Hi, I have no example at hand, but the usual way could be indeed to calculate a time series of an adequate statistic (e.g. spatial statistic) from both, the observed and the simulated data and then to apply standard model fitting. Thomas