| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Univ. de Lyon, 69003 Lyon, France, Laboratoire des Sciences de l'Environnement, E.N.T.P.E., rue M. Audin, 69518 Vaulx-en-Velin, France
b Dep. Génie de la Construction, Ecole de Technologie Supérieure, 1100 rue Notre-Dame Ouest, Montréal, QC H3C 1K3, Canada
c Laboratoire de Geophysique Interne et Tectonophysique, UMR CNRS 5559, BP 53, 38041 Grenoble Cedex 9, France, and Laboratoire Régional des Ponts et Chaussées, Boulevard de l'Industrie, B.P. 141, 71404 Autun Cedex, France
d Centre des Sciences de la Terre, UMR CNRS 5561 Biogéosciences, Université de Bourgogne, 6 Bd Gabriel, 21000 Dijon, France
e Dep. des Génies Civil, Géologique et des Mines, Ecole Polytechnique de Montréal, C.P. 6079, Succ. Centre-Ville, Montréal, QC H3C 3A7, Canada
* Corresponding author (goutaland{at}entpe.fr or winiarski{at}entpe.fr).
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Received 8 January 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
Abbreviations: BM, bimodal gravels hydrofacies • GPR, ground penetrating radar • M, massive gravels hydrofacies • OW, open-framework gravels hydrofacies • S, sand hydrofacies • TDR, time domain reflectometry
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
The prediction of these mechanisms requires reliable knowledge of the spatial distribution of hydrogeologic properties. Such knowledge cannot be provided by classical investigation techniques like boreholes (which often have wide spacing, making lateral interpolation difficult) or pumping tests (which provide effective parameters at a scale much larger than the typical length of structures in a heterogeneous aquifer; Beres et al., 1999; Klingbeil et al., 1999; Regli et al., 2002, 2003). In recent years, hydrogeophysics has developed to provide quantitative information about subsurface hydrogeologic parameters or processes that can be used as input for flow and transport models (Hubbard and Rubin, 2005).
A first widely used approach in hydrogeophysics consists of direct hydraulic parameter estimation using geophysical measurements and petrophysical relationships, which relate hydrogeologic properties and geophysical attributes. Hubbard and Rubin (2005) listed the most common geophysical methods used for this approach, depending on the estimation objective (e.g., surface and crosshole electrical resitivity tomography, seismic method, surface and cross-hole GPR). Few studies have been performed, however, to improve characterization of the vadose zone. In particular, estimation of the water content in the vadose zone can be performed using GPR (van Overmeeren et al., 1997; Grote et al., 2003; Huisman et al., 2003; Lunt et al., 2005), electrical resistivity (Banton et al., 1997; Yeh et al., 2002), or a combination of both (Dannowski and Yaramanci, 1999). This hydrogeophysical approach often presents the problem of the uncertain and potentially non-unique relation between geophysical and hydrogeologic properties (Hubbard et al., 1997; Hyndman and Tronicke, 2005).
Hubbard and Rubin (2005) defined a second hydrogeophysical approach known as hydrogeologic mapping, which infers information about the geometry of subsurface units or interfaces from geophysical measurements. Sedimentological studies provide information about the depositional paleosystem, from which the spatial continuity of stratigraphic units and connected hydrogeological properties across a range of spatial scales can be inferred. As highlighted by Fraser and Davis (1998), sedimentological information is an essential key to solve hydrogeologic problems. Sedimentologists can develop conceptual models that can be used to accurately portray the geometry, boundary characteristics, and heterogeneity of aquifer or vadose zone systems. They can provide more accurate, continuous, and detailed delineation of the geometry of internal sedimentary facies and bedding architecture of hydrostratigraphic units. Known or inferred spatial variation of facies parameters, from direct observations or from geophysical measurements, can then be used as a surrogate for defining the spatial distribution of hydrologic parameters. Moreover, the experience and observations of sedimentologists, notably from modern analogs, are useful in calibrating or adjusting a conceptual model to match the field observations and to provide information that can be used in flow and transport models within aquifers and vadose zones. In this study, we used this second mapping hydrogeophysical approach. To define a reliable three-dimensional hydrostratigraphic model, we extrapolated direct and detailed information about subsurface sedimentology and hydrogeology from less invasive and more laterally continuous geophysical data.
Among the geophysical methods used in the aforementioned type of studies, GPR has been extensively used to study alluvial deposits, notably because of its particular suitability for the characterization of highly resistive sediments, which are mostly composed of sands and gravels (Bristow and Jol, 2003). Several examples of GPR measurements on quaternary deposits show the ability of this hydrogeophysical approach to define three-dimensional hydrostratigraphic models (Huggenberger et al., 1994; Asprion and Aigner, 1999; Beres et al., 1999; Heinz and Aigner, 2003b; Lunt et al., 2004; Kostic et al., 2005; Huggenberger and Regli, 2006).
We studied a glaciofluvial deposit underlying a stormwater infiltration basin. Infiltration basins represent alternative techniques to classical drainage systems, which consist in rapid stormwater evacuation via a sewer network in the direction of the hydrologic network, and strongly modify the natural water cycle. Infiltration basins are frequently used if the hydrologic network is too far away and if the underlying subsoils are highly permeable. In urban areas, this system may concentrate diffused pollution contained in urban runoff water and then endanger the quality of subsoils and water resources. The aim of our study was to assess the impact of a stormwater infiltration system on heterogeneous deposits and underlying aquifers (Barraud et al., 2002; Winiarski et al., 2006).
In our study, we followed an approach similar to that described by Bridge and Hyndman (2004) and used by Lunt et al. (2004) to define a three-dimensional model of a braided river deposit—an approach that can be used as a model for aquifer characterization. Our approach consisted in relating GPR reflection features to sedimentary structures of the glaciofluvial deposit under study. This relation allowed interpretation of the radar stratigraphy and definition of a spatial distribution of lithofacies. Lithofacies were then related to corresponding hydrofacies (i.e., homogeneous, isotropic or anisotropic, sedimentological units hydrogeologically relevant; Anderson, 1989) by assigning them hydraulic properties (in this study, saturated hydraulic conductivity and porosity).
Our work was conducted in four steps, as shown in Fig. 1 . First, a sedimentological analysis of the trench walls was performed to describe the glaciofluvial lithofacies, depositional units, and ancient depositional environments corresponding to the studied glaciofluvial deposits. The second step consisted in the calibration of GPR reflections with the sedimentary structures of the trench walls to set up a classification of GPR stratigraphy as a function of lithofacies, depositional units, and system tracts. The third step consisted in a GPR investigation of an area of the infiltration basin for which the sedimentary structure had not been characterized beforehand (i.e., no trench excavation). The classification of GPR stratigraphy was used to interpret GPR reflections. A two-dimensional lithological model was defined. Hydraulic properties were assigned to each lithofacies to propose a hydrostratigraphic model of the glaciofluvial deposit. Finally, this interpretation was compared with water content variations measured using time domain reflectometry (TDR) probes in a measurement well. The probes were placed in a vertical section corresponding to the two-dimensional hydrostratigraphic model to evaluate the compatibility of the hypotheses about hydraulic behavior made from this model with experimental water content measurements.
|
|
Field Site and Installations |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
|
|
|
|
Characterization of the Glaciofluvial Sedimentary Architecture |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
Both lithofacies and depositional units of the studied glaciofluvial deposit were characterized. The lithofacies description was performed using the sedimentological code of Miall (1978) extended by Heinz et al. (2003). It allowed us to describe, in the field, lithological units using the main grain size fraction, component fabric (clast- or matrix-supported), sedimentary structures, and other characteristics (Table 1 ). This description was supplemented by subsequent grain size analysis (NF P94-056 standard, Association Française de Normalisation, 1996) of samples taken on the exposed wall of the trench (86 samples). Changes in lithofacies dip were analyzed to determine the boundaries of depositional units.
|
|
Facies 2 and 3: Massive Gravel and Sand Mixtures Gcm and Gcm,b
These lithofacies are composed of poorly sorted clast-supported massive sands and gravels. The gravel grain size fraction is about 85%. The grain size distribution ranges from fine sands to granules or pebbles. The difference between Gcm and Gcm,b is based on grain size distribution; Gcm has a broader grain size distribution than Gcm,b, whereas Gcm,b grain size distribution is bimodal, with a coarse mode corresponding to the gravel fraction (mainly granules) and a finer moderately well sorted mode corresponding to the sand matrix (mean grain size of about 325 µm). This larger amount of sand matrix in Gcm,b relative to Gcm may be due to a late filling of sand in a gravelly lithofacies (Heinz et al., 2003).
Facies 4: Matrix-Free Lithofacies Gcg,o or Gcx,o
This lithofacies is composed of poorly to moderately well sorted clast-supported gravels without a sandy matrix. Coarser grain size is centimetric and no clay or silt matrices are present. Their thickness ranges from centimeters to decimeters. Their lateral extension ranges from decimeters to meters. Their sedimentary structure is oblique (planar or tangential) or horizontal (base of a gravel sheet).
Matrix-free lithofacies sometimes present a normal grading (Gcg,o). In this case, they are associated with a lower zone composed of Gcm,b lithofacies. This alternating unit has been described in the literature, notably as gravel couplets (Siegenthaler and Huggenberger, 1993; Jussel et al., 1994) or alternating gravel (Heinz et al., 2003). The Gcm,b–Gcg,o association can be the result of different mechanisms. This alternation could be initiated by negative steps in the glaciofluvial outwash (scour pool or gravel dunes), leading to grain size segregation and turbulence for winnowing or infiltration of the sand matrix in the basal zones (Heinz et al., 2003). Planar strata composed of Gcx,o can be formed by the migration of bedload sheets (Lunt et al., 2004).
Depositional Units
Four depositional units were described on Section A. They are named Units 1, 2, 3, and 4, from the base to the top of the trench wall, corresponding to the basin surface (Fig. 6a
).
|
Unit 2 is composed of an alternation of planar Gcg,o and Gcm,b lithofacies. It corresponds to a dip change as Gcg,o lithofacies dip southward within the Gcm,b lithofacies. The Gcg,o beds have an average thickness of 5 cm.
Unit 3 is made up of a thick, lower S-x lithofacies (1 m at its base) and an upper Gcm lithofacies with intercalated Gcx,o lithofacies. Sedimentary structures in the sand evolve from tangential oblique laminations (dipping northward) at the base to climbing ripples at the top of the sandy layer. These sedimentary structures could be due to an overflow from a paleochannel located in a southerly direction. Northward-dipping Gcx,o lithofacies within the upper Gcm lithofacies confirms the hypothesis of a deposit phase identical to that of the sand.
Unit 4, located just beneath the basin surface, has a thickness of about 25 to 30 cm. This unit seems to have been deposited in a high-energy braided river system, according to the wide grain size distribution of its Gcm lithofacies in which Gcx,o beds are intercalated. This unit is disturbed by the roots of vegetation present at the surface of the basin.
Goutaland et al. (2005) provided a sedimentological description of Section B, where three depositional units were identified: a lower prograding unit of Gcm,b and Gcg,o lithofacies alternation below 2 m, an intermediate trough-fill sequence with associated S-x, Gcg,o, and Gcm,b lithofacies between depths of 0.7 and 2 m, and an upper unit composed of lithofacies Gcm with some inserted Gcx,o fine beds and a greater silt grain size at its base.
System Tracts
The sedimentary environment of the studied glaciofluvial deposit can be divided into two system tracts corresponding to the genesis of the deposit. At the base, we observed depositional units linked to a braided-stream paleoenvironment with dominant trough-fill sedimentation. These "cut-and-fill" structures, described by Heinz et al. (2003) in the Rhine valley, are mainly composed of sands and gravels. Lithofacies described are Gcm, Gcm,b, Gcx,o, Gcg,o, and S-x. The base boundary was not detected.
The upper level (Unit 4 in Fig. 6) corresponds to a high-flow system according to the wide grain size distribution of the Gcm lithofacies, in which some Gcx,o beds are inserted. This upper structure is present in both trench walls. This high-energy system could be an allocyclic event, explaining why the upper structure is present on both trench walls and probably beneath the whole area of the basin.
An interpretation of the tangential dipping Gcm,b–Gcg,o alternation (observed in the northern part of Unit 3) was given by Huggenberger and Regli (2006). This alternation developed simultaneously with the overlying gravel sheet (Unit 4). During the depositional process, the abrupt decrease in topology from a gravel sheet into a trough-shaped structure may have produced a flow separation zone, which would have favored sediment sorting and the development of gravel couplets. Tangential sets of alternating Gcm,b–Gcg,o lithofacies may be explained by gravel dune migration across lateral trough walls.
Finally, the upper horizon corresponds to the root zone of the vegetation present on the surface of the infiltration basin. This level presents a greater amount of silt particles likely to increase high-amplitude GPR reflections under dry conditions (van Dam et al., 2002).
|
Relating Radar Stratigraphy to Glaciofluvial Architecture |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
All GPR measurements were performed in the unsaturated zone of the basin (the water table is located 13 m below the surface). We translated two-way travel times (in nanoseconds) to actual depth (in meters) by direct calibration of GPR profiles on trench walls: we correlated characteristic GPR reflections with sedimentary interfaces (Fig. 6). Although such calibrations are time dependant (Lunt et al., 2005), the time period during which GPR surveys and trench excavations were performed (<24 h) was short enough to assume that water content and electromagnetic (EM) wave velocity remained constant during the studies. The evaluation of EM wave velocities ranged from 0.094 to 0.111 m ns–1, with a mean value of 0.102 m ns–1. We used a velocity value of 0.1 m ns–1 to convert the radar profiles, recorded in travel time, into depth profiles. This value is similar to estimated velocities measured by common midpoint analysis within the unsaturated zone of analogous quaternary deposits (Beres et al., 1999; Jakobsen and Overgaard, 2002; Heinz and Aigner, 2003b; Kostic et al., 2005). The time range used in this survey provides good quality data up to 80-ns two-way travel time, allowing an investigation depth >4 m. The highest vertical resolution that could be theoretically achieved is one-quarter of the wavelength, 
, of the received dominant frequency, which is about 260 MHz in our case. This theoretical resolution is about 10 cm. In practice, we considered a vertical resolution between /2 and /3 according to Beres and Haeni (1991). Therefore, the vertical resolution was determined to be in the range 12 to 20 cm.
Data acquisition was performed on two grids of orthogonal two-dimensional GPR lines (Fig. 3). One outline of each grid approximately corresponds to the layout of a trench wall. The acquisition grids associated with Sections A and B were named Grid A and B, respectively. Grid A was 15 m north–south by 7 m west–east, with line spacing of 1 m in each direction. Results obtained on Grid B were presented in Goutaland et al. (2005).
Data were collected in common-offset mode by using a constant 30 scans s–1 acquisition rate. A bandpass filter was applied; upper and lower cutoff frequencies were 100 and 800 MHz, respectively. A marker was set at each meter. Data processing was performed using the GSSI software Radan 6.5. Distance normalization was performed from markers, with a trace spacing of 40 scans m–1. A static time shift was used to obtain zero-offset traces, i.e., to align the direct ground wave arrival with a two-way travel time equal to 0 ns. A two-dimensional constant-velocity Kirchoff migration was performed using a constant wave velocity evaluated at 0.1 m ns–1. A background removal filter was used to eliminate the high-amplitude direct ground wave. As no other horizontal reflections due to real sedimentary contrasts were present on the profiles, the background removal was defined for all of the samples of the traces. All GPR profiles shown here were processed using the above sequence. In addition, we computed the instantaneous phase to emphasize the continuity of events. This attribute is commonly used in seismic processing (Yilmaz, 1987) to make weak coherent events more prominent. Although this processing helped us to define radar surfaces, GPR profiles processed with the instantaneous phase are not shown here due to space constraints.
Interpretation was performed using the terminology proposed by Neal (2004). This terminology distinguishes radar facies (shape, dip, continuity, and the relationship between reflections), radar surfaces (reflection geometries), and radar packages (three-dimensional external form). Three-dimensional mapping of depositional units was performed by outlining radar surfaces from profile to profile on each grid and by a horizontal interpolation between each profile. This interpolation was performed by kriging the radar surface. Kriging was chosen due to its flexibility in taking into account the spatial anisotropy of the sedimentary structure. We fitted each experimental variogram with an anisotropic linear model with a nugget effect (unless otherwise stated). Azimuths corresponding to the longest correlation length are reported below.
Calibration of GPR Reflections
The GPR reflections were calibrated in two different ways. First, a direct comparison between sedimentological description of the trench wall and corresponding two-dimensional GPR stratigraphy was performed to correlate reflection patterns with glaciofluvial lithofacies and depositional units. Second, an indirect calibration from adjacent trench wall profiles, sampled on grids of orthogonal GPR lines, allowed us to reconstruct the three-dimensional architecture of the depositional units as well as to relate lithofacies and reflection types to a depositional environment.
Figure 6 shows the direct comparison between the orthonormal picture of Section A and the corresponding GPR profile. The analysis of the GPR pattern, surface, and facies (in the sense of Neal [2004]), allows us to distinguish four different units that are well correlated with the four sedimentary depositional units described above.
Unit 1 presents continuous parallel reflections dipping northward and overlapping the lower boundary. They correspond to an alternation of Gcg,o and Gcm,b lithofacies. Only one Gcg,o lithofacies in the sandy-gravelly lithofacies was described during the sedimentological description. The GPR results provided us a better characterization of the sedimentary structure within Unit 1.
The boundaries of Unit 2 were observed with the GPR. This unit is characterized by high-amplitude subparallel reflections, dipping toward the south and concordant with the lower boundary. These reflections are related to the Gcg,o and Gcm,b alternance.
Unit 3 is characterized at depths between 0.3 and 1 m by subparallel reflections dipping northward, with occasional concave reflections at the base of the unit (near a distance of 12 m in the profile). These reflections correspond to the alternating Gcx,o and Gcm,b progradation. On the southern side, similarly dipping reflections are present between 1 and 2 m. Small subhorizontal or slightly oblique reflections are present at the same level. Oblique reflections are due to the sandy laminations. The boundary between sand and upper gravelly lithofacies is characterized by a continuous sinuous reflection dipping slightly southward. It is not possible, however, to characterize the lithology of the sandy facies.
Unit 4 presents continuous, subhorizontal, and slightly wavy reflections. Small wavelets are due to topography variations. Reflections are concordant with the lower boundary. This radar facies is present near the basin surface. The high amplitude of the reflections could be due to a greater quantity of silt particles. Reflections correspond to the Gcm gravels and sands mixture, in which fine Gcx,o gravel beds are inserted.
Characterization of the Depositional System
Figure 7
shows the interpretation of the pseudo-three-dimensional GPR block corresponding to Grid A. The profiles of Grid A near the trench wall were interpreted from the direct calibration on the trench wall. The interpreted radar reflection surfaces (hereafter named radar surfaces) are drawn in three dimensions.
|
Small Unit 2 has its longest range oriented at 172° N. A horizontal slice at a depth of 1.5 m shows the orientation of this unit (dark zone on the upper horizontal slice of Fig. 7). This unit has a lens-like external shape and is characterized by high-amplitude reflections corresponding to Gcg,o beds parallel to the unit boundaries (4° southward dip).
Unit 3 has a lower external elongated trough shape characteristic of a paleochannel. The direction of its longest range corresponds to the bottom of the trough shape, i.e., 160° N (orientation of the paleoflow). The northern side presents a 10° northeastward dip, corresponding to the Gcg,o–Gcm,b progradation.
Units 1, 2, and 3 were deposited in a braided system environment. The main orientation of the paleoflow seems to be southeast–northwest. This direction is fully consistent with the glaciofluvial corridor orientation.
The lower boundary of Unit 4 has a rather flat topography (an isotropic linear model was chosen to fit the experimental variogram), except for a north–south-oriented central depression corresponding to an isolated scour structure. The general shape is that of a sheet. Mean depth is about 20 to 30 cm. The base radar surface is an erosive truncation generated during a high-energy event in a braided river system.
Classification of Radar Facies
A classification of radar facies was set up from the calibration performed on Section A and the previous results from Section B (Goutaland et al., 2005). The characterization of the three-dimensional architecture of the glaciofluvial deposit was performed following GPR prospecting with the 400-MHz antenna. Three main radar facies (Fig. 8
) can be described using the terminology of Neal (2004). This classification is not exhaustive, but shows correlations between Sections A and B. Thus, extrapolations to the whole deposit underlying the basin could be made. Radar facies are listed below:
|
Radar facies f2: This facies is composed of concave or wavy reflections. They are trough cross-stratified. The general organization is a trough form. The reflections correspond to contrasts between minor lithofacies (S-x) with major lithofacies Gcm or Gcm,b. This facies corresponds to trough-fill deposits.
Radar facies f3: This facies presents continuous dipping reflections, subparallel, with occasional cross-cutting between reflections. This facies corresponds to progradations of Gcx,o or Gcg,o in sand and gravel mixtures (Gcm, Gcm,b).
Concerning the radar surfaces, contacts between upper units (Unit 1 for Section A) and lower units are erosive truncations due to the erosion of the deposits during a high-energy event. Radar surfaces between lower units are external surfaces of more or less preserved trough structures.
|
Definition of an Interpretative Hydrostratigraphic Model |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
Definition of the Lithofacies Distribution
A third grid of two-dimensional orthogonal GPR lines located on the north side of the measurement well was surveyed. This grid, named Grid C, was located perpendicular to the TDR probes (Fig. 3). Data were collected using a GSSI SIR 3000 system with a 400-MHz antenna. The grid area was 15 m west–east by 6 m north–south, with line spacing of 1 m in each direction. Its orientation was 115° N. The processing sequence described above was used. A three-dimensional model of depositional units was proposed. The classification of GPR reflections was used to interpret the radar stratigraphy. A two-dimensional interpretative model of lithofacies distribution was set up, which corresponded to the profile y = 3 m of Grid C (in bold in Fig. 3).
Four main depositional units were characterized: Units 5, 6, 7, and 8 from the base to the top of the profile (Fig. 9 ). Unit 5 is characterized by long continuous reflections dipping westward. They are mainly concave. Unit 6 presents planar oblique high-amplitude reflections dipping eastward. The reflections are long and continuous, parallel to each other. The termination of the reflections overlaps the lower boundary. Unit 7 is characterized by wavy and concave reflections. Reflections are short and discontinuous. The general shape is concave (channel form). The upper unit (Unit 8) is composed of long continuous reflections. They are subhorizontal or dip westward, generally parallel to each other but with some curved reflections. These reflections have high amplitude. Reflections are concordant with the erosive lower radar surface.
|
|
|
|
 |
w is the density of water, DR is the specific weight of solids defined as DR = s/w with s the density of solids (2.65 g cm–3), and C is a constant factor taking into account the shape and tortuosity.
Chapuis and Aubertin (2003) gave the following expression for the specific surface S:
 |
 |
Saturated hydraulic conductivity values vary over several orders of magnitude, from mean values of about 10–4 m s–1 for sandy and bimodal lithofacies (BM and S hydrofacies) to mean values of 10–1 m s–1 for open-framework matrix-free gravels (OW hydrofacies). Calculated Ks values are in agreement with literature values.
Matrix-free open-framework gravels have the highest Ks. Under fully saturated conditions, their hydraulic conductivity is at least two orders of magnitude higher than that of other glaciofluvial lithofacies. Thus, they may act as preferential flow paths during stormwater infiltration. These fast flow paths can convey dissolved or colloid-bound contaminants. Under unsaturated conditions, the coarse grain size of these gravels may be at the origin of capillary barrier effects at the contact of sand matrix lithofacies, causing water ponding above them (Kowalsky et al., 2004). Therefore, lithofacies located just above the OW hydrofacies can remain in contact with pollutants contained in stormwater for long periods and thus be more exposed to contamination. Carbonated lithofacies, for instance, are known to retain significant amounts of Pb, Cd, and Zn (Plassard et al., 2000).
Comparison between M and BM hydrofacies shows that the sand matrix fraction is a determinant parameter in the control of hydraulic conductivity. Saturated hydraulic conductivities of bimodal and sandy hydrofacies are similar. Massive hydrofacies have a greater saturated hydraulic conductivity. At low suction, we can suppose that, due to their finer mean grain size, BM hydrofacies will be more conductive to water flow than M hydrofacies.
Sandy lithofacies represent, with bimodal lithofacies, the least conductive hydrofacies under fully saturated conditions. Their fine grain size, in comparison with the three other hydrofacies, can lead to a higher degree of saturation under variably saturated conditions. They are, therefore, more exposed to contamination: dissolved contaminants can have a longer residence time in these hydrofacies, and retention mechanisms may have a greater probability of occurrence.
Under fully saturated conditions, OW hydrofacies may act as preferential flow paths. Under transient conditions, hydrofacies with a sandy matrix (M and BM) and sandy hydrofacies (S) may be more conductive than OW hydrofacies. Capillary barriers may develop and cause water ponding above OW hydrofacies, leading to a greater exposure of glaciofluvial deposits to contaminants.
|
Evaluation of Presumed Hydraulic Behavior from Hydrostratigraphic Model |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
Hydraulic Behavior of the Glaciofluvial Deposit
Figure 12
shows the water content variations at the surface (TDR Probe 1), at –0.50 m (TDR Probe 2), and –1.15 m (TDR Probe 3) during Infiltration Test A (Fig. 12a) and Infiltration Test B (Fig. 12b). Figure 11b shows the location of the three probes on the hydrostratigraphic model. Probe 2 is placed in the upper part (high-energy deposit), dominated by Gcm lithofacies (M hydrofacies) in which Gcx,o lithofacies (OW hydrofacies) are inserted according to the interpretative model. Probe 3 is located in the lower level, corresponding to trough-fill deposits.
|
, for both infiltration tests at the surface (Test A: = 3.9%; Test B: = 4.4%) and –0.50 m (Test A: = 6.4%; Test B: = 6.7%). Only the initial water content measured with TDR Probe 3 at –1.15 m significantly differed between the infiltration tests (Test A: = 12.5%; Test B: = 19.2%). During the infiltration phase of Test A (infiltration with initially "dry" conditions), volumetric water content reached values between 41.8 and 43.8% at the surface of the basin (TDR Probe 1). The elapsed time corresponding to these values is the time of constant hydraulic head imposition at the surface. The water infiltration began at 29 min. A significant increase in water content was observed at 35 min at TDR Probe 2. This increase corresponds to the appearance of the infiltration front at –0.50 m, with a delay of 6 min with respect to the beginning of water infiltration. No significant increase was measured at –1.15 m. Measured values ranged initially between 12.5 and 13%.
During Infiltration Test B (infiltration with initially "humid" conditions), similar water content variations were measured at TDR Probes 1 and 2. The delay between the arrival of the infiltration front at TDR Probe 1 and TDR Probe 2 was 7 min 30 s. Concerning the water content variations measured at –1.15 m, a significant increase was observed at 38 min, i.e., 26 min after the beginning of water infiltration. This water content increase was slower than those measured at TDR Probes 1 and 2. Water content reached a stage of 33.3%. The decrease in water content following the end of water infiltration was very slow and quasilinear.
It can be assumed that the significant difference in the water content measurements at –1.15 m for both infiltration tests was due to a capillary barrier effect occurring between –0.50 and –1.15 m. Under initially "humid" conditions, the propagation of the infiltration front at this depth could be due to a larger water conductivity of the lower level than under the initially "dry" conditions. This interpretation is in agreement with the interpretative hydrostratigraphic model, in which the upper level, composed of M hydrofacies, could be less conductive under unsaturated conditions than the S or BM hydrofacies constituting the lower part and having a lower mean grain size.
Measurements at the surface and at –0.50 m show values ranging from 39 to 44%. These values are large for M hydrofacies according to the porosity data of Table 2. At the surface, the presence of a larger sand fraction may explain these differences. At –0.50 m, larger values may be explained by the presence of a sand pack made of glaciofluvial facies sieved at 5 mm at the end of the subhorizontal pipes to improve the contact with the TDR probe. This sieved material may explain an increase in porosity and thus a larger water content than expected with regard to the data of Table 2. Water content variations measured at –1.15 m during Infiltration Test B are in the range of values of S hydrofacies according to the porosity data of Table 2. Thus, water content variations at the three levels seem to be in agreement with the proposed interpretative hydrostratigraphic model.
Our study suggests that stormwater infiltration in the vadose zone of such heterogenous quaternary deposits may be significantly influenced by the sedimentary heterogeneity and the initial water content. This assumption will be validated by a complete hydraulic characterization of each hydrofacies.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
The use of GPR in this hydrogeophysical approach is promising for the study of the unsaturated water flow mechanisms in underlying subsoils of infiltration basins. The high resolution obtained with the GPR method is suitable for characterization of the geometries of depositional units and dips of the internal lithofacies. Realistic interpretative stratigraphic models may also be defined at the lithofacies scale. Moreover, the vadose zone underlying infiltration basins is often comprised of highly permeable materials without a fine grain size fraction (clays and silts), allowing good penetration depth of radar signals.
To better characterize the impact of stormwater on the heterogeneous vadose zone underneath infiltration basins, a full characterization of hydraulic properties (water retention and hydraulic conductivity curves) is needed. Future research will be focused on the determination of the hydraulic properties of hydrofacies, which can be input into water flow numerical modeling. We showed that analogous glaciofluvial hydrofacies formed in similar depositional systems have similar saturated hydraulic conductivity. Future research has to assess the reliability of evaluating hydraulic properties of hydrofacies on modern analogous glaciofluvial deposits. Indeed, the easy access to the glaciofluvial lithofacies on the surface of these modern analogous deposits (in comparison with the difficulty of accessing quaternary lithofacies by trenching) would allow simple measurement of the hydraulic properties of these lithofacies. The numerical modeling performed from the interpretative hydrostratigraphic model and integrated hydraulic characteristic curves will improve our understanding of unsaturated hydraulic behavior and, therefore, transfer mechanisms in the vadose zone. In particular, this numerical modeling will help us to understand preferential flow mechanisms occurring in the vadose zone of an infiltration basin.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONField Site and InstallationsCharacterization of the...Relating Radar Stratigraphy to...Definition of an Interpretative...Evaluation of Presumed Hydraulic...ConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Lambot, A. Binley, E. Slob, and S. Hubbard Ground Penetrating Radar in Hydrogeophysics Vadose Zone J., February 25, 2008; 7(1): 137 - 139. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
