Sedimentology (1998) 45, 397±410

Dynamics of the turbidity maximum zone in a micro-tidal estuary: ,

M. G. HUGHES*, P. T. HARRIS and T. C. T. HUBBLE* *Department of Geology and Geophysics (FO5), University of , Sydney NSW 2006, Australia Antarctic CRC, Australian Geological Survey Organization, University of Tasmania, GPO Box 252C, Hobart TAS 7001, Australia

ABSTRACT Bed sediment, velocity and turbidity data are presented from a large (145 km long), generally well-mixed, micro-tidal estuary in south-eastern Australia. The percentage of mud in the bed sediments reaches a maximum in a relatively narrow zone centred 30±40 km from the estuary mouth. Regular tidal resuspension of these bed sediments produces a turbidity maximum (TM) zone in the same location. The maximum recorded depth-averaged turbidity was 90 FTU and the maximum near-bed turbidity was 228 FTU. These values correspond to suspended particulate matter (SPM) concentrations of roughly 86 and 219 mg l)1, respectively. Neither of the two existing theories that describe the development and location of the TM zone in the extensively studied meso- and macro-tidal estuaries of northern Europe (namely, gravitational circulation and tidal asymmetry) provide a complete explanation for the location of the TM zone in the Hawkesbury River. Two important factors distinguish the Hawkesbury from these other estuaries: (1) the fresh water discharge rate and supply of sediment to the estuary head is very low for most of the time, and (2) suspension concentrations derived from tidal stirring of the bed sediments are comparatively low. The ®rst factor means that sediment delivery to the estuary is largely restricted to short-lived, large-magnitude, ¯uvial ¯ood events. During these events the estuary becomes partially mixed and it is hypothesized that the resulting gravitational circulation focuses mud deposition at the ¯ood-determined salt intrusion limit (some 35 km seaward of the typical salt intrusion limit). The second factor means that easily entrained high concentration suspensions (or ¯uid muds), typical of meso- and macro-tidal estuaries, are absent. Maintenance of the TM zone during low-¯ow periods is due to an erosion-lag process, together with a local divergence in tidal velocity residuals, which prevent the TM zone from becoming diffused along the estuary axis.

INTRODUCTION imum (TM) zone (see Dyer, 1986; Eisma, 1993). In recent times the TM zone has emerged as an A large number of studies over the past 25 years important focus for estuarine research, because it have investigated suspended sediment dynamics is frequently located near the fresh±salt water in estuaries. Many of these studies report the interface and consists of high concentrations of existence of a zone toward the head of the estuary suspended particulate matter (SPM). These con- where the turbidity of the water is markedly ditions lead to strong spatial and temporal gradi- higher than that observed further landward or ents in geochemical processes, which can play a seaward. This zone is termed the turbidity max- determining role in the dynamics and fate of

Ó 1998 International Association of Sedimentologists 397 398 M. G. Hughes et al.

Table 1. Some previously reported values for SPM concentration in the TM zone.

Spring SPM concentration Estuary range TM zone* Reference

Meso- Macro-tidal Weser River 3á8 m 1á5 g l)1 Grabemann & Krause (1989) Fly River 4á0 m 5±30 g l)1 Wolanski et al. (1995) Tamar River 4á5 m 10 g l)1, 26 g l)1 Uncles & Stephens (1993) Gironde River 5á0 m 10 g l)1 Allen & Castaing (1973) Severn River 12á3 m 20 g l)1 Kirby (1988) Micro-tidal James River 0á7 m 100±270 mg l)1 Nichols (1993) Varde A Estuary 1á6 m 100±1000 mg l)1 Bartholdy (1984) Cooper River 2á0 m 40±100 mg l)1 Althausen & Kjerfve (1992)

* These ®gures are near bed values, except in the case of the Tamar River, where the ®rst is the depth-averaged value and the second is the near bed value. anthropogenic inputs to estuarine systems (e.g. net transport of sediment in a landward direction Morris et al., 1978; Morris et al., 1986; Turner results. A null point exists where the seaward et al., 1994). directed river ¯ow has a transport competency Most research on the TM zone has focused on equal to that of the ¯ooding tide. It is at this null meso- and macro-tidal estuaries, where tidal point, located somewhere landward of the point currents are strong (>1 m s)1) and SPM concen- where the tide becomes signi®cantly distorted, trations are large (Table 1). In many cases, the that ®ne sediment accumulates to produce a TM suspensions are well in excess of 10 g l)1, classi- zone (Dyer, 1986). ®ed as `¯uid mud' (see Einstein & Krone, 1962). In general, gravitational circulation is most Two mechanisms have been proposed for the important in meso-tidal estuaries, whereas tidal development of a TM zone in meso- and macro- asymmetry is of primary importance in macro- tidal estuaries. The ®rst involves the residual tidal estuaries. It is noteworthy, however, that in currents associated with gravitational circulation many of the meso- and macro-tidal estuaries in partially mixed estuaries. In these estuaries the studied to date, both residual circulation and salinity distribution drives a residual bottom ¯ow tidal asymmetry are coincident phenomenon. The directed landward along the estuary axis and a balance between the two can change along the residual surface ¯ow directed seaward. SPM estuary and between neap and spring , so supplied to the estuary head travels seaward in that the causal mechanism for the focused accu- the surface waters until it begins to ¯occulate and mulation of ®ne sediment in a TM zone is not settle deeper into the water column. Then it is always straightforward (Dyer, 1986). In both caught in the landward ¯owing bottom current cases, however, tidal resuspension of ®ne sedi- and transported back towards the estuary head. A ment in the zone of accumulation leads to null point exists where the velocity of the maximum turbidities. In this regard, the presence landward ¯owing estuarine bottom water equals of readily entrainable slackwater ¯uid muds, the seaward ¯owing river water. This null point which are characteristic of these meso- and typically occurs near the landward limit of the macro-tidal estuaries (Kirby, 1988), clearly con- salt intrusion, and it is here that ®ne sediment tributes to the existence of a TM zone. accumulates and undergoes tidal re-suspension Investigations of TM zone suspended sediment to produce a TM zone (Dyer, 1994). dynamics in micro-tidal estuaries are signi®cant- The second mechanism proposed for the de- ly fewer in number than for meso- and macro- velopment of a TM zone in meso- and macro-tidal tidal estuaries. Of the few published studies, all estuaries involves the distortion of the tide wave report SPM concentrations in the TM zone of less associated with non-linear interactions between than 1000 mg l)1 (Table 1). In particular, the the tide and channel morphology. Distortion origin, persistence and dynamics of the TM zone causes ¯ood currents to be stronger and of shorter in micro-tidal estuaries is not well established. duration than ebb currents. This asymmetry Gravitational circulation and tidal asymmetry do increases towards the head of the estuary and a occur in these estuaries, but it is not known if the

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 399

Fig. 1. Map of the Hawkesbury River showing locations of perma- nent tide gauges and instrument deployment sites. Tide Gauges 1±6 refer, respectively, to the following site names: Spencer, Gunderman, , Sackville, Ebenezer and Windsor. resulting sediment transport is suf®cient to lead Table 2. Summary statistics of fresh water discharge to continuous deposition of ®ne sediments within measured on the at Penrith weir during a narrow zone in the estuary. Moreover, the lack the period 1960±1992 (raw data provided by NSW De- partment of Land and Water Conservation). of regularly occurring, high-concentration, slack- water suspensions (or ¯uid muds) in these Discharge estuaries also means that the mechanism and Percentile (m3 s)1) role that tidal resuspension plays will be some- what different to that observed in meso- and 10 0á67 macro-tidal estuaries. The purpose of this paper 50 2á85 is to increase our knowledge of sediment dynam- 90 79á53 ics in the TM zone of micro-tidal estuaries, by Mean 46á48 presenting the results from a study of the micro- Maximum 11 267á19 tidal Hawkesbury River estuary.

THE HAWKESBURY RIVER Tides at the coast are mixed semi-diurnal with a mean spring range of 1á32 m and a mean neap The Hawkesbury River drains into the Tasman range of 0á78 m. The maximum tidal range is Sea on the eastern coast of , 1á92 m. The tide-affected part of the Hawkesbury Australia (Fig. 1). The total catchment area is River is 145 km long, with the tidal limit about 23 000 km2. Fresh water discharge into the located at the con¯uence (Fig. 1). estuary is modest. Analysis of 32 years of records This section can be divided into two reaches: (a) a from Penrith weir (provided by NSW Department ¯uvio-tidal reach, where the water is fresh and of Land and Water Conservation), located 20 km the ¯ow is mostly tidal, except during large ¯oods upstream of the tidal limit, indicates that the and (b) an estuarine reach, where the water is mean fresh water discharge into the estuary is brackish to saline and the ¯ow is always tidal, 46á48 m3 s)1 whereas the median is only although it is modi®ed considerably during large 2á85 m3 s)1 (Table 2). The fact that the median ¯oods. The estuarine reach occupies from c. 0± is so much smaller than the mean indicates that 75 km from the estuary mouth and the ¯uvio-tidal ¯uvial ¯ow into the estuary is highly skewed reach occupies 75±145 km from the estuary towards small discharges. The general pattern of mouth. fresh water discharge is one of extended low-¯ow The estuarine reach occupies a drowned river conditions punctuated by aseasonal, short-lived, gorge carved into Triassic Hawkesbury sandstone large-magnitude ¯oods. Floods often have dis- (Roy et al., 1980). The channel is ¯anked locally charge rates up to 3 orders of magnitude larger by steep bedrock valley walls, talus slopes or lat- than the mean. erally restricted back-plain swamps with fringing

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 400 M. G. Hughes et al. mangrove stands. Channel morphology includes 126 km from the estuary mouth (Fig. 1). A tide tidal levees, inter-tidal and shallow sub-tidal bars gauge in the entrance to Sydney Harbour, 30 km and point bars attached to channel margins and to the south of the Hawkesbury River, was used to limited inter-tidal and shallow sub-tidal mud represent the tide at the coast. The records are ¯ats located mid-channel (Hubble & Harris, 1994). continuous time series with a 15 minute sampling The ¯uvio-tidal reach consists of a channel that is interval. Twenty-nine days of record were select- moderately incised into bedrock or Tertiary sed- ed from each location for harmonic analysis using iment and is abutted by levee banks and ¯ood the method of Foreman (1977). The amplitude plains of variable width (Hubble & Harris, 1994). and phase of 33 tidal constituents were resolved. Channel morphology mostly appears to be of a ¯uvial nature, shaped by the infrequent ¯ood Salinity, turbidity and current velocity events rather than by the tide (Hubble & Harris, 1994; Hughes & Callaghan, 1995). Longitudinal pro®les of salinity and turbidity The channel width decreases at an exponential were measured on two occasions (26 March 1993 rate, from roughly 3500 m at the entrance to and 7 April 1993) by vertical sampling of the Broken Bay to 150 m at the tidal limit (Fig. 1). At water column at 3 km intervals along the estuary 35 km inland the channel width has decreased to as near as possible to the time of local high water  10% of its width at the entrance and at 75 km slack. The water samples were collected with a inland it has decreased to 5%. Beyond 75 km Niskin Bottle from the water surface, mid depth the channel width remains nearly constant at and 1 m above the bed. Salinity, temperature and 100±200 m. The thalweg depth displays a steady, turbidity of the samples were determined imme- gradual decrease from 15 to 20 m in Broken Bay diately on deck using a WTW LF196 temperature/ to roughly 10 m at 100 km inland. Beyond conductivity meter and a Hach 2100P turbidity 100 km it decreases more quickly to < 2 m depth meter. at the tidal limit (Hughes & Callaghan, 1995). Long-term records of the temperature, salinity, current speed and turbidity were obtained using Aanderaa RCM7 current meters ®tted with Sea- FIELD AND ANALYTICAL METHODS tech transmissometers. The instruments were deployed in taught-line moorings with the sen- Field data reported here was collected between sors situated 1 m above the bed. Data was logged May 1992 and June 1994. Throughout this period at 10 min intervals and the instruments were the fresh water discharge was indicative of serviced fortnightly. This was necessary to avoid extended, low-¯ow conditions. problematic soiling of the transmissometer win- dows. A total of 19 sites were occupied to provide Bed sediment representative coverage of the longitudinal and Bed sediments were sampled from channel cross- cross-channel dimensions of the estuary. Some sections spaced at intervals of 5 km along the areas of the estuary are, however, under-repre- estuary. A total of ®ve samples were collected sented due to the inability to moor instruments in from each cross-section: three samples from the established trawling grounds. All sites were sub-tidal channel and one from the inter-tidal occupied for at least one spring±neap tide cycle, zone on each bank. The samples were wet sieved, with most sites occupied for up to eight cycles. dried and weighed to determine the percentage of Because it is the longitudinal estuary dimension mud (<63 lm), sand (63 lm to 2 mm) and gravel that is of interest here, the data from only eight of (>2 mm). The equivalent size distribution of the the deployment sites are presented. At these sand and mud fractions were determined using a particular sites the instruments were moored in settling column and laser diffraction particle sizer the channel thalweg. Site locations are shown in (Hubble & Harris, 1994). Fig. 1 and time periods chosen for analysis are listed in Table 3. Although the time periods analysed do not correspond between all sites, Water level the pattern of velocity and turbidity at all sites is Water level data from seven permanent tide highly repetitive in time from one spring±neap gauges on the Hawkesbury River were provided tide cycle to the next (Hughes & Callaghan, 1992, by the NSW Public Works Department. The 1993, 1994). Hence, it is reasonable to compare gauges are located at 28, 49, 61, 97, 112 and data from different time periods provided the

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 401

Table 3. List of instrument deployment sites and the as is the large subaqueous ¯ood tide delta that deployment period reported here. occupies the ¯oor of Broken Bay and extends Distance from landward into the estuary (Roy et al., 1980). Site estuary mouth Deployment Between Broken Bay and 14 km from the estuary Number (km) period mouth the intertidal zone consists almost entirely of bedrock valley walls, thus sediment was only HR3 58á8 03/06/92±18/06/92 sampled from cross-sections landward of this HR5 55á0 03/06/92±18/06/92 point. Sub-tidal mud is typically dark olive grey HR6 51á5 07/05/92±21/05/92 and organic rich (Roy, 1983) with a mean grain HR8 49á5 02/12/92±17/12/92 size ranging from 7á2±13á9 lm (medium silt) at HR9 42á0 02/12/92±17/12/92 the current meter sites based on laser particle size HR10 39á0 02/12/92±17/12/92 HR11 34á5 02/12/92±17/12/92 analyses. The sand fraction was composed of HR19 21á5 24/05/94±09/06/94 angular to moderately rounded, quartzose, medi- um-well-sorted, ®ne to coarse sand, with a mean grain size ranging from 0á2±0á6 mm based on settling column analyses. comparison is based on a record length incorpo- Most of the inter-tidal mud deposited in the rating at least one spring±neap tide cycle. estuary occurs in a zone located 14±60 km from To permit comparison with previous studies, the estuary mouth (Fig. 2a). The proportion of the transmissometers were calibrated for turbidity mud in samples collected from this zone is using 10 formazin calibration standards ranging generally greater than 50%, and reaches a maxi- between 2 and 200 formazin turbidity units mum of greater than 95% in a narrower zone (FTU). The instrument response was linear over located 36±50 km from the estuary mouth. In the the calibration range and was found to be consis- ¯uvio-tidal reach, 75±145 km from the estuary tent between calibrations. Least squares regres- mouth, the mud content of the inter-tidal zone is sion equations were ®tted to the calibration data generally less than 30%, although some isolated to determine transform equations relating turbid- cases reach 40±95% (Fig. 2a). ity to light transmission. These transform equa- Sub-tidal mud content exceeds 40% in the 2 tions all had r values of 0á99 or greater. seaward section of the estuary (Fig. 2b) and It was not feasible to undertake an in situ calibration of the transmissometers for SPM concentration, because of the long deployment periods. For this reason only turbidity is reported here. It is worth noting, however, that the general repeatability of the turbidity records from one spring±neap cycle to the next at all sites strongly suggests that the relationship between turbidity and SPM concentration is reasonably consistent through the estuary and through time (Hughes & Callaghan, 1992, 1993, 1994). To enable an approximate comparison to be made between the data reported here and other studies the SPM concentration (mg l)1) is equal to roughly 0á96 times turbidity (FTU). This is based on laboratory calibrations of the transmissometers with bed sediment collected adjacent to the instrument deployment sites (Hughes & Callag- han, 1995).

RESULTS

Distribution of bed sediments Fig. 2. Cross-sectionally averaged composition of (a) The beaches lining much of the foreshore of inter-tidal and (b) sub-tidal sediments as a function of Broken Bay are composed of clean marine sands, distance from the estuary mouth.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 402 M. G. Hughes et al.

Fig. 3. (a) Spring tide range, and ¯ood and ebb duration as a function of distance from the estuary mouth. Amplitude of the important (b) di- urnal (c) semi-diurnal and (d) quar- ter-diurnal tidal constituents as a function of distance from the estu- ary mouth. reaches a maximum of 80% at 32 km from the al effects landward of 60 km from the estuary estuary mouth. In the ¯uvio-tidal reach the mud mouth caused non-linear tide behaviour to de- content of the sub-tidal zone is generally less than velop. The resulting tidal asymmetry is manifest 10%, although some isolated cases reach 40%. by a weaker, longer-duration ebb and a stronger, The presence of gravel size material in the shorter-duration ¯ood (Fig. 3a). samples was rare (Fig. 2a and 2b). Where it The tidal ampli®cation and damping observed occurred, it was composed entirely of shell at the seaward and landward ends of the estuary detritus in the estuarine reach of the river and largely results from the shoaling and frictional entirely of lithic material further landward. damping of the diurnal and semi-diurnal tide

constituents, particularly K1 and M2 (Fig. 3b and 3c). The importance of friction and other non- Tidal water levels and velocities linear effects on the tide as it propagates beyond The water level records indicate that a change in 50 km from the estuary mouth is demonstrated by tide behaviour occurs at c. 50 km from the estuary the strong growth in the quarter-diurnal, shallow mouth. Ampli®cation of the tide range in the water tide constituents (Fig. 3d). The values of direction of tide propagation occurs seaward of two parameters commonly used to characterize this point and damping occurs landward non-linear tide behaviour (see Friedrichs & Au- (Fig. 3a). The maximum spring tide range increas- brey, 1988) are listed in Table 4. The relative es from 1á92 m at the estuary mouth to a maxi- phases of M2 and M4, 2GM2±GM4, measured at mum of 2á10 m at Gunderman (49 km from the Spencer and Gunderman are between 180° and estuary mouth), and then decreases to zero at the 360°. This indicates that the seaward reach of the tidal limit (145 km from the estuary mouth). estuary has tidal asymmetry favouring shorter ebb Hughes (1992) showed that this behaviour is durations, however, the small values for the consistent with that expected for a damped, amplitude ratio of M2 and M4, aM4/aM2, indicates linear progressive wave propagating through a that the degree of this tidal asymmetry is small. channel of exponentially decreasing cross-sec- Landward of 50 km from the estuary mouth the tional area. He also found, however, that friction- degree of tidal distortion grows steadily due to

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 403

Table 4. Indicators of non-linear tide behaviour.

Distance from Station estuary mouth

Name (km) aM2/h aM4/aM2 2GM2±GM4(deg)

Spencer 28 0á033 0á0072 304 Gunderman 49 0á039 0á0045 341 Webbs Creek 61 0á040 0á0137 48 Sackville 97 0á039 0á0494 65 Ebenezer 112 0á037 0á0916 66 Windsor 126 0á053 0á1156 68

Table 5. Summary statistics for peak tidal velocities measured over a spring±neap tide cycle.

Distance from Peak ¯ood velocity (m s)1) Peak ebb velocity (m s)1) estuary mouth Site Number (km) Mean Max Min Mean Max Min.

Sackville* 97á0 0á67 0á79 0á55 0á30 0á42 0á13 Webbs Cr* 64á0 0á55 0á70 0á39 0á28 0á42 0á12 HR3 58á8 0á48 0á60 0á40 0á34 0á56 0á16 HR5 55á0 0á39 0á55 0á35 0á43 0á68 0á25 HR6 51á5 0á39 0á52 0á34 0á22 0á42 0á11 HR8 49á5 0á36 0á43 0á29 0á42 0á50 0á34 HR9 42á0 0á25 0á31 0á19 0á48 0á68 0á27 HR10 39á0 0á35 0á41 0á23 0á44 0á62 0á25 HR11 34á5 0á40 0á51 0á26 0á41 0á58 0á19 HR19 21á5 0á51 0á75 0á33 0á37 0á53 0á23

*Unpublished data provided by NSW Public Works Department's Manly Hydraulics Laboratory.

Fig. 4. Longitudinal salinity pro- ®le during (a) neaps and (b) springs. Results from the surface, mid depth and 1 m above the bed are indicated by a dashed, thin and thick line, respectively.

the combined effect of a growing aM4 and a velocities are signi®cantly stronger and of shorter dampening aM2 (Table 4; Fig. 3c and 3d). More- duration than ebb velocities. over, as the values of 2GM2±GM4 lie between 0° and 180° the sense of this distortion switches to Salinity favour shorter ¯ood duration. Summary statistics of the peak tidal velocities The longitudinal salinity pro®les surveyed once are consistent with the tide behaviour inferred during neaps and once during springs are shown from the water level data (Table 5). In the ®rst in Fig. 4. The total fresh water discharge rate into 50 km from the estuary mouth ebb velocities are the estuary from all major tributaries amounted to marginally stronger and of shorter duration than 6á4 m3 s)1 and 3á4 m3 s)1 for the neap and spring ¯ood velocities, whereas further landward ¯ood survey, respectively; thus the data is indicative of

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 404 M. G. Hughes et al.

Fig. 5. Semi-diurnal salinity range during springs as a Fig. 6. Depth-averaged longitudinal turbidity pro®le function of distance from the estuary mouth. measured during springs (solid line) and neaps (dashed line). Squares with error bars show means and standard errors of the daily averaged turbidities measured at the instrument deployment sites. The turbidity time series low fresh water ¯ow conditions. On both occa- at each site displayed a high degree of autocorrelation, sions the water column was well mixed with relating to tidal periodicity; thus the following proce- differences in salinity between surface and bot- dure was employed to obtain a satisfactory measure of tom waters typically less than 1 p.p.t. The limit of the mean and standard error: a 24 h 50 min running mean was calculated for the time series from each de- salt intrusion into the estuary was 73 and 75 km ployment site, which was then sub-sampled at noon of from the estuary mouth for neaps and springs, each day to provide a daily averaged turbidity. respectively. Intra-tidal variation in salinity at a ®xed point in an estuary arises largely from advection of the The daily averaged turbidity at Sites HR11 and longitudinal salinity pro®le by the ¯ood and ebb HR10 (34á5 and 39 km from the estuary mouth) is of the tide. Maximum values of the semi-diurnal statistically signi®cantly higher than sites located salinity range recorded during the instrument either seaward or landward (Fig. 6). This indi- deployments is shown in Fig. 5. The maximum cates that the TM zone identi®ed in the one-off range observed was 10 p.p.t. at Site HR10 (39 km longitudinal surveys is a persistent feature in this from the estuary mouth). part of the estuary.

Longitudinal turbidity pro®le Turbidity±velocity patterns The depth-averaged longitudinal turbidity pro®le Three-day records of turbidity and velocity dur- was surveyed once during neaps and once during ing springs for each of the instrument deployment springs (Fig. 6). During springs a clear maximum sites are shown in Fig. 7. During the ¯ooding tide in the depth-averaged turbidity, reaching 90 FTU, beginning late 4 June and continuing into 5 June, was measured at a distance of 33 km from the the turbidity at Site HR3 increased from a estuary mouth. The near bed turbidity at the same background value of 5 FTU to reach a maximum location was 228 FTU. During neaps the turbidity of 15 FTU (Fig. 7a). The peak turbidity occurred pro®le was ¯at and featureless (Fig. 6). when the velocity reached its maximum and an The turbidity time series at each long-term elevated turbidity was sustained through the instrument deployment site displayed a high remainder of the ¯ood. Shortly before high water degree of autocorrelation, relating to tidal period- slack the turbidity decreased, but only to 10 FTU. icity, thus the following procedure was employed The largest ebb of the day followed the largest to obtain a satisfactory measure of the mean and ¯ood and produced a narrow peak in turbidity standard error of turbidity: a 24 h 50 min running that occurred at the time of maximum ebb mean was calculated for the time series from each velocity. Unlike the ¯ood, the turbidity level deployment site, which was then sub-sampled at during the ebb was not sustained and fell rapidly noon of each day to provide a daily averaged to the background value at low water slack turbidity. The means and standard errors of the (Fig. 7a). The second ¯ood of the day was daily averaged turbidities obtained from the marginally smaller than the ®rst. The peak instrument deployment sites are consistent with turbidity again reached 15 FTU but, in contrast the results of the longitudinal surveys (Fig. 6). to the ®rst ¯ood, this value was not sustained and

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 405

Fig. 7. Simultaneous time series of velocity (thick line) and turbidity (thin line) measured during springs at Sites (a) HR3 (b) HR5 (c) HR6 (d) HR8 (e) HR9 (f) HR10 (g) HR11 and (h) HR19. Positive velocities are in the ¯ood direction and negative ve- locities are in the ebb direction. fell rapidly to the background value at high water ®rst is that the turbidity level at this site is slack. The following ebb was the smallest of the relatively high (exceeding 50 FTU), probably day and was not capable of entraining bed caused by the larger amount of ®ne grained sediment to an elevation of 1 m above the bed. sediment available for suspension in this section This pattern, of strong ¯ood currents associated of the estuary. The second difference is that the with high and sustained turbidity levels but weak peak turbidity relating to the largest ebb of the day ebb currents associated with low turbidity levels, was roughly equal to or greater than the turbidity is repeated on successive tides (Fig. 7a). associated with the largest ¯ood of the day. The turbidity±velocity pattern at Site HR5 is Notice, however, that high turbidities were not almost the same as that described above for Site sustained during the ebb to the same degree that HR3, but in this case there was elevated turbidity they were during the ¯ood. during the second ebb of the day (Fig. 7b). Note The turbidity±velocity pattern at Site HR8 is that the ebb velocities at Site HR5 are greater than very similar to Site HR5 (Fig. 7d), except that the those at Site HR3 (Fig. 7b). turbidity level at this site was generally higher Turbidity±velocity patterns at Site HR6 and there was a clear peak in turbidity with every (Fig. 7c) differ from Site HR3 in two ways: the ¯ood and ebb tide. Note that for comparison

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 406 M. G. Hughes et al. purposes the largest ¯ood velocities occur on the lation and the location of the TM zone should be morning of 12 December at Sites HR8 to HR11 some tens of kilometres further landward of (Fig. 7d,e,f and g). where it is actually found in the Hawkesbury At Site HR9 the turbidity during the largest River estuary. The gravitational circulation model ¯ood of the day increased from the background would place the TM zone at a null point located value gradually with increasing velocity, reaching near the limit of salt intrusion and for most of the a peak at the time of maximum velocity (Fig. 7e). time this occurs between 60 and 80 km from the The turbidity level was sustained during the estuary mouth. The tidal distortion model would remainder of the ¯ood, with only a small decrease place the TM zone at a null point located in turbidity at high water slack. In contrast, somewhere landward of the point where non- during the ebb the turbidity quickly reached a linear tide behaviour and velocity asymmetry ®rst maximum together with the velocity, then for the become apparent; this would be somewhere remainder of the ebb the turbidity decreased landward of 50 km from the estuary mouth. steadily to the background value at low water It is interesting to note that the location of the slack (Fig. 7e). This pattern was repeated nearly TM zone coincides with the largest semi-diurnal exactly every semi-diurnal tide cycle (thus there salinity range, which is consistent with observa- was little diurnal variation at this site). tions reported by Althausen & Kjerfve (1992) in The pattern in turbidity±velocity at Sites HR10 the micro-tidal Cooper River estuary. These and HR11 is very similar to Site HR6 (Fig. 7f and authors proposed that ¯occulation/de¯occulation 7 g) except that turbidity levels during the largest processes, in association with the large salinity ebb were not equivalent to those on the largest range, might be important in producing a TM ¯ood. At Site HR19 (Fig. 7h) turbidity increased zone. This explanation cannot be applied to the rapidly on the ¯ood, reaching a peak before the Hawkesbury River, however, since salt-induced maximum velocity was attained, then it de- ¯occulation is complete at 8 ppt or less (Eisma, creased steadily to the background value at high 1993) and the salinity in the TM zone is 15±25 water slack. In contrast, during the ebb the p.p.t. What process, then, has given rise to the turbidity increased slowly from the background Hawkesbury TM zone? value to reach its peak after the velocity reached While it seems clear that the location of the TM its maximum, then the turbidity dropped to the zone is linked closely to the focus of net mud background value again at low water slack. This accumulation through tidal resuspension pro- pattern is almost the reverse of that described for cesses, the mechanism responsible for focusing Site HR9 (Fig. 7e; see above). mud accumulation in the ®rst place is still uncertain. Probably no process acts to focus net mud accumulation during periods of extended DISCUSSION low ¯uvial ¯ow conditions. During these periods the estuary is well mixed so that vertical gravita- A TM zone exists in the micro-tidal Hawkesbury tional circulation will be either weak or com- River estuary and it is persistently located be- pletely absent. Moreover, the supply of sediment tween 30 and 40 km from the estuary mouth. The to the head of the estuary during these periods is TM coincides with an inter-tidal and sub-tidal also very low; values of SPM concentration in the mud deposition zone, located between 20 and ¯uvio-tidal reach are typically less than 5 mg l)1 50 km from the estuary mouth, which has expe- (Hooper & Humphreys, 1993). It is therefore rienced signi®cant shoaling (i.e. net mud accu- proposed that the mechanism for focusing mud mulation) in the past 60 years (Gardiner, 1993). accumulation in the TM zone operates during Two points of discussion arise from these obser- infrequent, moderate to large ¯uvial discharge vations. The ®rst concerns the processes respon- events. sible for focusing mud accumulation and creating During such events the estuary becomes par- a TM zone (i.e. what is the origin of the TM tially strati®ed, gravitational circulation is well zone?). Secondly, what are the tidal resuspension developed and the saline intrusion limit is processes responsible for maintenance of the TM pushed tens of kilometres seaward from its usual zone over time? location (Fig. 1; Wolanski & Collis, 1976; Kjerfve et al., 1992; Hughes & Callaghan, 1995). More- Origin of the TM zone over, substantial amounts of ®ne suspended The existing models for meso- and macro-tidal sediment are delivered to the head of the estuary estuaries predict that the focus for mud accumu- during these events. Williams et al. (1993) report

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 407 a suspended sediment load of 210 000 t delivered suspension. Entrainment is complicated by a to the estuary during a single ¯ood in August number of factors peculiar to cohesive sediments. 1991. No ®gures are available for the annual load Mehta et al. (1989) describe three modes of of suspended sediment delivered to the estuary, cohesive sediment entrainment: surface erosion, but the typical situation of extended low ¯ow mass erosion and resuspension of a stationary conditions together with very low SPM concen- suspension (or ¯uid mud). Given the fact that trations in the river suggests that signi®cant SPM concentrations in the Hawkesbury River sediment supply to the estuary is restricted to never reach values suf®cient to develop ¯uid low-frequency, moderate to large-magnitude, dis- muds, the latter mechanism cannot be important. charge events. In the cases of surface erosion and mass erosion We hypothesize that it is the gravitational the threshold of entrainment for the sediment is circulation established in the estuary during strongly dependent on the character of the bed as moderate to large discharge events that serves to a whole, in particular its water content and focus mud accumulation at the TM zone. The compaction (Mehta et al., 1986). Mud deposits fresh water discharge rate required to set up a de-water and compact with time until the excess gravitational circulation suitable for focusing shear stress exerted by the current is suf®cient to mud accumulation at c. 35±40 km from the re-initiate erosion. estuary mouth is unknown, but Wolanski & Collis The more `aged' (de-watered and compacted) (1976) present data that shows the near bed salt the bed is the more sustained the current shear intrusion limit was pushed seaward to a position needs to be before erosion and resuspension of 38 km from the estuary mouth when the fresh wa- the sediment is initiated. The unsteady nature of ter discharge into the estuary reached 73 m3 s)1. tidal currents and the variable resistance of `aged' Although this discharge rate is not substantially beds results in phenomena known as threshold greater than the mean, for nearly 90% of the time and erosion lags (Dyer, 1994), which serve to river discharge is signi®cantly less than this value produce a phase shift between turbidity and (Table 2). In this context therefore it represents a velocity that will vary between sites in an estuary signi®cant perturbation from the norm. and with diurnal and spring±neap variations in Given that the hypothesized mechanism for tidal energy. The fact that virtually no sediment is focusing mud accumulation in the Hawkesbury suspended in the estuary during neaps (Fig. 6) River estuary is inoperative for most of the time suggests that the bed of the estuary becomes and the fresh water discharge into the estuary is `aged' to some degree, at least fortnightly. Fur- generally episodic (aseasonal), there is not ex- thermore, `ageing' of the bed probably takes place pected to be any seasonal variability in the TM also on a semi-diurnal time scale, since low tidal zone like that described for the Tamar River current speeds mostly correspond with low tur- estuary (Uncles et al., 1994). Indeed, it appears bidity levels (Fig. 7). that in the Hawkesbury River the position of the This `ageing' of the bed effects tidal resuspen- TM zone is remarkably stable, migrating no more sion of sediment in the TM zone during springs. than a semi-diurnal tidal excursion length from Here, sediment is mainly eroded on the ¯ood tide, the zone of maximum mud content in the bed when peak velocities are sustained for a period of sediments. The question thus arises ± what time, whereas minimal resuspension occurs dur- prevents the mud being dispersed from the ing the relatively short-duration peak ebb ¯ows ¯ood-determined focus of accumulation by tidal (Fig. 7f and g). This is despite the fact that peak processes operating during the extensive inter- ebb velocities are larger than peak ¯ood velocities vening (low-¯ow) periods? at these sites (Table 4). This may be considered as a type of erosion lag process (terminology of Dyer, 1994). An important consequence of this is that Maintenance of the TM zone downstream and within the TM zone the net Although patterns in the turbidity time series are suspended sediment transport vector is directed highly repeatable through several tide cycles, landward (see below). there is no direct and consistent relationship The times of maximum turbidity and maximum between the magnitudes of turbidity and water velocity on the ¯ooding tide coincide at the TM velocity (Fig. 7). The reason for this must be due zone sites (Fig. 7f and g), suggesting that the in part to the fact that the measured turbidity at a source of the turbidity is local resuspension. site represents both entrainment of local bed Further landward, however, there is a phase lag sediment and advection of sediment already in between maximum turbidity and velocity that

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 408 M. G. Hughes et al. increases with distance along the channel, sug- velocities are the greatest and peak ¯ood veloci- gesting that suspended sediment is advected ties are the weakest (Table 5). The end result of landward from the TM zone (Fig. 7d and e). Note this turbidity±velocity pattern is that the net SPM that these sites are within one tidal excursion transport vector at this site is directed seaward, length of the TM zone (see below). This explana- creating a barrier to effective transport and tion implies that most of the SPM (turbidity) landward dispersion of the TM zone. observed at sites outside the TM zone represents The combined effect of the erosion lag process advection from the TM zone rather than tidal and local divergence in tidal current velocity resuspension of local bed sediment. A short residuals is summarized in Fig. 8. The following distance upstream of the TM zone (c. 42 km from are shown in order from top to bottom: (1) net the estuary mouth) the peak in turbidity coincides direction of suspended sediment transport (2) with the last portion of the ¯ood tide during tidal excursion length (3) daily averaged turbidity springs (Fig. 7e). Turbidity remains relatively (4) percentage of mud in bottom sediments (5) high during high water slack and a second maximum ¯ood directed tidal current velocity turbidity peak coincides with the ®rst portion of and (6) maximum ebb directed tidal current the ebb (Fig. 7e). This site is also where peak ebb velocity. (1) was calculated from the product of the velocity and turbidity records integrated over the duration of a spring±neap tide cycle and (2) was determined from the semi-diurnal salinity range and local gradient of the longitudinal salinity pro®le. (3) to (6) represent data already shown in previous ®gures and tables. The locality of the TM zone is maintained by net landward transport of suspended sediment in the lower estuary (an erosion lag mechanism) coupled with net seaward transport in the middle estuary where there is a local divergence in tidal velocity residuals. Together, these processes serve to trap the TM zone to well within one tidal excursion length of the ¯uvial ¯ood determined point of maximum mud accumulation in bottom sedi- ments.

SUMMARY

The origin of the TM zone in the micro-tidal Hawkesbury River estuary cannot be explained purely by either of the two available theories: gravitational circulation and tidal asymmetry. The predicted location of the TM zone, based on these two theories, is some tens of kilometres landward of the actual location for probably 90% Fig. 8. Summary diagram showing daily averaged tur- of the time. It is hypothesized here that the bidity, cross-sectionally averaged percentage mud in location of the TM zone is controlled largely by bottom sediments, maximum ¯ood tidal current speed and maximum ebb tidal current speed, relative to dis- episodic ¯uvial ¯ood events, which shift the near tance from the estuary mouth. The approximate loca- bed salt intrusion limit to about 35 km seaward of tion of the TM zone, tidal excursion length and net its normal (low ¯uvial ¯ow) position and cause direction of suspended sediment transport at each in- gravitational circulation within the estuary. Mud strument deployment site are indicated. The tidal ex- is deposited at this location during the ¯uvial cursion length was estimated from the semi-diurnal ¯ood and at other times it is resuspended by tidal salinity range and local gradient of the longitudinal salinity pro®le. The net sediment transport directions processes. The data presented here shows that were derived from the product of the turbidity and during the long intervening periods between velocity time series integrated over the duration of a ¯uvial ¯ood events the location of the TM zone spring±neap tide cycle. is trapped at its ¯ood-determined position. This

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 Dynamics of turbidity maximum zone in a microtidal estuary 409 appears to be caused by the coexistence of an Bartholdy, J. (1984) Transport of suspended matter in a erosion lag process, which favours landward bar-built Danish estuary. Est. Coast. Shelf Sci., 18, transport of SPM, and a local divergence in tidal 527±541. velocity residuals, which favours seaward trans- Dyer, K.R. (1986) Coastal and Estuarine Sediment Dy- namics. John Wiley & Sons, Chichester. port of SPM. This trapping mechanism prevents Dyer, K.R. (1994) Estuarine sediment transport and mud from being transported to the modal null deposition. In: Sediment Transport and Depositional point location (which is related to low rather than Processes (ed. K. Pye), pp. 193±218. Blackwell Sci- average ¯uvial ¯ow conditions in this case), enti®c Publications, Oxford. where we might normally expect the TM zone Einstein, A.J. and Krone, R.B. (1962) Experiments to to be located. It appears that the TM zone in this determine modes of cohesive sediment transport in micro-tidal estuary is not nearly as mobile as its saltwater. J. Geophys. Res., 64, 1451±1461. Eisma, D. (1993) Suspended Matter in the Aquatic En- meso- and macro-tidal counterparts. vironment. Springer-Verlag, Berlin. There remain several important issues that Foreman, M.G.G. (1977) Manual for Tidal Heights require further investigation. The mechanics of Analysis and Prediction. Paci®c Marine Science, the hypothesized gravitational circulation during Report no. 77±10. moderate±large discharge events needs to be Friedrichs, C.T. and Aubrey, D.G. (1988) Non-linear established by ®eld observation and the modal tidal distortion in shallow well-mixed estuaries: a position of the ¯ood-determined salt intrusion synthesis. Est. Coast. Shelf Sci., 27, 521±545. Gardiner, B. (1993) Hawkesbury-Nepean River Sedi- limit (null point) needs to be veri®ed. Further- ment Dynamics Study: Sediment Accumulation and more, the relative amounts of SPM derived from Sediment Cycling in Two Estuarine Depositional local resuspension and advection need to be Environments in the Foul Weather Reach and Spen- determined for sites located in the TM zone (e.g. cer Areas. Water Board, AWT Science and Environ- Sites HR10 and HR11) and for sites located ment, Report no. 93/133. immediately landward (e.g. Site HR9), in order to Grabemann, I. and Krause, G. (1989) Transport pro- establish the importance of erosion lag (`aging' of cesses of suspended matter derived from time series in a tidal estuary. J. Geophys. Res., 94, 14373± the bed) as a mechanism for trapping the TM zone. 14379. Hooper, C. and Humphreys, S. (1993) Water Quality in the Hawkesbury-Nepean River and its Tributaries: ACKNOWLEDGMENTS Maldon Weir to Flint and Steel Point July 1991 to June 1992. Water Board, AWT Science and Envi- Financial support and permission to publish data ronment, Report no. 92/56. was granted by Water Resources Planning Branch, Hubble, T.C.T. and Harris, P.T. (1994) Hawkesbury- Sydney Water Board through their Hawkesbury- Nepean River Sediment Dynamics Mapping Study. Nepean River Sediment Dynamics Study. David University of Sydney, Ocean Sciences Institute, Re- port no. 53. Brown (Manly Hydraulics Laboratory) provided Hughes, M.G. (1992) Preliminary Investigation of Tidal the tidal velocity data from Sackville and Webbs Dynamics in the Hawkesbury River and Suspended Creek and Kerryn Stephens and Simon Williams Sediment Transport Near Gunderman. Water Board, provided the river discharge data from Penrith Scienti®c Services, Report no. 92/12. weir. Special thanks to John Watkins and Ro- Hughes, M.G. and Callaghan, R.L. (1992) Hawkesbury- chelle Callaghan who assisted during all of the Nepean River Sediment Dynamics Study: long hours spent in the ®eld ± their companion- Hydrographic and Sediment Transport Data From to Gunderman. Water Board, AWT ship and good humour contributed greatly to the Science and Environment, Report no. 92/85. successful completion of the ®eld programme. Hughes, M.G. and Callaghan, R.L. (1993) Hawkesbury- Nepean River Sediment Dynamics Study: Hydro- graphic and Sediment Transport Data From REFERENCES Gunderman to Spencer. Water Board, AWT Science and Environment, Report no. 93/36. Allen, G.P. and Castaing, P. (1973) Suspended sedi- Hughes, M.G. and Callaghan, R.L. (1994) Hawkesbury- ment transport from the Gironde estuary (France) Nepean River Sediment Dynamics Study: Hydro- into the adjacent continental shelf. Mar. Geol., 14, graphic Data From Selected Channel Cross-Sections 47±53. Between Wisemans Ferry and Spencer. Water Board, Althausen, J.D. and Kjerfve, B. (1992) Distribution of AWT Science and Environment, Report no. 94/85. suspended sediment in a partially mixed estuary, Hughes, M.G. and Callaghan, R.L. (1995) Sediment Charleston Harbor, South Carolina, U.S.A. Est. Coast. Dynamics in the Hawkesbury River: Final Report. Shelf Sci., 35, 517±531. Water Board, AWT EnSight, Report no. 94/94.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410 410 M. G. Hughes et al.

Kirby, R. (1988) High concentration suspension (¯uid Roy, P.S. (1983) Quaternary Geology. In: Geology of the mud) layers in estuaries. In: Physical Processes in Sydney 1: 100 000 Sheet 9130 (ed. C. Herbert), Estuaries (eds J. Dronkers and W. van Leussen), pp. pp. 41±91. Geological Survey of New South Wales. 463±487. Springer-Verlag, Berlin. Turner, A., Millward, G.E. and Tyler, A.O. (1994) The Kjerfve, B., Seim, H.E., Blumberg, A.F. and Wright, L.D. distribution and chemical composition of particles in (1992) Modelling of the residual circulation in Bro- a macrotidal estuary. Est. Coast. Shelf Sci., 38, 1±17. ken Bay and the lower Hawkesbury River, NSW. Uncles, R.J. and Stephens, J.A. (1993) Nature of the Aust. J. Mar. Freshw. Res., 43, 1339±1357. turbidity maximum in the Tamar estuary, U.K. Est. Mehta, A.J., Hayter, E.J., Parker, W.R., Krone, R.B. and Coast. Shelf Sci., 36, 413±431. Teeter, A.M. (1989) Cohesive sediment transport 1: Uncles, R.J., Barton, M.L. and Stephens, J.A. (1994) Process description. J. Hydraulic Eng., 115, 1076± Seasonal variability of ®ne-sediment concentrations 1093. in the turbidity maximum region of the Tamar estu- Morris, A.W., Mantoura, R.F.C., Bale, A.J. and How- ary. Est. Coast. Shelf Sci., 38, 19±40. land, R.J.M. (1978) Very low salinity regions of es- Williams, S., Badcock, C. and Roberts, E. (1993) Water tuaries: Important sites for chemical and biological Quality in the Hawkesbury-Nepean River and Trib- reactions. Nature, 274, 678±680. utaries, 1977±92. Water Board, AWT Science and Morris, A.W., Bale, A.J., Howland, R.J.M. et al. (1986) Environment, Report no. 93/20. Sediment mobility and its contribution to trace metal Wolanski, E. and Collis, P. (1976) Aspects of aquatic cycling and retention in a macrotidal estuary. Water ecology of the Hawkesbury River 1: Hydrodynamical Sci. Technol., 18, 111±119. processes. Aust. J. Mar. Freshw. Res., 27, 565±582. Nichols, M.M. (1993) Response of coastal plain estu- Wolanski, E., King, B. and Galloway, D. (1995) Dy- aries to episodic events in the Chesapeake Bay re- namics of the turbidity maximum in the Fly River gion. In: Nearshore and Estuarine Cohesive Sediment estuary, Papua New Guinea. Est. Coast. Shelf Sci., 40, Transport, Coastal and Estuarine Studies 42 (ed. A. J. 321±337. Mehta), pp. 1±20. American Geophysical Union, Washington, DC. Roy, P.S., Thom, B.G. and Wright, L.D. (1980) Holocene sequences on an embayed high energy coast: An Manuscript received 22 July 1996; revision accepted 29 evolutionary model. Sed. Geol., 26, 1±19. July 1997.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 397±410