Discovering and understanding the geomorphological processes operating on the physical environment is a current focus in geological, environmental, ecological, and archaeological research. The mechanisms driving these physical processes are often complicated, involving many components that operate on diverse scales varying from macro- to micro-regional. Several events have contributed to the formation of the estuarine landscape of the lower Saint John River Valley and Grand Lake Meadows (Fig. 1) including riverine, marine, lacustrine, and terrestrial processes, as well as glaciation. Both the river and Meadows systems are dynamic features of the landscape, whose characteristics varied over time and space due to their associated link with geological and climatic controls affecting the region.

The study site afforded a unique location (Figs. 1, 2) situated on the northeast side of the Saint John River at the confluence of Grand Lake and the Saint John River. Drill logs from 34 holes drilled for the New Brunswick Department of Transport (DOT) related to Trans Canada Highway (TCH) and bridge construction (Washburn and Gillis Associates 1996, 2000), were provided by Maritime Road Development Corporation and used to construct a stratigraphic profile for sediments to about 70 m thick overlying bedrock at the Meadows (Figs. 3, 4). During June of 2006, GLM-01 was drilled to a depth of 67.5 m at 45° 50’ 14” N and 66° 10’ 33” W (NAD830CSRS); a location correlative with previous DOT drill log data (Figs. 2, 3). Analytical results of GLM-01 and interpretations are discussed herein. To minimize confusion with references citing published non-calibrated radiocarbon dates (yBP), calibrated dates are used sparingly and are denoted as calyBP.

A stratigraphic profile for the Meadows was delineated from TCH drill logs supplied by Maritime Road Development Corporation (Fig. 3), which confirmed visual observations of GLM-01 core samples (Fig. 4). A mixture of very compact, fine-grained, grey-brown, sandy silt with subangular to subrounded striated pebbles and cobbles of mixed lithologies was recovered from the lower 64–60 m overlying bedrock. The sediment is interpreted as till and is similar to that recognized locally (Rampton et al. 1984; Seaman et al. 1993). The predominance of non-local lithologies suggest that the till is englacial or supraglacial (Dreimanis 1976), most likely deposited from stagnant or retreating ice.

Overlying till at a depth of 60 m, poorly sorted silty sand grades upward into centimetre-thick layers of silt and sand, becoming 2–3 mm thick couplets of sandy silt and clay laminae at 58 m to 56 m (Fig. 5). The poorly sorted sand grains are angular to sub-angular in shape when viewed through a magnifying glass, and may indicate deposition proximal to a melting ice-front with variable current strengths, grading upwards into material deposited under a low-energy regime.

Clay layers increase in frequency upwards from 59 m (Fig. 5) as sediment becomes finer and well-laminated showing ‘varve’-like couplets of alternating grain size and colour. Between 56 and 52 m clay content increases, with a high of 78.3% clay at 53 m. This part of the core commonly demonstrates deformed layering and fluid-escape structures causing couplets to be less distinct. This zone (56–52 m) also includes the highest plastic index values from 7 to 12 PI (Dickinson 2008). Deformation is restricted to this interval of the samples, which may have been due to collection and transportation of samples containing zones of material sensitive to disturbance. Natural water content percent is approximately 30% in most zones and exceeds 40% in some deformed samples (Dickinson 2008).

Trace fossils Cochlichnus anguineus (Hitchcock) was found at 58.5 m (Fig. 6a) and Planolites were found at 48.5 m and 44 m (Fig. 6b). Both organisms are infaunal and occur across many facies. However, their occurrence indicates that this interval was not representative of an isolated proglacial lake, devoid of life.

Above 56 m the sediments are highly laminated with layers as thin as 2 mm occurring between 54–51 m (Fig. 7a). Lamina thickness was 3 to 5 mm with rare lamina up to 7 mm (i.e., at 33 m) possibly due to turbation or storm events causing erosion and re-sedimentation. Several core samples demonstrated deformation and mixing of layers in the upper part of the cores due to disturbance caused by impact of the sampling tube during collection or extraction of the samples from the sampling tubes (Fig. 7b). With the exception of the lowermost sandy layers (<58 m), natural water content was commonly found to be above 30% and >40% in some areas (Dickinson 2008). From 58 to 15 m a total of 3472 couplets were counted, with more than 70% of the couplets occurring below the level of the oldest dated material of 11 340 ± 210 yBP at 28.68 m (Table 1). The couplet data are not considered to provide an exact representation of the age of the unit as some couplets may not be true varves. Also, some laminations have been obscured by deformation and some material has likely been eroded during the onset of fluvial conditions at the top of the unit (at 15 m). Nonetheless, the data is consistent with the area becoming ice free sometime between 13 000 and 14 000 yBP.

Distinct visual laminations within the sediment were not clearly visible between 28 m and 22 m depth. The 8150 yBP and the 9600 yBP dated materials were recovered at 25.6 m and 22.68 m, respectively, suggesting a potential dating error associated with mixing, possibly due to subaqueous slumping, bioturbation, or some other catastrophic event. Re-sedimentation can be caused by strong storms or liquefaction due to seismic disturbance and these phenomena could explain the mixed dates and lack of preserved laminae.

Small, bright blue concretions were observed that X-ray diffraction confirmed to be the mineral vivianite (Fe²⁺₃ (PO ₄)₂·8H₂O). Results of X-ray diffraction analysis of 52 stratified samples collected between 58–18 m indicated a mineralogical composition consisting mostly of quartz, muscovite, albite, clinochlore, calcite, and orthoclase (details in Giudice 2005; Dickinson 2008). Vivianite was not found in samples collected below 42 m in depth (Fig. 8a) but increased in frequency of occurrence and size of structure upward, occurring most frequently in samples collected between 25 and 18 m. Vivianite was found as 2 mm fragments at lower depths and occurring as fissile nodules up to 1 cm in size in samples above 25 m.These concretions were observed surrounding or in association with some of the organic fragments (Fig. 8b), an association often recognized by others (Fagel et al. 2005).

Vivianite is a widespread mineral in lacustrine sediments (see samples of world occurrences in Fagel et al. 2005) and can be formed by either authigenic or diagenetic processes related to a chemical, physical, or biological change after its initial deposition. The vivianite concretions are likely related to authigenic processes as they are associated with areas of high biological productivity where organic matter was able to fall out of suspension with the fine silt and clay. Additionally, the vivianite was not associated with an oxidized layer or crust, but found only as small concretions. Deike et al. (1997) noted that vivianite is most likely to be formed just below a thick oxidized layer if related to a diagenetic formation.

Small wood fragments were also frequently encountered above 42 m to surface. Of the 37 samples of sufficient size, nine samples were selected for ¹⁴C/¹³C analysis by IsoTrace Radiocarbon Laboratory (See Table 1 for yBP and calyBP results). Eight of the nine dates were found to decrease in age sequentially from 11 340 ± 210 yBP at 28 m to 110.36 ± 70 yBP at 0.71 m. A sample recovered from 22.68 m returned a date of 9600 ± 110 yBP that is older than would be expected by examination of the dates obtained from samples immediately overlying and underlying this depth (Table 1). It may be that bioturbation has resulted in older material being moved up in the sequence or, more likely, mixing and re-sedimentation due to a turbation event from some other cause. As samples for dating were collected from material in the centre of cores, it is unlikely that younger wood fragments could have been dislodged from a higher position during drilling. Nonetheless, a linear regression analysis of both raw and calibrated data gives a best fit line equal to, or better than R² = 0.94. These results suggest that the dates, considered in toto, represent an accurate estimate of the ages of the upper 30 m of sediment at GLM-01. A tenth sample from a depth of 41 m was too small to return a date but the occurrence of wood fragments at this elevation suggests that the region was ice free much earlier than 11 000 yBP (13 000 calyBP, Table 1).

Further observations and additional analytical data are reported in Dickinson (2008) with a selection of tests presented in plots of value with depth of sample (Fig. 5). Granulometric analysis demonstrates the change of percent sand, silt and clay with depth and indicates higher silt and clay content for the lower half of the sediment core. Sand content increases throughout the middle portion of the core with gravel present from ~15 to 5 m depth. A unit of silt and clay occurs within the top 5 m of the core. The differences in grain size are interpreted to reflect current-deposited sand and gravel versus suspension-deposited silt and clay, although water depth and size of the water body may also affect grain size occurrence.

LOI is a method that has been used by others in the region to measure the organic content of sediments for the delineation of past fluctuations in climate (Levesque et al. 1993; Mayle and Cwynar 1995a). However, it is important for the lower Saint John River Valley that LOI organics, mainly pollen, may not only have come from local terrestrial sources but may also have been introduced to the system through the inundation of marine water. Therefore, using samples from a closed system as opposed to an open system like the lower Saint John River Valley would potentially allow for a more unbiased conclusion when correlating climatic conditions with LOI. High LOI is typically correlative with warm temperatures, but within GLM-01 samples, high concentrations are related to low-energy environments and low LOI concentrations are associated with major fluvial aggregation as between 15 and 4 m or proglacial environments as with samples from below 59 m (Fig. 5).

The percent of LOI in the samples, from bedrock to surface, fluctuates with grain size and depth (Fig. 5). The sandy sediments deeper than 60 m are characterized by the lowest percentage of organic material, commonly <2%. In material above 59 m LOI increases to between 3% and 4.7%, falling to between 2.5% to 3.5% for the silt-rich layers above 45 m. Between 15 and 6 m LOI decreases to less than 1.7% likely due to an increase in velocity of flow accompanying a change to a fluvial environment as indicated by the occurrence of gravel layers at that level (Figs. 4, 5). Above 6 m, organic content increases as does the occurrence of silt and clay deposits, likely reflecting deposition in a low- energy environment by flooding related to development of the modern lake and river system. These interpretations are supported by a strong Spearman Correlation of r_s = 0.61 significant at the 99% level between LOI and clay% deposited from suspension.

Variation in Br and Cl concentrations are interpreted to reflect changes between dominantly freshwater and saline water conditions (Gieskes et al. 1991; Davis et al. 1998, 2000, 2001; Broster et al. 2013). Both Cl and Br demonstrate similar trends in distribution and are supported by a strong positive Spearman Correlation of r _s = 0.7 significant at the 99% level. The absorption of chloride and bromide onto the surface of the clay minerals was used as a possible indication of brackish or marine water deposition (Burt 2004).

Davis et al. (1998, 2000, 2001) have indicated that the chemical ion ratios of chloride/bromide are especially useful as geochemical indicators in assessing marine inputs to a region. A large portion of the sediment core has low Cl/Br ratios suggesting ratios similar to those that can be generated in brackish water. The highest Cl/Br ratios were found in the lower half of the core; decreasing in concentration upward and fluctuating in value throughout the middle and upper portion of the sediment core (Fig. 5). The overall variations in Cl and Br concentrations with depth indicate initial low values increasing above 58 m to an area of highest values (i.e., Cl = 300 mg/l, Br = 2.0 mg/l) between 50–25 m and decreasing to the lowest concentrations between 20–10 m depth, with undetectable levels of Br and Cl in samples located between 16 and 9 m. The fluctuations in Br and Cl concentrations in the silt/clay unit between 58 and 20 m are likely due to dilution of the saline water by pulses of fresh meltwater from glacier melting (cf. Broster et al. 2013). Undetectable amounts of Br and Cl between 16–9 m indicate a net outflow across the area with no influx of brackish water. This change is also coincident with coarse-grained sediment and gravel suggesting higher flow velocities due to significant fluvial activity and possible erosion of upper parts of the marine/lacustrine unit. An increase in Br, Cl and Cl/Br ratio values in the upper 6 to 3 m (Fig. 5) is likely due to the initiation of modern estuarine conditions.

Analysis of drill logs indicates the stratigraphy can be described by six lithostratigraphic units (Fig. 4). Sandstone bedrock (unit 1) underlies the study site and is overlain by englacial and ablation till (unit 2).The till changes to interbedded sand and pebbly sand in places and is interpreted as an ice-contact/glacifluvial deposit (unit 3), most likely originating as a subaqueous outwash from melting of a glacier ice front. Till and glacifluvial sediments represent a glacial retreat sequence that grades upward into layered clayey silt at the base of a thick waterlain deposit (unit 4) informally referred to here as the marine/lacustrine unit.

Clay-silt sediments, exposed locally in raised terraces, were mapped by Rampton et al. (1984) as of undifferentiated marine and lacustrine origin. As well, Broster et al. (2013) drilled and sampled the clay-silt unit at Fredericton (Fig. 1) and found both attributes of lacustrine sedimentation, such as zones with distinct varve-like laminae, as well as poorly stratified zones of low-strength sediments characteristic of marine deposits. Differences and oscillations in Cl mg/l concentrations, within the marine/lacustrine unit at Fredericton were attributed to deposition in an estuarine environment, experiencing occasional dilution of brackish conditions by pulses of fresh water from melting glaciers (Broster et al. 2013).The marine/lacustrine unit sampled in GLM-01 demonstrates Cl changes similar to those found by drilling at Fredericton, except Cl concentrations here (exceeding 300 mg/l) are an order of magnitude higher than those at Fredericton. The higher concentrations are likely due to a greater marine influence at the location of the study site, being closer to the central, and likely deeper, parts of the marine/lake basin than is Fredericton, located 40 km farther up-river and possibly closer to a melting ice front.

The marine/lacustrine unit is interpreted to represent two phases of development of the Grand Lake area: (1) a lower part representing initial deposition in an open water marine environment, changing upward into an inland sea likely from isostatic rebound and captured marine water, and (2) water in the basin becoming increasingly fresher and lacustrine-like. Overlying units represent a fluvial phase (unit 5), which in turn is overlain by the establishment of the present lake and fluvial floodplain conditions (unit 6). The waterlain deposits of units 4, 5 and 6 represent four distinct phases of evolution of the lower Saint John River Valley and Grand Lake Meadows area including, (I) an open DeGeer Sea, (II) an inland DeGeer Sea and Lake Acadia, (III) a fluvial regime, and (IV) development of present conditions.

The Grand Lake area has experienced four significant periods of change since the Late Wisconsinan Glaciation as follows, using a calibrated time scale:

1. Phase I occurred more than 15 000 ya and represents deglaciation in a proglacial marine environment and formation of the DeGeer Sea by a marine incursion that occupied isostatically depressed portions of southeastern New Brunswick.

2. Phase II, between ~14  000 and 8000 calyBP, accompanied the end of the major isostatic readjustment through the region, and the development of an inland sea and its transformation into the predominantly freshwater Lake Acadia. Vegetation reclaimed the area as temperatures increased and remnant ice masses slowly disappeared from the headwaters of tributaries to the Saint John River.

3. Phase III occurred between ~8500 to 3000 calyBP. Sediments in this interval indicate major draining of the extended Lake Acadia and return of the lower Saint John River Valley to a fluvial-dominated system, likely initiating the beginning of the modern Grand Lake Meadows complex. This phase is associated with the down-cutting of the Reversing Falls gorge, capture of new tributaries, abandonment of raised terraces and construction of new alluvial landforms along the Saint John River.

4. Phase IV began after ~3000 calyBP and denotes the establishment of estuarine conditions in the Saint John River and an increase in brackish water to the area from breaching of the Reversing Falls gorge by marine water from rising sea level and higher tides in the Bay of Fundy.

From analyses of the sedimentary record in GLM- 01 it is clear that glaciers did not re-occupy the area after 13  000 calyBP (including the Younger Dryas cooling event). Occurrence of wood fragments in the upper half of unit 4 (waterlain suspension deposits) indicates that the area was likely vegetated prior to 13 000 calyBP. Flora and fauna adapted to local conditions even though original Cl content within the DeGeer Sea Phase, exceeded 300 mg/l. The modern Grand Lakes Meadows complex has existed for less than 3000 years and has contributed to preservation of Grand Lake as a freshwater resource.

The data examined here indicates that brackish water has occupied the area for approximately fourteen millennia and that the present supply of potable water may not be a sustainable resource. While Grand Lake and local estuaries have become less brackish over the last 8000 years, these freshwater resources should be considered as finite and susceptible to salinization from climate change and anthropogenic over-extraction. Sea level is predicted to rise with future climate warming, thus increasing saline intrusion from an accompanying rise in the saline groundwater table and from higher tides and further upriver migration of brackish water along estuaries. If local surface and groundwater extraction remains unchecked, brackish water may be drawn up from pockets of relict (glacial-age) saline-rich water in some areas (Giudice and Broster 2006) or from stratified lake and groundwater sources containing brackish water at depth. While the human taste-threshold for chlorine in water is between 200 mg/l and 300 mg/l depending upon associated cations (Health Canada 1996), some fauna and flora demonstrate a greater sensitivity to Chlorine. A program to identify and monitor areas were potable water resources could be threatened should be identified is a priority for research and litigation.