|
The comparison between the output of climate models and empirical data is often hampered by the scale mismatch of both reconstruction approaches, namely the large scale for the general circulation models [ GCMs ] and the local scale for the empirical data.
As an example of how to establish a conceptual and practical modeling chain allowing for this comparison on a common basis, in a first step the large-scale output of a GCM is statistically downscaled for the location of the empirically derived data in southern South America to obtain hydrological changes for the mid-Holocene. In this setup the statistical models have been setup by means of reanalysis data of the present day climate for co-located with the empirical data over south-eastern Patagonia. The statistical downscaling approach implicitly assumes that the statistical transfer functions linking the large scale circulation to the local hydrological climate are constant over time. To i) test the hypothesis of this assumption and ii) obtain a more spatially detailed pattern of hydrological changes in southern South America during the Holocene, numerical simulations with the regional climate model CCLM have been carried out. Results of the regional climate model simulation show that i) the basic assumption of a constant circulation-climate relationship is valid for the mid-Holocene and ii) hydrological changes are in line with results obtained for southeastern Patagonia by means of statistical downscaling. The spatial pattern of hydrological changes is however far more complex than previously assumed. |
|