Sure there are a lot of ways to do this!  My guess is the median is not that different from the mean in part because when you look at seasonal (3-month) averages most variables converge to a Gaussian (or nearly Gaussian) profile.  But certainly that can be done.