Wetlands Ecology and Management 12: 277–283, 2004.                             277
   # 2004 Kluwer Academic Publishers. Printed in the Netherlands.




Fine sediment trapping in two mangrove-fringed estuaries exposed to
contrasting land-use intensity, Palau, Micronesia

S. Victor1,*, Y. Golbuu1, E. Wolanski2 and R.H. Richmond3
1
Palau International Coral Reef Center, P.O. Box 7086, Koror, Palau 96940, Micronesia; 2Australian
Institute of Marine Science, PMB No. 3, Townsville MC, Qld. 4810, Australia; 3Kewalo Marine
Laboratory, University of Hawaii, 41 Ghui Street, Honolulu H1 96813, USA; *Author for
correspondence (e-mail: svictor@picrc.org)

Received 22 March 2003; accepted in revised form 15 September 2003



Key words: Land use, Mangroves, Palau, Reefs, Sediment


Abstract

A comparative study was undertaken of the fate of fine sediment in the Ngerikiil and Ngerdorch mangrove-
fringed estuaries in Babeldaob Island, Palau, Micronesia, in 2002. The mangroves comprised 3.8% of each
catchment area, and in both systems, they trapped about 30% of the riverine sediment. Mangroves are
important buffers protecting fringing coral reefs from excessive sedimentation. The sediment yield was
significantly higher in the Ngerikiil River catchment (150 tons kmÀ2 yrÀ1) that has been extensively cleared
and farmed, than in Ngerdorch River catchment (1.9 tons kmÀ2 yrÀ1) that was still relatively pristine during
the study period.


                                  including forest and mangrove clearing, land
Introduction
                                  grading, road construction, farming and housing
Rates of soil erosion are rapidly increasing on          development.
                                   The Ngerdorch River drains a 39 km2 mountai-
many tropical islands as a result of land clearing
and poor land use practices. Increased erosion is         nous catchment located northeast and adjacent to
threatening estuaries and coastal coral reefs           the Ngerikiil River catchment (Figure 1 and
(Dubinsky and Stambler 1996; Meade 1996;              Table 1). These two catchments have similar geol-
Edinger et al. 1998; Wolanski and Spagnol 2000;          ogy, topography and rainfall (USDI 1997), as well
Fortes 2001; McCook et al. 2001; Wolanski et al.          as an identical tidal range within their estuaries.
2003). Golbuu et al. (2003) reported that soil ero-        Both rivers flow into coral-fringed lagoons through
sion is presently affecting the Ngerikiil River, its        mangrove swamps, which comprises 3.8% of each
estuary and its fringing coral reefs in Airai Bay,         catchment area and fringe the whole length of the
Babeldaob Island, Palau, Micronesia (7 220 N,           estuary to the tidal excursion limit.
134 340 E; Figure 1). There are no pre-development          In 2000, the ‘Babeldaob Compact Road’
data on soil erosion rates in the area to help assess       (a 52-mile road around Babeldaob) and the
the changes in sediment yield during the recent          ‘Capital Relocation’ (a new capital building in
peak development period. However, local residents         Melekeok) projects were initiated. A portion of
attribute the recent increased erosion in the catch-        the Compact road runs through the Ngerdorch
ment, the reported rapid siltation of the estuary,         catchment and the new Capital Building is being
and the die-off of coral reefs in Airai Bay to           constructed on one hillside of the catchment.
unplanned development activities in the catchment         These developments will lead to large-scale land
278




Figure 1. (a and b) Map of the Ngerdorch and Ngerikiil River catchments, Babeldaob Island, Palau, Micronesia. (c) Ngerdorch estuary
showing the sampling sites described in the text. Abbreviations are T1, T2 . . . ¼ location of sediment traps and S0, S1. . . ¼ locations of
moored instruments.

clearing and an anticipated increase in soil               farms within the catchment while the Ngerikiil
erosion in the near future. In 2002, the Ngerdorch            catchment has experienced significant housing
River catchment was still relatively pristine,              and farming developments during the last few
being largely forested with only three small               years.
                                                           279

Table 1. Characteristics of the Ngerikiil and Ngerdorch river  in-situ using water samples brought to the laboratory
catchments, Palau.                       and filtered on 0.45 m filters, which were dried at
                                60  C in a drying oven for 24 h and weighed.
                         Ngerikiila
                  Ngerdorch
                                  At station S2, at 6 m depth, the vertical profile at
Catchment size (km2)         39       19     0.5 m intervals of horizontal currents was mea-
Mean estuarine SSC (mg lÀ1)     19      500
                                sured at 5 min intervals using a bottom-mounted
Peak estuarine SSC (mg lÀ1)     160     >1500
                                RDI Workhorse ADCP. In addition, the vertical
Mangrove area/catchment area     3.8%      3.8%
                                profile of salinity, temperature and SSC were mea-
Trapping in mangrove         28–44%     15–30%
Sediment yield (tons kmÀ2 yrÀ1)    1.9     >150     sured at stations S0–S3 with a ship-born YSI CTD
                                profiler-cum nephelometer. These measurements
a
  Source: Golbuu et al. (2003).
                                were carried out daily following a flood event in
                                October 2002 for one week and occasionally for the
  To assess the human impact on soil erosion rates,
                                duration of the study. Salinity is expressed in psu,
we undertook a field study to quantify and com-
                                which for our study sites is practically equivalent
pare the riverine sediment load and role of man-
                                to ppt.
groves in determining the fate of fine sediment in
                                  Single, bottom-mounted sediment traps, with a
the two tropical estuaries. Observations showed
                                diameter of 5.08 cm, were mounted at the edge of
that (1) the Ngerdorch catchment is relatively pris-
                                the mangroves on the river bank between stations
tine with mature vegetation, extensive ground
                                S0 and S3 and in the mangroves along a transect
cover and forested river banks; (2) the Ngerikiil
                                perpendicular to the river bank at 10, 20 and 30 m
catchment has been heavily impacted by poor
                                from station S1. The sediment traps were deployed
land use practices resulting in elevated rates of
                                on 12 October 2002 and recovered 120 days later.
soil erosion; and (3) mangroves comprises 3.8% of
                                  The National Weather Service provided daily
each catchment area. Based on these observations,
                                rainfall data at Koror, located about 12 km south-
we expected the following: (1) the sediment load in
                                west of study sites. The rainfall data were used to
Ngerdorch River will be less than in Ngerikiil River
                                correlate sedimentation rate and SSC values.
and (2) the holding capacity of the mangroves will
be the same for both estuaries.
                                Results
Methods
                                Semi-diurnal, meso-tides prevailed with a pro-
Four oceanographic moorings were deployed at          nounced diurnal inequality, and a strong spring–
stations S1–S3 (Figure 1) from September to           neap tide cycle. The tidal range was about 2 m at
October 2003. These stations formed an along-          spring tide and 1 m at neap tide. A salt wedge
channel transect. Salinity, temperature and           prevailed throughout the field study (Figure 2)
suspended sediment concentration (SSC) were           and the isohalines were nearly horizontal. The
measured at stations S1 (at 1 m above the bottom        brackish water plume lifted off the bottom between
at 4 m depth) and S3 (at 2 m below the surface at 6 m      stations S0 and S1.
depth), using self-logging Analite nephelometers,         The SSC contour lines were not parallel to the
Dataflow salinometers and an YSI self-logging          isohalines; instead they sloped upwards toward the
CTD-cum nephelometer. The instruments were           river mouth. A turbidity maximum zone existed
attached onto 1 m long steel star pickets (rebar)        near the lift-off point (Figure 2). Hence most of
driven into the substratum. The Analite nephelo-        the suspended sediments fall out of suspension in
meters and YSI instrument were equipped with          the estuary before the plume exits the mouth of the
wipers that cleaned the sensor every 30 and 10 min,       estuary (Figure 2b). Some of that sediment may get
respectively. The instruments logged data at 10 min       re-suspended during incoming tides (Figure 2c) and
intervals. The data were sampled at 0.5 s intervals       may be brought back into the inner estuary and the
and averaged over 1 min for all sensors except         mangroves.
the YSI, which logged data continuously without          The salinity in the salt wedge at site S1 fluctuated
averaging. The nephelometers were calibrated          as a result of both the tides and the rainfall (Figure 3).
280




Figure 2. A long-channel distribution of salinity (ppt) and SSC (mg LÀ1) in the Ngerdorch estuary during 17–19 October 2002. Station
locations are shown in Figure 1.

The highest values of salinity were found at high          in the estuary. The salt wedge exists as a result of the
tide, and the lowest values at low tides. The SSC          small tidal currents and the large depth of the estu-
values in the salt wedge at site S1 also fluctuated at        ary. The riverine fine sediment did not follow the
the tidal frequency (Figure 3). They were the high-         freshwater flow. Instead, the suspended sediment
est at high tide and following large rainfall events.        settled out of the brackish, surface plume and was
  The low-frequency currents at site S2 in the estu-        re-entrained towards the head of the estuary by the
ary were predominantly landward in the salt             baroclinic currents in the salt wedge. A turbidity
wedge, and seaward in the brackish water plume            maximum zone prevailed near the plume lift-off
near the surface (Figure 3). The currents also varied        point.
at the tidal frequency, though the tidal currents            Occasional aerial observations (P. Colin, pers.
were usually smaller than the low-frequency             comm.) suggest that the river plume was deflected
currents.                              alongshore, and this may explain why in our obser-
  During the field study, neither the freshwater          vations the river plume generally did not reach
plume nor the riverine fine sediments reached off-          offshore waters (S3).
shore waters (S3) in quantity, except once for about          It was possible to estimate the freshwater flow
20 min following a short river flood (Figure 3). The         rate, Qf, from the classical two-layer estuarine
siltation rates in the mangroves were 65, 12 and           equation (Pritchard and Burt 1951)
9 mg cmÀ2 dayÀ1 at 10, 20 and 30 m inside the                         Qf ¼ Qin
                                                        ,
mangroves from the banks of the estuary. Hence
                                             ½Sout =ðSin À SoutÞ Š
most of the suspended sediments were deposited
within 50 m of the edge of the river, in agreement          where Qf is the freshwater flow rate, Qin is the
with findings in other macro-tidal mangroves             inflow rate in the salt wedge, Sin is the salinity in
(Wolanski et al. 2001).                       the bottom layer and Sout is the salinity in the salt
                                   wedge. Qin was calculated as follows:

Discussion                                          Qin ¼ uA,

The observations suggest a permanent freshwater           where u was the velocity in the salt wedge as mea-
inflow that formed a permanent salt wedge circulation        sured by the ADCP and A was the cross-sectional
                                                                      281




Figure 3. Time-series plot of daily rainfall, salinity and SSC in the salt wedge at site S1 and near the surface at site S3, and velocity (>0 if
seaward; <0 if landward) in the brackish water plume and in the salt wedge at site 2. Station locations are shown in Figure 1.


area of the salt wedge. The brackish water outflow,            the brackish water plume and the salt wedge,
Qout, was calculated as:                          respectively. The latter values were taken from the
                                      daily CTD casts. The net estuarine sediment
           Qout ¼ Qf þ Qin :
                                      export, Qnet, was calculated as:
The net fine sediment fluxes in the river, in the
                                                 Qnet ¼ Qout À Qin :
brackish water plume and in the salt wedge were
then calculated as QfCf, QoutCout and QinCin, where            For the period of September–October 2002, the net
Cf, Cout and Cin were the SSC values in the river,             fine sediment fluxes in the river and out of the
282

estuary were 3.2 ± 1.915 and 2.3 ± 1.044 g sÀ1,     watershed. The implication is that while Ngerdorch
respectively. This suggests that the mangroves      watershed is still relatively pristine, coral reef con-
may trap about 0.9 g sÀ1, i.e., about 28% of the     servation and management effort may not be pos-
riverine fine sediment inflow. However the possible   sible without proper land management in the
error of this estimate is large. The sediment traps in  surrounding catchment.
the mangroves suggest a mean settling rate of        The result of this study may have broad applica-
1.4 g sÀ1, or about 44% of the riverine fine sediment  tions to coastal coral reef ecosystem worldwide.
flux. Interestingly, the mangroves trap a similar    Sedimentation associated with poor land manage-
fraction of the fine sediment in the Ngerikiil estu-   ment has been identified as a dominant problem by
ary, although the riverine fine sediment yield in this  the US Coral Reef Task Force. While physical and
estuary is 10–19 times higher (Table 1). In both the   biological characteristics may vary among coral
Ngerdorch and the Ngerikiil estuaries, the man-     reef sites, the outcome of poor land use will be the
groves comprise about 3.8% of the catchment.       same: accumulation of sediment that will prevent
This suggests that the sediment trapping efficiency   coral larval recruitment and recovery of corals
of mangroves is a function of tidal dynamics in the   (Golbuu et. al. 2003).
mangrove wetlands, and not of riverine suspended
sediment concentration.                 Acknowledgements
  The Ngerdorch estuary is still relatively pristine
compared with Ngerikiil estuary (Table 1). In the    The Palau International Coral Reef Center, the
context of the whole world, the Ngerdorch estuary    University of Guam, the Australian Institute of
is pristine compared with a similar sized river in    Marine Science, the US-EPA STAR program
Southeast Asia and Oceania (Milliman and Meade      (grant R 82-8008) and the NOAA Coastal Oceans
1983) and of a small mountainous catchment with     Program (grant NA16OP2920) supported this
elevation <100 m in Oceania (Milliman and        study. The authors gratefully acknowledge
Syvitski 1992). The sediment yield in the        the assistance and support of Arius Merep,
Ngerdorch River (1.9 tons kmÀ2 yrÀ1) was slightly    Masao Udui and Kenjo Yamashiro. Katharina
smaller than that (2.4 tons kmÀ2 yrÀ1) predicted by   Fabricius, Peter Houk and Fleming U. Sengebau
Milliman and Styvitski (1992).              reviewed and improved an earlier version of this
  Thus, mangroves play an important role in redu-    manuscript.
cing coastal erosion (Mazda et al. 2002) and pro-
tecting fringing coral reefs from sedimentation.
The Ngerdorch and Ngerikiil estuaries mangroves     References
flood semi-diurnally, and they may trap up to 44%
of the riverine fine sediment. As suggested by      Dubinsky Z. and Stambler N. 1996. Marine pollution and coral
Golbuu et al. (2003) for the Ngerikiil River, this     reefs. Global Change Biology 2: 511–526.
                             Edinger E.N., Jompa J., Limmon G.V., Widjatmoko W. and
trapping efficiency, while helpful, is not sufficient
                              Risk M. 1998. Reef degradation and coral biodiversity in
to prevent degradation of coastal coral reefs from
                              Indonesia: effects of land-based pollution, destructive fishing
excessive sedimentation resulting from extensive      practices and changes over time. Marine Pollution Bulletin
land clearing and poor farming practices.         36(8): 617–630.
Siltation of the Ngerdorch estuary and coral reef    Fortes M. 2001. The effects of siltation on tropical coastal
                              ecosystems. In: Wolanski E. (ed.), Oceanographic Processes
degradation in the lagoon waters is likely to occur
                              of Coral Reefs. Physical and Biological Links in the Great
if land clearing and poor farming practices are not
                              Barrier Reef, CRC Press, Boca Raton, FL, pp. 93–111.
regulated as they were the last few years in the     Golbuu Y., Victor S., Wolanski E. and Richmond R.H. 2003.
Ngerikiil catchment.                    Trapping of fine sediment in enclosed bay, Palau, Micronesia.
                              Estuarine, Coastal and Self Science 57: 941–949.
                             Mazda Y., Magi M., Nanao H., Kogo M., Miyagi T.,
                              Kanazawa N. and Kobashi D. 2002. Coastal erosion due to
Conclusion
                              long term human impact on mangrove forests. Wetlands
                              Ecology and Management 10: 1–9.
The sediment yield in Ngerikiil watershed is 10–19    McCook L.J., Wolanski E. and Spagnol S. 2001. Modelling and
times higher than in the less developed Ngerdorch     visualizing interactions between natural disturbances and
                                                               283

 eutrophication as causes of coral reef degradation. In:     USDI (United States Department of the Interior), 1997.
 Wolanski E. (ed.), Oceanographic Processes of Coral Reefs.    Environmental impact statement for construction of the
 Physical and Biological Links in the Great Barrier Reef, CRC   Palau compact road, Babeldaob Island, Republic of Palau.
 Press, Boca Raton, FL, pp. 113–125.               vol. II, Project No. 296-001, Washington D.C.
Meade R.H. 1996. River-sediment inputs to major deltas. In:    Wolanski E. and Spagnol S. 2000. Environmental degradation
 Milliman J.D. and Haq B.U. (eds), Sea Level Rise and       by mud in tropical estuaries. Regional Environmental Change
 Coastal Subsidence, Kluwer, Dordrecht, pp. 63–85.        1(3–4): 152–162.
Milliman J.D. and Meade R.H. 1983. World-wide delivery      Wolanski E., Mazda Y., Furukawa K., Ridd P., Kitheka J.,
 of river sediment to the oceans. Journal of Geology 91(1):    Spagnol S. and Stieglitz T. 2001. Water circulation through
 1–21.                              mangroves and its implications for biodiversity. In: Wolanski E.
Milliman J.D. and Syvitski J.P.M. 1992. Geomorphic/tectonic    (ed.), Oceanographic Processes of Coral Reefs: Physical and
 control of sediment discharge to the ocean: the importance    Biological Links in the Great Barrier Reef, CRC Press, Boca
 of small mountainous rivers. Journal of Geology 100:       Raton, FL, pp. 53–76.
 525–544.                            Wolanski E., Richmond R.H., Davis G. and Bonito V. 2003.
Prichard D.W. and Burt W.V. 1951. An inexpensive and rapid     Water and fine sediment dynamics in transient river plumes in
 technique for obtaining current profiles in estuarine waters.  a small, reef-fringed bay, Guam. Estuarine, Coastal and Shelf
 Journal of Marine Research 10: 180–189.             Science 56: 1029–1040.