Fine-sediment dynamics in the Mekong River estuary, Vietnam

Estuarine, Coastal and Shelf Science (1996) 43, 565–582 Fine-sediment Dynamics in the Mekong River Estuary, Vietnam Eric Wolanskia, Nguyen Ngoc Huanb, Le Trong Daob, Nguyen Huu Nhanc and Nguyen Ngoc Thuyd a Australian Institute of Marine Science, PMB No. 3, Townsville M.C., Queensland 4810, Australia, bHydrometeorological Service, 4 Dang Thai Than Street, Hanoi, Vietnam, cHydrometeorological Service, 19 Nguyen Thi Minh Khai Street, 1st District, Ho Chi Minh City, Vietnam and dVietnam Marine Science & Technology Association, 36 Hoang Dieu Street, Hanoi, Vietnam Received 24 January 1995 and accepted in revised form 11 September 1995 Keywords: Mekong Delta; sediments; salt wedge; turbidity maximum; siltation; salinity intrusion Field studies of fine-sediment transport were carried out in the Mekong River estuary, Vietnam, in the high-flow season of November 1993. In the freshwater region, erosion and deposition of suspended sediment occurred at tidal frequency with a strong down-river transport at a mean velocity of 1 m s "1 and mean suspended-solid concentration of about 250 mg l "1. Also, the suspended sediment was mostly fine silt, and the clay fraction accounted for only 15%. The suspended sediment was transported either as individual particles or agglomerated with organic detritus. At the mouth of the estuary, a salt wedge was present but was flushed out of the estuary at low tide. Flow reversal occurred across the pycnocline. A turbidity maximum zone was present at the toe of the salt wedge at flood tide. Most of the suspended sediment was coagulated with little organic matter. The bulk of the suspended sediment was exported to coastal waters, but some sediment returned to the estuary in the salt wedge. Data for the low-flow season are sparse but suggest that partially well mixed estuarine conditions prevail, with salinity penetrating about 40 km inland, carrying fine-sediment up-river to a turbidity maximum zone. Most of this sediment would have been deposited in shallow coastal waters in the previous high-flow seasons. The construction of large hydro-electric dams further upstream on the Mekong River may exacerbate siltation patterns in the ? 1996 Academic Press Limited estuary. Introduction The Mekong River (Figure 1) is 4200 km long, it drains an area of 0·79#106 km2 with a mean annual water discharge of 470 km3 (Borland, 1973; Milliman & Meade, 1983). The discharge of the lower Mekong River at Phnom Penh, Cambodia, varies seasonally (United Nations, 1957; Gagliano & McIntire, 1968); it is smallest (typically 1700 m3 s "1) in May and largest in October (typically 39 000 m3 s "1). The estuary has a few major channels, shown in Figure 1, and a large number of smaller channels (not shown), and these form a vast delta (Figure 1). The thalweg in the estuary is typically 0272–7714/96/050565+18 $25.00/0 ? 1996 Academic Press Limited 566 E. Wolanski et al. 106°E 106°E 0 Phnom Penh 50 km Can Tho Mekong River Delta 11 Vung Tau South China Sea 9°N 20 ea 10 South China Sea Ca Loc 9°N am tn 6 10 China e Vi 6 10° Thailand 6 10 Can Tho 10 km 20 21 22 23 24 S Sou th China 106°E Figure 1. Location map and sampling sites ( ), Mekong River estuary. 10 m deep in the freshwater region between 30 and 140 km from the mouth; closer to the mouth in the dry season saline-intrusion region, the depth decreases to 5 m typically; even shallower waters are found at the mouth (Figure 2). Coastal waters are shallow, the 20-m depth contour is located 30 km from the coast (Figure 1). Mixed, macro-tides prevail with a strong diurnal inequality (Nguyen Ngoc Thuy, 1979). At the mouth, the mean maximum and average tidal range are, respectively, c. 3·2 m and 2·2 m. The average tide range decreases with distance upstream (Gagliano & McIntire, 1968; Nguyen Ngoc Thuy, 1988a,b). Indeed at Can Tho (123 km), it is only 1·9 m in the low-flow season and 0·7 m in the high-flow season; at Chau Doc (228 km) it is only 0·49 m in the low-flow season and 0·0 m in the high-flow season. The tidal dynamics can be modelled as friction-damped, progressive waves in branched, onedimensional channels, with the tidal amplitude decreasing and the time lag increasing with distance from the river mouth (Nguyen Ngoc Huan, 1987a,b; Tingsanchali & Lien, 1987; Nguyen Ngoc Thuy, 1989). Data on sediment discharge in the lower Mekong River are even more sparse, especially over the last 30 years. The sediment discharge may be about 160#106 tonnes year "1 (Milliman & Meade, 1983; Milliman & Syvitski, 1992). To place the Mekong in perspective with other major rivers in the world, the Mekong River has (see Table 1) a smaller drainage area than the Yangtze (41%), the Amazon (12%), the Mississippi (24%), and the Ganges/Brahmaputra (53%) Rivers. The sediment load is about the same as that of the Mississippi River; however, its sediment yield is about twice that of the Mississippi. The Mekong River sediment yield is equal to about one-seventh that of the Fly and Ganges/Brahmaputra Rivers, 85% that of the Yangtze River, and it is about 12% larger than that of the Amazon River. Data on salinity are sparse and have been collected mainly near the surface and in the dry season, in view of the importance of salinity for irrigation. The surface salinity varies seasonally, being maximum in the low-flow season and minimum in the high-flow season. In the low-flow season, the maximum salinity at 21 km is 20 and at 45 km is 6. In the high-flow season, the water is fresh nearly up to the mouth of the main channels 567 Fine-sediment dynamics 0 240 200 Distance (km) 80 160 120 40 0 Depth (m) 10 20 River mouth 30 40 (a) 1 7 7 6 No data 6 5 1 2 4 3 2 No data 1 km (b) Figure 2. Depth in m (a) in the thalweg (adapted from Gagliano & McIntire, 1968) and (b) at the mouth (Nguyen Ngoc Thuy, unpubl. data). of the estuary (Gagliano & McIntire, 1968; Nguyen Ngoc Thuy, 1988a). Data on suspended sediment are sparse in the estuary (Gagliano & McIntire, 1968; Nguyen Ngoc Thuy, unpubl. data) and in coastal waters (Anikiyev et al., 1986). About 32#106 people live in the delta in Vietnam and are engaged mainly in rice farming, artisanal fishing and aquaculture. The estuary is important for transport and shipping, including the export of rice through ports such as Can Tho but this is hindered by siltation especially in the saline-water region (Figure 2). Irrigation for rice farming is hindered by salt intrusion in the estuary, particularly in the low-flow season. The delta is an important mangrove area (Phan Nguyen Hong, 1991; Le Duc An & Phan Trung Luong, 1993), and these mangroves are used extensively for artisan fisheries and wood. Important questions have been asked about the feasibility of dredging for a seaport (e.g. at Can Tho), the effect of a sea-level rise on the delta, and the effects on the delta from the construction of proposed, major hydro-electric dams in Cambodia, Laos and Thailand (Lohmann, 1991). 568 E. Wolanski et al. T 1. Comparison of the drainage area, sediment load and yield for the Yangtze, Amazon, Mississippi, Ganges/Brahmaputran, Yangtze, Mekong and Fly Rivers (adapted from Milliman & Meade, 1983; Milliman & Syvitski, 1992; Wolanski & Gibbs, 1995) River Yangtze Amazon Mississippi Ganges/Brahmaputra Mekong Fly Area (106 km2) 1·9 6·1 3·3 1·48 0·79 0·076 Load (106 tonne year "1) Yield (tonne km "2 year "1) 480 1200 210 2180 170 116 252 190 120 1670 215 1500 The available hydrodynamic data necessary to help answer some of these questions remain inadequate. In view of the sparse baseline data, the authors undertook this study to gain some insight on currents, salinity and suspended-sediment dynamics in the estuary in the high-flow season, to help answer ultimately some of these important management questions. Methods During 15–20 November 1993, vertical profiles of temperature, salinity and suspendedsediment concentration were obtained at stations shown in Figure 1, using a Seabird CTD. This CTD was fitted with an Analite optical fibre nephelometer which is more sensitive than the Seabird nephelometer. Position fixing was by dead reckoning. At Stations 11 and 24, a 13-h station was maintained from a small, wooden vessel with hourly profiles using the CTD. In addition at these two stations, vertical profiles of currents were measured at about 1 m intervals using a Vertusca current meter suspended from a vessel. The nephelometers were calibrated for suspended-sediment concentration using water samples taken in situ and later filtered through 0·45-ìm filters. None of the water collected using a 5-l Niskin bottle contained sand, although sand was the dominant sediment on the bottom, so that the nephelometer data were not aliased (Ludwig & Hanes, 1990); for the observed variation in floc size, the nephelometer calibration is expected to vary by 20% only (Gibbs & Wolanski, 1992). In situ floc cameras were unavailable and could not have been used because the strong tidal currents caused floc breakage around windows, and the high concentration caused excessive floc overlap (Eisma et al., 1990). Instead, the technique of Gibbs and Konwar (1986) and Gibbs et al. (1989), modified by Wolanski and Gibbs (1995), was used to measure floc size. In particular, a specially modified, wide mouth 5-l Niskin bottle was used, the water was sampled directly in the bottle using a slide with well, avoiding floc breakage in other sampling techniques such as pipettes, pumps, analysers and Niskin bottle ports. The sample was capped underwater by a glass, and viewed and photographed using an Olympus inverted microscope with a range of magnification from 4 to 800#. For each water sample, several images at different magnifications were used to sample the full range of floc sizes. The photographs were scanned on an IBM-compatible computer and the digitized data were used to calculate the floc-size distribution using an image analysis software package. The system was calibrated using ragweed pollen and latex particles of 569 Fine-sediment dynamics Sea level (m) 5 4 3 2 1 0 15 16 17 18 19 Date (November 1993) 20 21 Figure 3. Time-series plot of the sea level at Vung Tau during the field study. median size 17·5 and 40 ìm, respectively. Most samples were taken from 1 m above the bottom, and a few samples were taken from near the surface. Water samples were also collected from the Niskin bottles and used to measure the primary (not flocculated) particle-size distributions using a Horiba CAPA-300 gravitational/centrifugal particlesize analyser after pre-processing the samples by treatment with an ultrasonic bath and Calgon-T dispersant and visual examination on a microscope. Results Freshwater region The freshwater region starts at c. 10 km from the mouth. Spring tides prevailed (Figure 3). Freshwater was found throughout the estuary except at stations 21–24 near the mouth. The tides reversed the currents at least up to Can Tho [123 km; Figure 4(a)]. A strong inequality of the tidal currents was apparent, the ebb and flood currents peaking at 1·2 and 0·4 m s "1, respectively. A friction-induced vertical velocity shear existed in the bottom 3 m of the water column. The temperature (not shown) was homogeneous vertically, varying in time between 29·4 and 29·6 )C. The suspended-sediment concentration varied with tidal frequency, and was higher near the bottom where it reached 0·6 g l "1 at peak ebb currents [Figure 4(b)]. There also appeared to be a background concentration of 0·15 g l "1. The reason for this background concentration is to be found in the nature of the suspended sediment in the freshwater region. The particle-size distribution of the suspended sediment in the freshwater region of the estuary (Figure 5) varied little with site, depth and tide phase, the median particle size, d50, varying only between 2·5 and 3·9 ìm, characteristic of fine silt. The clay fraction (particle size <2 ìm) accounted for 15–20% (by volume) of the suspended sediment typically, a result in agreement with data taken at Phnom Penh (338 km; Gagliano & McIntire , 1968). On the other hand, the bottom sediment was mostly sandy. Some of the suspended sediment was coagulated on organic detritus [Figure 6(a)], while the rest of the sediment, particularly near the surface, remained uncoagulated and was transported as a ‘ wash load ’ of individual particles [Figure 6(b)]. The flocs were quite large, with a floc d50 of about 40 ìm, were most numerous near the bottom, and took the form of loose, seemingly unstructured, filamentous mixtures of transparent, organic detritus and mineral particles [Figure 6(a)]. Superficially, these flocs resemble those found in organic-rich rivers elsewhere (e.g. Van Leussen, 1988; E. Wolanski et al. –2 –4 –6 –8 –10 –12 –14 –16 (b) 0.4 0 0. 0.8 0.0 0.0 0.0 0.2 1.2 1.0 0.8 0.6 0.4 2 0. 2 0. 0.2 0.2 0.3 0.4 0.3 20 21 22 23 5 0. 0.6 24 25 26 Time (h) 0.4 19 1.2 0 1. 0 .2 (a) 0 .2 0 .4 0.68 0. 1 .0 –2 –4 –6 –8 –10 –12 –14 –16 –0.2 Depth (m) 570 27 28 29 30 31 Figure 4. Time-series plot of (a) velocity (positive if down-river) distribution (m s "1) and (b) suspended-sediment concentration (g l "1) at Station 11 on 17–18 November 1993. No saltwater was found. The difference in the depth for the velocity and suspended-sediment concentration plots is due to the presence of a weight below the current meter. Uiterwijk Winkel, 1975). The present data show that the flocculated sediment settled out at slack tide and was re-entrained for current speeds >0·5 m s "1, thus the suspended-sediment concentration fluctuated with the tidal currents; on the other hand, the non-flocculated sediment, being very fine, had no time to settle at slack tide, hence the presence of a background concentration. No evidence for a net (after tidal-averaging) along-channel gradient of suspendedsediment concentration was found, any such gradients being much smaller than the temporal changes due to erosion and settling at tidal frequency. Hence, no turbiditymaximum zone was present in the freshwater region of the estuary. This finding and the presence of a deep (10 m, excluding many shoals, in the Can Tho area) thalweg in the freshwater region of the estuary suggest that during high-flow conditions, little siltation occurs in the main channel. Saltwater region This region extends only over 10 km or so from the mouth. Brackish water was only found near the mouth and then only near the bottom, around high tide and in the thalweg, respectively (Figure 7); this water was also colder (by 0·4 )C) and more turbid (by 0·2 g l "1), than in the freshwater region. The top 5 m of the water column were freshwater, and the pycnocline below was sharp. The estuary was thus highly stratified in the bottom 3 m, the situation approaching that of a salt wedge. Time-series data over one tidal cycle on 19–20 November 1993, at Station 24, show tidally-reversing currents throughout the water column, with peak flood and ebb tidal velocities of comparable magnitude [Figure 8(a)]. However, near 29 h in the time series (the morning flood tide 571 Fine-sediment dynamics 100 (a) 80 60 20 0 500–1000 250–500 125–250 62.5–125 31.3–62.5 15.6–31.3 7.81–15.6 3.91–7.81 1.95–3.91 0.98–1.95 0.49–0.98 (b) 0.24–0.49 70 60 50 40 30 20 10 0 0–0.24 Frequency (%) 40 Size (µm) Figure 5. For the suspended sediment in the freshwater region of the Mekong River estuary, typical distribution of (a) particle size at various locations, depths and tide phases, and (b) floc size at the surface ( ) and near the bottom ( ). of 20 November 1993), a strong velocity shear was present [Figure 8(a)] at the elevation of a sharp pycnocline as evidenced by saltier [by 15; Figure 8(b)] and colder [by 0·5 )C; Figure 8(c)] water near the bottom. A salt-wedge estuarine situation was present. Coastal waters are not much saltier than the near-bottom values observed at Station 24 at 30–31 h (Anikiyev et al., 1986). While a salt wedge was present, it was unsteady, moving back and forth with the tides; at low tide, it was pushed out of the estuary completely [Figure 8(b)]. The suspended-sediment concentration varied with tidal frequency, again with a background of about 0·15 g l "1 [Figure 9(a)]. When the salt wedge was absent, the salinity stratification was weak and a strong relationship was apparent between velocity and concentration, with a short phase lag due to erosion and deposition. However, when the salt wedge was present, the salinity stratification, by inhibiting turbulence, limited the resuspension to below the pycnocline. Most noticeable was the high concentration at the toe of the salt wedge (27–29 h). The near-bottom particle size-distribution [Figure 9(b)] fluctuated slightly with the tides; the median particle size, d50, was slightly larger than in the freshwater region, varying between 5 and 3·5 ìm [Figure 10(b)], with an unexplained occurrence of a very small d50 of 2 ìm at 21 h. Except for that event, the clay fraction (size <2 ìm) accounted for between 20 and 30% of the sediment (by volume), i.e. somewhat larger than its value (15–20%) in the freshwater region. Most of the near-bottom sediment was flocculated. The flocs generally comprised a membrane of clay surrounding coarser particles [Figure 6(b)], with no evidence of biological detritus forming filamentous structures on which particles were aggregated 572 E. Wolanski et al. Figure 6. Typical microphotographs of suspended sediment flocs near the bottom in (a) the freshwater region and (b) the saltwater region. such as in the freshwater. Due to their high concentration of fine silt and their small fraction of clay, the flocs were similar in structure to those of the silt-dominant Fly River estuary (Wolanski & Gibbs, 1995). These silt-dominant flocs appear structurally different from, and smaller than, the more homogeneous, clay-dominant flocs of the Amazon and Gironde River estuaries (Gibbs & Konwar, 1986; Gibbs et al., 1989). The floc size distribution varied with the tides [Figure 9(c)], the median floc size varying between 50 and 200 ìm. However, the relationship between tides and floc size is not clear. The flocs were larger in freshwater, because of their long, filamentous structure, than in seawater. Discussion Erosion and settling laws In both saline and freshwater regions of the estuary, deposition and erosion were the dominant mechanisms controlling fluctuations in suspended-sediment concentration. 573 Depth (m) Fine-sediment dynamics –1 –2 –3 –4 –5 –6 –7 –8 (a) –1 –2 –3 –4 –5 –6 –7 –8 (b) –1 –2 –3 –4 –5 –6 –7 –8 (c) 0.0 29.4 29.3 29.4 29.5 29.4 29.3 29.1 29.2 29.0 1 2 5 10 15 15 0.2 0.3 0.2 0.4 0.6 0.8 Distance (nm) 1.0 0.4 1.2 Figure 7. Cross-channel transect of (a) temperature, (b) salinity and (c) suspendedsolid concentration (g l "1) at the mouth through Station 24, 05.00–05.30 h on 20 November 1993. The tidal station was located at about 1 nautical mile. Thus, at zero-order approximation, advection may be neglected. The time-series data enabled the authors to derive empirical constants in the sediment transport laws for the Mekong River Delta. The vertically-averaged sediment mass conservation equation is (Wolanski et al., 1995): )(HC)/)t=E "D (1) where t is time, C is the depth-averaged suspended-sediment concentration, and H is the depth. The source/sink terms are erosion (E) and deposition (D): E=á((u/uc)2 "1), if u>uc =0, if u<uc (2) D=Cws(1"(u/uo)2), if u<uo =0, if u>uo (3) and: E. Wolanski et al. 0.2 0.2 0.4 0.0 –0.2 0.0 (b) –1 –2 –3 –4 –5 –6 –7 –8 (c) 0.2 –0.4 0.6 –1 –2 –3 –4 –5 –6 –7 –8 19 0.4 0.2 –0.6 (a) 1 2 1 10 5 2 15 5 29 .5 .6 29 Depth (m) –1 –2 –3 –4 –5 –6 –7 –8 –0.4 574 29.4 .3 29 .0 29.5 20 21 22 23 29 24 25 26 Time (h) 27 28 29 30 31 Figure 8. Time-series plot of the distribution of (a) velocity (m s "1, positive if down-river), (b) salinity and (c) temperature on 19–20 November 1993, at station 24. where u is the velocity, uc and uo are the threshold velocities for, respectively, erosion and settling, á is an empirical constant, and ws is the settling velocity. These equations account for erosion and settling lags between velocity and suspended-sediment concentration (West & Sangodoyin, 1991). The value of ws was determined from the speed of descent of the centre of mass of the suspended sediment at slack tide, and this estimate compared favourably with the empirical data (Gibbs, 1985) as a function of floc size. The threshold velocities were determined from a visual examination of the field data. An important parameter, determined from the field data, is the relative elevation, â, of the centre of mass of the suspended sediment above the bed; this is important because the suspended sediment only has to travel a fraction, â, of the water column to either reach the bottom during deposition or reach equilibrium height during erosion. The model also accounts for the ‘ wash load ’ in the freshwater region. The adopted values for the parameters are detailed in Table 2; these values appear reasonable when compared with those from other estuaries (e.g. Fly River estuary; Wolanski et al., 1995). The high values of uo (i.e. sediment settles out even at relatively high water velocities) in freshwater seems reasonable in view of the large floc size. 575 Depth (m) –1 –2 –3 –4 –5 –6 –7 –8 –9 (a) 0.22 Fine-sediment dynamics 0.2 0.22 2 0.18 0.26 0.26 0.14 0.18 0.22 0. 26 0.30 0.14 0.30 Volume (%) (b) 100 90 80 70 60 50 40 30 20 10 0 8.0 4.0 2.0 0.98 0.49 Volume (%) (c) 100 90 80 70 60 50 40 30 20 10 0 19 500 63 125 250 31 16 8 20 21 22 23 24 25 26 Time (h) 27 28 29 30 31 Figure 9. Same as Figure 7 for (a) suspended-solid concentration (g l "1), (b) particle-size distribution (ìm, % by volume) and (c) floc-size distribution (ìm, % by volume). T 2. Parameters in the sediment erosion/deposition model (MKS units) Freshwater Saltwater á uc uo ws â 0·00001 0·000008 0·5 0·4 1·0 0·55 0·0004 0·00025 0·1 0·2 See text for definitions of parameters. The comparison between observed and predicted depth-averaged suspendedsediment concentration is encouraging for the freshwater region [Figure 11(a)], suggesting that the model has captured the essential dynamics. The comparison for the saltwater region is also encouraging [Figure 11(b)], except at 28 h which corresponds to the toe of the salt wedge; the model does not account for internal dynamics of a salt wedge. The high concentration at that time was probably due to the high turbulent intensity in the 576 E. Wolanski et al. 0.8 0.6 (a) 0.4 0.2 0.0 –0.2 Ebb Flood –0.4 –0.6 –0.8 6 (b) 5 4 3 2 1 19 20 21 22 23 24 25 26 Time (h) 27 28 29 30 31 Figure 10. Time-series plot of the depth-averaged velocity (m s "11, positive if down-river) and the median particle size, d50 (ìm), near the bottom at Station 24, 19–20 November 1993. toe of the salt wedge and to a convergence of bottom currents from either side of the toe as has been demonstrated in laboratory experiments (Simpson, 1972; Jirka, 1990). This process is quite different from the increased turbulence in a hydraulic jump after river plume lift-off, such as occurred in the Mississippi River mouth (Wright & Coleman, 1974; Eisma, 1993), and does not appear to have been studied previously in the field. Fate of fine sediment in the Mekong River estuary The present data are sparse but nevertheless unique for the Mekong River delta. They suggest that, in the high-flow season, the bulk of the fine sediment is made of fine silt and is carried right through the freshwater region of the estuary. A reasonably deep thalweg is maintained (Figure 2). The saline region is restricted to the mouth, which is where siltation and navigation problems are most severe. Calculations of the depthintegrated sediment flux suggest that at least 95% of that sediment may be exported to the sea. This sediment is deposited within 20 km of the coast (Anikiyev et al., 1986). During the high-flow season, a measurable fraction of that sediment (possibly 5%) returns to the estuary with the salt wedge, the majority of this sediment entering the estuary with the toe of the salt wedge at flood tide. The suspended sediment was finer (d50~3 ìm) at ebb tide than at flood tide (d50~5 ìm; Figure 10); since the clay fraction of the suspended sediment is larger (20–30%) in the saline region of the estuary than in the freshwater region (15–20%), a clay-enrichment sorting mechanism operates which appears to be similar to that in the Fly River estuary (Wolanski & Gibbs, 1995). In the low-flow season, saltwater penetrates about 40 km up-river in this channel, and even more in the other channels, and partially well-mixed estuarine conditions prevail Fine-sediment dynamics 577 (a) 0.3 –1 Suspended-sediment concentration (gl ) 0.2 0.1 0.0 (b) 0.3 0.2 0.1 19 20 21 22 23 24 25 26 27 28 29 30 31 Time (h) Figure 11. Time-series plot of the observed ( ) and predicted ( ) depth-averaged suspended-sediment concentration at (a) Station 11 and (b) Station 24. (Figure 12). Field data on suspended-sediment concentrations in the low-flow season are sparse, with occasional spot measurements of near-bottom concentration reaching 1 g l "1 (Gagliano & McIntire, 1968; Nguyen Ngoc Thuy, unpubl. data). Also, there is evidence of the lower region of the estuary being slightly flood-dominant (Nguyen Ngoc Thuy, 1989). These findings suggest the presence of a turbidity maximum zone (Eisma, 1993). Marine sediment is thus presumably pumped up-river, and much of this sediment may be Mekong River sediment that was deposited in shallow coastal waters during the previous high-flow season. This intrusion of marine sediment in the estuary will be facilitated by the effects of waves in coastal waters (Kendrick & Derbyshire, 1983). This up-river pumping of sediment in the dry season explains the smaller depth in the lower 20 km of the estuary (Figure 2). Conclusion The pathways of fine-sediment in the Mekong River estuary are sketched in Figure 13. In the high-flow season, the sediment is deposited and eroded periodically with the tides in the freshwater region, but is still transported down-river at mean velocities of 0·5–1 m s "1. A sizable fraction of the fine suspended sediment in freshwater is 578 E. Wolanski et al. Mouth Distance (km) 40 30 20 (a) 1 Depth (m) 0 10 0 5 5 15 10 2 10 0 6 Can Tho 10° 10 April November 10 9°N 6 10 South China Sea (b) 6 106°E Figure 12. (a) Typical salinity distribution in the dry season (adapted from Nguyen Ngoc Thuy, 1988b) and (b) distribution of the 0·4 surface salinity (a salinity value critical for irrigation; maximum salinity intrusion occurs in April and minimum in November) in waterways in the Mekong River delta (adapted from Gagliano & McIntire, 1968). aggregated in loose, filamentous flocs comprising organic detritus aggregating inorganic sediment. The fraction (c. 15 to 20%) of clay particles (size <2 ìm) is smaller than that in muddy rivers such as the Amazon and the Gironde Rivers (30%), but is comparable to that (~20%) in silty (as opposed to muddy) rivers such as the Fly and Jiaojiang Rivers (Gibbs et al., 1989; Li et al., 1993; Gibbs & Konwar, 1994; Wolanski & Gibbs, 1995). Most of the sediment is flocculated in brackish water. In the high-flow season, most of this sediment deposits in shallow (depth <20 m) coastal waters less than 20 km from the coast (Anikiyev et al., 1986). A small fraction (possibly 5%) of this sediment returns to the estuary in the salt wedge. In the low-flow season, classical, partially well-mixed estuarine dynamics prevail and entrains sediment 579 Fine-sediment dynamics Estuary (a) South China Sea Salt wedge 0 5 10 15 Deposition Deposition ~ 5 km Wind (b) 0 1 5 10 20 25 Deposition ~ 30 km Figure 13. Conceptual model of the pathways of fine sediment in the Mekong River estuary in (a) the high-flow season and (b) the low-flow season. The low-flow season model is still very much speculative in view of the paucity of field data in that season. Stippled area, turbidity maximum zone. from the coastal zone into the estuary, to a turbidity maximum zone, [Figure 13(b)] This sediment was deposited there during previous high-flow seasons. This up-river pumping of sediment will be facilitated by the action of waves in shallow coastal waters (Kendrick & Derbyshire, 1983), the north-east monsoon having a strong cross-slope wind component. The return to the estuary in the low-flow season of sediment deposited in shallow coastal water in the previous high-flow seasons, is made possible by the oceanographic conditions on the inner shelf, the wind driving southward currents in winter and northward currents in summer (Gagliano & McIntire, 1968; Anikiyev et al., 1986; Nguyen Ngoc Thuy, 1988a). Thus, there is a reversal of the littoral drift, and sediment exported longshore northward in summer returns in winter, although there is a suggestion of a net southward drift (Gagliano & McIntire, 1968). Sediment exported to coastal waters during the high-flow season thus becomes available for entrainment into the estuary during the low-flow season. The seasonal reversal of the littoral drift is qualitatively similar to that affecting the Yangtze River (Lazure & Girardot, 1990); it contrasts the Mekong River from the Amazon and the Mississippi Rivers where the sediment travels unidirectionally longshore away from the estuary (Wright & Coleman, 1974; Curtin & Legeckis, 1986; Geyer et al., 1991) and the Zaire and Ganges-Brahmaputra Rivers, where some of the sediment 580 E. Wolanski et al. sinks in the deep ocean via a canyon (Kuehl et al., 1989; Barua et al., 1994). The Mekong River estuary differs also from the Mississippi River in that the Mississippi River estuary has a salt wedge throughout the year while in the Mekong River estuary, a salt wedge is only present in the high-flow season, and partially well-mixed estuarine conditions prevail in the low-flow season (Wright & Coleman, 1974). The regularization of the discharge of the Mekong River following the construction of large, proposed hydro-electric dams further upstream in Cambodia, Laos and Thailand, may have significant effects on the Mekong River estuary. In particular, with decreased peak discharges, less riverine sediment will be exported to the sea during the high-flow season. Also, with decreased high-flows, the salt wedge will be replaced by partially well-mixed estuarine conditions within the estuary. This will lead to enhanced tidal pumping of sediment into the estuary from the coastal zone. There is sufficient (old riverine) sediment available in coastal waters (Anikiyev et al., 1986) to ensure that the Mekong River estuary will continue filling with coastal sediment even if the hydroelectric dams trap much of the (new) riverine sediment. Increased siltation in the Mekong River estuary is thus likely following the construction of large hydro-electric dams further upstream. At longer time scales, the coastal waters may become depleted of sediment in view of a possible net southward drift (Gagliano & McIntire, 1968) and this can be expected to lead to coastal erosion. At even longer time scales, salinity intrusion and siltation/erosion problems in the delta may be exacerbated by the relative sea-level rise estimated to be about 0·002 m year "1 (Nguyen Ngoc Thuy et al., 1994). In addition, the rapid economic developments in the Mekong riparian countries, as well as those in the delta itself, will probably result in increased discharges of pollutants. While a fast degradation of organic pollutants is possible (e.g. the Yangtze River estuary, see Yu et al., 1990), heavy metals and pesticides may accumulate in the mud of the estuary because of their cohesive nature. In the north-east monsoon, the wind has an inshore component and should bring shoreward oil slicks from spills at oil wells on the continental shelf offshore from the Mekong River delta; this is of concern because oil has a long residence time (years) when mixed in mangrove sediments (Burns et al., 1994). Thus an environmental management plan needs to be established for the Mekong River delta. Due to the importance of the delta, detailed oceanographic, chemical and biological field studies should be encouraged to provide data to enable wise management of this important resource. More detailed modelling studies are also needed and some will be undertaken, made possible by a grant from the IBM International Foundation. Acknowledgements This study was supported by the Hydrometeorological Service of Vietnam, the Marine Hydrometeorological Center (HYDROMET) and the Australian Institute of Marine Science. This study, undertaken in difficult field conditions, would not have been possible without the help, support and generosity of Dr Nguyen Duc Ngu, Dr Phan Van Hoac, Mr Ngo Huu Tang, colleagues at the HYDROMET office at Can Tho, the villagers at Ca Loc and Mr Duncan Galloway; it is a pleasure to thank them all warmly. References Anikiyev, V. V., Zaytsev, O. V., Trinh The Hieu, Savil’Yeva, I. I., Starodubtsev, Y. & Shumilin, Y. N. 1986 Variation in the time-space distribution of suspended matter in the coastal zone of the Mekong River. Oceanology 26, 725–729. Fine-sediment dynamics 581 Barua, K., Kuehl, S. A., Miller, R. L. & Moore, W. S. 1994 Suspended sediment distribution and residual transport in the coastal ocean off the Ganges-Brahmaputra river mouth. Marine Geology 120, 41–61. Borland, W. M. 1973 Pa Mong phase II. Supplement to Main Report (Hydraulics and Sediment Studies). U.S. Bureau Reclamation, Volume 1, 282 pp., Volume 2, 304 pp. Burns, K. A., Garrity, S. D., Jorissen, D. et al. 1994 The Galetta oil spill. II. Unexpected persistence of oil trapped in mangrove sediments. Estuarine, Coastal and Shelf Science 38, 349–364. Curtin, T. B. & Legeckis, R. V. 1986 Physical observations in the plume region of the Amazon River during peak discharge—I. Surface variability and II. Water Masses. Continental Shelf Research 6, 31–71. Eisma, D. 1993 Suspended Matter in the Aquatic Environment. Springer-Verlag, Berlin, 315 pp. Eisma, D., Schumacher, T., Boekel, H. et al. 1990 A camera and image-analysis system for in-situ observations of flocs in natural waters. Netherlands Journal of Sea Research 27, 43–56. Gagliano, S. M. & McIntire, W. G. 1968 Reports on the Mekong River delta. Coastal Studies Institute Technical Report No. 57, Louisiana State University, 143 pp. Geyer, W. R., Beardsley, R. C., Candela, J. et al. 1991 The physical oceanography of the Amazon outflow. Oceanography 4, 8–14. Gibbs, R. J. 1985 Estuarine flocs: their size, settling velocity and density. Journal of Geophysical Research 90, 3249–3251. Gibbs, R. J. & Konwar, L. 1986 Coagulation and settling of Amazon River suspended sediment. Continental Shelf Research 6, 127–149. Gibbs, R. J., Tshudy, D., Konwar, L. & Martin, J. M. 1989 Coagulation and transport of sediments in the Gironde estuary. Sedimentology 36, 987–999. Gibbs, R. J. & Wolanski, E. 1992 The effects of flocs on optical backscattering measurements of suspended material concentration. Marine Geology 107, 289–291. Jirka, G. H. 1990 Circulation in a salt wedge estuary. In Residual Currents and Long-term Transport (Cheng, R. T., ed.). Springer-Verlag, New York, pp. 223–237. Kendrick, M. P. & Derbyshire, D. V. 1983 Factors affecting the supply and distribution of sediment in some tropical rivers. Canadian Journal of Fish Aquatic Science 40 (Suppl.), 35–43. Kuehl, S. A., Hariu, T. & Moore, W. 1989 Shelf sedimentation off the Ganges-Brahmaputra river system: Evidence for sediment bypassing to the Bengal fan. Geology 17, 1132–1135. Lazure, P. & Girardot, J. P. 1990 Hydrodynamics of the Changjiang estuary area. Proceedings of the International Symposium on Biogeochemical Study of Changjiang Estuary and Adjacent Coastal Waters of the East China Sea. China Ocean Press, Hangzhou, pp. 38–61. Le Duc, An & Phan Trung Luong 1993 Environmental change in southern part of Mekong River delta and problems of territorial rational use. In Deltas of the World (Kay, R., ed.). American Society of Civil Engineers, New York, pp. 114–121. Li, Y., Wolanski, E. & Xie, Q. 1993 Coagulation and settling of suspended sediment in the Jiaojiang River estuary, China. Journal of Coastal Research 9, 390–402. Lohmann, L. 1991 Engineers move in on the Mekong. New Scientist 1777, 38–41. Ludwig, K. A. & Hanes, D. M. 1990 A laboratory evaluation of optical backscatterance suspended solids sensors exposed to sand-mud mixtures. Marine Geology 94, 173–179. Milliman, J. D. & Meade, R. H. 1983 World-wide delivery of river sediment to the oceans. Journal of Geology 91, 1–21. Milliman, J. D. & Syvitski, J. P. M. 1992 Geomorphic/tectonic control of sediment discharge to the ocean: The importance of small mountainous rivers. Journal of Geology 100, 525–544. Nguyen Ngoc Huan 1987a Water Regime of Tidal Mekong Delta. Ph.D. dissertation, Moscow University (in Russian). Nguyen Ngoc Huan 1987b Application of mathematical models for computation of water distribution of Mekong delta. Journal of Moscow University 3 (in Russian). Nguyen Ngoc Thuy 1979 Tides in the Gulf of Thailand and in the coastal regions of the Mekong delta. Symposium, Can-Tho Institute (in Vietnamese). Nguyen Ngoc Thuy 1988a Hydrometeorology of Vietnam Sea Waters Volume 1. Marine Hydrometeorological Center, Science & Technical Publishers, Hanoi (in Vietnamese). Nguyen Ngoc Thuy 1988b Tides in the Vietnam Estuaries. Marine Hydrometeorological Center, Science & Technical Publishers, Hanoi, 20 pp (in Vietnamese). Nguyen Ngoc Thuy 1989 Hydrodynamic modeling of the propagation of tidal waves in the river system of the Mekong delta. State Oceanographic Institute, GIMIZ Publ. No. 188, Moscow (in Russian). Nguyen Ngoc Thuy, Pham Hoang Lam & Bui Dinh Khuoc 1994 Global warming, El-Nino phenomenon and the sea level change in Vietnam and neighboring countries. IOC/WESTPAC Symposium, Bali, Indonesia, 22–26 November 1994, 15 pp. Phan Nguyen Hong 1991 Status of mangrove ecosystems in Vietnam: some management considerations. Proceedings of Mangrove Genetic Resources Centers project Formulation Workshop, Madras, India, 15–19 January 1991. M.S. Swaminathan Research Foundation, Madras, pp. 53–63. 582 E. Wolanski et al. Simpson, J. H. 1972 Effects of the lower boundary on the head of a gravity current. Journal of Fluid Mechanics 53, 759–768. Tingsanchali, T. & Lien, N. D. 1987 Flood simulation in the tidal delta of the Mekong River by SSARR model. Water, Asian Institute of Technology 11, 117–126. Uiterwijk Winkel, APB 1975 Microbiologische aspecten en het sedimentatie gedrag van rivierslib. Rijkwaterstaat, Direktie Waterhuishounding en Waterbeweging. District Zuidwest, Report No. 44.006.001, 60 pp. United Nations 1957 Development of water resources in the lower Mekong basin. United Nations Economic Commission for Asia and the Far East. Flood Control Series No. 12, Bangkok, 75 pp. Van Leussen, W. 1988 Aggregation of particles, settling velocity of mud flocs. A review. In Physical Processes in Estuaries (Dronkers, J. & Van Leussen, W., eds). Springer-Verlag, Berlin, pp. 347–403. West, J. R. & Sangodoyin, A. 1991 Depth-mean tidal current and sediment concentration relationships in three partially mixed estuaries. Estuarine, Coastal and Shelf Science 32, 141–159. Wolanski, E. & Gibbs, R. J. 1995 Flocculation of suspended sediment in the Fly River estuary, Papua New Guinea. Journal of Coastal Research 11, 754–762. Wolanski, E., King, B. & Galloway, D. 1995 Dynamics of the turbidity maximum in the Fly River estuary, Papua New Guinea. Estuarine, Coastal and Shelf Science 40, 321–337. Wright, L. D. & Coleman, L. M. 1974 Mississippi River mouth processes: effluent dynamics and morphologic developments. Journal of Geology 82, 751–778. Yu, G., Zhou, J. & Liao, X. 1990 The biogeochemical behaviors of pollutants and nutrients in the Changjiang estuary and its adjacent shelf areas. Proceedings of the International Symposium Biogeochemical Study of Changjiang Estuary and Adjacent Coastal Waters of the East China Sea. China Ocean Press, Hangzhou, pp. 1–15.