Modelling 50,000 Years of Landscape Evolution in the Upper Thames Valley, UK: Preliminary Results from PARALLEM

Abstract:

The last glacial-interglacial fluvial archive of the Upper Thames catchment, UK has received much attention over the past 50 years. Extensive exposures within the Pleistocene sequence have yielded a vast array of field observations, constrained by a substantive, but still inadequate, geochronological dataset. This catchment was not glaciated during this period, lying well beyond the Devensian (MIS2) ice margin, and thus landscape morphological evolution progressed largely in response to variation in fluvial and slope processes forced primarily by climate-driven hydrological and vegetation change. Furthermore, there is no evidence to support any significant tectonic-driven change, albeit fluvial system evolution progressed within the context of long-term regional uplift. The relative simplicity of process interactions within this catchment over the timeframe of interest, taken together with its comprehensive sedimentary record, makes the Upper Thames archive a good candidate for comparison with landscape evolution model outputs.

In this paper we present a ‘first look’ at the performance of PARALLEM, a new, parallel-coded, landscape evolution model, with results presented from an initial 50,000-year simulation. In this instance PARALLEM was deployed in annual, steady-flow mode. The model is driven by changing precipitation and temperatures. A pre-determined precipitation input data series was used, based upon a simple transfer function which utilises an up-sampled North Atlantic SST record, together with a modified temperature series based upon palaeoecological data. PARALLEM explicitly models mass-movement processes (i.e. creep, gelifluction) and surface erosion by flowing water (i.e. diffuse and concentrated erosion) using formulae derived from empirical modern-day process data. The model also deploys a simple logistic growth/die-back model for vegetation driven by hydrological and temperature change. Sediment transport is modelled using a modified transport-distance approach.

Comparison of the two datasets is problematical as there are few, if any, attributes that can directly be compared. As a consequence we will discuss some of the obstacles which limit comparison of the field and model data and suggest how some of these might be overcome.