Data volume and data reduction

When working at the ice-sheet scale, a necessary step in the analysis is data reduction, meaning aggregation of thousands of individual landforms, striae and till-fabric analyses into a modest number of coherent 'packages', where each package reflects a flow system that is spatially and temporally coherent. The data reduction is far more critical in subcontinental-scale reconstructions than in local-regional reconstructions, owing to the amount and complexity of the data, and also the presence of significant within-system age gradients, which increase in importance with the size of the system studied. The most appropriate way to achieve data reduction is through a landscape-level classification scheme, where main types of glacial landscapes are classified on the basis of genetic conditions. These landscapes can then be grouped together based on linkages in the time-space domain. Only if such a classification system is based on genetic conditions and time-space domains can inversion from geomorphology and geology into palaeoglaciological parameters, such as basal thermal regime, flow pattern and configuration (palaeogeography), be performed.

0 0

Post a comment