Sea-level change and paleogeographic reconstructions, southern Vancouver Island, British Columbia, Canada

Sea-level change and paleogeographic reconstructions, southern Vancouver Island, British Columbia, Canada

Quaternary Science Reviews 28 (2009) 1200–1216 Contents lists available at ScienceDirect Quaternary Science Reviews journal homepage: www.elsevier.c...

2MB Sizes 0 Downloads 20 Views

Quaternary Science Reviews 28 (2009) 1200–1216

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Sea-level change and paleogeographic reconstructions, southern Vancouver Island, British Columbia, Canada Thomas James a, b, *, Evan J. Gowan b, a, Ian Hutchinson c, John J. Clague d, J. Vaughn Barrie a, b, Kim W. Conway a a

Pacific Geoscience Centre, Geological Survey of Canada – Pacific, 9860 W. Saanich Road, Sidney, BC, Canada V8L 4B2 School of Earth and Ocean Sciences, University of Victoria, P.O. Box 3055 STN CSC, Victoria, BC, Canada V8W 3P6 Department of Geography, Simon Fraser University, 8888 University Drive, Burnaby, BC, Canada V5A 1S6 d Department of Earth Sciences, Simon Fraser University, 8888 University Drive, Burnaby, BC, Canada V5A 1S6 b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 June 2008 Received in revised form 10 December 2008 Accepted 21 December 2008

Forty-eight new and previously published radiocarbon ages constrain deglacial and postglacial sea levels on southern Vancouver Island, British Columbia. Sea level fell rapidly from its high stand of about þ75 m elevation just before 14 000 cal BP (12 000 radiocarbon yrs BP) to below the present shoreline by 13 200 cal BP (11 400 radiocarbon years BP). The sea fell below its present level 1000 years later in the central Strait of Georgia and 2000 years later in the northern Strait of Georgia, reflecting regional differences in ice sheet retreat and downwasting. Direct observations only constrain the low stand to be below 11 m and above 40 m. Analysis of the crustal isostatic depression with equations utilizing exponential decay functions appropriate to the Cascadia subduction zone, however, places the low stand at 30  5 m at about 11 200 cal BP (9800 BP). The inferred low stand for southern Vancouver Island, when compared to the sea-level curve previously derived for the central Strait of Georgia to the northwest, generates differential isostatic depression that is consistent with the expected crustal response between the two regions. Morphologic and sub-bottom features previously interpreted to indicate a low stand of 50 to 65 m are re-evaluated and found to be consistent with a low stand of 30  5 m. Submarine banks in eastern Juan de Fuca Strait were emergent at the time of the low stand, but marine passages persisted between southern Vancouver Island and the mainland. The crustal uplift presently occurring in response to the Late Pleistocene collapse of the southwestern sector of the Cordilleran Ice Sheet amounts to about 0.1 mm/yr. The small glacial isostatic adjustment rate is a consequence of low-viscosity mantle in this tectonically active region. Crown Copyright Ó 2008 Published by Elsevier Ltd. All rights reserved.

1. Introduction Perched deltas and marine deposits indicate that relative sea level was higher than present during terminal Pleistocene deglaciation in coastal areas of British Columbia and Washington state (Easterbrook, 1963; Mathews et al., 1970; Thorson, 1980; Armstrong, 1981; Clague et al., 1982; Domack, 1983; Dethier et al., 1995). The Lateglacial marine limit in southwest British Columbia and

* Corresponding author. Pacific Geoscience Centre, Geological Survey of Canada – Pacific, 9860 W. Saanich Road, Sidney, BC, Canada V8L 4B2. Tel.: þ1 250 363 6403; fax: þ1 250 363 6739. E-mail addresses: [email protected] (T. James), [email protected] (E.J. Gowan), [email protected] (I. Hutchinson), [email protected] (J.J. Clague), [email protected] nrcan.gc.ca (J.V. Barrie), [email protected] (K.W. Conway).

northwest Washington ranges from more than 200 m elevation in the east side of the Strait of Georgia to about 25 m on northwest Vancouver Island (Clague et al., 1982). Eustatic sea-level change is the average change in sea level over all ocean basins arising from a change in the total volume of water in the oceans. Eustatic sea level was significantly lower than present during the Late Pleistocene and early Holocene because large quantities of water resided in the continental ice sheets (Fairbanks, 1989; Hanebuth et al., 2000; Yokoyama et al., 2000). Because local relative sea level was higher than present during deglaciation in coastal British Columbia, the Cordilleran ice sheet, which covered most of British Columbia and parts of northern Washington (Fig. 1), must have isostatically depressed the surface of the Earth. The magnitude of the isostatic depression more than compensated for the lower eustatic sea level and generated local relative sea level that was higher than present.

0277-3791/$ – see front matter Crown Copyright Ó 2008 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.quascirev.2008.12.022

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

1201

Fig. 1. Location map showing the northern Cascadia subduction zone (after James et al., 2005) and the study area on southern Vancouver Island. Also shown are locations of previously constructed sea-level curves: central Strait of Georgia (Hutchinson et al., 2004a) and northern Strait of Georgia (James et al., 2005). Dashed contour lines show the depth to the subducting oceanic lithosphere. The thick line shows the maximum extent of the Cordilleran ice sheet (after Clague, 1983).

This paper presents new and previously published radiocarbon ages that constrain relative sea-level change in and around the city of Victoria on southern Vancouver Island. We infer the isostatic response from the sea-level curve to determine the rate and magnitude of postglacial uplift and the probable position of the sea-level low stand. Previously published curves for the central and northern Strait of Georgia provide context for the Victoria-area curve. Paleogeographic maps generated from the sea-level curve show inundated and subaerially exposed land in deglacial and postglacial times. 2. Glacial history and tectonic setting The Cordilleran ice sheet nucleated in the British Columbia Coast Mountains during the Late Pleistocene and flowed onto coastal lowlands and the interior plateaux (Clague, 1989). The glacier flowed along the Strait of Georgia and split into two lobes at the north end of Puget Sound – the Puget and Juan de Fuca lobes. The south-flowing Puget lobe reached its maximum limit in southern Puget Sound about 14 000 radiocarbon years ago (BP) (Thorson, 1989; Porter and Swanson, 1998). The west-flowing Juan de Fuca lobe extended onto the continental shelf off northernmost Washington state (Herzer and Bornhold, 1982). Ice was about 2000 m thick in the Strait of Georgia (James et al., 2000) and overtopped the mountainous spine of Vancouver Island (Alley and Chatwin, 1979). Glacial striations in Victoria record southward flow, although flow was probably to the west, and confined to Juan de Fuca Strait, during retreat (Alley and Chatwin, 1979). The southwest margin of the ice sheet rapidly thinned and retreated during Lateglacial time. Juan de Fuca Strait was deglaciated by about 13 600 BP (Mosher and Hewitt, 2004), and the Strait of Georgia was icefree before 12 000 BP (Barrie and Conway, 2002). Modern ice cover in British Columbia was attained shortly after 10 000 BP (Clague, 1981). The study area lies above the northern Cascadia subduction zone, where the Juan de Fuca plate subducts beneath North

America at a rate of about 4 cm/yr (Fig. 1). The Juan de Fuca plate is young (6–9 Ma at the trench, Wilson, 1993), and the shallow mantle beneath the subduction zone has a low viscosity (James et al., 2000, 2005). Spatial differences in Holocene sea level in the study region of a few metres amplitude may result, in part, from tectonic activity (e.g., Clague et al., 1982; Kelsey et al., 2004). 3. Previous sea-level work Mathews et al. (1970) constructed the first radiocarbon-constrained sea-level curve for the Victoria area on southern Vancouver Island. They argued that the Colwood Delta (Fig. 2), at 65–80 m elevation, marks the Lateglacial marine limit, and that sea level fell from about 75 m to its present position within a 2000-year period near the end of the Pleistocene. They cited Lamplugh (1886), who noted leached marine shells and hardened sediments in excavations at depths of 9–11 m in Esquimalt Harbour, to argue that the sediments had been subaerially exposed and that sea level had dropped below 11 m elevation. Clague et al. (1982) commented that during the first half of the Holocene, sea level remained below 4 m elevation. They also stated that sea level never rose above þ1.5 m elevation during the Holocene. Linden and Schurer (1988) collected cores and seismic reflection profiles in Juan de Fuca Strait near Victoria. The seismic profiles revealed a regional unconformity to a depth of 70 m, developed on an acoustically transparent unit composed of glaciomarine sediments. The unconformity is covered by coarser postglacial sediments. Linden and Schurer (1988) suggested that the unconformity was produced by subaerial exposure and by erosion in a shallow marine environment when sea level was about 50 m lower than today. James et al. (2002) and Gowan (2007) presented preliminary results from cores collected for this study. James et al. (2002) noted an inconsistency in bulk basal gyttja ages, a problem addressed by Hutchinson et al. (2004b). They concluded that sea level fell from

1202

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

determine the dominant diatom species and infer the environment of deposition. Radiocarbon dating was done at IsoTrace Laboratory (University of Toronto) by the accelerator mass spectrometry (AMS) method. It is now standard practice to normalize radiocarbon ages to a d13C value of 25&, but previously published radiocarbon ages from the Geological Survey of Canada (GSC) Radiocarbon Laboratory that are pertinent to our study were not normalized. Five unnormalized marine shell ages were reported without a d13C measurement. Marine shells usually have a d13C value near 0& so the GSC shell ages were normalized by adding 400 years (Stuiver and Polach, 1977) to the reported ages. As well, a GSC age on freshwater shells with a reported d13C value of þ1& (GSC-1130, Lowdon et al., 1971) was normalized by adding 420 years to the reported age (Stuiver et al., 2005). A reservoir correction of 950  50 years was applied to marine shells older than 10 000 radiocarbon years, whereas a correction of  720  90 years was applied to younger marine shells (Hutchinson et al., 2004b). Bulk organic ages on basal gyttja were adjusted by  625  60 years to account for a reservoir effect that may arise from leaching of old carbon into groundwater immediately after deglaciation or emergence from the ocean (Hutchinson et al., 2004b). The ‘‘old carbon’’ effect decreases rapidly after deglaciation or emergence, and the correction is only applied to gyttja directly above the marine–freshwater contact. Calibration of the radiocarbon ages was carried out using the program Calib 5 (Stuiver and Reimer, 1993). Terrestrial samples were calibrated using the Intcal04 calibration data set (Reimer et al., 2004), and marine samples were calibrated using the Marine04 data set (Hughen et al., 2004). 5. Results 5.1. Stratigraphy and radiocarbon ages Fig. 2. Site locations of samples used to constrain the sea-level curve.

above 60 m to below the present datum between 12 500 BP and 11 500 BP, and that the rate of sea-level fall decreased after 12 000 BP. Mosher and Hewitt (2004) conducted multibeam and seismic reflection surveys and collected cores to map glacier retreat in Juan de Fuca Strait and to infer the sea-level low stand offshore southern Vancouver Island. The multibeam and reflection surveys revealed terraces and ridges on the seafloor at 15, 35, 50 and 65 m. Mosher and Hewitt (2004) interpreted the terraces to be erosional features cut by waves when sea level was lower than today. They also noted features at 80 to 90 m depth and attributed those to shallow-water erosion. Assuming an erosional wave base of about 25 m, they estimated that the sea-level low stand was between 55 and 65 m. 4. Methods We cored isolation basins to retrieve and radiocarbon-date sediments marking the transition from marine to freshwater conditions or, in the case of marine basins, from freshwater to marine conditions. Six sites above sea level and two sites below sea level were cored. We determined site elevations by GPS occupations or with an altimeter calibrated to the high tide line; elevations are estimated to be accurate to 2 m. Percussion cores (Reasoner, 1993) were collected from five lakes and two marine basins, and a vibracore was recovered from one bog. The environment of deposition was inferred from the texture and color of the sediments and from the macrofossils contained within them. Diatom analysis was carried out on selected segments of cores to

We defined a sea-level curve for southern Vancouver Island with 48 radiocarbon ages (Table 1). Twenty-four of the 48 ages are from isolation basins cored during this study, and 24 are previously published radiocarbon ages. The stratigraphy of cores collected for this study and sites described by previous workers is shown in Figs. 3 and 4. Most cores collected from freshwater lakes and bogs below the marine limit contain basal marine or glaciomarine clastic sediments sharply overlain by gyttja and peat. Marine cores consist of organic-rich marine sediments underlain by glaciomarine sediments (Linden and Schurer, 1988; Mosher and Hewitt, 2004). At shallow depths in protected inlets, basal glaciomarine sediments are commonly overlain by early Holocene freshwater organic sediments capped by mid- and late-Holocene marine sediments (Foster, 1972). 5.1.1. Cusheon Lake A core collected from Cusheon Lake at 95 m elevation on Saltspring Island comprises about 2 m of freshwater gyttja overlying sandy clay (Fig. 3). The sandy clay lacks marine shells and is barren of diatoms, suggesting that it is not a marine deposit. The age of the basal gyttja is 9415  100 BP, which is a minimum age for deglaciation of the site. 5.1.2. Colwood Delta A piece of wood collected at þ76 m elevation from horizontally bedded topset beds of the Colwood Delta (Monahan et al., 2000; V. Levson, personal communication, 2002) yielded a radiocarbon age of 12 360  70 BP, which is a maximum age for the sea-level high stand in the Victoria area.

Table 1 Radiocarbon ages of samples used for constraining postglacial sea level in the Victoria region. Locationa Cusheon Lake Colwood Delta1 O’Donnell Flats O’Donnell Flats O’Donnell Flats Pike Lake Pike Lake Pike Lake Maltby Lake Maltby Lake Maltby Lake Prior Lake Prior Lake Prior Lake Gardner Pond2 Blenkinsop Lake3 McKenzie Ave.4 Matheson Lake Matheson Lake Patricia Bay5 Saanichton5 Rithets Bog4 Cook St.6 Cook St.6 Cook St.6 Cook St.6 Portage Inlet7

Site (Fig. 2) Latitude ( N)

Longitude ( W)

Elevation (m)

Material dated

Lab no.

Radiocarbon ageb Corrected agec Calibrated age (1 S.D.)

Sea-level position

Basal gyttja Wood Plant detritus Basal gyttja Shell (Nuculana?) Plant fragments Plant detritus Shell (Nuculana? valves) Plant fragments Basal organic mud Shell fragments Twig Twig Shell fragments Bison skull Shell Shell Plant fragments Plant fragments Shell Pelecypod fragment Gyttja Shell Freshwater shell Plant material Gyttja Peat

TO-10851 B-109128 TO-9193 TO-9194 TO-9195 TO-9190 TO-9191 TO-9192

10 040  80 12 360  70 11100  80 12 620  90 13 170  80 10 890  330 12 280  120 13 240  80

9415  100 12 360  70 11100  80 11 995  108 12 220  94 10 890  330 12 280  120 12 290  94

10 498–10 786 14 128–14 566 12 938–13 083 13 754–13 970 13 941–14 204 12 397–13 197 13 982–14 458 14 007–14 392

Above Marginal Below Marginal Above Below Above Above

TO-9181 TO-9182 TO-9183 TO-9186 TO-9187 TO-9189 SFU-549 GSC-246 GSC-763 TO-9184 TO-9185 GSC-418 GSC-398 GSC-945 GSC-1114 GSC-1130 GSC-1131 GSC-1142 GSC-4830

10 600  140 12 620  90 13 320  90 11 540  330 12 320  100 13 070  90 11750  110 13 060  80 13 120  80 12 210  100 12 120  100 13 150  85 12 840  115 11 400  95 12 500  80 11 620  85 11 500  80 11 200  95 6220  80

10 600  140 11 995  108 12 370  103 11 540  330 12 320  100 12 120  103 11750  110 12 110  94 12 170  94 12 210  100 12 120  100 12 200  99 11890  125 10 775  112 11 550  94 11 620  85 11 500  80 11 200  95 6220  80

12 395–12 804 13 754–13 970 14 129–14 596 13 118–13 735 14 046–14 489 13 845–14 073 13 472–13 720 13 844–14 052 13 904–14 133 13 925–14 206 13 852–14 076 13 919–14 179 13 606–13 891 12 709–12 876 13 275–13 474 13 366–13 580 13 269–13 413 13 020–13 193 7015–7247

Below Marginal Above Marginal Marginal Above Below Above Above Marginal Marginal Above Above Below Above Below Below Below Below

1 2 3 3 3 4 4 4

48.817 48.455 48.541 48.541 48.541 48.488 48.488 48.488

123.468 123.540 123.416 123.416 123.416 123.468 123.468 123.468

90 75 65 65 65 60 60 60

5 5 5 6 6 6 7 8 9 10 10 11 12 13 14 14 14 14 15

48.497 48.497 48.497 48.476 48.476 48.476 48.683 48.475 48.461 48.361 48.361 48.658 48.592 48.483 48.413 48.413 48.413 48.413 48.463

123.449 123.449 123.449 123.466 123.466 123.466 123.433 123.350 123.443 123.597 123.597 123.433 123.392 123.383 123.353 123.353 123.353 123.353 123.423

53 53 53 38 38 38 30 27 26 23 23 20 18 15 1 1 1 1 2

Locationa

Site Latitude (Fig. 2) ( N)

Longitude ( W)

Altitude (m)

Material dated

Lab no.

Radiocarbon ageb

Corrected agec Calibrated age (1 S.D.)

Sea-level position

Helmcken Park8 Portage Inlet9 Portage Inlet9 Portage Inlet9 Portage Inlet 01-01 Portage Inlet 01-01 Portage Inlet 01-01 Anderson Cove 01-03

15 15 15 15 15 15 15 16

48.460 48.463 48.463 48.463 48.459 48.459 48.459 48.361

123.428 123.422 123.422 123.422 123.422 123.422 123.422 123.659

2 2 2 2 2 2 2 4

GSC-4731 I-3673 I-3676 I-3675 TO-9885 TO-9886 TO-9887 TO-9888

8580  65 5470  115 9250  140 11700  170 4010  50 11170  80 13 140  80 4430  50

8580  65 5470  115 9250  140 11700  170 3290  103 11170  80 12 190  94 3710  103

9438–9739 6032–6404 10 250–10 575 13 389–13 727 3461–3727 12 972–13 140 13 921–14 159 4009–4311

Below Below Below Marginal Above Below Above Above

Anderson C. 01-03 16 Anderson Cove 01-04 16

48.361 48.361

123.659 123.660

4 4

TO-9889 TO-9890

6900  60 5100  70

6900  60 5100  70

7673–7792 5749–5917

Below Above

Anderson C. 01-04 Anderson C. 01-04 Anderson C. 01-04 Juan de Fuca Strait10 Juan de Fuca Strait10 Juan de Fuca Strait10 Juan de Fuca Strait10 Juan de Fuca Strait10 Juan de Fuca Strait11 Juan de Fuca Strait10 Juan de Fuca Strait10

48.361 48.361 48.361 48.420 48.415 48.415 48.415 48.415 48.398 48.400 48.400

123.660 123.660 123.660 123.430 123.427 123.427 123.426 123.426 123.372 123.414 123.414

4 4 4 32.8 41.7 42.5 42.7 44 55 60.5 61.3

Peat Peat Peat Phragmites frag. Shell (Ostrea lurida) Peat Shell fragments Shell (Saxidomus giganteus) Peat Plant and wood fragments Wood fragments Bark fragments (?) Peat Shell Shell Shell Shell Shell Shell Shell Shell

8160  80 7760  80 9010  80 8910  50 8490  50 13 370  50 10 640  50 9880  50 9670  140 10 720  60 13 690  50

8160  80 7760  80 9010  80 8190  103 7770  103 12 420  71 9690  71 9160  103 8950  166 9770  78 12 740  71

9011–9248 8434–8600 9941–10 248 9118–9382 8535–8841 14 212–14 620 11 029–11196 10 296–10 519 9949–10 399 11114–11 229 14 912–15 168

Below Below Below Above Above Above Above Above Above Above Above

16 16 16 17 17 17 17 17 17 17 17

TO-9891 TO-9892 TO-9893 CAMS-62767 CAMS-62533 CAMS-62534 CAMS-58684 CAMS-58685 RIDDL-265 CAMS-58695 CAMS-58696

a All dates are from this study unless noted: 1Monahan et al. (2000), 2Hebda (1988), 3Dyck et al. (1965), 4Lowdon and Blake (1970), 5Dyck et al. (1966), 6Lowdon et al. (1971), McNeely and Jorgensen (1993), 8McNeely and Jorgensen (1992), 9Buckley and Willis (1970), 10Mosher and Hewitt (2004) and 11Linden and Schurer (1988). b Ages are normalized to d13C ¼ 25.0& and uncertainties are 1s. c Corrections applied include: 950  50 yr for marine samples >10 000 yr BP; 720  90 yr for marine samples <10 000 yr BP; 625  60 yr for bulk organic ages from basal gyttja.

7

5.1.3. O’Donnell Flats A 7.8-m vibracore from a bog at 65 m elevation at O’Donnell Flats consists mainly of peat. The peat grades down into a thin gyttja, which sharply overlies a clayey silt with some very fine sand (Fig. 3). Diatom analysis indicated that the clayey silt is marine in origin (Fig. 5). The gyttja and basal peat were deposited, respectively, in a shallow freshwater pond and a boggy pond. Marine shell fragments, possibly Nuculana sp., from the top of the mud yielded a corrected radiocarbon age of 12 220  94 BP. Bulk organic sediment from the base of the gyttja layer gave a corrected age of 11 995  108 BP. Plant detritus from the base of the peat yielded an age of 11100  80 BP. The three radiocarbon ages indicate that sea level dropped below 65 m sometime between 12 200 and 12 000 BP.

5.1.4. Pike Lake A 5.45 m core recovered from Pike Lake (elevation 60 m) consists of clayey silt overlain by gyttja (Fig. 3). The two units are separated by a 4-cm-thick layer of organic-rich mud and a 3-cmthick layer of weakly laminated gyttja. The contacts between the layers are sharp. The gyttja above the weakly laminated layer is dark brown and massive and becomes less dense upward in the core. Mazama tephra (6800 BP) occurs at 2.01 m depth. Plant detritus from the clayey silt returned an age of 12 280  120 BP. An articulated marine shell, possibly Nuculana sp., from the same level yielded a reservoir-corrected age of 12 290  94 BP. These ages indicate that sea level was above w60 m at 12 300 BP. Plant fragments from a depth of 3.32–3.33 m at the transition between clayey silt and gyttja gave an age of

1204

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Fig. 3. Stratigraphy of cores collected for this study. The stratigraphy of a section at the Cook Street site is based on a written description in Lowdon et al. (1971). Ages are in corrected radiocarbon years BP. The symbols S and G indicate sharp and gradual contacts, respectively.

10 890  330 BP, indicating that sea level had dropped below 60 m by 10 900 BP. 5.1.5. Maltby Lake A 4.1 m core recovered from Maltby Lake (elevation 53 m) consists of mud sharply overlain by gyttja at 2.94 m (Fig. 3). The mud becomes increasingly laminated and organic-rich above 3.19 m. Mazama tephra occurs at 1.58 m depth. Marine shell fragments recovered from mud near the bottom of the core gave a reservoir-corrected age of 12 370  103 BP. A bulk sample of laminated organic mud at 3.13–3.18 m yielded a corrected age of 11 995  108 BP. Plant fragments from mud just below the contact with gyttja gave an age of 10 600  140 BP. The laminated mud likely represents the initiation of freshwater sedimentation, indicating that sea level fell below 53 m elevation after 12 400 BP but before 12 000 BP. 5.1.6. Prior Lake A 7.3 m core from Prior Lake (elevation 38 m) comprises mud gradationally overlain by gyttja between 2.01 and 2.34 m depth (Fig. 3). Mount Mazama tephra occurs at 1.58 m. A marine shell, possibly Nuculana sp., in mud at 4.25 m depth yielded a reservoir-corrected age of 12 120  103 BP. A twig at 2.40 m, near the top of the mud, gave an age of 12 320  100 BP. A

second twig at 2.26 m, in the transition from mud to gyttja, yielded an age of 11 540  330 BP. It dates the transition from marine to freshwater conditions. Sea level thus probably fell below 38 m elevation between 12 100 BP and 11 500 BP. 5.1.7. Gardner Pond A bison skull was excavated from marl overlying marine clay at an elevation of about 30 m at Gardner Pond (Mackie, 1987). The skull yielded a radiocarbon age of 11750  110 BP (Hebda, 1988), suggesting that sea level dropped below 30 m before that time. 5.1.8. Blenkinsop Lake Marine shells (Mya truncata) at an elevation of 27 m at Blenkinsop Lake gave a reservoir-corrected age of 12 110  94 BP (Dyck et al., 1965). Sea level at this site thus fell below 27 m sometime after 12 110 BP. 5.1.9. McKenzie Ave. Whole shells (Hiatella arctica) were collected from a shell-rich layer at 26 m elevation in a drill hole on McKenzie Ave. in Victoria (Lowdon and Blake, 1970). The shelly layer overlies silty clay and is overlain by peat. The corrected age of the sample is 12 170  94 BP. Sea level thus dropped below 26 m at this locality sometime after 12 170 BP.

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

1205

Fig. 4. Stratigraphy of cores collected from Portage Inlet and Anderson Cove. The interpreted stratigraphy of a trench dug at Portage Inlet (Foster, 1972) is also shown. Ages are in corrected radiocarbon years BP. The symbols S and G indicate sharp and gradual contacts, respectively.

5.1.10. Matheson Lake A 4.4 m core collected from Matheson Lake (elevation 23 m) consists of mud and sand overlain by gyttja (Fig. 3). The lowest 0.65 m of the core is massive mud with scattered pebbles. Above this (3.56–3.75 m) is medium sand grading up into silty, very fine sand with rip-up clasts of clay, plant detritus and shell fragments. The sand, in turn, grades upward into mud and organic-rich mud (3.28– 3.56 m). Above this (3.13–3.28 m) is a transitional layer separating the mud from overlying gyttja. Mazama tephra occurs at 1.37 m depth. Two radiocarbon samples were taken from mud overlying the sand layer. Plant fragments at 3.52 myielded a radiocarbon age of 12120  100 BP. Plant fragments 5 cm higher in the core gave an age of 12 210  100 BP. The sand layer records relatively high-energy conditions and may indicate that Matheson Lake was within the tidal range when it was deposited. The mud layer directly above the sand may record a brackish phase or initiation of freshwater sedimentation when sea level dropped below the elevation of Matheson Lake at, or shortly before, 12 200 BP.

11 890  125 BP (Dyck et al., 1966). The delta is graded to a shoreline at 18 m elevation. Sea level thus was above 18 m at this site until after 11 900 BP. 5.1.13. Rithets Bog A sample of gyttja from Rithets Bog (elevation 15 m) gave a corrected age of 10 775  112 BP (Lowdon and Blake, 1970). The sample was 5–8 cm above the contact between the gyttja and underlying marine clay, suggesting that sea level fell below þ15 m after 10 800 BP.

5.1.11. Patricia Bay Marine shells (Saxidomus sp.) collected from a gravelly shore deposit overlying marine clay at Patricia Bay yielded a corrected age of 12 200  94 BP (Dyck et al., 1966). The top of the gravel unit is 24 m above sea level, and the shells likely date to the time sea level was at this elevation.

5.1.14. Cook Street An excavation on Cook Street in Victoria exposed freshwater sediments overlying marine clay at about 1 m elevation (Lowdon et al., 1971, Fig. 3). A marine shell (Saxidomus giganteus) 0.45 m below the contact yielded a corrected age of 11 550  94 BP. Freshwater shells above the contact returned an age of 11 620  85 BP. Bulk plant material enclosing the freshwater shells gave an age of 11 500  80 BP, and black organic muck 15–18 cm above the freshwater shells yielded an age of 11 200  95. If all four radiocarbon ages are considered, the freshwater shell age appears to be slightly too old. Lowdon et al. (1971) noted that the sample was not preleached owing to its small size. The lack of preleaching may explain the discordant age. Based on the other three ages, we conclude that sea level fell below 1 m elevation after 11 550 BP but before 11 200 BP.

5.1.12. Saanichton Marine shell fragments recovered from clay underlying gravelly deltaic deposits at a gravel pit in Saanichton gave a corrected age of

5.1.15. Portage Inlet/Helmcken Park A 2.76 m percussion core was collected from Portage Inlet, which has a bedrock sill at an elevation of about 2 m. The core

1206

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Fig. 5. Dominant diatom species and inferred paleo-environments are given for selected intervals of cores from O’Donnell Flats and Anderson Cove.

consists of two mud units separated by peat (Fig. 4). The lower mud unit contains marine shells and black organic streaks, and grades up into olive-grey organic-rich mud (1.63–2.31 m). It is overlain by reddish-brown to reddish-grey muddy peat. A grey to black mud with abundant marine shells overlies the peat. Marine shell fragments near the base of the core gave a corrected age of 12 190  94 BP. Peat at 1.60–1.62, just above the lower mud, returned an age of 11170  80 BP. A marine shell (Ostrea lurida) at 1.29 m in the upper mud gave a corrected age of 3290  103 BP.

A trench was excavated on the north shore of Portage Inlet in a previous study (Fig. 4, Foster, 1972). A plant fragment (Phragmites sp.) from the basal organic-rich mud in the trench yielded an age of 11700  170 BP (Buckley and Willis, 1970). Peat directly overlying the mud gave an age of 9250  140 BP, and higher peat associated with Mazama tephra was dated at 6670  120 BP. Peat directly underlying the upper marine mud yielded an age of 5470  115 BP. The base of a freshwater peat at Helmcken Park on the west side of Portage Inlet gave an age of 8580  65 BP (McNeely and

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Jorgensen, 1992). Peat underlying silty sand on the north shore of Portage Inlet yielded an age of 6220  80 BP (McNeely and Jorgensen, 1993). The data from Portage Inlet indicate that sea level dropped below the elevation of the bedrock sill (2 m) after 12 200 BP but before 11 200 BP. Sea level rose above the sill elevation sometime after 5500 BP but before 3300 BP. 5.1.16. Anderson Cove Two cores were collected from Anderson Cove (Fig. 4), which has a sill at about 4 m. One of the two cores is 2.28 m long and consists of peat sharply overlain by dark grey, fine sandy silt. Peat just below the contact with the sandy silt gave an age of 6900  60 BP. A marine shell (S. giganteus) near the base of the sandy silt yielded a corrected age of 3710  103 BP. The second core is 2.92 m long and comprises peat and muddy peat sharply overlain by muddy sand at 1.90 m depth. Peat at 2.85 m, near the base of the core, gave an age of 9010  80 BP. A piece of bark from the top of the muddy peat at 1.94 m yielded an age of 7760  80 BP. Wood fragments in muddy fine sand at 1.70 m depth gave an age of 8160  80 BP. This age seems too old, possibly indicating that the wood fragments are derived from an older unit and have been reworked. Plant and wood fragments at 1.2 m depth, in a silty fine sand containing marine shells, gave an age of 5100  70 BP. Diatom analysis (Fig. 5) indicates that the muddy peat was deposited in a shallow freshwater marsh. The ages from the peat layers indicate that sea level remained below 4 m between 9000 and 6900 BP, but rose above this elevation before 5100 BP. 5.1.17. Juan de Fuca Strait Linden and Schurer (1988) and Mosher and Hewitt (2004) sampled marine sediments offshore of Victoria (Fig. 2, Table 1). Eight radiocarbon ages on marine shells collected at depths ranging from 32.8 to 61.3 m loosely constrain the sea-level low stand. The ages range from 12 740 to 8190 BP. Four of the cores described by Mosher and Hewitt (2004) have a basal grey or green silty clay that is sharply overlain by coarser sediments (Tul99-18, 42 m; Tul97-02, 44 m; Tul99-17, 46.7 m; and Tul97-12, 60.4 m; see their Fig. 10). The lithology of the clay, its location beneath a coarse lag deposit, and two radiocarbon ages from contained shell material (12 420  71 BP; 12 740  71 BP) suggest a glaciomarine origin. The presence of shell material and the grey and green coloration suggest that subaerial weathering did not extend below about 40 m depth. The lack of weathering in these four cores contrasts with conspicuous oxidation at shallower depths. 5.2. Sea-level curve for Victoria The stratigraphy and radiocarbon chronology of the cores provide tight constraints on the sea-level history of southern Vancouver Island at elevations above 4 m (Figs. 6 and 7). The time of the high stand is provided by the radiocarbon age on the wood sample at þ76 m, which is near the maximum elevation of the Colwood Delta (80 m). Sea level was near 70 m at 14 000 cal BP (12 000 BP), but had fallen to about 30 m by 13 700 cal BP (11750 BP). The rate of sea-level fall then may have slowed slightly, but the sea was below its present-day level around 13 200 cal BP (11 400 BP). A slowing in the rate of sea-level fall at lower elevations is consistent with the development of a coarse marine facies at Matheson Lake, which contrasts with the presence of only fine sediments at higher sites. Matheson Lake may have spent a longer time in shallow water, where wave and tidal action would produce coarser sediments.

1207

The inferred sea-level curve is consistent with most of the radiocarbon ages, but some of them require comment. Four dated samples, inferred to closely mark the transition from marine to freshwater environments, lie below the inferred sea-level curve, suggesting that they may be too old or that their environmental interpretation is incorrect. A twig from Prior Lake (þ38 m) is slightly older than a marine shell that lies below it in the core. It may be older than the sediments from which it was recovered; that is, it may be reworked. Two dated samples from Matheson Lake (þ23 m) are unidentified plant material collected above sand interpreted to have been deposited in the intertidal zone. The plant material could be marine in origin, in which case it would require a reservoir correction, but no identification was made. Alternatively, the sand may have been deposited in deeper marine water than we have inferred. The Phragmites fragment recovered from clay underlying peat in Portage Inlet suggests a marsh at 1 m at 11700  170 BP. The Cook Street data suggest, however, that sea level dropped below 1 m elevation between 11 550 BP and 11 200 BP. The standard deviation of the Phragmites age is large, and at two sigma, the age is consistent with the interpretation that sea level fell below 1 m elevation around 11 400 BP. An alternative explanation for the four discordant ages is that the marine reservoir correction of 950  50 years is too large. A reservoir correction of 850 years would shift the sea-level curves shown in Figs. 6 and 7 100 years to the left and reduce the slight discrepancy between some of the dated samples and their inferred relation to sea level. Given the alternative explanations discussed above, and noting that the statistics of the radiocarbon age distributions permit some samples to fall outside the one-sigma measurement uncertainty, we retain the 950  50 years marine reservoir correction proposed by Hutchinson et al. (2004b) for samples older than 10 000 BP. Our data do not strongly constrain the low stand off southern Vancouver Island, however, the sea rose above 4 m elevation by about 6000 cal BP (5000 BP). The rate at which the sea returned to near its present level is uncertain because it depends on the magnitude of the low stand. Sea level has remained near its present position for the past few thousand years (Hutchinson, 1992). 5.3. Sea-level low stand and analysis of isostatic depression Few radiocarbon ages exist at elevations between 4 and 40 m. Subaerial exposure and weathering of glaciomarine sediments to least 11 m (Lamplugh, 1886; Mathews et al., 1970) provide a minimum elevation for the low stand. If, however, the rate of sealevel fall that we have documented from þ30 m to 0 m is extrapolated to the end of the Pleistocene, the low stand could be about 40 m. A deeper low stand would require acceleration in the fall of sea level below 4 m. This scenario is unlikely because the sea falls most rapidly immediately after deglaciation when the rate of surface unloading is the largest (James et al., 2005). Furthermore, a low stand below 40 m, if it existed, could only have been shortlived, because ages on marine shells require sea level to lie above 40 m at 11 000 cal BP (9500 BP) and above 30 m at about 9000 cal BP (8000 BP). An analysis of the crustal isostatic depression supports our preferred low stand at about 30 m at 11 200 cal BP (9800 BP) (Figs. 6 and 7). Crustal isostatic depression, or vertical crustal displacement, can be derived from relative sea-level data provided global eustatic sea level is known or can be approximated (James et al., 2005). Using the Barbados sea-level curve (Fairbanks, 1989), James et al. (2005) derived the isostatic depression for the northern Strait of Georgia during deglacial and postglacial time by determining the difference between observed relative sea level and global eustatic

1208

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Fig. 6. Sea-level curve for southern Vancouver Island (timescale in radiocarbon years). The minimum and maximum low stands are based on subaerially exposed sediments at 11 m depth and constraining marine shell ages, respectively. The preferred curve is derived from an analysis of the crustal isostatic depression (Section 5.3).

Fig. 7. Inferred sea-level curve for the Victoria area (timescale in calibrated years). Calibrated ages are shown with their probability density functions. The preferred curve is derived from an analysis of the crustal isostatic depression (Section 5.3).

sea level. The isostatic depression was fit with two exponential functions with decay times of 500 and 2600 years. The short and long decay times were tentatively correlated with a shallow, lowviscosity mantle layer, and deeper, higher-viscosity mantle, respectively. In the first thousand years after deglaciation, the two decay functions combine to give an apparent decay time of 1200 years. At later times, the longer 2600-year decay function dominates. We repeated the analysis of James et al. (2005) with our southern Vancouver Island data set, subject to the caveat that significant uncertainty in the magnitude and time of the low stand precludes independent determination of decay times. Instead, the isostatic depression, like the relative sea-level curve, has an envelope of possible values in the Late Pleistocene and early Holocene, constrained by the minimum and maximum low-stand values (Fig. 8a). A graph of the isostatic depression on a semi-logarithmic plot (Fig. 8b) reveals that the early, rapid decrease in the depression had an apparent decay time of about 1000 years, which is similar to the 1200 years inferred for the northern Strait of Georgia. The crustal response of southern Vancouver Island thus is similar to that in the northern Strait of Georgia. The isostatic depression is well constrained in the first thousand years following deglaciation and again in the middle Holocene when the sea returned to near its present level. Applying two

exponential functions with decay times determined for the northern Strait of Georgia (500 and 2600 years) provides a good fit to the isostatic depression (Fig. 8c). The 500-year exponential curve has an initial amplitude of about 90 m and decays very rapidly. The 2600-year exponential curve has an initial amplitude of about 65 m. The exponential fit to the isostatic depression falls in the envelope determined by the minimum and maximum low-stand curves (Fig. 8d). Adding eustatic sea level to the exponential fit of the isostatic depression generates a relative sea-level curve (Fig. 9). This curve replicates the observed early, rapid fall of sea level and the return to near-present levels in the middle Holocene. It features a low stand of 30 m at about 11 200 cal BP (Fig. 9a), which is the preferred low stand shown in Figs. 6, 7 and 10. Uncertainty in the elevation of the low stand arises from uncertainties in the exponential fit to the isostatic depression and in the eustatic sea-level curve. In fitting the isostatic depression to the northern Strait of Georgia data, James et al. (2005) found that a 20% difference in decay times yielded fits nearly as good as optimal values, but that the fit degraded quickly for larger differences. The Victoria preferred low-stand value differs by 3 m for a 20% difference in assumed decay times. Decay times appropriate for the northern Strait of Georgia are unlikely to differ significantly from those for southern Vancouver

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Island. Both sites are located above the subducting Juan de Fuca plate. The top of the subducting plate is slightly shallower beneath Victoria than beneath the northern Strait of Georgia (Fig. 1), but this difference should not affect mantle viscosity beneath the subducting oceanic lithosphere. The age of the subducting plate at the trench west of the northern and southern Strait of Georgia differs slightly, but again the effect on the deeper mantle viscosity, which dominates the decay of the isostatic depression following the initial rapid uplift, would be minor. We thus adopt a value of 3 m as the uncertainty in the low stand arising from the exponential fit. Sea-level curves for localities far from former ice sheets (‘‘far-field curves’’) exhibit systematic differences. Some of the differences may be due to spatial differences in the gravitational effects of shrinking ice sheets (Clark et al., 2002), and some may arise from local hydroisostatic loading. Lambeck and Chappell (2001) corrected for the latter effect and commented that far-field curves give concordant results, considering observational and modeling uncertainties. Their

1209

stack of corrected far-field sea-level curves falls in an envelope with a height of about 15 m. Assuming that 15 m is the two-sigma (95%) range, 4 m is the one-sigma uncertainty in eustatic sea-level estimates. The decay time (3 m) and eustatic sea-level uncertainties (4 m) are independent and it is therefore appropriate to add them geometrically (square root of the sum of the squared uncertainties). This gives a net uncertainty of 5 m. In summary, our preferred low stand for southern Vancouver Island is 30  5 m. 6. Discussion 6.1. Comparison with previous work Fig. 10 compares the sea-level curve for southern Vancouver Island with previously published curves. Early researchers (Mathews et al., 1970; Clague et al., 1982; Linden and Schurer, 1988) did not use marine reservoir corrections, thus the sea-level fall from the high-stand position at þ75 to þ80 m occurs up to 1000 years earlier than in this study. The high-stand position was based on the maximum elevation of the Colwood Delta (65–80 m). The lowstand elevations of Mathews et al. (1970) and Clague et al. (1982) are higher than ours, but both groups of researchers recognized the constraint on the minimum low-stand position provided by Lamplugh’s (1886) observations. The falling limb of Linden and Schurer’s (1988) sea-level curve is similar to the early part of our curve, but has a much deeper low stand. They interpolated between the observed high stand at þ75 m and the assumed maximum depth of subaerial exposure (50 m) using the timing of Clague et al. (1982) for the Victoria area and Peterson et al. (1984) for Alsea Bay, Oregon. The sea-level curve of Mosher and Hewitt (2004, their Fig. 14) features a high stand of þ90 m. The elevation corresponds to the marine limit about 45 km north of the Colwood Delta (Huntley et al., 2001, their Fig. 4b). In their discussion of sea levels in eastern Juan de Fuca Strait, Mosher and Hewitt (2004) place the high stand at þ75 m, consistent with the elevation of the Colwood Delta. They argued for a deep low stand based on their interpretation of ocean-floor geomorphic features. Sea level fell from its upper limit earlier, and reached a substantially lower elevation (60 m), than favoured here because their curve passes through marine shell ages rather than above them. 6.2. Comparison to sea-level records from the Strait of Georgia

Fig. 8. (a) Derivation of crustal isostatic depression by subtracting observed relative sea level from assumed eustatic sea level. (b) Analysis of the decay time of the early segment of the isostatic response. (c) Exponential fit to the isostatic response using 500 and 2600-year decay times derived from a data set from the northern Strait of Georgia (James et al., 2005). (d) The exponential fit of the isostatic response falls inside the envelope defined by the minimum and maximum sea-level low-stand curves.

Comparison of the Victoria sea-level curve with previously published sea-level curves for the central and northern Strait of Georgia (Hutchinson et al., 2004a; James et al., 2005) reveals significant differences in the magnitude and timing of deglacial and postglacial sea-level change (Fig. 11a). The high stand at Victoria is þ75 m, much less than the þ150 to 175 m observed farther north and east. Sea level at Victoria dropped below the present datum about 1000 years earlier than in the central Strait of Georgia and more than 2000 years earlier than in the northern Strait of Georgia. The low stand was reached at about the same time at Victoria and in the central Strait of Georgia, but is about 15 m deeper at Victoria. The sea rose to within 2 m of its present level at Victoria at least 2000 years after doing so in the central and northern Strait of Georgia. The difference between the Victoria and central Strait of Georgia relative sea-level curves gives the amount of differential isostatic depression between the two locations (Fig. 11a). Unlike the calculations of total isostatic depression shown in Figs. 8 and 9, the differential isostatic depression does not require assumptions about imperfectly known eustatic sea-level positions because eustatic sea level is common to both relative sea-level curves. Taking the difference of two relative sea-level curves eliminates the eustatic signal and generates differential isostatic depression.

1210

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Fig. 9. (a) Derivation of the preferred low stand, obtained by adding the exponential fit to the crustal isostatic depression to assumed eustatic sea level. (b) The preferred sea-level low stand falls within the envelope defined by the minimum and maximum low stands.

In the first thousand years after sea level started to fall, the amount of differential isostatic depression decreased from 75 m to about 40 m. In other words, when sea level started to fall, the Earth’s surface was depressed 75 m more in the central Strait of Georgia than at Victoria. The distance between the two sites is about 140 km, thus the initial crustal tilt was about 0.55 m km1; this tilt decreased to 0.3 m km1 1000 years later. These values are similar to the peak crustal tilt of 0.45 m km1 between the central and northern Strait of Georgia, but are less than the peak tilts of 0.8 and 1.15 m km1 in the Puget Lowland to the south (Thorson, 1989; James et al., 2000). The larger crustal tilts in Puget Lowland suggest that the ice sheet front was steeper there than farther north, which is consistent with the inferred low profile of the glacier in the Strait of Georgia as it rapidly thinned and retreated (Barrie and Conway, 2002). The calculated value of crustal tilt between Victoria and the central Strait of Georgia differs significantly depending on whether the minimum, preferred, or maximum Victoria low stand is chosen (Fig. 11b). Owing to the exponential decay of the isostatic depression at each location, the differential isostatic depression would be expected to decay smoothly from its peak value to zero at present. The rate of decay should be large at first and small later. The differential isostatic depression generated by the minimum and maximum low stands does not display the expected smooth decline from a peak value. In the case of the maximum low stand, the differential isostatic depression drops rapidly from 14 000 to 12 000 cal BP, is constant between 12 000 and 10 000 cal BP, and then begins to decline again, initially slowly, but more rapidly after about 8500 cal BP. Changes in differential isostatic depression for the minimum low stand deviate even further from the expected behavior. The differential isostatic depression decreases to zero by 11 500 cal BP, then increases to 10 m shortly before 8000 cal BP before commencing a slow decline. The differential isostatic depression generated by the preferred sea-level curve is more consistent with expected behavior. It declines monotonically from its peak value and does not show the excursions and still-stands characteristic of the other curves. The differential isostatic depression generated by the preferred curve displays a break in slope at about 12 000 cal BP that would not be

expected to arise from the difference between decaying exponential functions. It could be caused by inaccuracies of a few metres in the low-stand portions of the relative sea-level curves.

Fig. 10. Comparison of the sea-level curve of this study with previously published sea-level curves for southern Vancouver Island.

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

1211

Fig. 11. (a) Comparison of the preferred Victoria sea-level curve with sea-level curves for the central Strait of Georgia (Hutchinson et al., 2004a) and northern Strait of Georgia (James et al., 2005). (b) Differential isostatic depression between Victoria and the central Strait of Georgia generated from the minimum, preferred, and maximum Victoria curves. (c) The same as (b), but with a correction for possible vertical tectonic motion between the two regions.

To assess the possible magnitude of regional variations in vertical crustal motion from tectonic activity, we compared Holocene sea-level curves from the Victoria area of southern Vancouver Island and the east coast of Vancouver Island (Hutchinson, 1992). Constraints on middle- and late-Holocene sea-level change are relatively sparse, but the curves indicate that the central Strait of Georgia has risen about 2.5 m in the past 4000 years relative to Victoria. About half of this amount is glacio-isostatic adjustment, suggesting that tectonic activity may have generated a few metres of relative vertical motion since deglaciation. The correction for possible vertical tectonic motion between the central Strait of Georgia and Victoria is 4–5 m since 14 000 cal BP (1.25 m/(4000 years)  14 000 years). Assuming that the correction can be applied linearly over the entire time range, we generate a new set of differential isostatic depression curves (Fig. 11c). The curves are not significantly different from those produced without a tectonic correction. The differential isostatic depression generated by the preferred Victoria sea-level curve is consistent with the expected crustal response, whereas the minimum and maximum curves are not.

6.3. Magnitude of the sea-level low stand The preferred 30  5 m low-stand value proposed here is substantially shallower than the low-stand depths of 50 m and 55 to 65 m proposed by Linden and Schurer (1988) and Mosher and Hewitt (2004), respectively. Here we propose alternative explanations for the evidence they present to argue for a deep low stand. Linden and Schurer (1988) described an irregular unconformity between glaciomarine sediments and overlying postglacial sediments. They suggested that erosion in a subaerial environment generated the irregular topography and thus placed the sea-level low stand at 50 m. A subaerial environment is not required, however, for the inferred erosion. If sea level was 30 m, the erosional wave base would have extended to about 55 m (Mosher and Hewitt, 2004). Subaerial exposure above 30 m, and wave and tidal action below 30 m, may be responsible for erosion of the glaciomarine surface. An unconformity with considerable relief does not require subaerial erosion, as Linden and Schurer (1988) argued. The paleosurface on which glaciomarine sedimentation commenced

1212

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Fig. 12. Paleogeographic reconstructions showing the extent of inundation and subaerial exposure at times of (a) the Lateglacial high stand and (b) the low stand. The reconstructions do not include ice cover.

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

1213

along a high-resolution boomer profile (their Fig. 8) intersects the ‘‘postglacial’’ unconformity. However, based on the core description (their Fig. 10) and a radiocarbon age, the basal unit is glaciomarine. Our previous comments regarding Linden and Schurer’s (1988) interpretation of this unconformity also apply here. Sea-level lowering undoubtedly influenced the erosional and depositional environments of Juan de Fuca Strait in Late Pleistocene and early Holocene time. Some of the features discussed by Linden and Schurer (1988) and Mosher and Hewitt (2004), such as the bank spillover deposits, are a consequence of sea-level lowering but not a direct indicator of sea level. Other features, such as the glaciomarine–postglacial unconformity, have explanations that involve sea-level lowering but also involve other factors. Finally, the ‘‘terraces’’ may have no relation to the sea-level low stand at all. 6.4. Paleogeographic reconstructions

Fig. 13. Time evolution of isostatic depression and rate of crustal uplift derived from the exponential fit to that depression.

following deglaciation was irregular, similar to most formerly glaciated surfaces. An airgun profile (Linden and Schurer, 1988, their Fig. 3) shows glaciomarine sediments draped over this irregular topography. The glaciomarine depositional surface was thus itself irregular, with much relief. This primary relief, combined with enhanced erosion of the glaciomarine sediments to depths of about 50 m, explains the observations just as well as a scenario in which sea level dropped to 50 m. Mosher and Hewitt (2004) proposed a sea-level low stand of 55 to 65 m. Some of the features they discussed are equally well explained with a low stand of 30  5 m. Wedge-shaped deposits of sediment at the edges of banks at depths of up to 90 m may have formed when the banks were shallower (Mosher and Hewitt, 2004). The tops of most of the banks are above 30 m, and would have passed through the erosional wave base, become subaerially exposed, and then transgressed in the early Holocene (Fig. 12b). During this period, the banks were eroded by waves. Currents could have delivered the eroded sediment to greater depths to be deposited as bank spillover deposits (Shaw and Forbes, 1992). Mosher and Hewitt (2004) attributed shelf-like features (‘‘terraces’’) at depths of 15, 35, 50, 55, 65, 75 and 83 m to erosion at a time of lower sea level, but alternative explanations should be considered. The terraces could reflect underlying bedrock topography. Some of them could be depositional landforms consisting of ice-contact sediments deposited in deep water during glacial retreat. The terraces possibly were eroded by strong seafloor currents. Faulting and other tectonic processes can generate linear bathymetric features such as ridges, scarps and cliff edges. Mosher and Hewitt (2004) described an unconformity within postglacial sediments and suggested that it was produced by erosion during their large sea-level low stand. It is likely that this postglacial unconformity is the above-described unconformity between glaciomarine and postglacial sediments. A core (Tul97-12) collected

We generated paleogeographic maps using a digital elevation model of present-day topography and bathymetry. For specific times, we shifted sea level vertically to elevations given by our sealevel curve. The paleogeographic reconstruction at the time of the sea-level high stand requires consideration of the spatial extent of glacier ice, but here we limit the discussion to the sea-level aspect. Paleogeographic maps for the sea-level high stand (þ75 m, 14 250 cal BP) and the preferred low stand (30 m, 11 200 cal BP) are shown in Fig. 12. A complete set of maps is provided in Appendix A. At the time of the high stand, much of Saanich Peninsula and San Juan Island was below sea level. Sooke Basin was substantially enlarged and had additional connections to Juan de Fuca Strait. The coastal lowland around Duncan is below sea level. Glaciomarine sediments accumulated in parts of the inundated area. At the time of the low stand, a coastal plain extended east of Saanich Peninsula. Haro Strait was narrower than today, in some places to less than half of its present width, but the deep channel west of San Juan Island had water depths locally in excess of 200 m. A water connection existed through Haro Strait and Boundary Pass between Juan de Fuca Strait to the west and the Strait of Georgia and northern reaches of Puget Sound to the north and east. Steep-sided Saanich Inlet differed little from today. Sooke Basin was above sea level and may have hosted a freshwater lake. Most of the shallow banks south and southeast of Victoria were partially emergent. 6.5. Present-day vertical crustal motion The rate of change through time of the exponential fit to the isostatic depression provides the crustal uplift rate (Fig. 13). When sea level began to fall, the Earth’s surface was rising faster than 100 mm/yr and possibly faster than 200 mm/yr. The rate of uplift decreased rapidly – it had dropped to less than 10 mm/yr by 11 000 cal BP and below 1 mm/yr before 5000 cal BP. At present the crust is rising at about 0.1 mm/yr. The current rate of crustal uplift of southern Vancouver Island is about half that in the northern Strait of Georgia (James et al., 2005), consistent with the location of Victoria near the periphery of the former Cordilleran ice sheet. The Strait of Georgia is farther inland from the periphery of the ice sheet and experienced thicker ice at the maximum of the last glaciation (James et al., 2000). The thicker ice generated larger amounts of isostatic depression, leading to larger rates of isostatic recovery. Our analysis is based on an isostatic depression history derived from the observed early rapid sea-level fall and recovery to nearpresent levels in the mid-Holocene. It does not account for broad-scale processes such as the collapse of the Laurentide ice sheet and water loading of the ocean basins. Thus, the derived uplift rate is an estimate of the residual isostatic response of the Earth to local retreat and

1214

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

collapse of the Cordilleran ice sheet. Computer modeling could provide a comprehensive view of the contributions from other surface loads. 7. Conclusions Sea level in southern Vancouver Island fell from a high-stand elevation of about þ75 m shortly before 14 000 cal BP to below its present level at 13 200 cal BP. This rapid fall is similar to that observed in the central and northern Strait of Georgia, although the high stand was lower, and sea level fell earlier, at Victoria. Constraints on sea-level change are fewer below 4 m. Direct observations only constrain the low stand to below 11 m and above 40 m. Our preferred low stand of 30  5 m is based on fitting crustal isostatic depression to crustal response times appropriate to the Cascadia subduction zone. Our low stand is consistent with published relative sea-level observations in the central Strait of Georgia. Previously proposed sea-level low stands of 50 m and deeper are inconsistent with the expected isostatic response of the Earth and provide a cautionary note on interpretations of geomorphic features on the seafloor. The paleogeographic reconstructions presented here provide insight into the deglacial and postglacial landscape. At the time of the sea-level low stand (11 200 cal BP), coastal plains were larger and marine passages narrower. Submarine banks in eastern Juan de Fuca Strait were emergent, but there was no land bridge along Haro Strait and Boundary Pass between Vancouver Island and mainland North America.

The crustal uplift presently occurring in response to the Late Pleistocene collapse of the southwestern sector of the Cordilleran Ice Sheet amounts to about 0.1 mm/yr on southern Vancouver Island. It is two orders of magnitude smaller than the crustal uplift rate at Churchill on the shores of Hudson Bay (Lambert et al., 2006). The small glacial isostatic adjustment rate is a consequence of lowviscosity mantle in this tectonically active region. Acknowledgements This work was conducted under the Georgia Basin Geohazards Initiative and is a product of the Enhancing Resilience to Climate Change Program. We thank the property owners who provided access to the lakes and bogs cored in this study; Vic Levson and Richard Hebda for useful discussions; Kevin Telmer for permitting us to sample the Cusheon Lake core; Robert Kung for generating the paleogeographic maps; Richard Franklin for drafting the figures; and Bill Hill, Jessica Jorna, Paul Ferguson, Michelle Watson, Matt Plotnikoff, Mike Sanborn, and Karen Simon for assistance in the field. Ian Shennan and Gene Domack are thanked for very thorough reviews that helped to substantially improve the manuscript. Support for Evan Gowan came from the University of Victoria, the Natural Sciences and Engineering Research Council (NSERC) of Canada, and the Geomatics for Informed Decision Making (GEOIDE) Network of Centres of Excellence (NCE). This paper is Earth Sciences Sector contribution number 20080257.

Appendix A Paleogeographic reconstructions are provided here for selected times, as described in Section 6.4 of the text.

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

References Alley, N.F., Chatwin, S.C., 1979. Late Pleistocene history and geomorphology, southwestern Vancouver Island, British Columbia. Canadian Journal of Earth Sciences 16, 1645–1657. Armstrong, J.E., 1981. Post-Vashon Wisconsin glaciation, Fraser Lowland, British Columbia. Geological Survey of Canada Bulletin 322, 34. Barrie, J.V., Conway, K.W., 2002. Contrasting glacial sedimentation processes and sea-level changes in two adjacent basins on the Pacific margin of Canada. In: Dowdeswell, J.A., Cofaigh, C.O. (Eds.), Glacier-influenced Sedimentation on High-latitude Continental Margins. Geological Society of London, Special Publication, vol. 203, pp. 181–194. Buckley, J.D., Willis, E.H., 1970. Isotopes’ radiocarbon measurements VIII. Radiocarbon 12, 87–129. Clague, J.J., 1981. Late Quaternary Geology and Geochronology of British Columbia, Part 2: Summary and Discussion of Radiocarbon-dated Quaternary History. Geological Survey of Canada, Energy Mines and Resources Canada, Paper 80-35, 41 pp. Clague, J.J., 1983. Glacio-isostatic effects of the Cordilleran ice-sheet, British Columbia, Canada. In: Smith, D.E., Dawson, A.G. (Eds.), Shorelines and Isostasy. Academic Press, London, pp. 321–343. Clague, J.J., 1989. Cordilleran Ice Sheet. In: Fulton, R.J. (Ed.), Quaternary Geology of Canada and Greenland, Vol. 1. Geological Survey of Canada, Geology of Canada, Ottawa, Ontario, pp. 40-42 (also Geological Society of America, The Geology of North America, Vol. K-1) Boulder, CO. Clague, J.J., Harper, J.R., Hebda, R.J., Howes, D.E., 1982. Late Quaternary sea levels and crustal movements, coastal British Columbia. Canadian Journal of Earth Sciences 19, 597–618. Clark, P.U., Mitrovica, J.X., Milne, G.A., Tamisiea, M.E., 2002. Sea-level fingerprinting as a direct test for the source of global meltwater pulse IA. Science 295, 2438–2441. Dethier, D.P., Pessl Jr., F., Keuler, R.F., Balzarini, M.A., Pevear, D.R., 1995. Late Wisconsinan glaciomarine deposition and isostatic rebound, northern Puget Lowland, Washington. Geological Society of America Bulletin 107, 1288–1303. Domack, E.W., 1983. Facies of Late Pleistocene glacial-marine sediments on Whidbey Island, Washington: an isostatic glacial-marine sequence. In: Molnia, B.F. (Ed.), Glacial-marine Sedimentation. Plenum Press, New York, NY, pp. 535–570. Dyck, W., Fyles, J.G., Blake Jr., W., 1965. Geological Survey of Canada radiocarbon dates IV. Radiocarbon 7, 24–46. Dyck, W., Lowdon, J.A., Fyles, J.G., Blake Jr., W., 1966. Geological Survey of Canada radiocarbon dates V. Radiocarbon 8, 96–127. Easterbrook, D.J., 1963. Late Pleistocene glacial events and relative sea-level changes in the northern Puget Sound, Washington. Geological Society of America Bulletin 74, 1465–1484. Fairbanks, R.G., 1989. A 17,000-year glacio-eustatic sea-level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342, 637–642. Foster, H.D., 1972. Geomorphology and water resource management: Portage Inlet, a case study on Vancouver Island. Canadian Geographer XVI, 128–143. Gowan, E.J., 2007. Glacio-isostatic Adjustment Modelling of Improved Relative Sea-level Observations in Southwestern British Columbia, Canada. M.Sc. thesis, University of Victoria, Victoria, BC, 148 pp. Hanebuth, T., Stattegger, K., Grootes, P.M., 2000. Rapid flooding of the Sunda Shelf: a Late-glacial sea-level record. Science 288, 1033–1035. Hebda, R., 1988. Buffalo George update. The Midden 20, 2. Herzer, R.H., Bornhold, B.D., 1982. Glaciation and post-glacial history of the continental shelf off southwestern Vancouver Island, British Columbia. Marine Geology 48, 285–319. Hughen, K.A., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Kromer, B., McCormac, F.G., Manning, S.W., Bronk Ramsey, C., Reimer, P.J., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J.,

1215

Weyhenmeyer, C.E., 2004. Marine04: marine radiocarbon age calibration, 26-0 ka BP. Radiocarbon 46, 1059–1086. Huntley, D.H., Bobrowsky, P.T., Clague, J.J., 2001. Ocean Drilling Program Leg 169S: surficial geology, stratigraphy and geomorphology of the Saanich Inlet area, southeastern Vancouver Island, British Columbia. Marine Geology 174, 247–241. Hutchinson, I.,1992. Holocene Sea Level Change in the Pacific Northwest: a Catalogue of Radiocarbon Dates and an Atlas of Regional Sea-level Curves. Occasional Paper No. 1, Institute of Quaternary Research, Simon Fraser University, Burnaby, BC, 100 pp. Hutchinson, I., James, T.S., Clague, J.J., Barrie, J.V., Conway, K.W., 2004a. Reconstruction of late Quaternary sea-level change in southwestern British Columbia from sediments in isolation basins. Boreas 33, 183–194. Hutchinson, I., James, T.S., Reimer, P.J., Bornhold, B.D., Clague, J.J., 2004b. Marine and limnic radiocarbon reservoir corrections for studies of late- and postglacial environments in Georgia Basin and Puget Lowland British Columbia, Canada and Washington, USA. Quaternary Research 61, 193–203. James, T.S., Clague, J.J., Wang, K., Hutchinson, I., 2000. Postglacial rebound at the northern Cascadia subduction zone. Quaternary Science Reviews 19, 1527–1541. James, T.S., Hutchinson, I., Clague, J.J., 2002. Improved Relative Sea-level Histories for Victoria and Vancouver, British Columbia, from Isolation Basin Coring. Geological Survey of Canada, Current Research 2002-A16, 7 pp. James, T.S., Hutchinson, I., Barrie, J.V., Conway, K.C., Mathews, D., 2005. Relative sealevel change in the northern Strait of Georgia, British Columbia. Ge´ographie physique et Quaternaire 59, 113–127. Kelsey, H.M., Sherrod, B., Johnson, S.Y., Dadisman, S.V., 2004. Land-level changes from a late Holocene earthquake in the northern Puget Lowland, Washington. Geology 32, 469–472. Lambeck, K., Chappell, J., 2001. Sea level change through the last glacial cycle. Science 292, 679–686. Lambert, A., Courtier, N., James, T.S., 2006. Long-term monitoring by absolute gravimetry: tides to postglacial rebound. Journal of Geodynamics 41, 307–317. Lamplugh, G.W., 1886. On glacial shell-beds in British Columbia. Quarterly Journal of the Geological Society of London 42, 276–286. Linden, R.H., Schurer, P.J., 1988. Sediment characteristics and sea-level history of Royal Roads Anchorage, Victoria, British Columbia. Canadian Journal of Earth Sciences 25, 1800–1810. Lowdon, J.A., Blake Jr., W., 1970. Geological Survey of Canada radiocarbon dates IX. Radiocarbon 12, 46–86. Lowdon, J.A., Robertson, I.M., Blake Jr., W., 1971. Geological Survey of Canada radiocarbon dates XI. Radiocarbon 13, 255–324. Mackie, Q., 1987. Marly bison turns up in pond. The Midden XIX (4), 2. Mathews, W.H., Fyles, J.G., Nasmith, H.W., 1970. Postglacial crustal movements in southwestern British Columbia and adjacent Washington state. Canadian Journal of Earth Sciences 7, 690–702. McNeely, R., Jorgensen, P.K., 1992. Geological Survey of Canada Radiocarbon Dates XXX. Geological Survey of Canada, Paper 90-7, 84 pp. McNeely, R., Jorgensen, P.K., 1993. Geological Survey of Canada Radiocarbon Dates XXXI. Geological Survey of Canada, Paper 91-7, 85 pp. Monahan, P.A., Levson, V.M., Henderson, P., Klohn-Crippen, A.S., 2000. Relative Liquefaction and Amplification of Ground Motion Hazard Maps of Greater Victoria (TRIM 92B.043, 044, 053, 054). British Columbia Ministry of Energy and Mines, Energy and Minerals Division, Geological Survey Branch, Geoscience Map 2000-3, 25 pp. Mosher, D.C., Hewitt, A.T., 2004. Late Quaternary deglaciation and sea-level history of eastern Juan de Fuca Strait, Cascadia. Quaternary International 121, 23–39. Peterson, C.D., Scheidegger, K.F., Schrader, H.J., 1984. Holocene depositional evolution of a small active-margin estuary of the northwestern United States. Marine Geology 59, 51–83. Porter, S.C., Swanson, T.W., 1998. Radiocarbon age constraints on rates of advance and retreat of the Puget Lobe of the Cordilleran ice-sheet during the last glaciation. Quaternary Research 50, 205–213. Reasoner, M.A., 1993. Equipment and procedure improvements for a light-weight, inexpensive, percussion core sampling system. Journal of Paleolimnology 8, 273–281.

1216

T. James et al. / Quaternary Science Reviews 28 (2009) 1200–1216

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, F.G., Manning, S.W., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04: terrestrial radiocarbon age calibration, 26-0 ka BP. Radiocarbon 46, 1029–1058. Shaw, J., Forbes, D.L., 1992. Barriers, barrier platforms, and spillover deposits in St. George’s Bay, Newfoundland; paraglacial sedimentation on the flanks of a deep coastal basin. Marine Geology 105, 119–140. Stuiver, M., Polach, H.A., 1977. Reporting of 14C data. Radiocarbon 19, 355–363.

Stuiver, M., Reimer, P.J., 1993. Extended 14C data base and revised Calib 3.0 14C age calibration program. Radiocarbon 35, 215–230. Stuiver, M., Reimer, P.J., Reimer, R.W., 2005. CALIB 5.0. (WWW program and documentation). Thorson, R.M., 1980. Ice-sheet glaciation of the Puget Lowland, Washington, during the Vashon Stade (Late Pleistocene). Quaternary Research 13, 303–321. Thorson, R.M., 1989. Glacio-isostatic response of the Puget Sound area, Washington. Geological Society of America Bulletin 101, 1163–1174. Wilson, D.S., 1993. Confidence intervals for motion and deformation of the Juan de Fuca Plate. Journal of Geophysical Research 98, 16053–16071. Yokoyama, Y., Lambeck, K., De Deckker, P., Johnston, P., Fifield, L.K., 2000. Timing of the Last Glacial Maximum from observed sea-level minima. Nature 406, 713–716.