Page tree
Skip to end of metadata
Go to start of metadata

Medium Range Forecasting - HRES alone

Here it is assumed that the latest HRES high-resolution forecast material is available but ENS material is not available.  This should occur only in exceptional circumstances.

Using HRES alone is not recommended.  If there is ENS data available, it should be used as well.  The HRES is best considered as part of the ENS and not as a stand alone model since, when viewed in isolation, it cannot provide any estimate of forecast uncertainty or confidence.  However, it is the most accurate forecast for a certain period, but any individual HRES does not necessarily produce a “better” forecast itself and for any particular forecast it may not be the most skilful.   The higher resolution brings advantages and disadvantages - smaller scale atmospheric features are modelled and forecast, and look beguilingly realistic.  Development of these atmospheric systems often is in response to inherent numerical instability (which affects all numerical models) and reliance on detail is inappropriate.  

The main strategy to adopt when using HRES is to avoid over-interpreting non-predictable features.  Therefore the most recent forecast should not be used in isolation.   Run-to-run jumpiness in the HRES forecast can on the one hand be tackled as something negative that has to be dampened, but on the other hand as something positive which can enrich the forecast information by giving alternative scenarios. 

Assessment based on the latest HRES forecasts alone

Smaller scale features are less predictable than larger features (the scale-predictability relation).  Forecasters should disregard the smaller and unpredictable scales and concentrate on the larger and predictable ones.   Small baroclinic systems or fronts are well forecast up to around Day2, large cyclonic systems up to around Day4 and the long planetary waves, defining weather regimes, up to around Day8 (see scale and predictive skill).  Exceptions to these rules are meteorological features that are coupled to the underlying surface (e.g. lee-troughs or heat lows).  Thus, as the lead-time increases, a forecast should become less detailed and more imprecise in location and timing.

Some researchers have found that  "blurring out", in an objective way, the intensity and position of precipitation features can be moderately effective in creating probabilistic guidance from a single deterministic forecast.  Forecasters will benefit from being mindful of this when viewing deterministic forecasts in isolation.

Assessment based on the two latest forecasts

Normally use of the scale-predictability relation will highlight similarities in the latest two HRES forecasts, reduce the error, dampen any “jumpiness” and thereby make the final forecast more trustworthy.  The scale-predictability relation also applies when the latest two HRES forecasts are highly consistent.  Paradoxically, it is in cases of high consistency that forecasters might be lured into unfounded over-interpretation of non-predictable smaller synoptic features.

Assessment based on the last three or more forecasts

One way to take advantage of the skill of previous forecasts is to combine them, together with the latest forecast, into a consensus forecast.  Together, they can be regarded as a "mini-ensemble" that has started from slightly different initial conditions (“a lagged average forecast”).  A consensus forecast will preserve those synoptic features which the individual HRES forecasts have in common and can, therefore, be considered more reliable and predictable.  The spread of the lagged “mini-ensemble” will define the degree of uncertainty and indicate possible alternative developments.  It might even be possible to infer crude but realistic probabilities with respect to weather parameters.  The “mini-ensemble” technique will train forecasters to manage the fully-fledged ensemble forecasts, where these problems are more consistently addressed using ENS alone, or better, using ENS and HRES together. 


  • No labels