Ensembles are being recognised increasingly in the climate change adaptation and risk assessment literature as an appropriate method for managing uncertainty in GCM impacts on climate change risk analysis i.e., the range of possible outcomes based on the choice of GCM pattern in one’s analysis. The convergence of future climate simulations by different models for a given emissions scenario has to be assessed; the closer the results are together the more reliable the projection.

The ensemble option in SimCLIM enables a user to select a group of GCM patterns to carry out climate change scenario generations and impact studies. By default, the output shows the median value of the group GCMs selected. For example, if a total of 21 GCM patterns are selected, the one that has the value in the 11th place in terms of the magnitude is chosen as the median value. 

The value is defined by the GCM that has value position decided by: (21-1)*50%+1 = 11.

If a total of 22 GCM patterns are selected, the median value is calculated as the average of the GCM values in the 11th and 12th places I terms of the magnitude. 

The position is decided by: (22-1)*50%+1=11.5.

For the purpose of sensitivity analysis, a user can also select low and high percentiles in the ensemble option. If low percentile is selected and a percentage value is given, SimCLIM will generate a climate change value corresponding to the percentile value. Similar as the median value, the determination of the change value is based on the number of GCMs selected for ensemble, as well as the percentile value. For example, if a total number of 21 GCMs are selected and 10% as the low percentile value, the 3rd place GCM value in term of its magnitude is defined as the low percentile climate change value. 

The position is defined by: (21-1)*10%+1=3.

If a total of 22 GCM is selected, the change value is calculated as the combination of the 3rd and 4th place GCM value in terms of their magnitude in the total GCMs, i.e., 0.9*3rd GCM value + 0.1*4th GCM value.

The position is defined by: (22-1)*10%+1=3.1