Effects of semi-additive measures on other dimensions

Semi-additive measures are measures that don't use the same aggregation method along all of the cube's dimensions. In SSAS the time dimension plays an important role here. For example, if you choose the 'AverageOfChildren' aggregation method, the measure is averaged over the time but summed up over all other dimensions. When do we need this? Well, usually semi-additive measures are used when working with snapshot data like stock levels, balances of accounts etc.

For an example, let's look at a car park. We want to analyze the number of cars that are in one of our two park houses at a certain time. For our example we use a very simple fact table looking like this (I simplified the time to just 2 days, so imagine we're counting the cars at 1pm):