Ross Purves

The Last Glacial Maximum and deglaciation in southern South America

Abstract This paper models the extent of the Patagonian icesheet during the Last Glacial Maximum (LGM) and its subsequent deglaciation. It constrains a new coupled icesheet/climate numerical model with empirical evidence and simulates the icesheet at the LGM and at stages of deglaciation. Under LGM conditions an icesheet with a modelled volume slightly in excess of 500,000xa0km3 builds up along the Andes. There is a marked contrast between the maritime and continental flanks of the modelled icesheet, with positive mass balance exceeding 2xa0m in the west and declining tenfold to the east. Modelled ice velocities commonly reach 400xa0mxa0yr−1 in the western fjords. The model is most sensitive to variations in temperature and good agreement between modelled ice extent and empirical evidence was achieved by applying a temperature decrease of 6°C relative to present day temperatures with constant wind fields over the model domain. Assuming a stepped start to deglaciation, modelled ice volumes declined sharply, contributing 1.2xa0m to global sea level, 80% of it within 2000 years. The empirical record suggests that such a stepped warming occurred around 17,500–17,150xa0calxa0yr ago.

Modelling the inception of the Patagonian icesheet

Abstract We use a coupled climate/icesheet model to simulate the inception and growth of the Patagonian icesheet. Assuming a temperature lowering of 6°C and a uniform pressure gradient driving windfields, an icesheet builds up with dimensions and morphology similar to that of the Last Glacial Maximum. Analysis of the stages of growth reveals the following conclusions. Different sectors of the 1800xa0km long icesheet behave differently as a result of latitudinal variations in mass balance; the fastest growth occurs in latitudes 45–52°S where the North and South Patagonian Icefields merge and expand to the coast within 1500 model years. The most favourable topography for glacier initiation is the presence of high-elevation basins. Low-elevation summits in the mountain chain take a long time to become covered by ice because of the delay in ice spreading along the main mountain axis. Feedback between the increasing elevation of the icesheet and increasing mass balance on its windward flank enhances the rate of migration of the icesheet towards the precipitation source. There is a lee-side effect whereby easterly mountains are glaciated early on but are then starved of snow as ice builds up to the windward. Such locations should hold an empirical record of the early stages of a glacial cycle

An approach to modelling the impact of snow drift on glaciation in the Cairngorm Mountains, Scotland

This paper uses a Geographical Information System (GIS) to develop a rule-based model of the effect of snow drifting on glacier growth in the Cairngorm Mountains, Scotland during the Younger Dryas interval. We develop a qualitative model of snow drift that uses a number of simple heuristics to generate a map of relative accumulation at a 50-m resolution over the massif. Relative susceptibility to snow drifting is then compared for a number of different zones under varying wind directions. We conclude that, under the conditions tested in our model, glaciation of the Cairngorms during the Younger Dryas was confined to the high corries and that the pattern of snow drifting suggested by our model points to dominant southerly winds. Copyright

Automated Displacement for Large Numbers of Discrete Map Objects

Abstract. Effective displacement is achieved by sharing small movements among a group of objects in order to minimize loss of positional accuracy and maintain the gestalt and topology of the map objects. This research was driven by a desire to implement a simple but effective way of handling large numbers of map objects that may require displacement. The algorithm works by considering, for each object in turn, a number of alternate positions close to its current location. The location which minimizes overlap among the neighbouring objects is chosen. As the program iterates, the objects effectively migrate small distances within the map space in search of solutions. The idea is relatively simple but produces visually acceptable solutions to the displacement of large numbers of objects with very low processing overhead. This paper describes the algorithm in detail, illustrates its application, refinement and evaluation.

A methodology to allow avalanche forecasting on an information retrieval system

This paper presents adaptations and tests undertaken to allow an information retrieval (IR) system to forecast the likelihood of avalanches on a particular day. The forecasting process uses historical data of the weather and avalanche conditions for a large number of days. A method for adapting these data into a form usable by a text‐based IR system is first described, followed by tests showing the resulting system’s accuracy to be equal to existing ‘custom built’ forecasting systems. From this, it is concluded that the adaptation methodology is effective at allowing such data to be used in a text‐based IR system. A number of advantages in using an IR system for avalanche forecasting are also presented.