P. Thompson Davis

Mid-Pleistocene cosmogenic minimum-age limits for pre-Wisconsinan glacial surfaces in southwestern Minnesota and southern Baffin Island: a multiple nuclide approach

Abstract Paired 10 Be and 26 Al analyses (n=14) indicate that pre-Wisconsinan, glaciated bedrock surfaces near the northern (Baffin Island) and southern (Minnesota) paleo-margins of the Laurentide Ice Sheet have long and complex histories of cosmic-ray exposure, including significant periods of partial or complete shielding from cosmic rays. Using the ratio, 26 Al / 10 Be , we calculate that striated outcrops of Sioux Quartzite in southwestern Minnesota (southern margin) were last overrun by ice at least 500,000 years ago. Weathered bedrock tors on the once-glaciated uplands of Baffin Island (northern margin) are eroding no faster than 1.1 m Myr−1, the equivalent of at least 450,000 years of surface and near-surface exposure. Our data demonstrate that exposure ages and erosion rates calculated from single nuclides can underestimate surface stability dramatically because any intermittent burial, and the resultant lowering of nuclide production rates and nuclide abundances, will remain undetected.

Cosmogenic radionuclides from fiord landscapes support differential erosion by overriding ice sheets

The interpretation of differentially weathered mountainous areas along the fringes of Pleistocene ice sheets is fundamental for determining ice-sheet behavior and thickness during the last glaciation. Two existing interpretations are either that highly weathered uplands remained as nunataks while freshly eroded troughs held outlet glaciers during the last glaciation or that uplands and lowlands were equally covered by ice, but that it was differentially erosive as a function of its spatially variable basal thermal regime. Cosmogenic radionuclide measurements from 33 bedrock samples and 27 upland erratics from differentially weathered fi ord landscapes on northeastern Baffi n Island shed light on Laurentide Ice Sheet (LIS) dynamics and thickness. Tors on weathered upland surfaces have minimum 10 Be ages between ca. 50 and ca. 170 ka (n = 12), whereas the majority of erratics perched in the uplands range from ca. 10 to ca. 13 ka (n = 14), indicating that the whole landscape was glaciated during the Last Glacial Maximum (LGM). Glacially sculpted bedrock (n = 7) near sea level refl ects the age of deglaciation, and increasing amounts of isotopic inheritance in highelevation sculpted bedrock (n = 3), higher elevation intermediately weathered bedrock (n = 11), and highest elevation intensely weathered bedrock refl ect the weakening of erosive power as the ice sheet transitioned from fi ords to interfi ord plateaus. Cold-based glaciation of uplands was contemporaneous with warm-based glaciation in fi ords, suggesting the presence of ice streams. 26 Al and 10 Be concentrations measured in 10 bedrock samples indicate that tors have experienced a complex history of alternating periods of exposure, burial, and limited glacial erosion. The paired isotope data reveal a minimum duration of burial of 80‐500 k.y., presumably by cold-based ice, and indicate that the tors have an age of at least 150‐580 ka. These data, together with other reconstructions of ice streams in the eastern Canadian Arctic, suggest that ice streams were a large infl uence on northeastern LIS dynamics throughout the late Quaternary. Thus, the northeastern LIS was sensitively tied with North Atlantic thermohaline circulation and abrupt climate change.

Holocene glacier fluctuations in the American Cordillera

Abstract In many areas of the American Rocky Mountains (Colorado, Wyoming, Utah, Idaho, Montana), Cascade Range (Washington, Oregon), and the Sierra Nevada (California), radiocarbon ages suggest that ice receded to near present limits before 10 ka BP. A pre-Altithermal readvance or stillstand left moraines ca. 1–3 km beyond and ca. 50–300 m below present glacier margins. At one locality on Glacier Peak in Washington, these deposits are perhaps dated to the early Holocene, but in general these deposits are probably at least 10 ka old. Glacial advances during the Altithermal (ca. 8 to 5 ka BP) elsewhere are rare; radiocarbon evidence for Altithermal glacial advances in the Colorado Front Range is questionable. The earliest radiocarbon-dated Neoglacial advances occurred about 5 ka BP in the Washington Cascades; moraines and related deposits believed to be of early Neoglacial age (i.e. 5 to 3 ka BP) elsewhere are poorly dated. For example, moraines believed to date to the early Neoglacial in Colorado (Triple Lakes) and Wyoming (Temple Lake) are actually late Pleistocene age, based on radiocarbon ages derived from lake sediments at the type localities. Although relative-age data (i.e. lichens, rock weathering, soils) from many areas suggest a three-fold Neoglacial sequence, little supporting radiocarbon evidence is available. However, tephrochronology distinguishes three Neoglacial advances near Mount Rainier in Washington. In most mountain ranges of the western United States, fresh, unweathered, sharp-crested moraines usually adjacent to present ice margins, or near headwalls in empty cirques, date to the Little Ice Age of the last several centuries. However, detailed photographic or other historical records of glacial fluctuations during the last century are rare, and detailed mass balance studies rarer still. Whether or not pre-Little Ice Age glacial fluctuations in the American Cordillera are synchronous must await better radiometric dating of local moraine sequences.

Cosmogenic 10Be and 26Al ages for the Last Glacial Maximum, eastern Baffin Island, Arctic Canada

A chronology for the Last Glacial Maximum based on cosmogenic exposure dating in the Pangnirtung Fjord area, eastern Baffin Island, Arctic Canada, is at odds with many decades of land-based, glacial stratigraphic research. Most previous chronologies focused primarily on relative weathering criteria and suggested that ice extent was restricted during the late Wisconsinan (ca. 24–8 ka). In contrast, by directly dating glacial features, we conclude that late Wisconsinan ice was far more extensive than previously believed. There are 36 gneissic boulders and 8 samples of ice-molded gneissic bedrock that yield late Wisconsinan 10 Be and 26 Al exposure ages for the last glaciation of Pangnirtung Fjord. The prominent Duval moraines, which were previously interpreted to represent a significant early Wisconsinan (100–60 ka) ice advance on southern Cumberland Peninsula, were actually deposited between 24 and 9 ka. Two boulders from a raised glaciomarine delta, stratigraphically related to the Duval moraines, date to about 10 ka. Two recessional moraines and striated bedrock along Pangnirtung Fjord, as well as erratics on the floor of the Kolik River valley, a tributary to Pangnirtung Fjord, indicate that deglaciation began between 12 and 9 ka. In situ produced 10 Be and 26 Al abundances indicate that ice filled Pangnirtung Fjord for about 15 k.y. (either continuously or intermittently) prior to 10 ka, which is compatible with 14 C chronologies for adjacent Cumberland Sound. Thus, our data support other recent studies that suggest the northern and southern margins of the Laurentide ice sheet were generally in phase during the latest Wisconsinan, contrary to earlier interpretations.

Potential glacial evidence for the younger dryas event in the Cordillera of North and South America

There is debate as to whether or not the Younger Dryas (YD) climatic event affected areas outside of the North Atlantic region. The potential of alpine glacial deposits to record the YD is of interest because alpine glaciers would be likely to leave a morainal record even during minor, short-lived advances caused by a period of deteriorating climate. In the Cordilleran ranges of North and South America there are several candidate moraines, many bracketed between radiocarbon ages of ca. 12 and 10 ka, which may be YD in age. The best examples are in the Canadian and American Rockies, the Ecuadorian, Peruvian and Bolivian Andes, and around the Patagonian Ice Cap. In most cases temporal brackets are not narrow enough and some radiocarbon ages may not be accurate enough to allow correlation at the present time. There are two cases of Lateglacial readvance whose ages are interpreted to match or fall within the YD interval: an advance of the Crowfoot Glacier in Canada is interpreted to have occurred within the period ca. 11.3 and 10 14C ka BP, and an advance of the Reschreiter Glacier in Ecuador is interpreted to have culminated in the period ca. 11 to 10.6 14C ka BP. The latter interpretation depends on an assumption that lake-level variations were caused by glacier advances and retreats.

Evaluation of proposed early-Holocene advances of alpine glaciers in the North Cascade Range, Washington State, USA: constraints provided by palaeoenvironmental reconstructions

Many palaeoenvironmental reconstructions from across western North America indicate that the early to mid-Holocene was warmer and drier than present. The wide distribution of these records suggests that relatively mild and arid conditions were regionally ubiquitous during the early Holocene. In contrast, two recently proposed advances of alpine glaciers in the Cascade Range of Washington State, corresponding to equilibrium-line altitude depressions of 350 to 450 m, are placed in the early Holocene. The juxtaposition of the proposed major expansion of early-Holocene alpine glaciers with the constraint provided by a number of palaeoenvironmental reconstructions from directly adjacent areas results in palaeoclimatic gradients that are untenable. The apparent diachronous behaviour of adjacent palaeoenvironmental records may be an artifact of dating errors.

Reconstructing lake and drainage basin history using terrestrial sediment layers: analysis of cores from a post-glacial lake in New England, USA

Four sediment cores and twenty-five 14C ages from Ritterbush Pond in northern Vermont provide a detailed and continuous temporal record of Holocene lake and watershed dynamics. Using visual logs, carbon content, magnetic susceptibility, stable isotope signatures, and X-radiography, all measured at 1-cm scale, we identify and date discrete layers of terrestrially-derived sediment in the organic-rich, lacustrine gyttja. These inorganic layers range in thickness from <1 mm to >10 cm and range in grain size and sorting from homogeneous silt to graded sand. AMS radiocarbon ages both from macrofossils within the thickest layers, and gyttja bracketing these layers, provide the basis for correlation among the cores, the dating of 52 basin-wide sedimentation events, and the development of a detailed sedimentation chronology for the Holocene.Physical, chemical, and isotopic analyses suggest the inorganic layers are terrestrially derived and result from hydrologic events large enough to erode and transport sediment from the watershed into the pond. The temporal and spatial distribution of the inorganic layers suggests changing basin-wide sedimentation and thus erosion dynamics since deglaciation over 12,000 years ago. Specifically, for intervals lasting 400 to 1000 years, during the early (>8600 cal yBP), middle (6400 to 6800 cal yBP) and late Holocene (1800 to 2600 cal yBP), the Ritterbush Pond watershed eroded more rapidly than at other times and terrestrially derived material poured into the pond. Analysis of Ritterbush Pond sediments demonstrates the potential for North American lakes to preserve a record of drainage basin dynamics.

Cold-based Laurentide ice covered New England’s highest summits during the Last Glacial Maximum

To better understand glacial history and process in New England (northeastern United States), a mountainous area overrun by the Laurentide Ice Sheet, we measured three cosmogenic nuclides in nine upland samples. The concentrations of 10Be and 26Al in some samples collected near the summits of Katahdin (Maine) and Mount Washington and Little Haystack Mountain (New Hampshire) are 2–10 times higher than expected for a single exposure period, considering field evidence indicating that continental ice-covered all New England peaks during the Last Glacial Maximum. In-situ 14C exposure ages from the summits are much younger, suggesting that high-elevation sampling sites were ice-covered before and during the Last Glacial Maximum. Field and isotopic data are consistent with New England summits being covered in part by cold-based continental ice that did not erode much rock. The contrast in erosion rates between stable summits and deeply eroded valleys likely contributes to the development and maintenance of northern Appalachian topography.

Response to Discussion by Wolfe et al. on Bierman et al. (Geomorphology 25 (1999) 25–39)

On the basis of radiocarbon ages in lake sediments collected near our field site in the Pangnirtung Fjord area and elsewhere on Baffin Island, Wolfe et al. argue for continued acceptance of the Atraditional weathering zone modelB and the existence of glacier-free highlands that functioned as refugia Ž . through at least the last glacial maximum LGM . Ž . Although our original paper Bierman et al., 1999 considers both of these topics only in passing, we appreciate the opportunity to discuss Wolfe et al.’s ideas in light of new results, both ours and theirs, obtained andror published after our original paper was submitted in 1997. In contrast to the certainty with which Wolfe et al. declare the continued validity of the weathering zone model and the existence of extensive ice-free refugia throughout Wisconsinan time, we still believe that data pertinent to defining the extent and duration of ice cover on the highlands of Baffin Ž Island remain contradictory Steig et al., 1998; Bier. man et al., 1999; Marsella et al., 2000 . However, on

Glacial history and landscape evolution of southern Cumberland Peninsula, Baffin Island, Canada, constrained by cosmogenic 10Be and 26Al

Since the first application of cosmogenic nuclides to the study of glacial history and processes in 1990, increasing numbers of studies have used a variety of cosmogenic isotopes to quantify the exposure age and erosion rate of glaciated landscapes. However, obtaining chronological data from glaciated landscapes once covered by cold-based, nonerosive ice is challenging because these surfaces violate assumptions associated with simple cosmogenic exposure dating. Nonerosive glacial ice fails to completely remove nuclides produced during previous periods of exposure, leaving behind rock surfaces with complex, multigenetic nuclide inventories. Here, we constrain the glacial history, landscape evolution, and efficiency of subglacial erosion in the Pangnirtung Fiord region of southern Baffin Island using over 300 paired analyses of in situ cosmogenic 10Be and 26Al. Simple exposure ages are 6.3–160 ka for 10Be ( n = 152) and 4.3–124 ka for 26Al ( n = 153). Paired bedrock-boulder samples have discordant ages, simple exposure ages generally increase with elevation, 10Be and 26Al ages for the same sample disagree, and both boulders and bedrock yield multimodal age distributions—all patterns suggestive of limited subglacial erosion. Measured 26Al/10Be ratios indicate that about one third of the samples in the data set experienced at least one period of pre-Holocene exposure followed by burial with limited erosion. Modeled two-isotope minimum-limiting exposure durations are as high as hundreds of thousands of years, and minimum-limiting burial durations range up to millions of years, implying that parts of southern Baffin Island have been preserved beneath nonerosive glacial ice for much, if not all, of the Quaternary. A subset of the samples contains few nuclides inherited from prior periods of exposure and is thus useful for constraining the chronology of the most recent deglaciation. Using these samples, we infer that deglaciation of most of the landscape occurred ca. 11.7 ka and that the Duval moraines, a prominent feature of the last deglaciation, formed ca. 11.2 ka.