This research was developed with Funds of the Andalousian Gouvernement (Excellence Project RNM-03093 of the Andalousian Plan of Investigation). This is the contribution nº 64 from the CEIMAR Journal Series.We strongly acknowledge the help of J. Rangel, M. González and specially J. Galvez during the field surveys. The bedload traps could not be built without the special intervention of Dr. J. Guillen from the Institut de Ciències del Mar, CSIC (Barcelone).
1The Guadiana Estuary consists of a narrow channel, which presents a meandering morphology imposed by the hard geology of the substrate (Lobo et al., 2004; Morales et al., 2006). Nevertheless, the last kilometers of its channel, the valley becomes wider in coincidence with the presence of Cenozoic Guadalquivir Basin formations (Morales, 1993). In this area, the Guadiana develops a prograding system constituted by successive sandy barriers separated by salt marshes, which configure a wave-dominated delta (Morales, 1997). Its mouth is affected by a mesotidal regime, with tidal ranges comprised between 3 and 1 m (spring vs neap tides), where the mean tidal range is 2.1 meters and waves with a mean significant height of 0.5 meters (Morales et al., 1994).
2The marine sector of the estuarine channel consists of successive meanders, where the curved sectors present a section characterized by a pool in the concave margin and a lateral tidal bar in the convexe one. Between successive curved tracks, the inflection points present a symmetric profile showing a bed with an intermediate depth and without pools and bars (Morales, 1997).
3The present work studies a cross section of the channel located 6.5 km from the outer point of the estuary (fig. 1). This profile presents an asymmetric morphology with a deeper zone in the eastern margin and a flat bar in the western one (fig. 2). The tidal sand transport was measured across a transversal profile of this pool-and-bar system.
Fig. 1 – Location of the study area on the lower sector of the Guadiana estuary.
Fig. 1 – Localisation du secteur d'étude dans la partie inférieure de l'estuaire du Guadiana.
Fig. 2 – Flow section with indication of the measure station (ADCP and Sediment trap).
Fig. 2 – Section d'écoulement avec indication des points de mesure (ADCP et pièges à sédiment).
4To study the sedimentary beds of this bar located in the Guadiana Estuary mouth two marine geophysics surveys were done. The first survey was developed in July 2008. The aim of this campaign was the elaboration of a mosaic of images of Side Scan Sonar. The used equipment was the CMAX CM2000, employing the High Frequency mode (325 kHZ) and a work range of 50 meters per band. During the post-processing phase, a geo-referenced mosaic of acoustic images was displayed using ARCVIEW 3.2.
5The second survey was carried out between the last week of June and the two first weeks of July 2010 and the measurements were developed during two opposite tidal conditions (fig. 2). In a first phase, an acoustic scan of the estuarine channel allowed to get a detailed bathymetry. On the other hand, 13 surficial sediment samples were taken using a Van Veen drag.
6The bathymetry was done using a mono-beam echo sound Valeport Midas Surveyor, with a transducer of 210 kHz. The records were geo-referenced with a 12 channel dGPS EGNOS, to build accurate files XYZ. The depth was corrected using a tidal curve from the harbor of Ayamonte, the nearest to the studied area. The data were processed by means a triangulation interpolation from the points XYZ using the module 3D Terrain model from the program Hypack 2009 (Hidrographic Survey Software).
7The sediment samples were conserved in plastic bags and analyzed at the laboratory by means a Laser Difractometer Malvern Mastersizer 2000.
8The tidal currents at the surveyed zone were measured by an ADCP (Acoustic Doppler Currents Profiler) deployed at the deeper part of the surveyed section during two different tidal cycles: the first during a Mean Neap Tide (5ft of July) and the second during a Mean Spring Tide (15th of July).
9The potential bedload transport (Qb) was calculated using the equation suggested by Bagnold (1963):
Qb = (K1 δs U*3)/dg
10Coefficient K1 (in g s2 cm−4) was related to factors such as the drag coefficient and fluid density suggested that K1 is related to the grain size by Hardisty et al. (1983), but his equation to determine this factor was contested by Y.P. Wang and S. Gao (2001). These authors reanalyzed the data of for 0.19, 0.28, 0.45 and 0.93 mm sand, concluding that:
K1 = 0.10exp(0.17/D)
D is the percentile 50 in the grain size curve (0.47 mm)
D10 is the percentile 10 in the grain size curve (0.84 mm)
d is an adimensional parameter which put in relation the density of the sediment with the density of the water:
d = (δs-δ)/ δ
δs is the density of the sediment (2.53 g/cm3)
δ is the density of the water (1.03 g/cm3)
g is the value of the gravity acceleration (981 cm/s2)
U* is obtained from the measured Near Bottom Velocity for each interval of 5 minutes using the equation:
U* = (0.4 UZ )/ Ln (30Z/3D10)
Z is the distance from the current measurement to the bottom (blanking distance = 0.50 m).
11This equation was applied in intervals of 5 minutes, calculating and applying the average U* for each interval. The values of instantaneous Qb obtained for each interval were finally integrated to obtain the total Qb for the Ebb and Flood semicycles during both spring and neap tides.
12The values of Qb calculated using this formula were compared with measures of real bedload transport (Sb) obtained by using Poliakoff sediment traps (Shingal et al., 1981) which were installed at the bed of the estuary at same time and place where the ADCP was measuring the tidal currents. The first trap was installed the 5ft of July (neap tide) at 12:00 in the noon, just in the flood-to-ebb slack moment and extracted at 16:50 in the afternoon, so this trap was working all the ebb semicycle. The second trap was installed in the same place and day at 16:35, just before the ebb-to-flood slack and taken at 20:10 15th after the high tide; so this trap was working during the entire flood semicycle. A third trap was installed the 15th of July (spring tide) at 8:05 in the morning (flood-to ebb slack) and taken at 14:00 in the afternoon; this trap was working during the ebb semicycle. Finally, a fourth trap was installed at 13:50 (ebb-to flood slack) to be recovered at 19:50 in the evening, and work the entire flood semicycle.
13The isobaths distribution (fig. 3) show the presence of a deep channel in the eastern part of the study area (Spanish margin) which reach more than 6 meters deep under the Extreme Equinox Low Water level, whereas in the western area a near horizontal lateral tidal bar is developed between 2 and 4 meters deep. This bar is separated from the intertidal areas by a high slope step of 2 meters.
Fig. 3 – Bathymetric scheme of the studied area.
Fig. 3 – Bathymétrie du secteur d'étude.
A: View map; B: 3D view.
A :Vue en plan ; B : Vue en 3D.
14In the southern part of the study area the deep channel migrates to the west to be installed in the Portuguese margin, at same time another lateral tidal bar is installed in the Spanish margin.
15The transversal profile in which the currents and real bedload transport have been measured (fig. 2) is located in the southern zone of the inner bar (fig. 3), so the deeper channel is located at the East, just in front of the dock of Ayamonte.
16The grain size distribution was determined for the 13 samples taken in this estuarine area. In each sample some statistic parameters were determined (tab. 1): C (maximum percentile), M (median), σ (standard deviation), So1 (sorting), SK1 (skewness) and Kg (kurtosis).
Tab. 1 – Grain size parameters according to R.L. Folk and W.C. Ward (1957).
Tab. 1 – Paramètres granulométriques selon R.L. Folk (1972).
Sediment samples
|
Grain size parameters
|
C
|
M
|
σ
|
So1
|
SK1
|
Kg
|
AYN-1
|
-1,00
|
5,20
|
2,99
|
4,25
|
-0,26
|
0,67
|
AYN-2
|
-2,90
|
3,20
|
2,92
|
4,45
|
0,13
|
0,76
|
AYN-3
|
-1,00
|
1,00
|
0,79
|
1,20
|
-0,11
|
0,82
|
AYN-4
|
-1,00
|
1,10
|
0,78
|
1,25
|
0,14
|
0,85
|
AYN-5
|
-1,50
|
1,40
|
2,23
|
3,55
|
0,60
|
2,24
|
AYN-6
|
-2,00
|
3,00
|
2,93
|
4,40
|
0,21
|
0,78
|
AYN-7
|
-1,00
|
1,20
|
0,84
|
1,45
|
0,06
|
1,08
|
AYN-8
|
-1,00
|
1,30
|
1,08
|
2,25
|
0,15
|
1,68
|
AYN-9
|
-1,50
|
1,50
|
2,15
|
3,70
|
0,53
|
1,52
|
AYN-10
|
-1,00
|
1,50
|
0,87
|
1,40
|
-0,06
|
1,15
|
AYN-11
|
-2,00
|
0,70
|
0,69
|
1,20
|
0,12
|
1,23
|
AYN-12
|
-1,50
|
0,80
|
0,68
|
1,00
|
0,20
|
0,82
|
AYN-13
|
-1,50
|
0,80
|
0,60
|
0,90
|
0,23
|
0,74
|
C: centil; M: median; σ: graphic standard deviation; So1: Sorting; SK1: Skewness; Kg: Kurtosis.
C : centile ; M : médiane ; σ : déviation standard ; So1 Sorting ; SK1 : Skewness ; Kg : Kurtosis.
17Four types of beds have been distinguished in the studied area from the point of view of the grain size (fig. 4 and fig. 5). The first class of sediment (fig. 4A) is characterized by a mean grain size of coarse sand, but with an important population of medium sand, a moderate sorting, a mesokurtic shape and a lightly positive skewness (with tail towards the fine). The second class (fig. 4B) is characterized by a mean grain size of medium sand, but with abundant populations of coarse and fine sands, present also a moderate sorting and a positively skewed distribution (with tail towards the fine), but with a leptokurtic shape. The third class (fig. 4C) is a transition between the classes 1 and 2, similar to the second class except by the more abundance of fine than coarse sands as secondary population. The fourth class of sediment (fig. 4D) is completely different and presents a very high dispersion, an extremely poor sorting, a negative skewness (tail towards the coarse) and a platykurtic (plane) shape.
Fig. 4 – Types of grain size distribution.
Fig. 4 – Types de distribution granulométrique.
A: First class; B: Second class; C: Third class; D: Fourth class.
A :Première classe ; B : Deuxième classe ; C : Troisième classe ; D : Quatrième classe.
18The map of mean grain size distribution (fig. 5) shows a band of fine sands (class 2) separating the very fine sands (class 4) located on the northern area from the medium sands distributed on the lateral bar surface (class 3), these fine sands are also distributed in the bottom of the northeastern deep channel surrounding the bar. The coarse sands (class 1) are located covering the bed of the deep channel of the southern area. A band of muddy fine sand (class 4) extends from North-to-South the western margin of the lateral bar.
Fig. 5 – Surficial facies distribution.
Fig. 5 – Répartition des faciès granulométriques de surface.
Stars show the sediment sample position.
Les étoiles indiquent la position des échantillons.
19Four types of beds are distinguished in the studied area from the point of view of the observed bedforms (fig. 6): sands with high plane beds, sands with large dunes, sands with small dunes and ripples and muddy sand with low plane bed.
20Bed without visible bedforms (fig. 7A) located in the marginal shallowest areas, between 0 and 2 meters under the Extreme Equinox Low Water. The cohesive character of the bed, inhibit the possibility to develop migrating bedforms, but erosional features are visible in some tracks (fig. 7B).
Fig. 6 – Bedform field distribution and orientation of bedform crests.
Fig. 6 – Figures sédimentaires sur le fond et orientation des crêtes.
Fig. 7 – Bed types.
Fig. 7 – Types de lits.
21This bed is developed in zones deeper than 4 meters under the Extreme Equinox Low Water and is characterized to be formed in medium-to-coarse moderately sorted sands. The most characteristic feature is the absence of large bedforms presenting a high regime plane bed (fig. 7C).
22This type of bed is located in the low slope surface of the lateral tidal bars, between 2 and 4 meters deep. Near the entire surface is covered by ebb-oriented sand waves with sinuous crests (three-dimensional dunes). Two different bedform trains are present: 1) located between 2 and 3 meters deep, wavelengths comprised between 7 and 15 meters and heights of 0.50 m (fig. 7D) and 2) located between 3 and 4 meters deep in the transition with the high plane bed, wavelengths comprised between 25 and 50 meters and heights of 0.70 m (fig. 7E). The crests of the first set of forms are disposed 40º oblique to the channel margin, whereas the second and deeper set is perpendicular to the channel axis and experiments a rotation with him. In the southern part of the tidal bar, these two different bedform trains interact, forming an interference pattern (fig. 7F).
23Between the low plane cohesive shallow beds and the sands with large dunes, fields of minor bedforms are developed. These fields are characterized by medium sand, and display ebb-oriented sand waves and megaripples with sinuous crests (three-dimensional dunes). Wavelengths of almost 5 m with heights of 0.35 m are the most common dimensions (fig. 7G and fig. 7H).
24Two tidal current profiles were measured in the deepest point of the studied section (fig. 2). The used ADCP allows obtaining a measure of the Near-Bottom current, which is interesting to obtain the current values of the entire water column (fig. 8A and fig. 8B), but in this work we specially work the values of the near bottom current velocities obtained at a distance of 1 meter up the bottom (fig. 9). The first measure was done the 5ft of July, during a cycle of Mean Neap Tide (Coefficient 44 and tidal range 1.48 m) and the second the 15th of July, during a cycle of Mean Spring Tide (Coefficient 91 and tidal range 2.58 m).
Fig. 8 – ADCP profiles.
Fig. 8 – Profils ADCP.
A: Mean Spring tide; B: Mean Neap Tide.
A : Marée de vives-eaux moyenne ; B : Marée de mortes-eaux moyenne.
Fig. 9 – Profiles of near-bottom currents and tide height.
Fig. 9 – Profils des courants de fond et de la hauteur de marée.
A: During a Mean Spring Tide; B: During a Mean Neap Tide.
A : Durant une marée de vives-eaux ; B :Durant une marée de mortes-eaux.
25During the Mean Neap Tide (fig. 9A) the maximum measured ebb current velocity was 70.7 cm/s and the maximum flood was 54.6 cm/s. The ebb current flew during 5 hours and 20 minutes whereas the flood acts during 7 hours.
26During the Mean Spring Tide (fig. 9B) the maximum value of the ebb current was 108.4 cm/s whereas the maximum flood current was 90.3 cm/s. The ebb current flew during 5 hours and 57 minutes and the flood current flew during 5 hours and 47 minutes.
27The values of Near Bottom Velocity of these curves has been used as U(50) (velocity at 50 cm from the bottom) to obtain the values of U* (Current velocity on the interface flow-sediment).
28The potential bedload transport (Qb) was calculated using the equation suggested by Bagnold (1963).
29The obtained values are shown in the table 2. It shows as the values of capability of currents to transport sands are higher during the ebb than during the flood for all tides. Nevertheless, the unbalance is higher during the neap tides, when the ebb potential transport can be 8 times higher than the flood one.
Tab. 2 – Calculated potential bedload (Qb) for complete tidal current cycles.
Tab. 2 – Charge de fond potentielle (Qb) calculée pour des cycles complets de marée.
Qb (gr/cm)
|
Flood
|
Ebb
|
Balance
|
Neap Tide
|
11.31
|
85.95
|
-74.64
|
Spring Tide
|
319.22
|
355.38
|
-36.16
|
Qb (Tons)
|
Flood
|
Ebb
|
Balance
|
Neap Tide
|
0.60
|
4.56
|
-3.96
|
Spring Tide
|
16.92
|
18.83
|
-1.92
|
The total weight of sediment potentially transported was calculated for 530 m of flow section.
Le poids total du sédiment potentiellement transporté est calculé sur une section d'écoulement de 530 m.
30At same time and in the same point the ADCP was measuring the Near Bottom Velocity of the tidal currents a Poliakoff sediment trap (Shingal et al., 1981) was installed. This sediment trap was catching the bedload transported sand during both Ebb and Flood semicycles. The bed-load transport (Sb) can be determined as:
Sb = G/(bT)
In which:
G = dry mass of sediment catch (kg), b = width of intake opening (0.3 m), T = sampling period (s).
31The traps were installed by dives to accurate the correct orientation according the current sense, and were recovered also by dives to avoid the lost of sediment during the process.
32The values of real transport of sands (Sb) are shown in the table 3.
Tab. 3 – Measured bedload transported material (Sb) for complete tidal current cycles.
Tab. 3 – Charge de fond mesurée (Sb) pour des cycles complets de marée.
|
Flood
|
Ebb
|
Balance
|
Neap Tide
|
5.90
|
14.60
|
-8.70
|
Spring Tide
|
69.41
|
72.10
|
-2.69
|
|
Flood
|
Ebb
|
Balance
|
Neap Tide
|
0.31
|
0.77
|
-0.46
|
Spring Tide
|
3.68
|
3.82
|
-0.14
|
33The studied estuarine channel sector is constituted by a lateral tidal in the shallower zones and a bypassing channel in the deep area, after the criteria described by Morales et al. (2006). According to the surficial facies and bedform distribution, the flow regime diminishes from deep to shallow areas. So, the deep bypassing channel displays a high plane bed typical of a high flow regime, whereas the tidal bar presents medium dunes, small dunes, ripples and low plane bed, which indicate a decreasing of the flow regime from medium to low to the shallower areas.
34The upper surface of the lateral tidal bar is near horizontal and covered by dunes with different dimensions and crest orientations. The records of Side Scan Sonar show as the orientation of the bedform crests rotate with the channel margins and appear to follow flow trajectories (fig. 6). The dominant sense observed in the bedform asymmetry is the ebb, nevertheless, flood oriented dunes are also observed. This coexistence of opposite bedforms was previously reported by Lobo et al. (2004), however the origin of the disparity of orientation in the bedforms fields was not explained. The evidences presented in this paper suggest that the observed medium dunes correspond with the dominant flow. So, ebb dunes follow a North-to-South dominant migration, whereas the main part of flood dunes are oriented surrounding the bar. Minor flood forms are also observed in the top surface of the bar, in this case the minor dunes follow secondary trajectories, using the bar front as a flood ramp. At same time, the deeper part of the channel acts as a reversible bypassing channel but presents a dominant ebb transport balance as is observed in the current velocity curves.
35The ebb values of potential bedload transport (Qb) were calculated using the grain size properties of the bed and the near bottom current velocities measured in a section of the bypassing channel during neap and spring tides. Capacity of transport of sediment calculated for the ebb current is higher than the flood one, marking a tidal asymmetry in the ebb sense. This tidal asymmetry in the potential transport is higher during the neap tides.
36The real bedload transport was also measured using Poliakoff type sediment traps during the same moments the current were measured. The values of balanced Sb show a net transport in the seaward sense which is also higher during the neap tides than during the spring tides.
37The values of real bedload transported sediment (Sb) were in all the cases minor than the calculated potential transport (Qb). The rate between real (Sb) and potential (Qb) transport slightly surpass 20%. The absence of cohesive material makes sediments move easy, however the rate Sb/Qb is near 1/5, which indicates a high capacity to transport sands but a lower sand availability.
38The net weight of sediment transported across the flow section reaches 0.46 tons during neap tides and 0.14 during spring tides. This mark a clear deficit situation in the sedimentary supply and a reworking of sands previously deposited in inner areas of the estuary.
39The annual balance of sands towards the sea is 352.59 m3 in the outer estuary. These values are significantly lower than the obtained by Garel and Ferreira (2011) for the balanced potential bedload in the outer area of the estuary and this is interpreted as another evidence of a deficit in the sand availability.
40The recent construction of Alqueva dam, just in the head of the estuary, could be the responsible of a decreasing in the sand availability which diminished the real sediment transport in an 80%, but it is also possible that this relations between potential an real transport values were typical in the estuary before the building of the dam. Regrettably, data obtained before the dam presence are not available.