This thesis presents stratigraphic, sedimentological and geochronological results from valley fill and glacial moraines of the Hope-Waiau Valleys in North Canterbury, New Zealand. The findings demonstrate that a substantial portion of the modern valley fill comprises in-situ sedimentary sequences that were deposited during the penultimate glaciation (OIS 6), the last interglacial (OIS 5) and during the mid-late last glacial cycle (OIS 3/2). The sediments survived at low elevations in the valley floor despite overriding by later glacial advances. Sedimentologically, the fill indicates deposition in an ice marginal zone and consists of paraglacial/distal-proglacial aggradation gravels and ice-proximal/marginal-subglacial sediments. Deposition during glacial advance phases was characterized by the sedimentation of outwash gravels and small push moraines while glacial retreat phases are dominated by glaciolacustrine deposits which are frequently interbedded with debris flow diamictons. The overall depositional arrangement indicates that glacial retreat from the lower valley portion occurred via large scale ice stagnation. Results from infra-red stimulated luminescence (IRSL) dating gives evidence for five large aggradation and degradation phases in the Hope-Waiau Valleys over the last 200 ka. Combined with surface exposure dating (SED) of moraines the geochronological results indicate that glacial advances during OIS 6 were substantially larger in both ice extent and ice volume than during OIS 4-2. The last glacial maximum (LGM) ice advance occurred prior to 20.5 ka and glacial retreat from extended ice positions began by ~18 ka BP. A late glacial re-advance (Lewis Pass advance) occurred at ~13 ka BP and is probably associated with a regional cooling event correlated to the Antarctic Cold Reversal (ACR). The findings from the Hope-Waiau Valleys were integrated into a model for glaciations in the Southern Alps which uses data from a snow mass balance model to analyse the sensitivity of glacial accumulation to temperature forcing. Model results indicate that in the central hyperhumid sector of the Southern Alps ice would expand rapidly with minor cooling (2-4℃) suggesting that full glaciation could be generated with little thermal forcing. Some Quaternary glacial advances in the Southern Alps may have been triggered by regional climate phenomena (e.g. changes in ENSO mode) rather than requiring a thermal trigger from the Northern Hemisphere.

This thesis presents stratigraphic, sedimentological and geochronological results from valley fill and glacial moraines of the Hope-Waiau Valleys in North Canterbury, New Zealand. The findings demonstrate that a substantial portion of the modern valley fill comprises in-situ sedimentary sequences that were deposited during the penultimate glaciation (OIS 6), the last interglacial (OIS 5) and during the mid-late last glacial cycle (OIS 3/2). The sediments survived at low elevations in the valley floor despite overriding by later glacial advances. Sedimentologically, the fill indicates deposition in an ice marginal zone and consists of paraglacial/distal-proglacial aggradation gravels and ice-proximal/marginal-subglacial sediments. Deposition during glacial advance phases was characterized by the sedimentation of outwash gravels and small push moraines while glacial retreat phases are dominated by glaciolacustrine deposits which are frequently interbedded with debris flow diamictons. The overall depositional arrangement indicates that glacial retreat from the lower valley portion occurred via large scale ice stagnation. Results from infra-red stimulated luminescence (IRSL) dating gives evidence for five large aggradation and degradation phases in the Hope-Waiau Valleys over the last 200 ka. Combined with surface exposure dating (SED) of moraines the geochronological results indicate that glacial advances during OIS 6 were substantially larger in both ice extent and ice volume than during OIS 4-2. The last glacial maximum (LGM) ice advance occurred prior to 20.5 ka and glacial retreat from extended ice positions began by ~18 ka BP. A late glacial re-advance (Lewis Pass advance) occurred at ~13 ka BP and is probably associated with a regional cooling event correlated to the Antarctic Cold Reversal (ACR). The findings from the Hope-Waiau Valleys were integrated into a model for glaciations in the Southern Alps which uses data from a snow mass balance model to analyse the sensitivity of glacial accumulation to temperature forcing. Model results indicate that in the central hyperhumid sector of the Southern Alps ice would expand rapidly with minor cooling (2-4℃) suggesting that full glaciation could be generated with little thermal forcing. Some Quaternary glacial advances in the Southern Alps may have been triggered by regional climate phenomena (e.g. changes in ENSO mode) rather than requiring a thermal trigger from the Northern Hemisphere.

No description available.

The Lake Heron basin is an intermontane basin located approximately 30 kms west of Mount Hutt. Sediments within the basin are derived from a glacier that passed through the Lake Stream Valley from the upper Rakaia Valley. The lack of major drainage in the south part of the basin has increased the preservation potential of glacial phenomena. The area provides opportunities for detailed glacial geomorphology, sedimentology and micropaleontogical work, from which a very high-resolution study on climate change spanning the Last Glacial Maximum (LGM) through to the present was able to be reconstructed. The geomorphology reveals a complex glacial history spanning multiple glaciations. The Pyramid and Dogs Hill Advance are undated but possibly relate to the Waimaungan and Waimean glaciations. The Emily Formation (EM), previously thought to be MIS 4 (Mabin, 1984), was dated using Be10 to c. 25 ka B.P. The EM was largest advance of the Last Glacial Maximum (LGM). Ice during the LGM was at least 150m thicker than previously thought, as indicated by relatively young ages of high elevation moraines. Numerous moraine ridges and kame terraces show a continuous recession from LGM limits, and, supported by decreasing Be10 ages for other LGM moraines, it seems ice retreat was punctuated by minor glacial readvances and still-stands. These may be associated with decadal-scale climate variations, such as the PDO or early ENSO-like systems. There are relatively little sedimentological exposures in the area other than those on the shores of Lake Heron. The sediment at this location demonstrates the nature of glacial and paraglacial sedimentation during the later stages of ice retreat. They show that ice fronts oscillated across several hundred metres before retreating into Lake Heron proper. Vegetation change at Staces Tarn (1200m asl) indicates climate amelioration in the early Holocene. The late glacial vegetation cover of herb and small shrubs was replaced by a low, montane forest about 7,000 yrs B.P, approximately at the time of the regional thermal maxima. From 7,000 and 1,400 yrs B.P, temperatures slowly declined, and grasses slowly moved back onto the site, although the montane forest was still the dominant vegetation. Fires were frequent in the area extending back at least 6,000 years B.P. The largest fire, about 5,300 yrs B.P, caused major forest disruption. But full recovered occurred within about 500 years. Beech forest appears at the site about 3,300 yrs B.P and becomes the dominant forest cover about 1,400 yrs B.P. Cooler, cloudier winters and disturbance by fire promoted the expansion of beech forest at the expense of the previous low, montane forest. Both the increased frequency of fire events and late Holocene beech spread may be linked to ENSO-related variations in rainfall. The youngest zone is characterised by both a dramatic decline in beech forest and an increase in grasses, possibly representing human activity in the area.

The end-Ordovician mass extinction dramatically altered the course of conodont evolution. This extinction event is probably unique in that it can be strongly correlated with a glacial climatic control. This study has identified, through the application of high-resolution stratigraphy, events within the extinction and recovery intervals. Elements of the uppermost Ordovician Shelf-edge Biofacies were severely affected by the oceanic cooling and introduction of cold-water currents associated with the initiation of the glacial maximum. In contrast, elements of the Shelf Biofacies were more severely affected by the intense cooling, shallowing and overcrowding during the glacial maximum. A number of the Shelf-edge taxa that had survived the glacial maximum suffered extinction at the hands of increasing water temperatures, rising anoxia and/or the cessation of oceanic circulation during the post-glacial transgression. Recovery was initiated by the appearance of Crisis Progenitor Taxa within the glacial maximum in the Shelf Biofacies and during the post-glacial transgression in the Shelf- edge- Slope biofacies. The Shelf-edge Biofacies identified within the uppermost Ordovician is not recognised in the Lower Silurian. Two main biofacies occurred on the Shelf and Slope, which had directly evolved from their Upper Ordovician equivalents. The long-term recovery involved the evolution of Crisis Progenitor Taxa and Ecological Generalists within the Shelf and Slope Biofacies (autochthonous taxa). Punctuated equilibrium likely predominated in the Shelf Biofacies as a consequence of widely fluctuating physical conditions. In contrast, the more stable environments of the slope encouraged gradualistic evolution within the Slope Biofacies (Plus ça change Model). Transgressive episodes within the Llandovery, possibly linked to eccentricity cycles, caused the iterative appearance of Long-term Refugia Taxa (allochthonous taxa), sourced from a Pterospathodontid Biofacies. The transgressive episodes also drove elements of the Slope Biofacies onto the shelf. It has been observed that the mechanisms driving extinction, namely environmental disruption and temperature changes, were also responsible for fuelling the subsequent recovery.

Ice streams are the dominant drainage pathways of contemporary ice sheets and their location and behaviour are viewed as key controls on ice sheet stability. Identifying palaeo-ice streams is of paramount importance if we are to produce accurate reconstructions of former ice sheets and examine their critical role in the oceanclimate system. Many workers have invoked palaeo-ice streams from a variety of former ice sheets, despite a limited understanding of their glacial geomorphology. This thesis addresses the problem by predicting several diagnostic geomorphological criteria indicative of ice stream activity. These are developed objectively from the known characteristics of contemporary ice streams and can be summarised as: large flow-set dimensions (>20 km wide and >150 km long), highly convergent flow patterns, highly attenuated subglacial bedforms (length:width >10: 1), Boothia-type dispersal plumes, abrupt lateral margins «2 km), ice stream marginal moraines, evidence of pervasively deformed till, and submarine sediment accumulations (marine-terminating ice streams only). Collectively, the criteria are used to construct conceptual landsystems of palaeo-ice stream tracks. Using satellite imagery and aerial photography to map glacial geomorphology, identification of the criteria is used to validate the location of a previously hypothesised ice stream and identify a hitherto undetected palaeo-ice stream from the former Laurentide Ice Sheet. Implications for ice stream basal processes are explored and their ice sheet-wide significance is assessed. On Victoria Island (Arctic Canada) five of the geomorphological criteria are identified and the extent of the marine-based M'Clintock Channel Ice Stream is reconstructed at 720 km in length and 140 km in width. The ice stream (operating between 10,400 and 10,000 yr BP) was located within a broad topographic trough, but internal glaciological processes, rather than properties of the bed controlled the margin locations. It eroded into pre-existing unconsolidated sediments and left a spectacular pattern of subglacially-produced landforms, recording a snapshot view of the bed prior to ice stream shut-down. Sediment availability appears critical to its functioning (deformable bed?) and the debris flux of the ice stream is inferred to have been high. Frictional shut-down occurred once down-cutting through sediments reached hard bedrock close to the terminus. The presence of four of the geomorphological criteria are used to identify a terrestrial ice stream which drained the Keewatin Sector of the Laurentide Ice Sheet between ca. 10,000 and 8,500 yr BP. Its size is reconstructed at over 450 km in length and 140 km in width, and it left behind a subglacial bedform pattern consisting of highly attenuated drumlins (length:width ratios up to 48: 1) displaying exceptional parallel conformity. This represents an isochronous bedform pattern and variations in lineament elongation ratio are thought to be a useful proxy for ice velocity. Highest elongation ratios occur immediately downstream of a topographic step where the ice stream entered a sedimentary basin. It is inferred that the ice stream was triggered by climatic warming which altered the ice sheet configuration and the thermal state of the bed. A switch from cold to warm-based conditions probably triggered rapid basal sliding. The ice stream (and a tributary) shut down when it ran out of ice, causing widespread thinning of the ice sheet and subsequent deglaciation. These ice streams denote considerable ice sheet instability over both hard and soft (deformable) beds and emphasise the enormous effects that ice streams had in controlling the deglaciation of the Laurentide Ice Sheet.

In Iceland, glaciers cover approximately 11% of the land surface and comprise the country's largest reservoir of freshwater. Increases in summer temperatures since the mid-1980s have led to rapid mass loss from the Icelandic ice caps of 9.5-11.4 Gt/yr, and continuously operating GPS stations nationwide have recorded rapid and accelerating surface uplift. Understanding the behavior of Icelandic ice caps-and their relationship to surface uplift, which is modulated by the rheological structure of the crust and upper mantle-has important implications for water resource management and geohazards analysis. The goals of this study are twofold. First, we aim improve the current estimates of glacial isostatic adjustment (GIA)-related uplift in Iceland and to examine if and how uplift rates have changed over the last several decades. Secondly, we explore the utility of motion recorded by continuously operating Global Positioning System stations (cGPS) as an independent measure of ice cap mass variation over both decadal and annual time scales. We take advantage of the now longstanding cGPS network in Iceland and consider position time series from 62 stations across the entire island. Observations made by cGPS stations from 1995-2014.7 show a broad region of rapid uplift in central Iceland with near zero uplift observed along the coastal regions to the north and west. The most rapid uplift and uplift accelerations occur near the center of the island, between the Vatnajökull and Hofsjökull ice caps, with rates exceeding 30 mm/yr and accelerations of 1-2 mm/yr². Statistically significant uplift and uplift accelerations are recorded at 27 of the 62 cGPS stations, and estimates for the timing of uplift initiation correlate with Arctic warming trends and observations of increasing summer temperatures since the mid-1980s. These results represent a significant improvement over previous uplift estimates and indicate a likely relationship between accelerated ice cap melting and contemporaneous changes in uplift rates. Incorporating cGPS-recorded information about modern-day uplift rates affects estimates of mantle viscosity. Ice cap thinning rates are computed by a weighted least squares estimation scheme utilizing cGPS-derived secular uplift rates and accelerations and Green's functions for an elastic layer over a Maxwell viscoelastic half-space. We test a range of viscosities from 8 x 10¹⁷ and 1 x 10²⁰ Pa·s and find that thinning rates computed with low viscosities between 2 x 10¹⁸ and 1 x 10¹⁹ Pa·s match independently derived ice cap thinning rates best, in accordance with previous upper mantle viscosity estimates. Similar estimation techniques demonstrate the utility of cGPS to provide a seasonal mass variation time series as a potential low-cost compliment to traditional field-based mass balance measurements. We use estimates of secular site velocity and acceleration to reduce the time series and focus only on the annual periodic motion. The increased temporal resolution afforded by the daily cGPS position estimates recovers the interannual variability in the timing and magnitude of accumulation and melt seasons with a small RMS reduction relative to a sinusoidal model. We also find we are able to identify of the effects of both ice cap insulation as well as reduced surface albedo following volcanic eruptions.

"Since the publication of the work of Jehu and Charlesworth the distribution and significance of the glacial deposits of the Irish Sea have remained in a somewhat confused condition." So said Professor W.D. Evans in 1964. This thesis represents an attempt to remove the confusion from one small corner of the Irish Sea basin; it is concentrated largely in North Pembrokeshire, and aims to resolve some of the stratigraphic and chronological problems concerned with the Pleistocene deposits of the region. There is a long history of research into the glaciation of North Pembrokeshire. Most of the early work was undertaken in Dewisland (the St. David's Peninsula), but in the major contributions of Jehu (1904) ond Charlesworth (1929) something was revealed of the Pleistocene history of North Pembrokeshire as a whole. Subsequent work has been greatly influenced by these two authors, and there has been no detailed field study of the area since Charlesworth. As a result, many controversies have arisen in recent years; these range from the purely stratigraphic problems of the classic tripartite drift succession and the nature of the Upper Boulder-clay to the absolute age of the Pembrokeshire drifts. Again, there is doubt concerning the number of glaciations represented in North Pembrokeshire; about the existence of the South Wales End-moraine, the pro-glacial lakes, and the so-called overflow channels of Charlesworth; and about the age and origins of the coastal features of the county. These are among the problems that this thesis attempts to answer.

One of the profound realizations in Earth science during the last several decades has been that the solid earth and climate system interact through mountain belt evolution. Tectonic forces generate topography, and erosion, driven largely by the climate, destroys topography. Perturbations to the competition between these processes may, for example, have driven the transition from greenhouse to icehouse climate during the Cenozoic. Erosion is the ultimate connection between the climate and solid earth system, and because landscapes are shaped by erosion, they hold in their form information about climatic and tectonic forcings. Reading climatic and tectonic processes from the landscape requires an understanding of how these processes drive erosion. One way that climate influences erosion is by setting the elevation at which glaciation occurs. It has been thought for over a century that erosion by glaciers can limit the height of cold, heavily glaciated mountains. In this thesis, I argue that the prevalence of this phenomenon is underappreciated, and that glacial erosion has imposed an upper limit on the growth of warm, tropical mountains. The argument is premised on a combination of field observations from two (sub)tropical mountain ranges in Costa Rica and Taiwan (including 10Be and 3He surface exposure ages), a new method of topographic analysis that identifies previously unrecognized patterns of landscape rearrangement introduced by high elevation glaciation, and a study of ten tropical mountain ranges that reveals a widespread glacial control on their height. The results of this thesis demonstrate the efficacy of glacial erosion even in the warmest mountains, and challenge the hypothesis that quickly uplifting and eroding landscapes have approached a steady state balance between rock uplift and fluvial erosion during the Pleistocene.

Oceanic circulation remains one of the poorly understood elements of the global climate system, despite its importance to planetary heat redistribution and carbon cycling. The nature of deep-water formation and circulation in ancient oceans are even more poorly constrained. In order to understand climate dynamics of past and future climates we must have a better understanding of the role of deep-ocean circulation. In this dissertation I investigated changes in intermediate- to deep-water circulation in three different ocean basins during two different geologic eras. The first study focused on the late Pleistocene (~25 ? 60 ka) California margin to investigate the role of intermediate water circulation in abrupt climate fluctuations. The other two studies investigated deep-water circulation during the Late Cretaceous (~70 ? 100 Ma) greenhouse interval, to determine if deep waters formed in the southern Indian or Atlantic basins. The above studies employed neodymium isotopes preserved in biogenic apatite (fish teeth and bones) and foraminiferal calcite to reconstruct the provenance of intermediate- to deep-water masses. Here I present data from two sites located at intermediate depths on the late Pleistocene California margin as well as seven Deep Sea Drilling Project and Ocean Drilling Program Cretaceous aged sites; four in the South Atlantic Ocean, and three in the Indian Ocean. The new Pleistocene data rule out changes in the source of intermediate waters to the California margin, thus the recorded changes in seafloor oxygenation were caused by changes in sea surface productivity. In the Cretaceous, the spread of deep waters formed in the high-latitude South Atlantic was hindered by tectonic barriers until the mid Campanian when the subduction of Rio Grande Rise allowed for the continuous flow of deep waters from the Southern Ocean into the North Atlantic. The deep Cretaceous Indian Ocean was filled with deep waters formed in the high-latitude Indian Ocean, until being replaced with waters sourced in the Pacific from the late Cenomanian to early Campanian before a return to southern Indian-sourced waters for the remainder of the study interval.