M4=mean(data,4);
will do the first step, but it'll be a bugger for the next one it would appear.
Not at all clear to me how the rest must be correlated; they're divisible by 12 and/or 24 but where does the 4 come in --is that somehow related to the every fourth year is a leap year (excepting for those that aren't, of course), maybe?
I think you'd be well served to figure out the actual time storage and convert to a timetable -- then retime can do all the averaging over whatever groupings you want...plus any other analyses that come up.
