An approach to the multi-element and multi-scale classification of the limestone pavement environment of Hutton Roof and Farleton Fell, Cumbria, UK
Limestone Pavements are highly significant components of the physiographic and ecological landscapes of the UK. As relict glacial features they are subject to destruction by natural processes but also by human intervention. This thesis identifies the most effective methods to monitor such change at a variety of temporal and spatial scales, based on the Morecambe Bay pavements at Hutton Roof and Farleton Fell. The starting point for such a study is a methodology to define the baseline on which to base change detection and the key to this is the development of a. suitably detailed scene model. This must reflect the environment at the macro-, meso- and micro- scales and also incorporate considerations of the dynamics involved in the landscape evolution. The scene model (the land Surface Classification Hierarchy (lSCH)) was developed by field measurement of the reflectance spectra of the main elements, biotic and abiotic, with measurements of the pavement surface in terms of the scale of karren development and toe texture of the limestone itself. Study of the OEM allowed a fractal dimension to be established and also the nature of ice-flow and its contribution to pavement development, with extending flow, entraining fractured limestone blocks above a plastic, impermeable shale band, being the main mechanism. At the meso scale pavements were classified according to clint form derived from intra-pavement trends in grike direction calculated by Preferred Direction Analysis. Measurements of the key karren forms, runnels, solution pits and pipes and grikes allow assessment of their contribution to the variability of the pavement surface as an element of the scene model through the identification of solution domains.' Identification of different lithologies allowed an investigation of spatial variation across the study area, although lithological control on karren form and magnitude is weaker than variability from age of exposure as shown by statistical analysis of karren morphometry using univariate comparative methods and Link diagrams, bivariate and multivariate regression, discriminant analysis, cluster analysis, multi-dimensional scaling and star diagrams with the derived Star Index. Pavements were classified according to karren morphometry. The traditional view of pedestals asan indicator of solution rates, and hence the concentration of solution at the surface, is challenged through the investigation of water flow over the pavement surface and the consideration of the role of lichen as a protective agent as we" as the size of solution pits and grike width. It is suggested that only 10% of solution potential is achieved at the surface with 43% in the immediate epikarst. From this solution rate diagrams were developed, allowing the dating of exposure of pavements. These were shown to be within the period when human impact in the area was becoming significant and confirms an early anthropogenic impact on this element of the landscape. Further to this the development of grikes as emergent features was confirmed and this linked to' the concept of breakthrough, a"o~ing a model of grike development to be proposed, an important consideration in the dynamics of pavement change. At the micro scale texture analysis allowed the calculation of fractal measures which are related to variations in reflectance. The radiometric response of biotic and abiotic elements of the scene model was analysed confirming the facility of the baseline scene reflectance model of the pavement. Remotely sensed images from the Airborne Digital Camera were linked to ATM, CASI and TM images assessing the effect of scale on change detection and the evaluation of the pavement environment.