Unsicherheiten von atmosphärischen Indizes in Klimasimulationen
Modern climate models are able to simulate important climate parameters over a long period of time. One parameter is sea level pressure, which can be used to calculate atmospheric indices. The Southern Annual Mode (SAM) index describes the distribution of atmospheric pressure areas in the higher lat...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | German |
| Published: |
2018
|
| Subjects: | |
| Online Access: | https://oceanrep.geomar.de/id/eprint/46184/ https://oceanrep.geomar.de/id/eprint/46184/1/Bachelorarbeit_Moritz_Kossmann.pdf.pdf |
| Summary: | Modern climate models are able to simulate important climate parameters over a long period of time. One parameter is sea level pressure, which can be used to calculate atmospheric indices. The Southern Annual Mode (SAM) index describes the distribution of atmospheric pressure areas in the higher latitudes of the southern hemisphere, the North Atlantic Oscillation (NAO) index the distribution of pressure areas over the North Atlantic. Both indices provide information on weather conditions and are therefore an important indicator of temperature and precipitation in Australia/New Zealand, resp. Europe. Due to the imperfection of the climate models, the predictions of the SAM index and the NAO index contain uncertainties. The four different sources of uncertainty - internal variability, model uncertainty, scenario uncertainty and uncertainty resulting from model initialisation - were identified and quantified by using statistical methods. A selection of 7 models, each with 3 scenarios from the CMIP5, was examined for uncertainties in the 21st century. To estimate the role of model initialisation, an ensemble of 10 runs of the model CSIRO-Mk3-6-0 was examined. It turned out that models in both scenarios simulate a positive trend of the SAM index for the 21st century. In the period 2020-2080, the trend is bigger than the total uncertainty. At the beginning of the 21st century, internal variability is dominating, in the middle of the century model uncertainty is responsible for the biggest part and scenario uncertainty dominates the last two decades. Regarding the NAO index, no trend could be observed. Internal variability dominates the first half of the century, model uncertainty the second half, while scenario uncertainty is too small to be considered important. It turned out that model initialisation only plays a small role regarding the uncertainties of both indices. |
|---|