3 Fredericton is the capital city of New Brunswick with a population of about 55,000 inhabitants (Fig. 1). The Saint John River bisects the city at a location down river of the point of tidal influence that extends up-river to Mactaquac Dam (Fig. 2). At approximately 125 km upstream of its confluence with the Bay of Fundy, the elevation of the river at Fredericton averages around 2–3 m in elevation with the river bottom deepening to more than -5.5 m in places (Butler et al. 2004; Nadeau 2006). The city’s downtown urban centre is located on a wide floodplain mostly located along the south side of the river. The relief changes from the floodplain at an average of 10 m elevation to greater than 100 m in elevation at the top of the steep valley wall to the south of the river. The north side of the valley rises gradually in elevation to greater than 150 m, with relief largely controlled by the incision of small rivers locally flowing into the Saint John River (Fig. 2). The City’s urban centre overlays a deep bedrock valley that has been infilled with up to 60 m of glacial and post-glacial sediments. This ancient valley has been cut into bedrock comprised mainly of Carboniferous Pictou Group terrestrial sedimentary units, consisting of red and grey conglomerate, sandstone, siltstone, and shale (McLeod et al. 1994).

4 Several authors have contributed to the present-day understanding of the unconsolidated deposits in the area (e.g., Chalmers 1884; Kiewiet de Jonge 1951; Lee 1957; Gadd 1971; Kingston and Lohse 1975; Seaman and Thibault 1981; Rampton et al. 1984). During glaciation the bedrock surfaces were scoured by glacial ice and parts of the Saint John River Valley were deepened (Rampton et al. 1984). Locally, glacial flow direction was southeasterly (Lee 1957; Gadd 1971; Kingston and Lohse 1975; Seaman and Thibault 1981; Seaman 2004) concluding with deposition of ice-contact and glaciofluvial sediments as the glacier retreated inland.

5 During, or immediately following, deglaciation a large body of water inundated the Saint John River Valley, resulting in the deposition of a thick succession of clays and silts in the study area. This water body has been termed Lake Acadia by Chalmers (1902) and interpreted as an arm of the Late Glacial De Geer Sea by Kiewiet de Jonge (1951). Lee (1957) noted varve-like laminae in the clay and silt and identified small fragments from the bottom of this unit as the outer coating (periostracum) of Yoldia arctica (Gray), a marine species tolerant of muddy and brackish water conditions as are found at the mouths of glacial streams. Lee (1957) concluded that Lake Acadia was at first an icedammed glacial lake, possibly due to a residual ice cap near Saint John (Seaman 2004), and subsequently an estuary. However, burial of the outlet for the Saint John River at the Bay of Fundy (Fig. 1) likely facilitated capture of marine water and formation of a large brackish to freshwater lake in the area, prior to opening of a new outlet and return to a full estuarine environment (Seaman et al. 1993; Dickinson 2008).

6 Clay-silt sediments exposed locally were mapped by Rampton et al. (1984) as undifferentiated marine and lacustrine sediments. Daigle and Broster (2006) noted that the clay-silt unit at Fredericton demonstrates attributes of both lacustrine deposits such as zones with distinct varve-like laminae and poorly stratified zones of low strength sediments similar to marine deposits, both of which would be expected to occur in a glacial estuary. Investigation of the clay-silt unit revealed the presence of Cochlichnus anguineus Hitchcock traces on silt partings, a species reported from postglacial lacustrine deposits (Gibbard and Dreimanis 1978). These observations indicate that the unit was deposited under conditions ranging from freshwater to brackish water, associated with variation in glacier melting and marine incursion during deglaciation of the Saint John River Valley.

7 The stratigraphic sequence most commonly reported at Fredericton, includes: bedrock of red, grey or green sandstone and shale; till that serves as a lower aquitard; a sand and gravel unit occasionally containing cobbles and boulders or with sporadic silt and clay that serves as the main aquifer for the city (lower aquifer); a glacilacustrine/marine clay and silt unit with occasional sand lenses (middle aquitard, here informally termed the clay-silt unit); a glacifluvial sand to sand and gravel unit (upper aquifer); and a surface unit of alluvium/colluvium silty, fine-grained sand. Daigle and Broster (2006) characterized the stratigraphy as 11 hydrostratigraphic units including bedrock (Table 1). The 10 sediment units represent events from Late Wisconsinan glaciation to the present consisting of three till units, three intervening ice-contact and two glaciofluvial units, an estuarine clay and silt unit, and postglacial fluvial and modern anthroprogenic deposts.

8 Much of the information on the subsurface units at Fredericton have been derived from geotechnical drilling mainly concerned with foundation stability due to underlying sensitive clay-silts or protection of the underlying municipal groundwater supply aquifer. One challenge to engineering projects at Fredericton has been the occurrence of thixotropic, low-strength clays at depth reported by Alkins (1983) and Landva et al. (1988). At one site in the City centre, Landva et al. (1988) noted that light construction equipment traffic on the 40 cm sand layer above the clay, “caused the clay to change to a jelly-like consistency” and flow into the open excavation trench. Geotechnical investigations at locations, short distances from the City (Fig. 2), were under taken by Watson (1970) and Munro (1994) but neither reported the occurrence of thixotropic clays.

9 Violette (1990) studied the occurrence and distribution of discontinuities in the clay and silt unit, termed “windows” by ADI Limited (1982). The windows are considered to be hydrologic pathways into the underlying supply aquifer, resulting from the thinning of clay and silt deposits above an ‘esker-like’ feature that forms the bulk of the Fredericton aquifer (Violette 1990). Violette (1990) was first to create a digital Geographic Information System (GIS) database of the Downtown Fredericton area, called GEO.DAT. This database included 107 boreholes from south side of Saint John River overlying the supply aquifer and identified six main stratigraphic units. Peters (1992, 1995) enlarged the database by examining 539 borehole logs from drilling on both sides of the Saint John River at Fredericton, but due to changes in computer and soſtware technology the raw data is no longer accessible, although lithological descriptions and stratigraphic interpretations are documented and support the earlier findings of Violette (1990).

10 As part of the current project Daigle and Broster (2006) established a GIS-referenced data base that could be used to delineate recharge pathways into the underlying supply aquifer, consisting of: (1) three-dimensional stratigraphic models of the sediment architecture underlying Fredericton, (2) geo-reference borehole data from previous studies, and (3) a data base that can be modified as new borehole information is collected. Over 1500 drill logs were collected from a variety of sources and culled to the 602 records representing boreholes within city limits that extended to depths exceeding 1.5 m, contained lithological description of the material encountered and could be accurately geo-referenced. The resulting digital data set enables better delineation of units for engineering applications and assessment and can be used to view individual logs, stratigraphic cross-sections, or in its entirety as three-dimensional block diagrams that can be rotated and magnified as desired (Daigle and Broster 2006). While providing an area view of distribution of the units, only a general description of the lithology is presented without any detailed characterisation of engineering properties or delineation of zones of low strength materials.

12 Samples were initially stored in PVC sample tubes and returned to the Civil Engineering Soils Laboratory at the University of New Brunswick. Core samples were split longitudinally with one half preserved for detailed sediment logging, and the other half used for geotechnical analysis. The full length of the core was laid out in the University of New Brunswick, Department of Geology Sedimentology Laboratory, allowed to dry and visually examined to produce a detailed lithological record.

13 Natural moisture contents, wet densities and penetrometer estimates of variation in unconfined compressive strengths have been measured (Daigle 2005; Nadeau 2006). A total of 29 samples of the clay-silt unit were analysed for geotechnical properties and granulometric analysis. As the clay-silt unit was of particular interest only fine-grained particle analysis was conducted as set out in ASTM D 422-63, slightly modified (Bouyoucos 1962).

14 For 28 samples from the clay-silt unit, mineralogy was determined using X-Ray Diffraction (XRD) and chloride concentrations were determined using Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES). Samples were analysed using standard operating procedures by a Bruker D8 Advance Solid State Powder Diffraction XRD System at the University of New Brunswick.

15 A log of the UNB 1-03 core was constructed based on visual observations in its dried state (Fig. 3). The lowest potion of the core (41–37 m) consists of brown matrix-supported diamict containing a mixture of subangular to subrounded clasts, varying in size and lithology, with most lithologies from distant sources. Several of the clasts were subangular, striated and coated with silt. The presence of silt coatings on the clasts and predominance of lithologies that are not local, suggests that the sediment is englacial meltout till (cf. Dreimanis, 1976). Therefore, this lowermost unit likely represents deposition from a stagnant or retreating ice-lobe and is correlated with Unit 4 (Till 2, Table 1).

16 Above 37 m poorly sorted diamict occurs in beds separated by sporadic lenses of sand and sub-angular to sub-rounded gravel with traces of silt in the lower layers, becoming mainly sand and gravel from 33–31 m. The sand and gravel layers interfinger with diamict layers and are interpreted to be ice contact sediments (Unit 5), likely deposited at or very near the ice front with minor reworking from meltwater. Improved sorting and rounding of clasts at 31–30 m may represent a significant local retreat of the ice front prior to deposition of the uppermost till.

17 A grey-brown diamict, varying in compaction with sub-angular clasts and fragments of shale, possibly of local origin, occurs from approximately 31 to 28 m and is correlated with Unit 6. At 28 m the diamict (till) is divided by 1 m layer of poorly sorted sand and sub-angular to sub-rounded gravel with traces of silt and clay. The interfingering of till and poorly sorted sand and gravel suggest a gradation to an ice-contact facies likely correlative with Unit 7. Above 27 m the sand is course-grained, gravel clasts are rounded to sub-rounded and the sediment grades upward to fine sand more characteristic of glaciofluvial (Unit 8) deposits.

18 From 26 to 24 m sand layers are interbedded with beige to brown sandy-silt that becomes clay-rich towards 22 m (unit 9). At approximately 20 m the unit changes upward to a laminated silty-clay, grey and buff in colour. Rhythmically laminated grey and buff coloured clay and silt with occasional thin lenses of fine-grained sand millimetres thick occur from a depth of approximately 20 to 6.75 m below surface. Clay content increases upward from the base to a clay-rich zone (average 70% clay) at 20 to 17.5 m (Figs. 3, 4). At a depth of 22 m, a 1–2 cm wide lens of graded sand with small angular fragments was found (Fig. 5a). Such layers may represent resedimentation due to slumping, a tempestite, or deposition from ice raſting. Faulting in the clay-silt unit was noted in several locations (Figs. 3, 5b). In core, these appeared to be small normal faults with measured displacements along the fault planes ranging from a few millimetres to a few centimetres. Samples of grey/buff laminated clay-silts collected above 12 m were easily disturbed by sampling and lab treatments. These samples (note top of samples in Fig. 4b) demonstrate deformed layering, fluid expulsion features, but no brittle faults. Some minor faults may be from disturbance during drilling and sampling; in particular, those occurring near the ends of the sample cores. Other faults occurred in undisturbed portions of the core samples (Figs. 3, 5b; 17–12 m) and it was interpreted that these faults likely pre-dated sampling as the faults were brittle, showed clear vertical displacement of laminae, and the surrounding materials where undeformed. At approximately 21 m (69 ſt) depth, a 12 by 5 mm dropstone was found in a thin (5 mm) fine-grained sand seam (Figs. 5c, d) and the first of three trace fossils was found directly beneath this coarser lamina (Fig. 3). Nematode traces identified as Cochlichnus anguineus Hitchcock by R.K. Pickerill (personal communication, 2005), were found along silt partings at depths of and 21.20, 18.14, and 12.78 m from the surface (Figs. 5e, f).

19 At about 16.15 m (53 ſt) there is a marked colour variation in the laminations to a reddish-brown colour that grades to grey-buff lamina at 14.6 m (Figs. 3, 4), although intermittent groups of reddish-brown layers occur to about 13.7 m. Grey/buff laminated clay-silts occur to a depth of approximately 6 m. The sediments are highly deformed from 12.5–6 m and are intercalated with thin lenses of well-sorted fine sand, increasing in frequency upward. At 6 m sand (Unit 10) becomes coarse, pebbly sand, with interstitial silt and occasional lenses of silt, changing to layers of well sorted sand and pebbly sand above 3.8 m. The overlaying beds (Unit 11) consist of a faintly laminated oxidized orange-brown clay-silt and silt-fine sand, with sand and sub-rounded pebbles, with anthropogenic reworked material (fill) in the upper 0.5 m. The ground surface at the site of UNB 1-03 is a flat-lying floodplain at an elevation of 7.5 m adjacent to the present river. Materials described here represent depths below ground surface.

20 Grain size analysis of the clay-silt unit enables the lower 22 to 20 m (72 to 66 ſt) to be characterized as clayey-silt, grading upward into silty-clay from 20 to 14 m (66 to 46 ſt), and the upper 14 to 6 m (46 to 20 ſt) as clayey-silt. Generally there is an overall decrease in percent clay content in sediment above 17 m (Fig. 6).

21 Mineralogical analyses of the sediments in the area have been conducted by Allen and Johns (1960), Munro (1994), Giudice (2005), and Dickinson (2008). Munro (1994) reported the occurrence of a swelling clay, which was not found in UNB 1-03. Examination of UNB 1-03 revealed the clays to be predominantly clay-sized particles with few secondary clay minerals present (Giudice 2005). From the 28 samples analysed, the mineralogical composition of the clay-silt unit consisted of quartz, clinochlore, muscovite, albite, and calcite. Calcite was absent in four samples. Quartz, muscovite and calcite are common rock-forming minerals found in sedimentary rocks (Klein and Hurlbut 1993), although calcite can also precipitate from groundwater. Clinochlore is common in argillaceous sedimentary rocks such as the Carboniferous shale and sandstone units that underlie Fredericton. Albite is most commonly found in igneous and metamorphic rocks and less commonly in sedimentary rocks (Klein and Hurlbut 1993) and may be indicative of detrital contributions from glacigenic sources (i.e., till or outwash) rather than local bedrock.

22 Geochemical analyses of the clay-slit sediments have been conducted by Gudice (2005), Dickinson (2008) and Munro et al. (1997). For this study, chloride concentrations are reported for 28 samples (Fig. 6). Chloride concentrations are highest in the lower part of UNB 1-03 increasing upward from the base of the unit to a maximum of 34.5 ppm at 19.5 m. Samples above 17.5 m demonstrate an overall upward decrease in chloride ppm to a depth of 14 m (Fig. 6) coincident with the change in colour discussed earlier. At depths 13–7 m, chloride concentrations are below 10 ppm to undetectable and are similar to that reported by Munro (1994), except that in UNB1-03 values increase upward from undetectable at 8 m to 16 ppm at 6 m (Gudice 2005).

23 Plasticity characteristics of the UNB1-03 clay-silts indicate the sediments to be inorganic clays of low to medium plasticity (Figs. 6, 7). The range over which the UNB1-03 values plot is typical of illite clays (Stumpf et al. 1998), although none were found here. Skempton (1953) defined the activity of clay as the ratio of plastic index to percent clay fraction content. Clays with activities less than 0.75 are considered inactive clays, those with activities from 0.75 to 1.25 are considered normal, and those with activities greater than 1.25 are considered active (Skempton 1953). Calculated activities of clays in this study ranged from 0.17 to 0.64 (Fig. 6) and would therefore be considered inactive with medium compressibility (Fig. 7)

24 Sensitivity is defined as, “the loss of undrained strength that is observed when an undisturbed clay is remoulded” (Mitchell and Houston 1969, p. 845). The sensitivity of clays has been correlated to the liquidity index, such that clays with liquidity indexes greater than 1.0 can be transformed to a viscous state and flow like a liquid when remoulded (Das 2005). The samples analysed in this study display liquidity index ranges of 0.57 to 1.73 with the majority of the grey clay-silts above 18 m having values greater than, or approaching 1.0 (Fig. 6). These upper grey clay-silts are therefore considered to be highly sensitive, and appear to correlate to the sensitive soils reported by Landva et al. (1988) and Alkins (1983) near the Bank of Montreal site in Fredericton (site 4, Fig. 2).

25 Drilling indicates that underlying the floodplain at Fredericton, two water tables exist separated by the clay-silt unit serving as an aquitard where that unit forms an unbroken layer. An upper groundwater table is commonly found that occasionally represents a perched water table from ground water within the permeable near-surface sediments. A lower water table representing the potentiometric surface of a lower aquifer underlying the clay-silt sedments occurs where windows, gaps in the estuarine unit enabling linkage between the lower and upper water systems, are not present. The natural moisture contents of the clay-silts examined were found to increase at depth to a maximum of 42% at 19.5 m (Fig. 6). Landva et al. (1988) suggest that the high water content zone, within the clay-silt sediments, may be a result of some form of bonding developed in the clays during and aſter deposition. Another explanation could be that the increased moisture content at 7.5 and 19.5 m (Fig. 6) is a capillary fringe-effect in the clay (cf. Fetter 2001), drawing water downward from the overlying aquifer and upward from the underlying city-source aquifer (Fig. 3). The slight decrease in the density profile of UNB1-03 (Fig. 6, as determined by Nadeau 2006) is contrary to an expected increase in density with depth, but may be caused by increased moisture content in the clays.

26 Pocket penetrometer readings taken from fine-grained sediments show variable readings with depth. Sediments at 4 m depth, interlayered within the upper coarser grained sand beds, demonstrate the highest penetrometer and wet density values (Fig. 6). The high penetrometer and density readings occur in the upper layers above 6 M (20 ſt) and may be indicative of post-depositional minor cementation from precipitation of iron oxide (see Fig. 4a) or occasional subareal exposure at time of deposition. As the sediments are mostly laminated, variable results are dependent on the laminae being tested. Generally, the penetrometer readings do not demonstrate a general decrease in values with depth as was found for density or liquidity index, or general increase in values with depth as was found for percent clay, liquid limit, or plastic index.