located and provide an important source approximately 20 km off the coast of of protein to the local community

Corresponding Author: JQM E-mail: [email protected]

Volume 9 Final 20th Oct 2010.indd 119 10/27/2010 10:02:22 AM 116 J. Q. MAGGS ET AL.

(Everett et al., 2008; Reina, 1998). beach seines, gamboa traps and The ecological importance of the spearguns. Mozambican citizens living archipelago first received attention outside the BANP are only allowed in 1971 with three of the islands, to use handlines inside the park and Benguera, Magaruque and Bangue, are allowed to use beach-seines in a receiving national park status (area of small area south of Magaruque. Gill- protection ~600 km2); Bazaruto and netting is prohibited and no industrial Santa Carolina were only designated or semi-industrial fishing operations as ‘special surveillance zones’ (Reina, are allowed in the BANP. The no-take 1998). After many years, the Bazaruto zones are Two-mile Reef, Lighthouse Archipelago National Park (BANP) Reef, Santa Carolina and small rocky was proclaimed in 2001, protecting all outcrops on the inside and outside of five islands. The BANP was extended Bazaruto Island (Fig. 1). However, in 2003 to include the Cabo de São only Two-mile Reef, Lighthouse Sebastião peninsula in the south and Reef and Santa Carolina enjoy strong now covers 1430 km2. Mozambique’s compliance. There are no seasonal Ministry of Tourism is responsible for restrictions on fishing in the BANP management of the BANP. and recreational SCUBA diving and The coral and rocky reefs in the snorkelling is allowed in the no-take archipelago provide habitat for a zones. A new management plan is wealth of biodiversity, making it a being developed for the Park but, at popular tourist destination (Schleyer the time of writing, had not yet been & Celliers 2005). Visiting SCUBA implemented. divers and recreational anglers bring The livelihoods of the local in valuable revenue to the area and, communities depend, to a great extent, although fishing is allowed in the on marine and coastal resources BANP, it is regulated by means (Everett et al., 2008). An increasing of permits and no-take zones. All population characterises many coastal recreational fishing requires a permit communities in the WIO, placing and is mostly boat-based, emanating pressure on such resources (Lindén from a number of resorts scattered et al., 2002). Effective management through the islands. A number of is thus required, the success of which seasonal fishing competitions are depends on monitoring programmes hosted by the various resorts, bringing (Obura et al., 2002). These have been in foreign anglers. seen as a priority in the WIO region The artisanal fishery is the main since the 1998 mass coral bleaching economic activity for more than 70% event (Lindén et al., 2002). Monitoring of the local population (Everett et al., of a resource depends on a thorough 2008). Artisanal fishermen harvest knowledge of the biodiversity of an fish using dhows, pirogues, approved area. Species inventories comprise

Volume 9 Final 20th Oct 2010.indd 120 10/27/2010 10:02:22 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 117

a fundamental baseline in this MATERIALS AND METHODS regard, providing a foundation for an understanding of ecological processes Study Area and the effects of biodiversity loss The Bazaruto Archipelago is a chain on ecosystem function (Bellwood & of four islands, Bazaruto, Benguera, Hughes 2001; Gillibrand et al., 2007). Magaruque and Bangue, with a fifth Baseline fish community data are island, Santa Carolina, lying on the sparse for the WIO region but those inside of the island chain (Fig. 1). published include fish inventories for The reefs around the archipelago are the Glorieuses Islands (Durville et diverse and include rocky patch reefs, al., 2003), Mayotte (Chabanet, 2002), rocky massifs, fringing and barrier Andavadoaka (Gillibrand et al., 2007), coral reefs and deeply submerged Juan De Nova (Chabanet & Durville coral reefs (Schleyer & Maggs 2008). 2005), Tuléar (Harmelin-Vivien, The fish community assessment was 1979) Geyser and Zéléé (Chabanet et undertaken on all these reef types in al., 2002), Réunion (Chabanet, 1994), depths from 1-20 m. A range of reef Sodwana Bay (Chater et al., 1993; habitats within these reef types was 1995) and Bassas da India (van der sampled with varying coral cover and Elst & Chater 2001). topographic complexity. Pereira (2000) prepared a general Twelve-mile Reef is a submerged checklist of reef-associated fishes for sandstone coral reef roughly twelve Mozambique, Benayahu & Schleyer nautical miles (18 km) north of (1996) and Schleyer & Celliers (2005) Bazaruto’s northern point in the compiled coral inventories for the open sea. It is open to fishing and Bazaruto reefs, Motta et al., (2002) and SCUBA diving but is relatively Rodrigues et al., (2000) quantified fish inaccessible for artisanal fisherman communities at two sites off Bazaruto using traditional pirogues and dhows. and van der Elst & Afonso (2008) However, recreational fishers and compiled a fish inventory based on SCUBA divers, using large boats with work undertaken at Bazaruto in the outboard engines, are able to access late 1980s. However, there appears this reef. Relative to the other reefs to have been no further ichthyofaunal sampled in this study, Twelve-mile research in the Bazaruto Archipelago. Reef lies furthest from the waters This paper, therefore, presents a recent enclosed between the islands and the and more comprehensive inventory mainland. The reef was sampled at of the fish community as a precursor depths between 15-20 m. to further ecological studies on the Tubarão, Garoupa and Kingfish Bazaruto reefs and highlights the need Reefs are sedimented rocky patch for long-term monitoring. reefs which are open to fishing and

Volume 9 Final 20th Oct 2010.indd 121 10/27/2010 10:02:22 AM 118 J. Q. MAGGS ET AL.

Fig. 1. Map of the Bazaruto Archipelago, Mozambique. Study sites are indicated by (♦).

Volume 9 Final 20th Oct 2010.indd 122 10/27/2010 10:02:23 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 119

SCUBA diving. Garoupa lies nine km Data Collection north of Bazaruto and was sampled at Two fish surveys were conducted, one 16-20 m. Tubarão lies 19 km north- in February 2007 and one in November east of Inhassoro and was sampled 2007. Surveys were undertaken in the at 13-18 m. Kingfish Reef is 13.5 km late morning on a low to outgoing east of Inhassoro and was sampled at spring tide. Fish communities 6-11 m. High turbidity is common on were sampled on the reefs using an these reefs. underwater visual census technique Lighthouse Reef is a fringing coral adapted from Samoilys (1997) in reef located on the north-eastern tip of which three divers recorded the Bazaruto. It is a no-take zone and is presence of fish species on slates. closed to fishing (including artisanal A combination of diving methods fishing) but open to SCUBA diving was used with SCUBA being used and snorkelling. Only the inner lagoon for deeper locations and snorkelling was sampled and the depth ranged for shallow inner lagoons. Divers between 1-3 m. conducted a 45 minute timed swim Two-mile Reef is a barrier coral following a random path. Although reef which lies four kilometres the underwater visual census method out to sea between Bazaruto and is known to underestimate small and Benguera Islands. This reef is also a cryptic species, it was employed no-take zone being closed to fishing because it provides a means of (including artisanal fishing). Although sampling the fish community with little fishing is prohibited, Two-mile Reef disturbance (Fowler, 1987; Harmelin- is subjected to diver pressure and Vivien et al., 1985). Identification of anchor damage from visiting small species was confirmed after sampling craft, recreational dive operators and using appropriate reference books tourists. All reef habitats at Two-mile (King, 1996; King & Fraser 2001; Reef were sampled at depths between Lieske & Myers 1999; Smith & 1-18 m. Heemstra 1986). Amphitheatre and Camel’s Hump are submerged rocky massifs located Trophic Categorisation two kilometres seaward of Cabo de Fish species were assigned to one São Sebastião. They are open to fishing of ten trophic categories based on and SCUBA diving but turbidity classifications by Harmelin-Vivien is high on these reefs. Sampling at (1979); Hiatt & Strasberg (1960); Camel’s Hump ranged between 13- Hobson (1974) and Myers (1999) 16 m in depth and at Amphitheatre, as cited by Chabanet & Durville between 14-19 m.

Volume 9 Final 20th Oct 2010.indd 123 10/27/2010 10:02:23 AM 120 J. Q. MAGGS ET AL.

(2005); Durville et al., (2003) and Table 2. Summary of the number of Gillibrand et al., (2007). These studies species and families recorded on the used eight categories: herbivores, Bazaruto reefs. Shaded reefs are no-take omnivores, browsers of sessile zones where fishing (including artisanal invertebrates, diurnal carnivores, fishing) is prohibited, but SCUBA diving and snorkelling are allowed. Reefs are nocturnal carnivores, piscivores, ordered according to increasing latitude. diurnal planktivores and nocturnal planktivores. In the present study, some Reef No. of No. of Species Families of these categories were consolidated, Twelve-mile Reef 59 21 viz. general carnivores and general Tubarão 44 20 planktivores, as the diel preference of Garoupa 101 31 some species was unknown (Froese & Pauly 2009; Heemstra & Heemstra Kingfish 84 31 2004; King, 1996; King & Fraser 2001 Lighthouse Reef 103 25 and Smith & Heemstra 1986). Two-mile Reef 197 43 Camel’s Hump 45 21 RESULTS Amphitheatre 50 22

Species Richness with 101 species. Two-mile Reef also had the highest number of fish families A total of 249 species belonging to (43) but Lighthouse Reef, despite 50 families were recorded (Table having the second highest number of 1), of which six were cartilaginous species, had relatively few families fishes in four families and the (25). Conversely, Garoupa had a remaining 243 species were bony fish relatively high number of fish families in 46 families. The top five families (31). All the other reefs had relatively according to species count were the few fish families and species; only Labridae (37 species), Acanthuridae Kingfish was better represented by 84 (22 species), Chaetodontidae (22 species in 31 families. species), Pomacentridae (17 species) and Serranidae (13 species). These Trophic Structure five families contributed 45% to the When all the carnivorous categories species diversity (Table 2). Overall, were grouped (i.e. all groups except 19 families were represented by only herbivores and omnivores), they one species. constituted 76% of the species Fish families and species were not composition (Fig. 2). Herbivores, evenly distributed among all the reefs (mostly acanthurids) and omnivores (Table 2). The top three reefs according (mostly pomacentrids) each accounted to species richness were Two-mile for 12% of the species composition. Reef with 197 species, Lighthouse The largest group, diurnal carnivores Reef with 103 species, and Garoupa (27%), was dominated by labrids

Volume 9 Final 20th Oct 2010.indd 124 10/27/2010 10:02:23 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 121

Nocturnal General planktivore General carnivore planktivore 2% 6% 0% Herbivore Diurnal 12% Omnivore planktivore 12% 8%

Piscivore 5%

Nocturnal carnivore Browser of 17% sessile invertebrates 11%

Diurnal carnivore 27% Figure 2. Overall trophic structure of Bazaruto reef fish communities.

and the nocturnal carnivores (17%) DISCUSSION were dominated by larger lutjanids, lethrinids and serranids. Chaetodons Species Richness accounted for the majority of browsers This study yielded 249 fish species in of sessile invertebrates. None of the 50 families, a lower tally than other other categories was dominated by studies in the region. Durville et al., any specific family. The piscivores, (2003) recorded 332 fish species in 57 contributing 5% to the species families at the Glorieuses Islands, while composition, comprised mostly larger Chabanet & Durville (2005) listed predators such as Carcharhinus 299 species in 55 families for Juan De amblyrhynchos, Aprion virescens and Nova. Further south, Gillibrand et al., Scomberomorus commerson. (2007) counted 334 species of fish in

Volume 9 Final 20th Oct 2010.indd 125 10/27/2010 10:02:23 AM 122 J. Q. MAGGS ET AL.

Table 1. Species list of the Bazaruto Archipelago on a per reef basis (depth 1-20 m). Presence is indicated by (●). H, herbivores; O, omnivores; BSI, browsers of sessile invertebrates; DC, Diurnal carnivores; NC, Nocturnal carnivores; PI, Piscivores; DP, Diurnal planktivores; NP, Nocturnal planktivores; C, General carnivores; PL, General planktivores. Shaded reefs are no-take zones where fishing (including artisanal fishing) is prohibited, but SCUBA diving and snorkelling are allowed.

Trophic FAMILY species Category Twelve-mile Twelve-mile Reef Tubarão Garoupa Kingfish Lighthouse Reef Two-mile Reef Camel’s Hump Amphitheatre Acanthuridae Acanthurus dussumieri Valenciennnes, 1835 H ● ● ● ● ● ● Acanthurus leucocheilus Herre, 1927 H ● Acanthurus leucosternon Bennet, 1833 H ● ● ● ● ● ● ● Acanthurus lineatus (Linnaeus, 1758) H ● ● Acanthurus mata Russel in Cuvier, 1829 DP ● ● Acanthurus nigrofuscus (Forsskål, 1775) H ● ● ● ● ● ● ● Acanthurus tennenti Günther, 1861 H ● ● ● ● ● ● Acanthurus thompsoni Fowler, 1923 H ● ● ● Acanthurus triostegus triostegus (Linnaeus, 1758) H ● ● Ctenochaetus binotatus Randall, 1955 H ● ● Ctenochaetus strigosus (Bennet, 1828) H ● ● ● Naso annulatus (Quoy & Gaimard, 1825) H ● ● Naso brachycentron H ● (Valenciennes in Cuvier and Valenciennes, 1835) Naso brevirostris (Cuvier, 1829) H ● ● ● ● Naso hexacanthus (Bleeker, 1855) H ● ● ● Naso lituratus (Forster in Bloch & Schneider, 1801) H ● ● ● ● ● ● Naso unicornis (Forsskål, 1775) H ● ● ● Naso vlamingii (Valenciennnes, 1835) DP ● Paracanthurus hepatus (Linnaeus, 1766) DP ● ● ● Zebrasoma gemmatum (Valenciennnes, 1835) H ● ● Zebrasoma scopas (Cuvier, 1829) H ● ● ● ● Zebrasoma desjardinii (Bennet, 1836) H ● ●

APOGONIDAE Apogon aureus (Lacepède, 1802) DC ●

AULOSTOMIDAE Aulostomus chinensis (Linnaeus, 1766) PI ● ●

BALISTIDAE Balistapus undulatus (Mungo Park, 1797) DC ● ● ● ● ● Balistoides conspicillum (Bloch & Schneider, 1801) DC ● ● ● ● Balistoides viridescens (Bloch & Schneider, 1801) DC ● ● Odonus niger (Rüppel, 1836) DC ● ● ● ● ● ● Pseudobalistes fuscus (Bloch & Schneider, 1801) C ● Rhinecanthus rectangulus (Bloch & Schneider, 1801) O ● ● Sufflamen chrysopterus (Bloch & Schneider, 1801) DC ● ● ● Sufflamen fraenatum (Latreille, 1804) DC ● ● ● ● ● ●

BLENNIIDAE Ecsenius midas Stark, 1969 H ● ● Plagiotremus rhinorhynchos (Bleeker, 1852) NP ● ● Plagiotremus tapeinosoma (Bleeker, 1857) O ● ●

Volume 9 Final 20th Oct 2010.indd 126 10/27/2010 10:02:23 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 123

Trophic

FAMILY species Category Twelve-mile Reef Tubarão Garoupa Kingfish Lighthouse Reef Two-mile Reef Camel’s Hump Amphitheatre

CAESIONIDAE Caesio caerulaurea (Lacepède, 1801) DP ● ● Caesio lunaris Cuvier, 1830 DP ● ● Caesio sp. Lacepède, 1801 DP ● Caesio xanthonota Bleeker, 1853 DP ● ● ● ● ● Pterocaesio sp. Bleeker, 1876 DP ● Pterocaesio tile (Cuvier, 1830) DP ● ●

CARANGIDAE Carangoides fulvoguttatus (Forsskål, 1775) DC ● ● ● ● Caranx ignobilis (Forsskål, 1775) DC ● Caranx melampygus Cuvier & Valenciennes, 1833 DC ● ● ● ● Caranx papuensis Alleyne & MacLeay, 1877 C ● Elagatis bipinnulata (Quoy & Gaimard, 1825) DC ● ● Gnathanodon speciosus (Forsskål, 1775) DC ● Scomberoides lysan (Forsskål, 1775) PI ●

CARCHARHINIDAE Carcharhinus amblyrhynchos (Bleeker, 1856) PI ● ● Triaenodon obesus (Rüppel, 1837) DC ●

CHAETODONTIDAE Chaetodon auriga Forsskål, 1775 BSI ● ● ● ● ● ● ● ● Chaetodon blackburnii Desjardins, 1836 BSI ● Chaetodon dolosus Ahl, 1923 O ● ● Chaetodon falcula Bloch, 1793 BSI ● Chaetodon guttatissimus Bennet, 1832 BSI ● ● ● ● ● ● ● ● Chaetodon interruptus Ahl, 1923 BSI ● ● ● ● Chaetodon kleinii Bloch, 1790 BSI ● ● ● ● ● ● ● ● Chaetodon lineolatus BSI ● ● (Quoy & Gaimard, 1831 in Cuvier & Valenciennes) Chaetodon lunula (Lacepède, 1802) BSI ● ● ● ● ● ● ● Chaetodon madagaskariensis Ahl, 1923 BSI ● ● ● ● ● ● ● Chaetodon melannotus (Bloch & Schneider, 1801) BSI ● ● ● Chaetodon meyeri (Bloch & Schneider, 1801) BSI ● ● ● ● Chaetodon trifascialis Quoy & Gaimard, 1825 BSI ● ● Chaetodon trifasciatus (Mungo Park, 1797) BSI ● ● ● ● ● ● Chaetodon vagabundus Linnaeus, 1758 BSI ● ● ● ● ● ● ● Chaetodon xanthocephalus Bennet, 1832 BSI ● ● ● Chaetodon zanzibarensis BSI ● Playfair, in Playfair & Günther, 1867 Forcipiger flavissimus Jordan & McGregor, 1898 BSI ● ● ● ● Hemitaurichthys zoster (Bennet, 1831) DP ● ● Heniochus acuminatus (Linnaeus, 1758) BSI ● ● ● ● ● ● ● ● Heniochus diphreutes Jordan, 1903 PL ● Heniochus monoceros BSI ● ● Cuvier in Cuvier and Valenciennes, 1831

CIRRHITIDAE Cirrhitichthys oxycephalus (Bleeker, 1855) DC ● ● ● ● Paracirrhites arcatus DC ● ● ● ● Cuvier in Cuvier and Valenciennes, 1829 Paracirrhites forsteri (Bloch & Schneider, 1801) DC ● ● ● ●

Volume 9 Final 20th Oct 2010.indd 127 10/27/2010 10:02:23 AM 124 J. Q. MAGGS ET AL.

Trophic

FAMILY species Category Twelve-mile Reef Tubarão Garoupa Kingfish Lighthouse Reef Two-mile Reef Camel’s Hump Amphitheatre

DASYATIDAE Himantura gerrardi (Gray, 1851) C ● Taeniura lymma (Forsskål, 1775) NC ● DIODONTIDAE Diodon liturosus Shaw, 1804 NC ●

ECHENEIDAE Echeneis naucrates Linnaeus, 1758 NC ● ● ● EPHIPPIDAE Platax orbicularis (Forsskål, 1775) O ● ● ● ● Platax teira (Forsskål, 1775) O ● Tripterodon orbis Playfair, 1867 C ●

FISTULARIIDAE Fistularia commersonii Rüppel, 1838 DC ● ● ●

GOBIIDAE Valenciennea strigata (Broussonet, 1782) DC ● ● ●

HAEMULIDAE Plectorhinchus chubbi (Thunberg, 1792) DC ● ● ● Plectorhinchus plagiodesmus Fowler, 1935 NC ● ● Plectorhinchus flavomaculatus (Cuvier, 1830) NC ● ● ● ● ● ● ● Plectorhinchus gaterinus (Forsskål, 1775) NC ● ● ● ● Plectorhinchus gibbosus (Lacepède, 1802) C ● Plectorhinchus playfairi (Pellegrin, 1914) DC ● ● ● ● ● ● ● Plectorhinchus schotaf (Forsskål, 1775) C ●

HEMIRAMPHIDAE Hyporhamphus affinis (Günther, 1866) O ● ●

HOLOCENTRIDAE Myripristis botche Cuvier, 1829 NC ● Myripristis murdjan Forsskål, 1775 NP ● ● ● Neoniphon argenteus (Valenciennes, 1831) C ● Neoniphon sammara Forsskål, 1775 NC ● ● Sargocentron caudimaculatum Rüppel, 1838 NC ● ● Sargocentron diadema Lacepède, 1802 NC ● ● ● Sargocentron spiniferum Forsskål, 1775 NC ●

KYPHOSIDAE Kyphosus cinerascens Forsskål, 1775 H ● Kyphosus sp. Lacepède, 1801 H ● Kyphosus vaigiensis (Quoy & Gaimard, 1825) O ● ●

LABRIDAE Anampses caeruleopunctatus Rüppel, 1829 DC ● ● ● ● Anampses lineatus Randall, 1972 DC ● Anampses meleagrides Valenciennes, 1840 DC ● ● ● Anampses twistii Bleeker, 1856 DC ● ● Bodianus anthioides Bennet, 1832 DC ● Bodianus axillaris Bennet, 1832 DC ● ● ● Bodianus bilunulatus (Lacepède, 1801) DC ● ●

Volume 9 Final 20th Oct 2010.indd 128 10/27/2010 10:02:23 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 125

Trophic

FAMILY species Category Twelve-mile Reef Tubarão Garoupa Kingfish Lighthouse Reef Two-mile Reef Camel’s Hump Amphitheatre

Bodianus diana Lacepède, 1801 DC ● ● ● ● ● Cheilinus fasciatus (Bloch, 1791) DC ● ● Cheilinus trilobatus Lacepède, 1801 DC ● ● Cheilinus undulatus Rüppel, 1835 DC ● ● Cheilio inermis Forsskål, 1775 DC ● Cirrhilabrus exquisitus Smith, 1957 DC ● ● Coris aygula Lacepède, 1801 DC ● ● Coris caudimacula Quoy & Gaimard, 1834 DC ● Coris cuvieri (Bennett, 1831) DC ● ● ● Coris frerei (Bennett, 1830) DC ● ● ● ● Gomphosus caeruleus Lacepède, 1801 DC ● ● ● ● Halichoeres cosmetus Randall, & Smith, 1982 DC ● ● Halichoeres hortulanus Lacepède, 1801 DC ● ● ● ● Halichoeres iridis Randall, & Smith, 1982 DC ● ● ● Halichoeres scapularis Bennet, 1832 DC ● Hemigymnus fasciatus Bloch, 1792 DC ● ● ● Hologymnosus annulatus Lacepède, 1801 DC ● ● Hologymnosus doliatus Lacepède, 1801 DC ● Labroides bicolor Fowler & Bean, 1928 DC ● ● ● Labroides dimidiatus Valenciennes, 1839 DC ● ● ● ● ● ● ● Macropharyngodon bipartitus Smith 1957 DC ● ● ● Macropharyngodon cyanoguttatus Randall, 1978 DC ● Novaculichthys taeniourus (Lacepède, 1801) DC ● ● Pseudocheilinus hexataenia (Bleeker, 1857) DC ● ● Pseudodax moluccanus (Valenciennes, 1840) O ● ● ● Stethojulis interrupta (Bleeker, 1851) DC ● Thalassoma amblycephalum Bleeker, 1856 DC ● ● ● ● ● Thalassoma hardwicke Bennet, 1830 DC ● Thalassoma hebraicum Lacepède, 1801 DC ● ● ● ● ● ● Thalassoma lunare Linnaeus, 1758 DC ● ● ● ● ● ●

LETHRINIDAE Gnathodentex aureolineatus (Lacepède, 1802) NC ● ● ● Lethrinus crocineus Smith, 1959 NC ● Lethrinus harak (Forsskål, 1775) NC ● ● Lethrinus nebulosos (Forsskål, 1775) NC ● ● Lethrinus rubrioperculatus Sato, 1978 NC ● Lethrinus mahsena (Forsskål, 1775) NC ● Monotaxis grandoculis (Forsskål, 1775) NC ● ● ●

LUTJANIDAE Aphareus furca (Lacepède, 1801) PI ● Aprion virescens Valenciennes, 1830 PI ● ● ● ● ● ● Lutjanus argentimaculatus (Forsskål, 1775) NC ● Lutjanus bohar (Forsskål, 1775) NC ● ● ● ● ● Lutjanus fulviflamma (Forsskål, 1775) NC ● ● ● ● Lutjanus gibbus (Forsskål, 1775) NC ● ● ● ● Lutjanus kasmira (Forsskål, 1775) NC ● ● ● ● ● ● ● Lutjanus lutjanus Bloch, 1790 NC ● ● ● Lutjanus notatus (Cuvier, 1828) NC ● ● ● Lutjanus rivulatus NC ● (Cuvier in Cuvier and Valenciennes, 1828)

Volume 9 Final 20th Oct 2010.indd 129 10/27/2010 10:02:24 AM 126 J. Q. MAGGS ET AL.

Trophic

FAMILY species Category Twelve-mile Reef Tubarão Garoupa Kingfish Lighthouse Reef Two-mile Reef Camel’s Hump Amphitheatre Lutjanus sebae (Cuvier, 1816) NC ● Macolor niger (Forsskål, 1775) NC ● ●

MALACANTHIDAE Malacanthus brevirostris Guichenot, 1848 DC ● ●

MICRODESMIDAE Ptereleotris evides (Jordan & Hubbs, 1925) DP ● Ptereleotris heteroptera (Bleeker, 1855) DP ●

MOBULIDAE Manta birostris (Walbaum, 1792) DP ●

MONACANTHIDAE Amanses scopas (Cuvier, 1829) BSI ● Cantherhines pardalis (Rüppel, 1837) BSI ● Pervagor janthinosoma (Bleeker, 1854) BSI ●

MULLIDAE Mulloidichthys flavolineatus(Lacepède, 1801) NC ● ● ● Mulloidichthys vanicolensis NC ● ● ● Valenciennes in Cuvier and Valenciennes, 1831) Parupeneus barberinus (Lacepède, 1801) DC ● Parupeneus bifasciatus (Lacepède, 1801) C ● ● Parupeneus cyclostomus (Lacepède, 1801) PI ● ● Parupeneus indicus (Shaw, 1803) DC ● Parupeneus macronema (Lacepède, 1801) DC ● ● ● ●

MURAENIDAE Gymnothorax breedeni McCosker and Randall, 1977 C ● Gymnothorax favagineus Bloch & Schneider, 1801 NC ● ● ● Gymnothorax meleagris (Shaw, 1795) DC ● ● MYLIOBATIDAE Aetobatus narinari (Euphrasen, 1790) DC ●

NEMIPTERIDAE Scolopsis ghanam (Forsskål, 1775) DC ● ● ● ● ● ● Scolopsis vosmeri (Bloch, 1792) DC ● ●

OSTRACIIDAE Ostracion cubicus (Linnaeus, 1758) BSI ● ● ● Ostracion meleagris Shaw, 1796 BSI ● ●

PEMPHERIDAE Parapricanthus ransonneti Steindachner, 1870 NP ● Pempheris adusta Bleeker, 1877 NP ● ●

PINGUIPEDIDAE Parapercis hexophtalma (Cuvier, 1829) DC ● ● ● PLATYCEPHALIDAE Papilloculiceps longiceps (Cuvier, 1829) DC ● ● ●

POMACANTHIDAE Apolemichthys trimaculatus (Lacepède, 1831) O ● ● ● ● ●

Volume 9 Final 20th Oct 2010.indd 130 10/27/2010 10:02:24 AM Volume 9Final20th Oct2010.indd 131 Scomberomorus commerson Pomacentrus caeruleus(Quoy&Gaimard,1825) (Quoy&Gaimard,1825) Plectroglyphidodon lacrymatus Fowler&Ball,1924 Plectroglyphidodon johnstonianus Plectroglyphidodon dickii Dascyllus trimaculatus Dascyllus carneusFischer, 1885 Chrysiptera unimaculata Chromis weberi(Fowler&Bean,1928) Chromis opercularis (Günther, 1867) Chromis dimidiata(Klunzinger, 1871) Amphiprion allardi Klausewitz,1970 Amphiprion akallopisos Abudefduf vaigiensis Abudefduf sparoides (Quoy&Gaimard,1825) Abudefduf sordidus (Forsskål,1775) Abudefduf notatus Abudefduf natalensis Priacanthus hamrur(Forsskål,1775) Euthynnus affinis SCOMBRIDAE SCIAENIDAE SCARIDAE PSEUDOCHROMIDAE PRIACANTHIDAE POMACENTRIDAE (CuvierinCuvier& Valenciennes, 1831) Pomacanthus semicirculatus (Gilchristand Thompson, 1908) Pomacanthus rhomboides Pomacanthus imperator(Bloch,1787) Pomacanthus chrysurus(Cuvier, 1831) Centropyge multispinis Centropyge bispinosus(Günther, 1860) Centropyge acanthops(Norman,1922) FAMILY species REEFS BAZARUTO SELECTED ON COMMUNITIES FISH OF ASSESSMENT VISUAL RAPID Scarus tricolorBleeker, 1847 Scarus sordidus (Forsskål,1775) (Valenciennes inCuvierand Valenciennes, 1840) Scarus scaber Scarus rubroviolaceus Scarus ghobbanForsskål,1775 Scarus frenatus (Lacepède,1802) Pseudochromis dutoiti Umbrina robinsoni Gilchristand Thompson, 1908

(Cantor, 1849) Day, 1870

Hensley&Randall,1983 (Quoy &Gaimard,1825) Bleeker, 1847 Smith, 1955 (Rüppel,1829) (Playfair, 1867)

Bleeker, 1853 (Cuvier, 1830)

(Liènard,1839) (Lacepède,1800)

Category Trophic PI C O O O O DP O O DP DP DP O O O O O O O NC

BSI C BSI O O O O H H H H H H DC C

● ● ● ● ● ● ● ● ● ● Twelve-mile Reef ● ● ● ● ● ● ● ● Tubarão ● ● ● ● ● ● ● ● ● ● Garoupa ● ● ● ● ● ● ● ● ● ● ● Kingfish ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Lighthouse Reef ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Two-mile Reef ● ● ● ● Camel’s

10/27/2010 10:02:24 AM Hump 127 ● ● ● ● ● ● ● Amphitheatre Volume 9Final20th Oct2010.indd 132 Synodus dermatogenysFowler, 1912 Acanthopagrus bifasciatus SPARIDAE SIGANIDAE Cephalopholis miniata Cephalopholis argus Bloch&Schneider, 1801 Aethaloperca rogaa (Forsskål,1775) SERRANIDAE Scorpaenopsis venosa(Cuvier, 1829) Pterois miles(Bennet,1825) SCORPAENIDAE ZANCLIDAE TETRAODONTIDAE SYNODONTIDAE SPHYRAENIDAE Variola louti Pseudanthias squamipinnisPeters,1855 Plectropomus punctatusQuoy&Gaimard,1824 Nemanthias carberryiSmith,1954 Epinephelus tukula Epinephelus malabaricus Epinephelus macrospilos (Bleeker, 1855) Epinephelus lanceolatus Epinephelus flavocaeruleus Epinephelus fasciatus FAMILY species Zanclus canescens(Linnaeus,1758) Sphyraena jello Sphyraena barracuda(Walbaum, 1792) Sphyraena putnamaeJordanandSeale,1905 Arothron stellatus Arothron nigropunctatus Bloch&Schneider1801 Arothron hispidus(Linnaeus,1758) Canthigaster valentini Canthigaster solandri(Richardson,1845) Canthigaster smithae (Valenciennes inCuvier and Valenciennes, 1835) Siganus sutor Siganus luridus(Rüppell,1829) 128 (Forsskål,1775) Cuvierinand Valenciennes, 1829

(Bloch&Schneider1801) Morgans, 1959 Allen andRandall,1977 (Forsskål,1775) (Bleeker, 1853)

(Forsskål,1775)

(Bloch,1790) (Bloch&Schneider, 1801) (Forsskål,1775)

(Lacepède,1801) J. Q.MAGGS Category Trophic

PI DC PI NC C PI PI DP PI DP NC NC C NC PI NC NC BSI DC NC NC NC NC O O O NC H H

ET AL. ● ● ● ● ● ● Twelve-mile Reef ● ● ● ● ● Tubarão ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Garoupa ● ● ● ● ● ● ● ● Kingfish ● ● ● ● ● Lighthouse Reef ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Two-mile Reef ● ● ● ● ● ● ● Camel’s Hump ● ● ● ● ● ● Amphitheatre 10/27/2010 10:02:24 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 129

58 families at Andavadoaka (south- not comparable to our study, which west Madagascar) and van der Elst & used only UVC. Chater (2001), working at Bassas da The five reef types sampled in India, recorded 305 species. It is not this study comprised a submerged certain why Bazaruto’s reefs have a sandstone reef, sedimented rocky lower species richness and, since the patch reefs, a fringing coral reef, a other studies in the region are not barrier coral reef and two submerged directly comparable, it is difficult to rocky massifs. In terms of reef damage, place Bazaruto in a spatial or temporal diver and anchor damage were evident context. in the coral-covered inner lagoon of Glorieuses Islands, Juan de Nova Two-mile Reef where large areas of and Bassas da India are isolated Acropora were dead. Corallivorous coral atolls with no permanent crown-of-thorns (Acanthaster planci) human habitation, and consequently starfish were also observed on Two- experience low to negligible fishing mile Reef. These have been persistent, pressure, which has been reported to being first recorded at Bazaruto in 1994 reduce species richness (McClanahan, (Schleyer, 1998), providing further 1994; Wantiez et al., 1997). Their ecological pressure. Nevertheless, isolation from human disturbance and Two-mile Reef had the highest species consequent lack of fishing would make richness in our study. In other studies, a them suitable candidates for control reduction in hard coral cover resulting studies but their reefs are different from mechanical damage has been from those at Bazaruto. Andavadoaka linked to recreational SCUBA diving is more directly comparable with (Hawkins et al., 1999). This reduces Bazaruto, being exposed to fishing reef complexity which correlates pressure from a nearby fishing with species richness (Bell & Galzin village and located on the mainland 1984; Gratwicke & Speight 2005), of Madagascar. Andavadoaka is but without long-term monitoring, it also further south and therefore at is uncertain whether such an effect is a comparable latitude, negating the taking place on Two-mile Reef. latitudinal effect on biodiversity. Garoupa (open to fishing) is a However, the study at Andavadoaka small, flat, sandy ledge with low coral was conducted over one year cover and little physical complexity, compared to nine days at Bazaruto. yet its high fish species richness and A previous fish inventory of Bazaruto abundance was comparable to that of by van der Elst and Afonso (2008), Lighthouse Reef but with six more based on a study done in the late 80s, fish families. The reason for this rich yielded 269 species from 74 families, diversity and abundance is unclear; but their results included fishery- however, it is remarkably similar to dependent data and are consequently Stringer Reef, a small sandy ledge at

Volume 9 Final 20th Oct 2010.indd 133 10/27/2010 10:02:24 AM 130 J. Q. MAGGS ET AL.

Sodwana Bay in South Africa (29° The submerged rocky massifs; 31.784’ S; 32° 40.969’ E), which is Camel’s Hump and Amphitheatre closed to fishing. Both reefs are similar (both open to fishing) have high in terms of coral cover, structural vertical relief but are not structurally simplicity and high fish abundance complex and have minimal coral (pers. obs.). At Garoupa, the cover, probably due to high turbidity piscivorous lutjanid, Aprion virescens, (Rogers, 1990). Their relatively low and the scombrid (mackerel), fish diversity may be explained by Scomberomorus commerson, were these attributes, which may also be prevalent in the mid-water, suggesting the reason why the top five families a predator-dominated environment, (labrids, acanthurids, chaetodons, but further ecological investigation of pomacentrids and serranids) were this reef is warranted. underrepresented. The fish diversity at Twelve-mile Reef (submerged sandstone reef) Trophic Structure was expected to be high, given its Kulbicki (1988) suggested that trophic physical complexity, high coral cover structure is usually constant within a and relative inaccessibility, being region and this has been confirmed furthest offshore. Since the greatest in other studies in the WIO (Table 3). fishing pressure on the Bazaruto reefs Reef disturbances in the form of over- is believed to be caused by artisanal fishing, pollution or coral bleaching fishermen using non-motorised have been reported to cause a reduction dhows, Twelve-mile Reef should in the number of carnivores and an experience relatively little fishing increase in herbivores (Chabanet, pressure because of its remoteness. 2002; Harmelin-Vivien, 1992). While High fish diversity may be associated large carnivores are targeted by with low fishing effort (McClanahan, fishers (Chabanet & Durville 2005), 1994; Wantiez et al., 1997) but did not a reduction in coral cover caused prove the case on Twelve-mile Reef. by pollution (Rogers, 1990) and

Table 3. Comparison of trophic structure (%) in reef fish communities in the Western Indian Ocean (WIO). Location Reference Carnivores Omnivores Herbivores Tuléar Harmelin-Vivien, 1979 74 13.5 12.5 Réunion Chabanet, 1994 51 24 25 Mayotte Chabanet, 2002 69 12.5 18.5 Geyser and Zéléé Chabanet et al., 2002 69 16 15 Glorieuses Durville et al., 2003 73 12 15 Juan de Nova Chabanet & Durville 2005 73 11 16 Andavadoaka Gillibrand et al., 2007 76 11 13 Bazaruto This study 76 12 12

Volume 9 Final 20th Oct 2010.indd 134 10/27/2010 10:02:24 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 131

coral bleaching encourages the rapid to be affected by fishing pressure. growth of filamentous algae, which These smaller carnivores may even provides increased food for herbivores benefit from fishing because of (Chabanet, 2002). reduced predation, giving a false Reefs that are considered healthy impression of reef health. Although usually have carnivore levels of a thorough analysis of the carnivore between 60-80% (Harmelin-Vivien, hierarchy was lacking, the proportion 1979), as found on the Bazaruto and of herbivores on Bazaruto’s reefs other reefs in the WIO (Table 3). was low, consistent with other reef Bazaruto had a high proportion of environments in the region. This carnivores compared to these other indicates a relative measure of reef studies. If one compares the Bazaruto health where pollution and bleaching fish communities with those in are concerned. isolated environments with little or no human interference (Glorieuses CONCLUSION – Durville et al., 2003; Juan de The Bazaruto reefs are exposed to Nova – Chabanet & Durville 2005), fishing pressure, diving and anchor they appear, superficially, to be in a damage, and crown-of-thorns healthy state. However, amalgamation (Acanthaster planci) starfish, yet have of the carnivorous groups in this fish communities rich in diversity study and in many others yields an and a trophic structure similar to that oversimplification of the situation. of other reefs in the WIO which are Andavadoaka also has an abundance considered healthy. They endured of carnivores (Gillibrand et al., 2007) the 1998 bleaching event without similar to the Bazaruto reefs, yet substantial die-off (Schleyer & experiences fishing pressure and the Celliers 2005). However, without a reefs are reported to be in a degraded long-term quantitative monitoring state following broad-scale coral programme, it is difficult to place the bleaching (Gillibrand et al., 2007). It health of the Bazaruto reefs in context. is therefore difficult to link reef health This study, like others in the WIO, to a simple index such as carnivore presents a representative, updated abundance. inventory of the fish communities, Future studies should focus on providing a baseline for more detailed the true trophic hierarchy in the fish studies. However, further studies communities on the Bazaruto reefs, should analyse the trophic hierarchy differentiating between the higher and and include abundance measurements. lower carnivores, rather than between A lack of directly comparable results diel preferences. Small carnivorous became evident during our study, species (e.g., labrids, chaetodons) are highlighting a significant gap in the the most abundant and are unlikely regional understanding of the reef

Volume 9 Final 20th Oct 2010.indd 135 10/27/2010 10:02:24 AM 132 J. Q. MAGGS ET AL.

fish populations. It is suggested that Chabanet, P. (1994) Etude des relations entre les a long-term monitoring programme peuplements benthiques et les peuplements (Chabanet & Durville 2005), ichtyologiques sur le complexe récifal de St-Gilles La Saline à l’île de La Réunion. specifically adapted for Bazaruto’s Thèse Environ. Marin, Univ. Aix-Marseille multiple reef-types, would be suitable III, 235pp. + annexes for conservation planning in the area. Chabanet, P. (2002) Coral Reef Fish Communities of Mayotte (Western Acknowledgements: Sasol provided Indian Ocean) Two Years After the funding for this research. Eduardo Bleaching Event. Mar. Fresh. Wat. Res. 53: 107-113. Videira participated in data collection in the first survey. Bruce Mann is Chabanet, P. & Durville, P. (2005) Reef fish Inventory of Juan de Nova’s Natural thanked for advice in planning the Park (Western Indian Ocean). Western surveys, Fiona Mackay for assistance Indian Ocean Journal of Marine Science with data analysis, Taryn Winson and 4: 145-162. Derick Young for preparation of the Chabanet, P., Tessier, E., Durville, P., map and Pierre Pradervand and Sean Mulochau, T., Rene, F. (2002) Fish Fennessy for their general comments. Communities of Geyser and Zéléé Coral Ben Thompson, Vicky Page, John Banks (Western Indian Ocean). Cybium 26: 11-26. Cranswick and Nicolene Rossouw provided logistical support in the field. Chater, S.A., Beckley, L.E., Garrat, P.A., Ballard, J.A. & van der Elst, R.P. (1993) Anonymous referees are gratefully Fishes from Offshore Reefs in the St acknowledged for valuable input to Lucia and Maputaland marine reserves, the manuscript. South Africa. Lammergeyer 42: 1-17. Chater, S.A., Beckley, L.E., van der Elst, REFERENCES R.P. & Garratt, P.A. (1995) Underwater Visual Census of Fishes in the St Bell, J.D. & Galzin, R. (1984) Influence of Lucia Marine Reserve, South Africa. Live Coral Cover on Coral Reef Fish Lammergeyer 43: 15-23. Communities. Mar. Ecol. Prog. Ser. 15: Durville, P., Chabanet, P. & Quod, J.P. (2003) 265-274. Visual Census of the Reef Fishes in Bellwood, D.R. & Hughes, T.P. (2001) the Natural Reserve of the Glorieuses Regional-scale Assembly Rules and Islands (Western Indian Ocean). Western Biodiversity of Coral Reefs. Science Indian Ocean Journal of Marine Science 292:1532-1534. 2: 95-104. Benayahu, Y. & Schleyer, M.H. (1996) Corals Everett, B.I., van der Elst, R.P. & Schleyer, of the south-west Indian Ocean III. M.H. (eds) (2008) A natural history of Alcyonacea (Octocorallia) of Bazaruto Bazaruto Archipelago, Mozambique. Island, Mozambique, with a Redescription Oceanographic Research Institute, of Cladiella australis (Macfayden, 1963) Durban. Special Publication, 8. 118pp. and a Description of Cladiella kashmani Fowler, J. (1987) The Development of spec. nov. Oceanographic Research Sampling Strategies for Population Institute, Durban. Investigational Report, Studies of Coastal Reef Fishes. A Case 69. 22pp. Study. Coral Reefs 6: 49-58.

Volume 9 Final 20th Oct 2010.indd 136 10/27/2010 10:02:24 AM RAPID VISUAL ASSESSMENT OF FISH COMMUNITIES ON SELECTED BAZARUTO REEFS 133

Froese, R. & Pauly, D. (eds) (2009) FishBase. Hiatt, W.R. & Strasberg, D.W. (1960) World Wide Web Electronic Publication. Ecological Relationship of the Fish www.fishbase.org, Version (09/2009), Fauna on Coral Reefs of the Marshall November 2009. Islands. Ecological Monograph 30: 65- Gillibrand, C.J., Harris, A.R., Mara, E. (2007) 127. Inventory and Spatial Assemblage Hobson, E.S. (1974) Feeding Relationships of Study of Reef Fish in the Area of Teleostean Fish on Coral Reefs in Kona, Andavadoaka, south-west Madagascar Hawaii. Fish Bulletin 72: 915-1031. (Western Indian Ocean). Western Indian King, D. (1996) Reef Fishes and Corals: Ocean Journal of Marine Science 6: East Coast of southern Africa. Struik 183-197. Publishers, Cape Town. 128pp. Gratwicke, B. & Speight, M.R. (2005) The Relationship Between Fish Species King, D. & Fraser, V. (2001) More Reef Fishes Richness, Abundance and Habitat and Nudibranchs: East and South Coast Complexity in a Range of Shallow of southern Africa. Struik Publishers, Tropical Marine Habitats. Journal of Cape Town. 136pp. Fish Biology 66: 650-667. Kulbicki, M. (1988) Main Variation of the Harmelin-Vivien, M.L. (1979) Ichtyofaune des Trophic Structure of Fish Populations récifs coralliens en France Outre-Mer. in the SW Lagoon of New Caledonia. ICRI. Doc. Secrétariat d’Etat à l’Outre- Proc. 6th Coral Reef Symp. Townsville, Mer et Ministère de l’Aménagement du Australia (August 8-12). 2: 305-312. Territoire et de l’Environment. 136pp. Lieske, E. & Myers, R. (1999) Coral Reef Harmelin-Vivien, M.L. (1992) Impact des Fishes: Caribbean, Indian Ocean, and activités humaines sur les peuplements Pacific Ocean including the Red Sea. ichtyologiques des récifs coralliens de Princeton University Press, Princeton. Polynésie français. Cybium 16: 279-289. 400pp. Harmelin-Vivien, M.L., Harmelin, J., Lindén, O., Souter, D., Wilhelmsson, D. & Chauvet, C., Duval, C., Galzin, R., Obura, D.O. (eds) (2002). Coral Reef Lejeune, P., Barnabé, G., Blanc, F., Degradation in the Indian Ocean: Chevalier, R., Duclerc, J. & Lasserre, Status Report 2002. CORDIO, Kalmar, G. (1985) Evaluation visuelle des Sweden. 284pp. peuplements et populations de poissons: McClanahan, T. (1994) Kenyan Coral Reef méthodes et problèmes. Revue Ecologie Lagoon Fish: Effects of Fishing, (Terre Vie) 40: 467-539. Substrate Complexity and Sea Urchins. Hawkins, J.P., Roberts, C.M., Van’t Hoff, T., Coral Reefs 13: 231-241. De Meyer, K., Tratalos, J. & Aldam, C. Motta, H., Pereira, M.A.M. & Schleyer, (1999) Effects of Recreational Scuba M.H. (2002) Coral Reef Degradation Diving on Caribbean Coral and Fish in Mozambique, Results of Monitoring Communities. Conservation Biology 1999-2000. In: Lindén, O., Souter, D., 13: 888-897. Wilhelmsson, D. & Obura, D.O. (eds) Heemstra, P.C. & Heemstra, E. (2004) Coastal (2002). Coral Reef Degradation in sea Fishes of southern Africa. National the Indian Ocean. Status Report 2002. Inquiry Services Centre, Grahamstown. CORDIO, Kalmar, Sweden. pp55-60. 512pp.

Volume 9 Final 20th Oct 2010.indd 137 10/27/2010 10:02:24 AM 134 J. Q. MAGGS ET AL.

Myers, R.F. (1999) Micronesian Reef Fishes. Schleyer, M.H. (1998) Observations on the Coral Graphics, Barrigada, Guam, Incidence Of Crown-of-thorns Starfish 298pp. in the Western Indian Ocean. Reef Obura, D.O., Wanyonyi, I.N., Mwaura, Encounter 23: 25-27. J.M. (2002) Participatory Monitoring Schleyer, M.H. & Celliers, L. (2005) The of an Artisanal Fishery in Kenya. In: Coral Reefs of Bazaruto Island, Lindén, O., Souter, D., Wilhelmsson, Mozambique, with recommendations D. & Obura, D.O. (eds), Coral Reef for their management. Western Indian Degradation in the Indian Ocean: Status Ocean Journal of Marine Science 4 (2): Report 2000, CORDIO, Stockholm. pp 227-236. 70-82. Schleyer, M.H. & Maggs, J.Q. (2008) Pereira, M.A.M. (2000) Preliminary Surveys of Reef Benthos Conducted in Checklist of Reef-Associated Fishes of the Bazaruto Archipelago on Behalf of Mozambique. MICOA, Maputo. 21pp. Sasol in 2007. Oceanographic Research Reina, A. (1998) Bazaruto Archipelago: Institute, Durban. Unpublished Report, Protected Area Development 257. 9pp. and Management. Proceedings Smith, M.M. & Heemstra, P.C. (1986) Smith’s of International Tropical Marine Sea Fishes. Struik Publishers, Cape Ecosystems Management Symposium Town: 1047p. (ITMEMS), Townsville, November, Van der Elst, R.P. & Afonso, P.S. (2008) 1998. pp343–353. Fish and fisheries. In: Everett, Rodrigues, M.J., Motta, H., Pereira, B.I., van der Elst, R.P. & Schleyer, M.A.M., Gonçalves, M., Carvalho, M.H. (eds) A Natural History of M. & Schleyer, M.H. (2000) Coral Bazaruto Archipelago, Mozambique. Reef Monitoring in Mozambique: The Oceanographic Research Institute, Monitoring Programme and 1999 Durban. Special Publication, 8. pp 93- Report. MICOA-ORI-IIP, Maputo, 64 109. pp. Van der Elst, R.P. & Chater, S. (2001) The Rogers, C.S. (1990) Responses of Coral Reefs Ichthyofauna of Bassas da India atoll and Reef Organisms to Sedimentation. in the Mozambique Channel. 6th Fish Mar. Ecol. Prog. Ser. 62: 185-202. Indo-Pacific Conf.Durban, South Africa Samoilys, M.A. (1997) Underwater Visual (May 20-25) (abstract). Census Surveys. In: Samoilys MA Wantiez, L., Thollot, P., Kulbicki, M. (1997) (ed) Manual for Assessing Fish Stocks Effects of Marine Reserves on Coral on Pacific coral reefs. Department of Reef Fish Communities from Five Primary Industries, Training Series Islands in New Caledonia. Coral Reefs QE97009, Queensland, pp 16-29. 16: 215-224.

Volume 9 Final 20th Oct 2010.indd 138 10/27/2010 10:02:24 AM Instructions for Authors i

WESTERN INDIAN OCEAN JOURNAL OF MARINE SCIENCES

Instructions for Authors

Editorial Policy The Western Indian Ocean Journal of Marine Sciences is the research publication of the Western Indian Ocean Marine Science Association (WIOMSA). It publishes original research papers or other relevant information in all aspects of marine science and coastal management as articles, reviews, and short communications (notes). While submissions on tropical and subtropical waters of the western Indian Ocean and the Red Sea will be given primary consideration, articles from other regions of direct interest to the western Indian Ocean will also be considered for publication. All manuscripts submitted to the Western Indian Ocean Journal of Marine Sciences are accepted for consideration on the understanding that their content has not been published elsewhere and is not under consideration by any other journal. Manuscripts and all illustrations should be prepared according to the instructions provided below. Submissions will be subject to a pre-review by the Editor-in-Chief or a member of the Editorial Board and those that fall within the remit of the journal, make a substantial contribution to the field of research, and are in the correct style and format will be sent for review. Manuscripts that do not meet these criteria will be rejected. Every manuscript will be reviewed by at least two referees competent in the field of interest. The choice of reviewers is made by the Editor-in-Chief or the Editorial Board.

The Manuscript 1. Contributions must be written in UK English. Authors should submit an electronic version of the manuscript by e-mail to the Editor-in-Chief ([email protected]). If English is not your first language we suggest that the text is edited, before submission, by an English speaker. 2. The manuscript must be typed in Times Roman, font size 12 and double- spacing. The total number of pages, including all illustrations (graphs, tables, photos etc.), should not exceed 20 pages. Short communications must not exceed 8 pages. A separate sheet should be used for each table and figure. 3. Species names must be in italics; the genus is written in full at the first mention in the Abstract, again in the main text and the figure and table legends, and abbreviated thereafter.

Volume 9 Final 20th Oct 2010.indd 139 10/27/2010 10:02:25 AM ii Instructions for Authors

4. Illustrations (figures, tables) should be placed separately at the end ofthe manuscript. Originals of all figures should be in black and white; the lettering should be of a size readable after reduction for the final layout. Figure legends (captions) should be written on a separate page. Table legends must incorporate all the information needed and placed on the same page as the table. Authors are requested to indicate the recommended position of illustrations in the left- hand margin of the text. 5. The international system of units (SI Units) must be used throughout; abbreviations and acronyms should be identified where they first appear; mathematical symbols and formulae should be used only when absolutely necessary and should be clearly defined in the text. 6. A complete manuscript must include the following: title page, abstract, keywords, introduction, materials and methods, results, discussion, acknowledgements, references, tables and illustrations (with figure legends) in that order. a. Title Page: This should contain a concise title and the names of authors followed by affiliations and their complete postal addresses. The contact author and email address must be indicated. b. Abstract: The abstract should not exceed 200 words, and should be on a separate page. It should briefly describe the main points of the manuscript, i.e. the topic, the main findings and the conclusions. c. Keywords - four to six key words are required for indexing purposes. d. Introduction: A brief survey of relevant literature and objectives of the work should be given in this section. Thus, the introduction should largely be limited to the scope, purpose and rationale of the study. e. Materials and Methods: In this section, the methodology used should be clearly explained, including relevant references, such that another person can repeat the procedures. It should provide the framework to gain answers to the questions or problems identified. Sampling methods must be elaborated as well as analytical frameworks and model specifications. f. Results: Make the text as objective and descriptive as possible. Only material pertinent to the subject should be included. Avoid presenting the same information in both graphical and tabular form. g. Discussion: This section could be combined with the above to present “Results and Discussion”. It should interpret the results in view of the problems identified in the introduction, as well as in relation to other published work. The final paragraph of this section could include concluding remarks and recommendations for future work.

Volume 9 Final 20th Oct 2010.indd 140 10/27/2010 10:02:25 AM Instructions for Authors iii

h. Citations: Authors should be cited using their surnames, followed by the year of publication. Two authors should be separated by an ampersand. If there are more than two authors, only the first author, followed by et“ al.”, should be given. This and other Latin or foreign terms should be italicized. i. Acknowledgement/s: This section should be brief. Authors are advised to limit acknowledgements to substantial contributions to the scientific and technical aspects of the paper, financial support or improvements in the quality of the manuscript. j. References: The reference section must contain an alphabetical list of all references mentioned in the text of the manuscript. Limit punctuation and special fonts as indicated and give all journal names in full. Examples for citations from periodicals, books and composite works are given below: • Periodicals. Here the following should be sequentially listed: author’s name/s, initials, year of publication, full title of paper, periodical (in full), volume, first and last page numbers. Example: Richardson K, Beardall J, Raven J (1983) Adaptation of unicellular algae to irradiance: An analysis of strategies. The New Phytologist 93: 157-191 • Books. The following should be listed: author’s or editor’s name, initials, year of publication, full title, publisher, place of publication, total pages. Example: Kirk TJO (1983) Light and photosynthesis in aquatic ecosystems. Cambridge University Press, Cambridge. 401pp • Composite works or serials. The sequence should be as above, but also should include full title of paper followed by In: editor(s) if any, full title of publication, publisher, etc., and the first and last page numbers. Example: Sathyendranath S, Platt T (1993a) Remote sensing of water-column primary production. In: Li WKW, Maestrini SY (eds) Measurement of primary production from the molecular to the global Scale. ICES Marine Science Symposia, Vol. 97, Copenhagen. pp 236-243 • Articles with a Digital Object Identifier (DOI). Example: Gooseff MN, McKnight DM, Lyons HJ, Blum RJ (2002) Weathering reactions and hyporheic exchange controls on stream water chemistry in a glacial meltwater stream in the McMurdo Dry Valleys. Water Resources Bulletin 38 [doi: 10.1029/2001WR000834] k. Tables and illustrations: Each figure/table/photograph should be numbered consecutively, accompanied by a complete caption, and must be cited in the text. Figures should be of high quality to allow reproduction and reduction without loss of information. Photographs should be of excellent quality to avoid loss of contrast during printing.

Volume 9 Final 20th Oct 2010.indd 141 10/27/2010 10:02:25 AM