TKE Calculation on Resampled Timeseries

[tjwixtrom]

Currently the tke function is designed to perform the calculation as averaged over the entire input time series. Often it is desirable to average over a smaller time period. This is easy to do for calculation of something like variance with Xarray; var = data.resample(freq='1Min').var(). It would be useful to have a similar functionality within Metpy, where calculation like TKE and similar (I'm hoping to add some others) would be performed on a user-specified resample frequency, or be designed to accept an Xarray resample object. Below is a working example, though probably not robust enough yet.

def turbulence_kinetic_energy(*args, freq=None, dim=None):
    terms = []
    for arg in args:
        if freq is None:
            perturbation = arg - arg.mean()
            term = (perturbation**2).mean()
        else:
            perturbation = arg - arg.mean(dim=dim)
            term = (perturbation**2).resample(time=freq).mean(dim=dim)
        terms.append(term) 
    return 0.5 * (sum(terms))

And an example of a function that accepts resample objects:

def turbulence_intensity(ws, dim=None):
    sdev = ws.std(dim=dim)
    mean = ws.mean(dim=dim)
    return sdev / mean