By only considering records of climate and glaciers over the period that humans have been monitoring them, one might think that climate normally changes quite rapidly, and that glaciers have always been small. But in the not-so-distant past, an ice sheet covered the Southern Andes, flowing across the southern tip of the continent, and in some places, even terminating into the Atlantic Ocean. Glaciers rewrite the surfaces they inhabit, leaving behind indicators of their past behavior. By studying the landforms they create, we can reconstruct climates of the past. Here I present unique and novel glacier-climate reconstructions over southern and central Patagonia using a state-of-the-art dating approach tied to high resolution spatial mapping and glaciological modeling. The main goal of this thesis is to constrain the precise timing and character of the past advances of three glaciers in Patagonia. To this end, I present new precise 10Be surface exposure datasets from two paleo outlet glacier lobes (at 45°S and 53°S), totaling 71 new moraine boulder ages. In these two valleys, I am able to reconstruct the last three major glaciations (MIS 6, 4, 2), as well as provide a constraint for the last two terminations (T1, T2). At a third site, I create a novel dataset to reconstruct the behavior of the Calluqueo glacier (48°S) from ~7,000 years ago to the present. All three sites are eastwards of the main Andes mountain range, spanning from southernmost Patagonia (53°S) to central Patagonia (45°S). To achieve my thesis objectives, I employed recent improvements in the 10Be exposure dating method, and tied the geochronological studies to new, high resolution maps of the glacial geomorphology created by the former glaciers and associated processes. We find in central Patagonia, the Ñirehuao glacier lobe was most extensive potentially during Marine Isotope Stage (MIS) 8, but certainly prior to MIS 6, followed by a major advance during MIS 6. This study presents one of the first directly dated records of a MIS 6 glacier expansion in Patagonia at 153±5.1 ka, where the glacier may have been in retreat at 137±4.2 ka. During the last glacial cycle, the glacier was most extensive during the middle of MIS 2, at 23.6±0.9 ka. The southernmost section of the Ice Sheet, at Estrecho de Magallanes, was more extensive during Marine Isotope Stage 4 (MIS 4) than during MIS 2, representing the first direct dating of the MIS 4 glacier culmination in South America. Similar to the MIS 2 glacial maximum, within MIS 4 there were multiple advances that we date (6 samples) to between 67.5±2.1 and 62.1±2.0 ka. Inboard of the MIS 4 moraine complex, we date a sequence of geomorphically distinct MIS 2 moraines that represent separate major periods of glacial stability. The MIS 2 maximum extent occurred by 27.4±0.8 ka and was followed by at least four more full glacial culminations over a hundred miles beyond the Andes mountains. About 18 km inboard of the main MIS 2 landforms, the sequence is followed by smaller-scale recessional moraine crests that we date to 18.0±0.8 ka, indicating the glacier was in net retreat at this time. In order to estimate the climate conditions necessary to drive the glacier advances that we date and map, we apply the University of Maine Ice Sheet Model to the Estrecho de Magallanes and Ñirehuao records. Tentative results suggest that the Magallanes lobe may have reached mapped inner and outer MIS 2 moraines with a climate that had approximately 4.5°C and 5.5°C cooler summers, respectively, assuming about 25% less annual precipitation relative to modern conditions. A new record at Calluqueo, in central Patagonia, allows us to reconstruct Holocene (interglacial) glacier changes. Using 33 new 10Be ages with unprecedented precision, geomorphic mapping and historical imagery, we find that the Calluqueo glacier sat at its mid-Holocene maximum extent from ~6,900 until ~6,700 years before the present. Major moraine forming advances subsequently culminated at least seven more times, averaging every 500±31 years, between 5,620±203 and 3,120±106 years ago. A hiatus in moraine formation occurred from 3,120±106 until 1,160±50 years ago (860 CE). Major retreat occurred between 1600-1800 CE, followed by stability from 1800-1940 CE, and pronounced ongoing retreat since after 1940 CE. For the Holocene period, this record represents one of the first precise, directly-dated glacier histories from central Patagonia, and one of the few available for all of Patagonia. The timing of advances of the Calluqueo glacier has little in common with the glacial histories from the Northern Hemisphere, suggesting an inter-hemispheric asynchronicity. All together, we reconstruct the timing of glacial maxima at three sites in terrestrial Patagonia from 53°S to 45°S, with unprecedented precision, from pre-MIS 6 to the present day.

Effective coastal adaptation to sea-level rise requires an understanding of how much and how fast glaciers and ice sheets will melt in the coming decades, together with an understanding of the provenance of that ice melt. When land ice is lost to the oceans, sea-levels do not rise uniformly across the globe, but exhibit a “sea-level fingerprint” specific to the source of ice melt, posing an important question motivating this thesis: Which ice mass(es) will contribute the first 1m/3 feet of sea-level rise? The glacial-geologic record archives the vulnerability of ice sheets and their sub-sectors to past warming. To analyze this record of past glacial change, I develop and apply cosmogenic-nuclide techniques for investigating the climate sensitivity of four key ice sheets. The novel geochemical techniques described here also allow me to investigate processes of landscape evolution, including subglacial and subaerial erosion. Subglacial erosion dictates landscape development in glaciated and formerly glaciated settings, which in turn influences ice-flow dynamics and the climate sensitivity of ice masses, making it an important input in ice-sheet models. In Chapter 1, I use 10Be measurements in surficial bedrock and a 4-m-long bedrock near Jakobshavn Isbræ, to constrain the erosion rate beneath the Greenland Ice Sheet (GrIS) on historical and orbital timescales. 10Be concentrations measured below ~2 m depth in a 4-m-long bedrock core are greater than what is predicted by an idealized production-rate depth profile and I develop a model to utilize this excess 10Be at depth to constrain orbital-scale erosion rates. I find that erosion rates beneath GrIS were 0.4–0.8 mm yr-1 during historical times and 0.1–0.3 mm yr-1 on Pleistocene timescales. The broad similarity between centennial- and orbital-scale erosion rates suggests that subglacial erosion rates adjacent to Jakobshavn Isbræ have remained relatively uniform throughout the Pleistocene. In Chapter 2, I present cosmogenic 10Be and 3He data from Ferrar dolerite pyroxenes in surficial rock samples and a bedrock core from the McMurdo Dry Valleys, Antarctica, opening new opportunities for exposure dating in mafic rocks. I describe scalable laboratory methods for isolating beryllium from pyroxene, estimate a spallation production rate for 10Be in this mineral phase, referenced to 3He, of 3.6 ± 0.2 atoms g-1 yr-1, and present initial estimates for parameters associated with 10Be and 3He production by negative muon capture. I also demonstrate that the 10Be-3He pair in pyroxene can be used to simultaneously resolve exposure ages and subaerial erosion rates, and that the precision of my 10Be measurements in pyroxene enable exposure dating on Last Glacial Maximum to Late Holocene surfaces, including moraines, on a global scale. In Chapter 3, I apply exposure dating locally to investigate the Last Glacial Maximum (LGM) and initial deglaciation of the Laurentide Ice Sheet (LIS), the most dynamic continental ice sheet, in southern New England and New York City. I synthesize new and existing exposure age chronologies from moraines and other glacial deposits that span ~26 to 20.5 ka, and quantify retreat rates for the southeastern LIS margin. Initial retreat at <5 to 30 m yr-1 started within the canonical LGM period, representing the slowest LIS retreat rates of the entire New England deglacial record, which I relate to a slow rise in modeled local summer temperatures through the LGM. Employing similar exposure dating techniques in Chapter 4, I describe the first 10Be ages from nunataks of the Juneau Icefield (JIF), Alaska, that I collected through the Juneau Icefield Research Program (JIRP) in order to evaluate icefield thinning during the Late Glacial and Holocene. I find that the JIF was smaller-than-present under warm climate conditions during the early-to-mid Holocene, elucidating the sensitivity of the icefield to warming. Tackling the climate crisis more broadly and in turn, addressing pressing Earth science questions like those posed in this dissertation, requires diverse perspectives. Yet, the Earth sciences have historically been among the least diverse of the STEM disciplines. As one contribution to a comprehensive effort through JIRP to increase diversity in the geosciences pipeline, Chapter 5 details the curriculum for a two-week course titled ‘A Virtual Expedition to the Juneau Icefield’ that I co-designed and co-taught in 2021 to bring accessible polar science experiences to high school students.

Located in the far western equatorial Pacific, the Western Pacific Warm Pool (WPWP) is a greater than 10 million km² area of the warmest water on the planet. The WPWP therefore facilitates intense atmospheric convection and participates in coupled ocean-atmosphere climate phenomena such as El Niño Southern Oscillation, regional monsoons, and the shifting Intertropical Convergence Zone. The WPWP is also a water mass crossroads where thermocline-depth western boundary currents (WBCs) such as the New Guinea Coastal Undercurrent (NGCUC) facilitate the transfer of mass, heat, and nutrients vertically, zonally, and meridionally in the ocean. In this dissertation I focus mostly on reconstructing WPWP upper ocean temperature, salinity, nutrient, and productivity dynamics via a suite of physical and geochemical paleoclimate proxies. I apply these proxies in bulk sediments and planktic foraminifera from International Ocean Discovery Program (IODP) Site U1486 over the Pleistocene (2580 ka to 11.7 ka) and Holocene (11.7 ka to present). Site U1486 is located at 2°22’S, 144°36’E in the Bismarck Sea north of New Guinea in the southern WPWP, and is ideally situated to track changes to the WPWP upper water column forced by the South Pacific. The presence of glacial-interglacial (G-IG) variability within WPWP records is particularly important for determining local versus high-latitude climatic influences on the WPWP – with climate shifts such as the mid-Pleistocene Transition (MPT; ~1250 – 700 ka) and mid-Brunhes Event (MBE; ~430 ka) of particular interest in the long-term records I present. In chapter 1, I explore the paleoceanography of the low-latitude Pacific via upper ocean nitrate dynamics. I present a new bulk sediment ẟ¹⁵N record from Site U1486 that spans from 1420 to 0.67 ka – over a million years longer than any nearby records. Via analysis of orbital variability and secular trends at Site U1486 and in records directly along the equator in the Pacific, I find that nitrate dynamics were largely unrelated in the two regions in the Middle and Late Pleistocene. Whereas ẟ¹⁵N at Site U1486 is in line with patterns of eastern Pacific denitrification, increasing ẟ¹⁵N after the MPT at sites located directly along the equator appears linked to increasing Southern Ocean nitrate utilization. Enhanced nitrate utilization is an indicator of a strengthened biological pump – a major contributor to the reduction of atmospheric 𝑝CO₂ during the last glacial. A post-MPT increase in nitrate utilization may therefore point to the Southern Ocean biological pump as a driver for the deeper and longer glacial periods of the 100-kyr world after the MPT. In Chapter 2, I investigate changes in the vertical temperature and salinity structure of the southern sector of the WPWP in relation to the upper ocean’s response to climate change. When combined with Mg/Ca paleotemperatures and δ¹⁸O_sw, my 670-kyr record of Δẟ¹⁸O between the surface-dwelling foraminifera Globigerinoides ruber (sensu stricto) and the thermocline-dwelling foraminifera Pulleniatina obliquiloculata and Globorotalia tumida suggests enhanced thermocline shoaling and a progressively increasing vertical salinity gradient commencing near 240 ka. This secular change in upper water column dynamics does not appear to be associated with previously documented changes in G-IG variability such as the MPT or MBE. Via comparison to other records, I identify widespread cooling of the thermocline in the equatorial Pacific after ~240 ka. After combining these reconstructions with ²³⁰Th-derived focusing factors I validate previous model results indicating obliquity-driven strengthening of low-latitude ocean currents and extend this to imply the periodic increased transport of high-salinity thermocline water masses. These results strengthen previous evidence that the structure of the WPWP thermocline is relatively independent from the drivers of climate at the surface and support that variability in WPWP thermocline circulation is substantially influenced by obliquity. Because of the nitrate dynamics in the Bismarck Sea, bulk sediment ẟ¹⁵N cannot be used to reconstruct productivity. However, chapter 3 constrains variability in productivity via the analysis of new ²³⁰Th-normalized records of preserved biogenic flux and its components at Site U1486 over the last 138 kyr. Here, I assess the drivers of variability in paleo-productivity by reconstructing paleo-stratification, as in the modern Bismarck Sea productivity is stimulated by the delivery of nutrients to the surface during increased upwelling (reduced stratification). Paleo-stratification is approximated by calculating upper ocean density gradients between the calcification depths of G. ruber, P. obliquiloculata, and G. tumida using Mg/Ca temperatures and δ¹⁸O_sw-estimated salinity. Decreased paleo-stratification (a reduced vertical density gradient) was associated with increased productivity and is generally in phase with maximum orbital precession. Paleo-productivity therefore appears to respond to monsoonal increases in coastal upwelling when the Intertropical Convergence Zone (ITCZ) was at its southernmost extent. This illustrates that the unique and more direct method of constraining stratification presented here, which is subject to greater uncertainty, yields results consistent with our current understanding of upper ocean dynamics. I also identify a period between 100 and 60 ka during a potential reorganization of the upper water column in which variability in productivity occurs at a higher frequency than that of precession. Finally, while also related to ITCZ shifts, a nearby record closer to the equator is phase-lagged from Site U1486 – emphasizing the fine-scale regional differences in the drivers of primary productivity in the WPWP.