Sensitivity analysis of MAGICC/SCENGEN

Abstract

The dissertation was intended to investigate how the results obtained from MAGICC and SCENGEN are influenced by the input variables which are chosen. Both MAGICC and SCENGEN have a number of input variables which are set by the user. They also give different results, such as the rise in temperature and sea level for MAGICC, and the change in pressure and precipitation for SCENGEN. All the eleven outputs (five from MAGICC and six from SCENGEN) were noted while the inputs were changed so as to see which of the inputs influenced most that particular output. The sensitivity analysis was first carried out by varying one input variable at a time. This was important so that any change in the results can be attributed to one source. Then more variables were changed at a time to see if the combined effect of varying two or more outputs influenced considerably the results. Of all the input variables, some proved to be much more effective on the results than the rest. MAGICC showed to be very much influenced by the choice of the emission scenario and to a lesser extent on the climate sensitivity which was chosen. When using MAGICC, the two outputs which were affected the most were the global mean sea level and the rise in temperature. SCENGEN works by using the results from MAGI CC and giving a more detailed examination of the same results. Following the sensitivity analysis carried out on it, SCENGEN was found to be most sensitive to the emission scenario and also to the particular location or region chosen by the user. One other thing which also affected much the results of SCENGEN was the time frame taken into consideration, that is, whether the results were based on annual or seasonal averages. These conclusions and others were drawn by plotting the results graphically using Microsoft Excel.