Page History

...

Forecast error growth is, on average, largest at the beginning of the forecast.  At longer forecast ranges it levels off asymptotically towards the error level of persistence forecasts, pure guesses or the difference between two randomly chosen atmospheric states (see Fig4.1-1).  This error level is significantly higher than the average error level for a simple climatological average if used as a forecast.  Forecast verification is discussed in the annexe.

...

The typical predictability is currently approximately twice the timescale, but might ultimately be three times the timescale.  Small baroclinic systems or fronts are currently well forecast to around Day2, cyclonic systems to around Day4 and the long planetary waves defining weather regimes to around Day8.  As models improve over time these limits are expected to advance further ahead of the data time.  Features that are coupled to the orography (e.g. lee-troughs), or to the underlying surface (e.g. heat lows), are rather less consistently well forecast.  The predictable scales also show the largest consistency from one run to the next.  Fig4.1-3 shows 1000hPa forecasts from six sequential runs of HRES (identical to Ensemble Control (ex-HRES)) verifying at the same time. 

...

It can be seen from the above that some of HRES (identical to Ensemble Control (ex-HRES)) forecasts in Fig4.1-3 (T+96, T+108 and perhaps T+144) were quite good with respect to strong winds over Britain and Ireland but at the time the ensemble indicated that gale force winds were not certain. 

...

In order to estimate and compensate for any model drift the model output is compared with the corresponding model climates (M-climate for medium range, ER-M-climate for extended sub-seasonal range, S-M-climate for seasonal forecasting) for the current forecast date.  This is derived using the same model construction as the ensemble from a number of perturbed forecasts based on calendar dates surrounding the date of the current ensemble run using historical data from several years.  Systematic errors are then corrected during post-processing after the forecast is run.

...