Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 67 2016 NORTHEASTERN NATURALIST 23(1):67–87 Younger Dryas Paleoenvironments and Ice Dynamics in Northern Maine: A Multi-Proxy, Case History Ann C. Dieffenbacher-Krall1,*, Harold W. Borns Jr.2, Andrea M. Nurse2, Geneva E.C. Langley3, Sean Birkel2, Les C. Cwynar4, Lisa A. Doner5, Christopher C. Dorion6, James Fastook2, George L. Jacobson Jr.2, and Christopher Sayles7 Abstract - Geological evidence for modeled Younger Dryas ice expansion in northern Maine is assessed in conjunction with temperature and precipitation estimates from chironomids and pollen, and plant macrofossil and lake-level analyses from lake sediment. Pollen and chironomid temperature and precipitation transfer-function estimates for the Allerød warming period indicate colder winters, precipitation levels half that of modern times, and summer temperatures near modern levels. The combination of cold winters and low precipitation prevented forest establishment in northern Maine along the Maine/New Brunswick border. While winter temperatures and precipitation remained stable, summer temperatures decreased as much as 7.5 °C during the Younger Dryas stadial, forcing a shift from shrub-dominated to sedge-dominated tundra. Summer and winter temperatures, as well as annual precipitation, increased rapidly at the Holocene onset. Introduction A significant body of evidence firmly establishes the occurrence of Younger Dryas (YD) cooling in the Canadian Atlantic Provinces and adjacent northeastern United States between 13,000 and 11,600 calendar years before present (yrs BP; e.g., Borns et al. 2004, Cwynar and Levesque 1995, Mayle and Cwynar 1995b). Newman et al. (1985) and Borns et al. (2004) reported evidence for advance of ice in T9 R5 W.E.L.S., Aroostook County, ME (Oxbow region) during the YD. Using the University of Maine Ice Sheet Model (Fastook 1993, Fastook and Chapman 1989), Sayles (2004) modeled a persistent ice cap over the highlands of northern Aroostook County, ME, assuming YD temperatures were 8 °C lower than today’s average temperatures (Figs. 1, 2). The modeled ice cap receded during the warm Allerød period (13,700 to 13,000 yrs BP) and re-advanced during the colder YD. Despite undeniable physical evidence of YD ice advance in the Oxbow region (Fig. 1 insert), geographic and temporal distribution of resurgence of a residual ice cap over northern Maine is problematic. This paper presents physical evidence 1School of Biology and Ecology, 100 Murray Hall, University of Maine, Orono, ME 04469. 2Climate Change Institute, University of Maine, Orono, ME 04462. 3College of the Atlantic Herbarium, 105 Eden Street, Bar Harbor, ME 04609. 4Department of Biology, University of New Brunswick, Fredericton, NB, Canada E3B 5A3. 5Center for the Environment, Plymouth State University, NH 03264. 6CC Dorion Geological Services, 200 High Street, Suite 2D, Portland, ME 04101. 734 Middlesex Circle, Apartment 8, Waltham, MA 02451. *Corresponding author - annd@maine.edu. Manuscript Editor: Daniel Keppie Northeastern Naturalist 68 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 that constrains ice-mass persistence and movement to the Oxbow region as well as biological evidence for temperature, precipitation, and vegetation shifts during the late-glacial period in northern Maine. Following the last glacial maximum, portions of the eastern margin of the Laurentide Ice Sheet in contact with the sea on the outer edge of the continental shelf receded rapidly into the Gulf of Maine and the Gulf of St. Lawrence, producing calving embayments that progressed into central Maine, the Bay of Fundy, and the St. Lawrence Lowland. The embayment in the St. Lawrence Lowland reached the Quebec City area and produced the Champlain Sea about 14,400 yrs BP (Chauvin et al. 1985). To the south, the receding ice margin crossed the present position of the Maine, New Brunswick, and Nova Scotia coastlines around 15,500 Figure 1. Location of sites examined in the current study (site numbers correspond to Table 1): 2 = Matherson Pond (46°12'N, 68°12'W); 3 = Little Machias Lake (46°24'N, 68°18'W); 12 = Upper McNally Pond (46°24'N, 69°W); 17 = Island Pond (46°54'N, 68°42'W); 18 = Oxbow excavation site; and 19 = LaPomkeag Lake. Location of sites researched by other investigators and referenced in this study: 20 = Joe Lake (46°25'N, 66°40'5"W; Mayle et al. 1993a); 21 = Roulston Lake (46°53'40"N, 67°23'59"W; Mott et al. 1986); 22 = Splan Pond (45o51'55"N, 67o19'53"W; Mayle et al. 1993a); and 23 = Deep (Tilley) Lake (46°12'43"N, 67°23'59"W; Cwynar and Levesque 1995). Lakes containing YD-age lithic zone indicated by “+”; lakes lacking a clear YD lithic zone indicated by “o”. Enlarged area shows the Oxbow excavation site with direction of ice flow and associated mo raines. Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 69 Figure 2. Ice-cap model images by Sayles (2004) using University of Maine Ice Cap Model (Fastook and Chapman 1989, Fastook 1993). Color scale denotes ice thickness in centimeters. Aroostook highlands center around 46°35'N, 68°34'W. Site designations follow Figure 1 and Table 1. (a) By 13,000 yrs BP (end of Allerød warming) glacial ice retreated north of the St. Lawrence River.( b) Maximum ice advance at 11,600 BP (end of YD cold period) assuming mean summer temperatures 11 °C below current values. (c) Maximum ice advance at 11,600 BP assuming mean summer temperatures 8 °C below current values. Northeastern Naturalist 70 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 yrs BP (Borns 1985). The marine invasion of the St. Lawrence Lowland bisected the Laurentide Ice Sheet, leaving a residual ice plateau over most of west-central and northern Maine and establishing a new margin of the contiguous Laurentide Ice Sheet north of the St. Lawrence River. As the separated ice plateau over Maine began to melt and thin, it continued to flow toward its new margins (Lowell 1985). By 11,400 yrs BP, the large ice plateau wasted away to separate and stagnant ice masses in the lowest topographic areas (Davis and Jacobson 1985). We used temperature transfer-function models (inference models based on modern data sets) of chironomid (midge fly) larvae to evaluate late glacial through early Holocene temperature variations. These models assume that modern chironomid assemblages respond to temperature shifts the same way chironomids responded 13,000 years ago. Chironomids have been widely used to estimate late glacial temperature changes across northern United States and Canada (Cwynar and Levesque 1995; Levesque et al. 1993, 1997; Mayle and Cwynar 1995b; Walker et al. 1991a; Wilson et al. 1993). Pollen-inference models use statistical evaluations of hundreds of pollen records from the North American Pollen Database (NOAA 2003, 2013) to estimate temperature and precipitation (e.g., Fréchette et al. 2008, Juggins 2003, Peros and Gajewski 2008, Whitmore et al. 2005, Williams and Shuman 2008). We compared Conroy Lake temperature estimates from pollen inference models with Pennington Pond and Whitehead Lake temperature estimates from chironomid inference models. We then compared Conroy Lake pollen-inferred precipitation values with Whitehead Lake lake-level moisture-balance studies (Dieffenbacher-Krall and Nurse 2005). We also reviewed stratigraphic deposition data from 10 additional lakes in Aroostook County, ME (see Fig. 1 and Table 1 for site locations). Site Description H.W. Borns was present when Newman et al. (1985) discovered a convoluted peat layer sandwiched between layers of glacially deposited gravel in a freshly excavated roadside drainage ditch on the west side of Maine Rt. 11, 90 m south of the intersection with the Oxbow Road (Figs. 1, 3a). The 1-m-deep ditch exposed Figure 3. (a) Schematic interpretation of W–E view across the Oxbow site. (b) Exposed Oxbow site stratigraphy. Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 71 Table 1. Unpublished lake core dates for Cranberry, Conroy, and Pennington Ponds and basal dates (start of lake sediment accumulation) for lakes assessed for presence of a YD inorganic sediment interval. Radiocarbon dates (column 6) calibrated to calendar years (column 7) using Calib 5.0 (Stuiver et al. 2005) and IntCal04 (Reimer et al. 2004). Site numbers follow Figure 1. Depth in Radiocarbon Calendar Site #/Lake core Dated material Lab # age years BP Reference 1 Cranberry Pond 46°27'N, 68°52'W 405–400 Birch seeds, charcoal OS-5450 11,500 ± 60 13,330 This paper 430-425 Aquatic plant seeds OS-5449 12,050 ± 70 13,900 This paper 4 Whitehead Lake 46°27'N, 67°52'W Basal date Dryas leaves, stem OS- 38292 11,800 ± 65 13,800 Dieffenbacher-Krall and Nurse 2005 5 Conroy Lake 56°17’N, 67°53’W 595–597 Alnus (Alder) twig Beta-40101; 9090 ± 85 10,300 This paper, Doner 1995 ETH-7050 605 Plant fragment Beta-65419; 10,070 ± 60 11,580 This paper, Doner 1995 CAMS-8657 618 Woody twig TO-1722 10,450 ± 90 12,380 This paper, Doner 1995 633–636.5 Woody twig Beta-43723; 10,890 ± 90 12,900 This paper, Doner 1995 ETH-7829 680–684 Woody fragment Beta-40102; 11,135 ± 110 13,050 This paper, Doner 1995 ETH-7051 6 Caribou Lake 46°49'N, 68°4'W Basal date Terrestrial vegetation OS-5993 11,000 ± 160 12,915 Borns et al. 2004 7 Echo Lake 46°37'N, 68°0’W Basal date Terrestrial vegetation OS-3002 11,950 ± 190 13,800 Borns et al. 2004 8 Pennington Pond 46°56'N, 69°32'W 401.5 Stick OS-54672 9430 ± 55 10,650 This paper, Chase 2004 463–467 Dryas leaves, insect remains OS- 38208 10,700 ± 80 12,730 This paper, Chase 2004 468–472 Dryas leaves, insect remains OS- 38209 11,050 ± 45 12,980 This paper, Chase 2004 472–475 Dryas leaves, insect remains OS- 38289 11,350 ± 70 13,220 This paper, Chase 2004 476–480 Dryas leaves, insect remains OS- 38290 11,400 ± 65 13,260 This paper, Chase 2004 9 Young’s Lake 46°31'N, 67°57'W Basal date Terrestrial vegetation OS-5305 12,800 ± 100 15,100 Borns et al. 2004 10 Matthew’s Pond 46°19'N, 69°4'W 695–700 Bulk gyttja B-168755 11,470 ± 150 13,440 Dieffenbacher-Krall and Nurse 2005 11 Galilee Pond 46°57'N, 68°50'W Basal date Dryas leaves, woody stem B-222923 11,320 ± 40 13,210 Putnam and Putnam 2009 13 Black Lake 47°13'N, 68°29'W Basal date Terrestrial vegetation OS-4383 10,600 ± 30 12,685 Borns et al. 2004 14 First Pelletier Brook Lake Basal date Terrestrial vegetation OS-5303 11,500 ± 60 13,330 Borns et al. 2004 47°4'N, 68°55'W 15 Fischer Lake 46°43'N, 67°49'W Basal date Bulk sediment OS-4839 12,750 ± 40 15,060 Borns et al. 2004 16 Isie Lake 47°4'N, 68°39'W Basal date Moss OS-6683 10,900 ± 90 12,870 Borns et al. 2004 Northeastern Naturalist 72 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 peat masses within a poorly sorted glacial till. Four calibrated radiocarbon ages of the convoluted peat averaged 12,500 yrs BP; direction of ice flow was judged to be towards the south-southeast based on a strongly oriented cobble fabric in the till (Newman et al. 1985). Newman et al. (1985) concluded that the YD-age peat was overrun by glacier ice that expanded towards the south-southeast and incorporated the peat into the basal till (Fig. 3a). Dieffenbacher-Krall and Nurse cored Cranberry Pond in 2006. The pond is located 3 km east of the Oxbow site, and is bordered on the northwest by a low, lobate moraine. Lake cores consisted of 3 m of lacustrine sediment atop at least 2 m of grey, inorganic glacial lacustrine clay. C.C. Dorion cored over 40 lakes in Maine, including Caribou, Echo, Young’s, Black, First Pelletier Brook, Fisher, and Isie lakes (Fig. 1, Table 1). Lithology cited in this paper comes from Dorion’s field notes. Methods Lithology Because of the shallow depth of the Newman et al. (1985) exposure and in consideration of the site’s potential import to late-glacial history of the region, in 2004 we excavated and analyzed the stratigraphy of 2 deeper pits to verify the Newman et al. (1985) findings. We dug the first pit on the west side of Rt. 11, 0.24 km south of the intersection with the Oxbow Road. In the fall of 2004, we excavated a second pit on the west side of Maine Rt. 11, 0.32 km south of the Oxbow Road intersection. We applied a “lake-core method” for locating the YD-advanced ice margin in the northern Maine study area (Fig. 4). Sediment from biologically productive lakes is predominantly composed of organic material and is dark to olive brown in color. Sediment from cold, biologically inactive lakes has little organic content and is composed of fine sands and gray clays (lithic zone). Our method involved coring lake sediment down to glacial till, and determining whether or not the lake Figure 4. Lake sediment cores in lakes within the YD glacial advance lack an organic-poor lithic zone dating between 13,000 and 11,600 yrs BP, while lake sediment cores from lakes outside the glacial advance exhibit the organic-poor zone, e.g., Matherson Pond and Whitehead Lake. Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 73 contained an organic-poor lithic zone (first described by Deevey 1951) deposited during the cold YD interval between 13,000 and 11,600 yrs BP. Organic content was determined by loss-on-ignition analysis of lake sediment (550 °C burn for 2 hours to remove carbon-based organics following Bengtsson and Enell [2003]; LOI550). If glacial advance overran a lake basin during the YD, no characteristic lithic zone would be present and bottom sediment dates would be younger than 11,600 yrs BP (Fig. 4). In contrast, lake basins beyond the margin of the ice advance would contain the organic-poor lithic zone and bottom sediments would be older than 13,000 yrs BP. Chironomids C.C. Dorion collected a sediment core from Pennington Pond (T15 R6 W.E.L.S., Aroostook County, Maine) in 1996. The pond is 18.2 h in area with a maximum depth of 1.5 m. We retained the lowermost meter of sediment (400–500 cm; Fig. 5, Table 1) for this study. LOI550 was performed on contiguous, 2-cm3 samples along the entire meter section. To avoid “old” carbon contamination, radiocarbon dates were obtained on terrestrial plant and insect remains (Table 1). We processed sediment samples from Whitehead Lake (site 4; Fig. 1, Table 1; site and sediment cores described by Dieffenbacher-Krall and Nurse [2005]) and from Pennington Pond (site 8; Fig. 1, Table 1) for chironomid analysis according to methods of Walker et al. (1991b). We used CONISS programming (Grimm 1987) to define chironomid zones. We derived mean summer surface-water temperatures (Figs. 5, 6) from chironomid remains using the inference model developed by Walker et al. (1997). Pollen We selected a training subset of 1267 lacustrine sites (see Fig. 8A) and 82 pollen types from the modern pollen dataset compiled by Whitmore et al. (2005). Southern and western site boundaries were set at 39°48'N latitude and 104°W longitude to reduce influence of western and southern conifer species (Williams and Shuman 2008). Inference models predict past environmental parameters by comparing ancient species assemblages from within lake sediment with those found at modern sites; the models assume that similar environmental conditions result in similar assemblages. We generated pollen inference models (transfer functions) with the program C2 (Juggins 2003) for mean July air temperature (TJuly), mean January air temperature (TJan), and annual precipitation (Pann) using both the modern analog technique (MAT-5) and weighted averaging partial least squares, 2-component model (WAPLS-2). Square-chord distance dissimilarity coefficient averaging of the 5 closest analogs, performed with C2, tested whether the training set included samples analogous to each of the subfossil pollen samples. Performance statistics for all models were excellent (Table 2). The coefficients of determination (r2 boot), which compare measured and predicted values for the training set, ranged from 0.97 (MAT-5 TJuly) to 0.75 (WAPLS-2, Pann). We examined the performance statistics for all parameters in order to select the most reliable model. The MAT-5 model had a higher r2 boot , higher mean bias, lower or comparable maximum bias, and Northeastern Naturalist 74 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 Figure 5. Results of LOI550 and chironomid summary diagrams for Pennington Pond. Second graph from left y-axis is chironomid-inferred mean summer surface-water temperature in degrees Celsius. Chironomid x-axis shows species percent of total chironomid head capsules identified. Hatched areas indicate 10x exaggeration. See Table 1 for calibration of dates to calendar years before presen t (yrs BP). Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 75 Figure 6. Results of LOI550 and chironomid summary diagrams for Whitehead Lake. Second graph from left y-axis is chironomid-inferred mean summer surface-water temperature in degrees Celsius. Chironomid x-axis shows species percent of total chironomid head capsules identified. Hatched areas indicate 10x exaggeration. See Table 1 for calibration of dates to calendar years before presen t (yrs BP). Northeastern Naturalist 76 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 lower root mean square error of prediction (RMSEP). While MAT-5 performance statistics indicated greater reliability, we applied both models to Conroy Lake square root transformed pollen-percentage data because each model has different limitations (Birks 2005, Peros and Gajewski 2008). Borns et al. (2004) presented pollen analysis and radiocarbon results on Conroy Lake (Monticello, ME [site 5]; Fig. 1, Table 1). We reanalyzed the age-model for this site (Table 1), discarding 2 dates reported in Borns et al. (2004) in a slump zone indicated by a slanted, abrupt transition in sediment and multiple, abrupt, pollen transitions. In Conroy Lake, Pennington Pond, and Whitehead Lake sediment, we also used LOI925 (heating to 925 °C for 4 hours to burn off carbonate; Bengtsson and Enell 2003) to identify periods of increased carbonate deposition. Radiocarbon dates from Pennington Pond and Whitehead Lake (Dieffenbacher-Krall and Nurse 2005) constrain the Conroy Lake age-model. Results The first Oxbow pit we excavated in 2004 revealed a lower basal till overlain by 1.5 m of organic-rich sand layers. This unit, in turn, was overlain by basal till containing ripped-up masses of organic-rich sand. One calibrated radiocarbon date of 12,400 yrs BP from the organics in the sand layers (Borns et al. 2004) correlated with dates from Newman et al. (1985). In our second 2004 Oxbow pit, 45 cm of compact basal till was exposed at the bottom of the pit overlain by 90 cm of convoluted and sheared organic-rich sand with folds overturned to the south. Fivecm- thick layers of glacial till intercalated with the sand layers. Above the sand, a 90-cm layer of compact basal till contained a 100-cm elongated mass of rounded, ripped-up peat. Both the lower and upper tills were compact silts and clays with abundant glacially striated clasts whose long axes were strongly oriented WNW– ESE (Fig. 3b). Because the stratigraphy of both 2004 pits was consistent with that of the originally described road cut, we conclude that the organics exposed in all pits correlate and are YD in age. Three km east of the Oxbow site and along the ice-flow line at Cranberry Pond, the YD lithic zone is absent and dates of organic deposition initiation (13,280 and 14,000 BP) are older than YD cooling (Fig. 1, Table 1). Neither Galilee Pond Table 2. Performance statistics for pollen inference models generated by this study. Model r2 boot Mean bias Maximum bias RMSEP Mean July temperature (TJuly) Mat-5 analogs 0.97 0.07 2.61 1.1 WAPLS-2 component 0.92 0.03 3.62 1.6 Mean January temperature (TJan) Mat-5 analogs 0.91 0.41 3.23 2.8 WAPLS-2 component 0.82 0.10 4.49 3.5 Mean annual precipitation (Pann) Mat-5 analogs 0.83 15.54 414.12 144.5 WAPLS-2 component 0.75 -1.31 753.36 158.0 Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 77 (Putnam and Putnam 2009) nor First Pelletier Lake (Borns et al. 2004) exhibited the YD lithic zone but produced basal dates of 13,210 yrs BP and 13,330 yrs BP, respectively (Table 1). Cranberry Pond, Galilee Pond, and First Pelletier Brook Lake contain large accumulations of inorganic, glacial lacustrine clay suggesting a location proximal to melting glacial ice prior to onset of organic deposition. Whitehead Lake, Conroy Lake, Caribou Lake, Echo Lake, Pennington Pond, and Young’s Lake exhibit organic brown mud overlying glacial lacustrine clay. This deposit in turn is overlain by a low organic, YD lithic zone which is visually evident and clearly exhibited by LOI550 results (Borns et al. 2004, Dieffenbacher-Krall and Nurse 2005). An abrupt transition to gyttja (organic-rich lake sediment) ends the YD zone. Matthew’s Pond (Dieffenbacher-Krall and Nurse 2005) and Fischer Lake do not contain a visual YD lithic zone but have a high/low/high LOI550 section typical of YD deposition. Mathew’s Pond YD zone is 5 cm wide indicating that either the pond dried to puddle dimensions or was iced over during most of the YD. Basal dates older than the 13,000 yrs BP onset of YD cooling (Table 1) preclude YD ice scouring in these Aroostook lakes. The classic YD lithic zone is present further along the Oxbow flow line at Matherson Pond, 6 km south of the Oxbow site (Fig. 1). Rock outcrops north of Oxbow and northwest of Cranberry Pond display glacial striations representing flow in several directions, including lightly over-printed striations indicating flow generally towards the south and south-southeast. Low-lying end moraines east and south of the Oxbow site indicate an advancing, internally dynamic, ice margin. Borns identified a cluster of smaller moraines that cross-cut older, recessional moraines at LaPomkeag Lake, approximately 13 km west of the Oxbow site. Crosscutting moraines indicate either ice margin re-advance during late-glacial retreat or resurgence of a local ice mass. The distribution of lakes that do not contain the YD lithic zone, the moraine configuration around the Oxbow peat exposure (Fig. 1 insert), and basal dates for the onset of lake sedimentation at 3 strategically located lakes (Table 1) indicate that YD ice re-advance was not widespread. We conclude that ice advance was confined to re-activation of a large, isolated, local ice mass centered north-northwest of the Oxbow site. Whitehead Lake sediment older than 11,400 yrs BP consistently contained leaves of Dryas integrifolia Vahl (Dryas; Dieffenbacher-Krall and Nurse 2005), a dwarf, arctic shrub not found in Maine today, even in alpine locations. Carex (sedge) and Juncus (reed) seeds were also present, but there were few other terrestrial plant macrofossils. Pre-11,400 yrs BP sediment contained occasional shrub Betula spp. (birch) seeds and bracts, species common to cold, shrub-tundra environments. Dryas leaves disappeared in sediment after 11,400 yrs BP (Fig. 6, zone 4) and Larix laricina (Du Roi) K. Koch (Eastern Larch) and Picea (spruce) needles appeared (Dieffenbacher-Krall and Nurse 2005). LOI925 results for Conroy Lake, Pennington Pond (Fig. 5), and Whitehead Lake sediment demonstrate carbonate deposition prior to and following the YD interval, but very little carbonate deposition during the YD. The same pattern is apparent in Caribou, Echo, and Fischer Lakes. We consider this pattern to represent marl Northeastern Naturalist 78 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 deposition mediated by biological activity during warmer pre- and post-YD times, a process that ceased during the cold YD. Pennington Pond chironomids (Fig. 5) Basal sediment in zone 1 consisted of dense, gray clay with LOI550 < 4% (Fig. 5). Zone 2 sediment consisted of banded marl with several 0.5-cm bands of gray clay. LOI550 in this section was slightly higher, from 4 to 8%. YD sediment in zones 3a and 3b consisted of uniform, gray clay with LOI550 < 10%. LOI550 increased to 10–20% at the start of zone 4 with sediment of banded carbonate-rich marl. At the top of zone 4, early Holocene sediments transitioned to highly organic (LOI550 > 30%) gyttja. Chironomid zone 1 contained a high proportion of head capsules from the subtribe Tanytarsina as well as cold-tolerant genera Paracladius, Sergentia, Heterotrissocladius, Protanypus, and Corynoneura/Thienemanniella. Zone 2 was characterized by a near absence of cold-tolerant taxa, with the exception of Corynoneura/ Thienemanniella and a low level of Sergentia, a decline in subtribe Tanytarsina, and an increase in temperate taxa Psectrocladius, Chironomus, Microtendipes, Dictrotendipes, and tribe Pentaneurini. Zone 3 encompasses YD-age sediment. In zone 3a, beginning around 12,700 yrs BP, cold-tolerant genera Sergentia reached its highest level and subtribe Tanytarsina increased. Propsilocerus head capsules were present in sediment from this section, but were not present in any of the Walker et al. (1997) training sets; Propsilocerus was therefore not included in the chironomid inference model. Propsilocerus has been found only rarely in North America in subfossil form, but extant specimens were collected in 2009 in northern British Columbia (Cranston et al. 2011). Temperate taxa Chironomus, Microtendipes, Psectrocladius, Procladius, Dicrotendipes, and tribe Pentaneurini declined in abundance. Zone 3b assemblage composition was similar to that of zone 3a with a modest increase of Diamesa, Sergentia, and subtribe Tanytarsina, and decline of temperate Chironomus, Psectrocladius, and Dicrotendipes. Zone 4 is marked by the disappearance of all cold-tolerant chironomids and a rise in nearly all temperate taxa. Summer surface-water temperature estimates for Pennington Pond were 14 °C in the middle of zone 1 and rose to 18 °C by the end of the zone (Fig. 5). Temperature estimates averaged between 16.5 °C and 17 °C in zone 2. During zone 3a, the earlier part of the YD, summer surface-water temperatures declined to an average of 14 °C, reaching a low of 12.5 °C (5.5 °C below modern). Temperature estimates averaged close to 13 °C (6 °C below modern) during zone 3b, the later YD. The minimum YD-age temperature estimate, 9.7 °C, occurred at 455 cm (Fig. 5, zone 3a). Temperatures rose to just above 18 °C in zone 4, (temperatures equal to modern). Whitehead Lake chironomids (Fig. 6) Whitehead Lake zone 1 contained gray, low-organic material with LOI550 of 5%. Sediment in zone 2 was comprised of banded brown gyttja mixed with gray clay; LOI550 climbed from 6% at the start of the zone to a peak of 20%. Zones 3a and 3b contained gray clay with LOI550 dropping back to 13%. In zone 4, brownish gray Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 79 gyttja with LOI550 of 30% graded rapidly into organic, uniform, dark brown gyttja with LOI550 > 60%. Zone 1 contained a high proportion of cold-tolerant Sergentia, as well as head capsules from temperate genera Microtendipes and Cladopelma, and the ubiquitous subtribe Tanytarsina. The proportions of temperate taxa, including Chironomus, tribe Pentaneurini, Psectrocladius, Procladius, Dicrotendipes, Parakiefferiella cf. (comparable to) bathophila, and Polypedilum, as well as Ceratopogonidae increased in zone 2, while the proportion of Tanytarsina decreased in zone 2. Zone 3 encompasses YD-age sediment. In zone 3a beginning at 12,800 yrs BP, proportions of cold-tolerant taxa Sergentia, Protanypus, and Corynoneura/ Thienemanniella increased. This was the only zone in which Heterotrissocladius occurred in any abundance. Coincidentally, temperate taxa Chironomus, tribe Pentaneurini, Psectrocladius, Dicrotendipes, Cladopelma, and Parakiefferiella cf. bathophilia declined and Microtendipes, Cryptochironomus, Polypedilum, Endochironomus, and Ceratopogonidae disappeared. Zone 3b proportions of cold-tolerant Sergentia were similar to those of zone 3a, but Heterotrissocladius disappeared. Proportions of Chironomus declined, while proportions of tribe Pentaneurini and Psectrocladius were slightly higher than in zone 3a. Zone 4, beginning at 11,400 yrs BP, saw the disappearance of all cold-tolerant taxa and a resurgence of most temperate taxa. Summer surface-water temperature estimates for Whitehead Lake (Fig. 6) ranged from 13 °C to 15 °C during zone 1 and from 13 °C to 20 °C during zone 2. During zone 3a, the earlier part of the YD, summer surface-water temperatures ranged from 10 °C to 15.5 °C (mean of 12.5 °C was 6 °C below modern). Temperature estimates ranged from 10 °C to 14.5 °C during zone 3b (mean of 12 °C was 6.5 °C below modern). The minimum YD-age temperature estimate, 9.9 °C, occurred at the transition from zone 3a to zone 3b. In zone 4, temperatures rose from 15.5 °C to 21 °C (equal to modern temperatures). Conroy pollen Squared chord distance is a measure of how dissimilar the composition of pollen samples are to each other, with greater values indicating greater differences. Low squared chord distances indicate samples are more similar (analogous), implying that vegetation of the surrounding area was also similar. Squared chord distance of ≤0.15 (Overpeck et al. 1985) is a widely used threshold in North America for determining that pollen samples originated in the same vegetation zone. Frechette et al. (2008) used a squared chord distance threshold of ≤0.26 for Arctic biomes. Williams and Shuman (2008) determined squared chord distance thresholds of 0.188 and 0.194 for 64 taxa models for the Whitmore et al. (2005) training set. The majority of Conroy Lake fossil samples are analogous to samples within the modern data set (Fig. 7b). Only 6 of the 27 Conroy Lake samples have squared chord distances greater than 0.15, with a high of 0.19, still within Frechette et al. (2008) and Williams and Shuman (2008) guidelines. The average squared chord distance of the top 5 analogs (5 modern samples most similar to each fossil sample) for all Northeastern Naturalist 80 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 fossil samples is 0.136. We thus consider the MAT-5 models robust and more accurate than the WAPLS-2 models, which appear affected by “edge effects” (Peros Figure 7. (a) Pollen-inferred mean July and January temperatures for Conroy Lake and chironomid-inferred mean summer surface water temperatures for Whitehead Lake and Pennington Pond. Vertical lines indicate modern means for each site (BRIDGE 2008). (b) Mean annual precipitation for Conroy Lake from pollen inference models; Whitehead Lake multi-proxy lake-level curve (adapted from Dieffenbacher-Krall and Nurse 2005); Squared chord distance between Conroy Lake fossil pollen (Borns et al. 2004) and the nearest modern analog. Squared chord distances between 0.15 and 0.26 (bracketed by vertical dotted lines) are the range of values established for Arctic (Frechette et al. 2008) and North American (Williams and Shuman 2008) biomes. The majority of Conroy Lake pollen samples have square chord distances less than 0.15, indicating that a large number of samples in the modern data set are analogous to Conroy Lake samples during a given time period. Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 81 and Gajewski 2008), a tendency for weighted averaging-based methods to underestimate high values and overestimate low values. July temperatures for the earliest portion of the Conroy Lake record (Fig. 7a, late-Allerød) ranged from 15.1 °C to 18.2 °C, 0.5 °C to 3.5 °C below modern (Fig. 7a). Conroy Lake YD mean July temperature averaged 11 °C, a decline of 6 °C to 7 °C from late-Allerød temperatures, and roughly 7.5 °C lower than today. After 11,400 yrs BP, TJuly increased by 7.5 °C (1 °C to 2 °C below modern). Mean January temperatures, however, remained cold through the late-Allerød and YD periods, dropping by 2 °C at the start of the YD and increasing to 2.5 °C below modern immediately after the YD. The MAT-5 model indicates that Pann increased after the YD from 450–500 mm/ year to 950–1000 mm/year. The WAPLS-2 component model, on the other hand, indicates low precipitation at the beginning of the YD ranging 550–600 mm/year with steadily rising levels through the YD to a maximum of 700 mm/year. WAPLS-2 model predictions for Pann during the YD and at the YD/Holocene transition closely match the multi-core, multi-proxy record of lake-level change from Whitehead Lake (Fig. 7b; Dieffenbacher-Krall and Nurse 2005), located 18.5 km north of Conroy Lake (Fig. 1). Discussion Oxbow is unique because it contains the only documented evidence of YD ice re-advance in northern Maine and New Brunswick. Overall stratigraphy of the Oxbow pits (Fig. 3) indicates that prior to YD cooling an ice margin retreated north of the site, exposing the lower basal till. Sand containing sparse organic remains accumulated on the till, either directly from the ice margin or from subsequent slope-wash, along with a thicker layer of peat during YD time. During the late YD, ice advanced toward the south-southeast over the site deforming the sand and peat and incorporating peat lumps into the compact upper basal till. We reject the possibility that the upper till derived from earlier till moving down-slope due to cryogenic activity during the YD. The deposit is highly compacted, as is the lower till, and both the oriented stone fabric and the deformational structures in the sand demonstrate motion across, rather than down, the adjacent low-angle slope. This scenario is consistent with south-southeast flow directions recorded in both the lower and upper till and with the primary regional flow recorded by predominant striation directions on bedrock outcrops in the area (Fig. 1 insert). Down-wasting glacial ice left the Aroostook highlands predominantly ice free by 13,800 yrs BP while leaving behind large, stagnating ice blocks in low-lying areas (Borns et al. 2004). YD-age peat sandwiched between glacial basal tills at the Oxbow site indicate local presence of an active ice mass during the YD. Temperatures in northern Maine declined substantially from the warm Allerød to the cold YD (summer temperatures declined 5.5 °C to 7.5 °C, and winter temperatures declined 6 °C to 7 °C). However, temperature decline does not support modeled re-growth of an ice field over the Aroostook highland region bordered by the Oxbow site and Pennington and Galilee Ponds (minimum of 8 °C lower than today’s Northeastern Naturalist 82 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 average temperature required for ice cap re-advance; Figs. 1, 2c). Although classic YD sediment deposition is not present in Cranberry Pond or Galilee Pond (Putnam and Putnam 2009), dates at start of lake sedimentation (13,280 yrs BP and 12,880 yrs BP, respectively; Table 1) indicate that no YD ice scour occurred. The southern extent of a low-volume ice margin from an isolated ice mass that advanced over the Oxbow site may be defined by the low-lying moraines immediately south of the Oxbow site. A small, lobate moraine along the shore of Cranberry Pond on the east, and cross-cutting moraines along LaPomkeag Lake to the southwest, may further define the area of ice advance. Pollen diagrams from across northeastern United States and the Canadian Maritimes demonstrate vegetation shifts at the beginning and end of the YD (e.g., Cwynar and Spear 2001, Grimm and Maher 2002, Mott and Stea 1993, Mott et al. 1986, Newby et al. 2005). Mayle and Cwynar (1995a) provided pollen evidence from the Canadian Maritimes that forested sites reverted to shrub tundra and shrub-tundra sites reverted to herb tundra at the YD onset. Pollen grains are typically graphed as species percent of total pollen deposition or as species concentration. Shift in species composition can be subtle and difficult to discern with traditional graphing. When metrics of change, such as square chord distance or principal components analysis, are applied to pollen assemblages, assemblage shifts become more evident (Shuman et al. 2005, Yu 2007). Which climate reconstruction is most robust? Agreement between Conroy Lake pollen-reconstructed mean July air temperatures and chironomid-reconstructed mean summer water temperatures from Whitehead Lake, Pennington Pond, and Deep (Tilley) Lake (site 29; Fig. 1; Cwynar and Levesque 1995) supports the precision of pollen TJuly reconstructions. The case of TJan and Pann is less certain, particularly for the late-Allerød period. The pollen- reconstructed Pann curve for Conroy Lake and Whitehead Lake multi-proxy lake-level reconstruction (Fig. 7b) indicate that climate was considerably drier than today for the entire late-glacial through YD period. Conroy Lake pollen-derived Pann values show a rapid rise in precipitation at the beginning of the Holocene, while the Whitehead Lake multi-proxy lake-level curve shows a more gradual waterlevel rise at this time. The following modern climate values were derived from the Bristol Research Initiative for the Dynamic Global Environment (BRIDGE 2008). This interactive website interpolates mean climate values for locations between data-collection sites and is based on 30 years of climate data (New et al. 1999). Allerød pollen samples are most analogous to the pollen composition found today in the region west of Lake Winnepegosis in Manitoba (Fig. 8B), a dry location (Pann 450 mm) with cold winters (TJan mean = -22 °C) and relatively warm summers (TJuly = 17.3 °C). YD modern analog locations are located predominantly in far northern Labrador and southern Baffin Island, with single analog sites in southern Greenland and the Lake Winnepegosis area (Fig. 8C). These locations are dry (Pann = 350–500 mm) with cold winters (TJan = -20 °C to -25 °C) and cool summers (TJuly = 5.5 °C to 6.5 °C). Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 83 Examples of post-YD analog locations are found in southern Ontario and Quebec (Fig. 8D), moist areas (Pann = 850–1500 mm) with cool winters (TJan = -5 °C to -16 °C) and warm summers (TJuly = 17 °C to 18.5 °C). Vegetation reconstructions from previous studies support temperature and precipitation values modeled in this study. The pollen record from Conroy Lake (Borns et al. 2004) and the plant macrofossil record from Whitehead Lake (Dieffenbacher- Krall and Nurse 2005) indicate that open, tundra-like vegetation persisted along the Maine/New Brunswick border through the Allerød warming and into the YD cold period. Allerød (Fig. 6, zone 2) vegetation around Whitehead Lake was shrub-herb tundra dominated by Dryas, sedges, and rushes, with presence of shrub birch and Chamaedaphne calyculata (L.) Moench (Leatherleaf) (Dieffenbacher- Krall and Nurse 2005). The terrestrial plant macrofossil assemblage remained stable through the late-glacial period until the beginning of the Holocene. Conifer Figure 8. Location of modern sites with pollen analogous (squared chord distance ≤ 0.15) to Conroy Lake for time periods shown. Larger dots indicate a greater number of Conroy Lake samples for specified time periods with analogs at each location. (A) All modern sites (Whitmore et al. 2005) included in pollen models. (B) Allerød pollen samples are most analogous to the pollen composition found today in the region west of Lake Winnepegosis in Manitoba. (C) YD modern analog locations are located in northern Labrador and southern Baffin Island with single analog sites in southern Greenland and near Lake Winnepegosis. (D) Post YD/early Holocene analog locations are in southern Ontario and Quebec. Northeastern Naturalist 84 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 needles were absent until the start of the Holocene warming at 11,400 yrs BP (Dieffenbacher-Krall and Nurse 2005). Pollen from the YD at Conroy Lake indicates a predominance of sedges and terrestrial herbs with patchy shrub and early succession tree growth (Borns et al. 2004). Vegetation shifts at Whitehead and Conroy Lakes were consistent with changes observed at Joe (site 20; Fig. 1; Mayle et al. 1993a) and Roulston (site 21; Fig. 1; Mott et al. 1986) lakes in New Brunswick, Canada, i.e., shift from shrub- to herb-dominated tundra during the YD. The combination of cold winter temperatures (TJan; Fig. 7a) and low precipitation (Pann; Fig. 7b) could explain the lack of forest development during the Allerød period and the decline of woody shrubs during the YD. Cooler summers would further delay or reverse expansion of woody plants in the region during the YD. With a shift to considerably warmer summers and winters and a near doubling of precipitation, post-YD climate favored rapid establishment of forest dominated by birch, aspen, pine, and oak (Borns et al. 2004, Dieffenbacher-Krall and Nurse 2005). Based on both pollen and chironomid inference models, northern Maine experienced summer YD temperatures ranging from 7.5 °C to 5.5 °C lower than today’s mean July air temperature. This finding is consistent with pollen-model estimates of Viau and Gajewski (2007) from 15 sites in northeastern North America and multiproxy evidence from the Canadian Maritimes (Mayle and Cwynar 1995b). Pann for Conroy Lake is consistent with the multi-proxy lake-level reconstruction of Whitehead Lake (Fig. 7b; Dieffenbacher-Krall and Nurse 2005). Both studies indicate that the transition from warm Allerød to cold YD (Fig. 5, zone 2 to zone 3a) in this region was dry, with dry conditions persisting throughout the YD. Precipitation increased by 500 mm a year (Fig. 7b) and water level in Whitehead Lake increased 2 m at the cold YD to warm Holocene interface (Dieffenbacher-Krall and Nurse 2005). To date, investigation of lake cores and glacial features in the Aroostook highlands have found no additional evidence of residual ice or ice re-advance during the YD. Ponds may have been ice-covered for much of the YD, and glacial lacustrine clay deposition over that region is extraordinary (meters instead of centimeters). Oxbow presents an opportunity to better understand the complex ice-recessional processes that occurred in northern Maine and along the Boundary Mountains between Maine and the St. Lawrence lowlands. Further knowledge of the ice advance in the Oxbow region can be gained by dating the lobate moraine at Cranberry Pond and the cross-cutting moraines at LaPomkeag Lake. Lake-sediment analyses for pollen from LaPomkeag and Galilee Lakes will expand understanding of vegetation response to shifts in temperature and precipitation in northern Maine. Acknowledgments The authors gratefully acknowledge the efforts of Steve Barteaux, an extraordinary chironomid-extraction technician. We thank Steve Brooks and Ole Saether for confirming identification of Propsilocerus, David Putnam for geology field assistance, and Matts Lindbladh and Marcus Vandergoes for coring assistance. We are also grateful to Daniel Keppie Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 85 and an anonymous reviewer for their constructive reviews. This work was supported by the National Science Foundation, ATM-0082680. Literature Cited Bengtsson, L., and M. Enell. 2003. Chemical asnalysis. Pp. 423–454, In B.E. Berglund (Ed.). Handbook of Holocene Palaeoecology and Palaeohydrology. Blackburn Press, Caldwell, NJ. 869 pp. Birks, H.J.B. 2005. Quantitative palaeoenvironmental reconstructions from Holocene biological data. Pp. 107–123, In A. Mackay, R.W.Battarbee, H.J.B. Birks, and J.S. Oldfield (Eds.). Global Change in the Holocene. Arnold, London, UK. Borns, H.W., Jr. 1985. Changing models of deglaciation in northern New England and adjacent Canada. Pp. 135–138, In H.W. Borns, P. LaSalle, and W. Thompson (Eds.). Late Pleistocene History of Northeastern New England and Adjacent Canada: Geological Society of America Special Paper 197. 159 pp. Borns, H.W., Jr., L.A Doner, C.C. Dorion, G.L., Jacobson Jr., M.R. Kaplan, K.J. Kreutz, T.V. Lowell, W.B. Thompson, and T.K. Weddle. 2004. The deglaciation of Maine, USA. Pp. 89–109, In J. Ehlers and P.L. Gibbard (Eds.). Quaternary Glaciations: Extent and Chronology, Part II. Elsevier Science, Amsterdam, Netherlands. 452 pp. Bristol Research Initiative for the Dynamic Global Environment (BRIDGE). 2008. Earth systems modelling results. Available online at www.paleo.bris.ac.uk/ummodel/scripts/ html_bridge/clamp_UEA.html. Accessed 16 November 2015. Chase, G.E. 2004. Late-glacial climate as inferred from chironomid assemblages in lake sediments from Aroostook County, Northeastern Maine. M.Sc. Thesis. University of Maine, Orono, ME. 100 pp. Chauvin, L., G. Martineau, and P. LaSalle. 1985. Deglaciation of the lower St. Lawrence region, Quebec. Pp. 111–123, In H.W. Borns, P. LaSalle, and W. Thompson (Eds.). Late Pleistocene History of Northeastern New England and Adjacent Quebec. Geological Society of America Special Paper 197. Boulder, CO. 159 pp. Cranston, P.S., E. Barley, G.E.C. Langley, A. Dieffenbacher-Krall, A. Longmuir, and J. Zloty. 2011. Propsilocerus Kieffer (Diptera: Chironomidae) from the Nearctic Region. Aquatic Insects 33:343–350. Cwynar, L.C., and A.J. Levesque. 1995. Chironomid evidence for late-glacial climatic reversals in Maine, USA. Quaternary Research 43:405–413. Cwynar, L.C., and R.W. Spear. 2001. Late-glacial climate change in the White Mountains of New Hampshire. Quaternary Science Reviews 20:1265–1274. Davis, R.B., and G.L Jacobson Jr. 1985. Late-glacial and early Holocene landscapes in northern New England and adjacent areas of Canada. Quaternary Research 23:341–368. Deevey, E.S., Jr. 1951. Late-glacial and post-glacial pollen diagrams from Maine. American Journal of Science 249:177–207. Dieffenbacher-Krall, A.C., and A.M. Nurse. 2005. Late-glacial and Holocene record of lake levels of Mathews Pond and Whitehead Lake, northern Maine, USA. Journal of Paleolimnology 34:283–310. Doner, L.A. 1995. Late-Pleistocene environments in Maine and the Younger Dryas dilemma. M.Sc. Thesis. University of Maine, Orono, ME. 68 pp. Fastook, J.L. 1993. The finite-element method for solving conservation equations in glaciology. Computer Science Engineering 1:55–67. Fastook, J.L., and J. Chapman 1989. A map plane finite-element model: Three modeling experiments. Journal of Glaciology 35:48–52. Northeastern Naturalist 86 A.C. Dieffenbacher-Krall, et al. 2016 Vol. 23, No. 1 Fréchette, B., A. deVernal, J. Guiot, A.P. Wolfe, G.H. Miller, B. Fredskild, M.W. Kerwin, and P.J.H. Richard. 2008. Methodological basis for quantitative reconstruction of air temperature and sunshine from pollen assemblages in Arctic Canada and Greenland. Quaternary Science Reviews 27:1197–1216. Grimm, E. 1987. CONISS: FORTRAN 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computers and Geosciences 13:13–35. Grimm, E.C., and L.J. Maher. 2002. AMS radiocarbon dating documents climate events in the upper Midwest coeval with the Bølling/Allerød and Younger Dryas episodes. Abstract from Paper No. 159-5 of the 2002 Annual Meeting of the Geological Society of America. Available online at https://gsa.confex.com/gsa/2002AM/finalprogram/abstract_ 40518.htm. Accessed 15 November 2002. Juggins, S. 2003. User Guide C2: Software for ecological and paleoecological data analysis and visualization. Available online at https://www.staff.ncl.ac.uk/stephen.juggins/software/ code/C2.pdf. 73 pp. Accessed 20 October 2015. Levesque, A.J., F.E. Mayle, I.R. Walker, and L.C. Cwynar. 1993. A previously unrecognized late-glacial cold event in eastern North America. Nature 361:623–626. Levesque, A.J., L.C. Cwynar, and I.R. Walker. 1997. Exceptionally steep north–south gradients in lake temperatures during the last deglaciation. Nature 385:423–426. Lowell, T.V. 1985, Late Wisconsin ice-flow reversals and deglaciation, northwestern Maine. Pp. 71–83. In H.W. Borns, P. LaSalle, and W. Thompson (Eds.). Late Pleistocene History of Northeastern New England and Adjacent Canada: Geological Society of America Special Paper 197. Boulder, CO. 159 pp. Mayle, F.E., and L.C. Cwynar. 1995a. Impact of the Younger Dryas cooling event upon lowland vegetation of Maritime Canada. Ecological Monographs 65:129–154. Mayle, F.E., and L.C. Cwynar. 1995b. A review of multi-proxy data for the Younger Dryas in Atlantic Canada. Quaternary Science Reviews 14:813–821. Mayle, F.E., A.J. Levesque, and L.C. Cwynar. 1993a. Accelerator-mass-spectrometer ages for the Younger Dryas event in Atlantic Canada. Quaternary Research 39:355–360. Mayle, F.E., A.J. Levesque, and L.C. Cwynar. 1993b. Alnus as an indicator taxon of the Younger Dryas cooling in eastern North America. Quaternary Science Reviews 12:295–305. Mott, R.J., and R.R. Stea. 1993. Late-glacial (Allerød/Younger Dryas) buried organic deposits, Nova Scotia, Canada. Quaternary Science Reviews 12:645–657. Mott, R.J., D.R.G. Grant, R.R. Stea, and S. Occhietti. 1986. Late-glacial climate oscillation in Atlantic Canada equivalent to the Allerød-Younger Dryas event. Nature 323:247–250. National Oceanic and Atmospheric Administration (NOAA). 2003. North American pollen database. Available online at www.ncdc.noaa.gov/paleo/napd.html. Accessed 14 October 2015. NOAA. 2013. National Centers for Environmental Information: Pollen. Available online at http://www.ncdc.noaa.gov/data-access/paleoclimatology-data/datasets/pollen. Accessed 14 October 2015. New, M., M. Hulme, and P. Jones. 1999. Representing twentieth-century space–time climate variability. Part I: Development of a 1961–90 mean monthly terrestrial climatology. Journal of Climate 12:829–856. Newby, P., J. Bradley, A. Spiess, B. Shuman, and P. Leduc. 2005. A Paleoindian response to Younger Dryas climate change. Quaternary Science Reviews 4:141–154. Newman, W.A., A.N. Genes, and T. Brewer. 1985. Pleistocene geology of northeastern Maine. Pp. 59–70, In H.W. Borns, P. LaSalle, and W. Thompson (Eds.). Late Pleistocene History of Northeastern New England and Adjacent Canada. Geological Society of America Special Paper 19. Boulder, CO. 159 pp. Northeastern Naturalist Vol. 23, No. 1 A.C. Dieffenbacher-Krall, et al. 2016 87 Overpeck, J.T., T. Webb III, and I.C. Prentice. 1985. Quantitative interpretation of fossil pollen spectra: Dissimilarity coefficients and the method of modern analogs. Quaternary Research 23:87–108. Peros, M.C., and K. Gajewski. 2008. Holocene climate and vegetation change on Victoria Island, western Canadian Arctic. Quaternary Science Reviews 27:235–249. Putnam, A.E., and D.E. Putnam. 2009. Inactive and relict rock glaciers of the Deboullie Lakes Ecological Reserve, northern Maine, USA. Journal of Quaternary Science 24:773–784. Sayles, C. 2004. Modeling the glaciation of Maine. B.S. Honors Thesis. University of Maine, Orono, ME. 106 pp. Shuman, B., P.J. Bartlein, and T. Webb III. 2005. The magnitudes of millennial- and orbitalscale climatic change in eastern North American during the Late Quaternary, Quaternary Science Reviews 24:2194–2206. Stuiver, M., P.J. Reimer, and R.W. Reimer. 2005. Calib 5.0. Available online at http://radiocarbon. pa.qub.ac.uk/calib. Accessed 20 October 2015. Viau, A.E., and K. Gajewski. 2007. Comments on: “The magnitudes of millennial- and orbital-scale climatic change in eastern North America during the Late Quaternary” by Shuman et al. [Quaternary Science Reviews 24(2005):2194–2206]. Quaternary Science Reviews 26:264–267. Walker, I.R., R.J. Mott, and Smol J.P. 1991a. Allerød-Younger Dryas lake temperatures from midge fossils in Atlantic Canada. Science 253:1010–1012. Walker, I.R., J.P. Smol, D.R. Engstrom, and H.J.B. Birks. 1991b. An assessment of Chironomidae as quantitative indicators of past climate change. Canadian Journal of Fisheries and Aquatic Sciences 48:975–987. Walker, I.R., A.J. Levesque, L.C. Cwynar, and A.F. Lotter. 1997. An expanded surfacewater paleotemperature inference model for use with fossil midges from eastern Canada. Journal of Paleolimnology 18:165–178. Whitmore, J., K. Gajewski, M. Sawada, J.W. Williams, T. Minckley, B. Shuman, P.J. Bartlein, T. Webb III, A.E. Viau, S. Shafer, P.M. Andersen, and L.B. Brubaker. 2005. A North American modern pollen database for multi-scale paleoecological and paleoclimatic applications. Quaternary Science Reviews 24:1828–1848. Williams, J.W., and B. Shuman. 2008. Obtaining accurate and precise environmental reconstructions from the modern analog technique and North American surface pollen dataset. Quaternary Science Reviews 27:669–687. Wilson, S.E., I.R. Walker, R.J. Mott, and J.P. Smol. 1993. Climatic and limnological changes associated with the Younger Dryas in Atlantic Canada. Climate Dynamics 8:177–187. Yu, Z. 2007. Rapid response of forested vegetation to multiple climatic oscillations during the last deglaciation in the northeastern United States. Quaternary Research 67:297–303.