Downscaling climate change model output: Assessment of Mediterranean precipitation and temperature under increased greenhouse warming conditions (already finished)

Start date: 01.03.2001
End date: 31.05.2004
Funded by: Bay. Staatsmin. f. Wissenschaft, Forschung u. Kunst
Local project leader: Prof. Dr. J. Jacobeit


In recent years, different downscaling techniques have been developed to obtain information for localized areas from global general circulation models (see WILBY and WIGLEY, 1997). Many of these downscaling techniques are based on statistical relationships linking local variables to the large-scale atmospheric circulation. In view of expected changes in climate for the near future due to anthropogenic greenhouse forcing it is crucial to know about the thermal and hydrological changes on a regional scale. In climatological transition regions like the Mediterranean area, small changes in local climate can have major effects on the ecosystem e.g. the availability of water resources. Changes in temperature and precipitation of the past and future have been investigated by several scientists for the whole Mediterranean area or selected regions within (e.g. JACOBEIT, 2000, VON STORCH et al., 1993). Based on gridded data covering the Mediterranean area homogeneously (CRU05 dataset), precipitation and temperature regions are derived from s-mode principal component analysis (PCA) for the period 1948-1998. North-Atlantic-European pressure fields (1000hPa/500hPa) and specific humidity fields (data from NCEP/NCAR Reanalysis) as well as North-Atlantic SST and Mediterranean SST fields (data from REYNOLDS and British Atmospheric Data Centre) have been decomposed into so-called centres of variation (Fig. 1). These centres have also been determined by s-mode principal component analysis of the gridded data with spatial loadings reflecting geographical locations and time-dependent scores defining a set of partial circulation indices. All extracted PCA time coefficients are used as potential predictors for the local climate variables. Thus, it is attempted to improve the estimate of local temperature and precipitation by inclusion of the specific humidity and the sea surface temperatures. Previously, assessment of local variables was mainly based on downscaling from pressure fields only. As a result, the estimates can be enhanced by the use of several predictors, particularly by the specific humidity, whereas North Atlantic sea surface temperatures and Mediterranean sea surface temperatures can not contribute to such an improvement. For the prediction of Mediterranean temperature and precipitation multiple regression models and canonical correlation models are established. After verifying the quality of the models in an independent period of time, the statistical relationships are used to predict future precipitation and temperature in the Mediterranean region under increased greenhouse warming conditions. Tests with only one calibration period (1948-1977) and one period for verification (1978-1998) showed a considerable loss of correlation between the large-scale parameters and local variables. This points to a large degree of instationarity in the timeframe considered. To account for this attribute, ten different calibration periods have been defined and ten verification periods, respectively, thus trying to assess the temporal variability by a spectrum of statistical models. In a next step, appropriate models, defined by means of statistical testing, are applied to predict temperature and precipitation in the Mediterranean region for the 21st century. Output from transient ECHAM 4 model runs (data from Deutsches Klimarechenzentrum) for enhanced greenhouse warming scenarios are used as new predictor fields.

Hurrell, J. W., K. E. Trenberth (1999): Global Sea Surface Temperature Analyses: Multiple Problems and Their Implications for Climate Analysis, Modeling, and Reanalysis.- In: Bulletin of the American Meteorological Society Vol. 80/12, 2661-2678.
Jacobeit, J. (2000): Rezente Klimaentwicklung im Mittelmeerraum.- In: Petermanns Geographische Mitteilungen 144, 26-37.
Kalnay, E. et al. (1996): The NCEP/NCAR 40-Year Reanalysis Project.- In: Bulletin of the American Meteorological Society Vol.77, 437-471.
von Storch, H., Zorita, E. und Cubasch, U. (1993): Downscaling of Global Climate Change Estimates to Regional Scales: An Application to Iberian Rainfall in Wintertime.- In: J. Climate 6, 1161-1171.
Wilby, R. L. und Wigley, T. M. L. (1997): Downscaling general circulation model output: a review of methods and limitations.- In: Prog. Phys. Geogr. 21,4, 530-548.