Page History

Rationale for high resolution

The higher the numerical resolution, the more accurate the calculations becomeCalculations become more accurate with higher resolution.  A high spatial resolution also enables a better representation of topographical fields (mountains, coasts, islands) and the effects they have both on local weather and the on large-scale flow.  It also produces a more accurate description of horizontal and vertical structures, which facilitates allows better assimilation of observations.   However, the Increasing the resolution (i.e. making grid point spacing smaller) benefits the analyses and forecasts of both large-scale systems (such as large-scale blocking "omega" anticyclones, and "cut-off lows") and the small-scale systems often associated with severe weather.  The ability to accurately forecast the formation of large-scale blocking “omega” anticyclones and “cut-off lows” depends crucially on sufficient effective resolution.  However, a resolution of at least ~1km is required to capture mesoscale phenomena.  

The smallest atmospheric features that can be resolved by models using the Reduced Gaussian Octahedral grid (HRES and ENS) have wavelengths four to six times the nominal grid point distance.   Such Grid point spacing is currently about 9km in the medium range ensemble.  This implies features with wavelengths less than about 45km may not be captured.  Such small atmospheric systems have a predictability of relatively few hours but their representation is important for energy exchanges between different atmospheric scales.  Increasing the resolution (i.e. making grid point spacing smaller) benefits the analyses and forecasts of both large-scale systems (such as large-scale blocking "omega" anticyclones, and "cut-off lows") and the small-scale systems often associated with severe weather.  The ability to accurately forecast the formation of large-scale blocking “omega” anticyclones and “cut-off lows” depends crucially on sufficient effective resolution.  To capture mesoscale phenomena a resolution of at least ~1km is required.

An illustration of the effect of resolution upon forecast rainfall distribution and intensity. 

Fig2.1.51.2-1: Rainfall forecasts run for the same period (24hr rainfall to T+69 verifying at 06UTC 6 Dec 2015) at several resolutions.    Resolutions: 80km , 36 km (used by system 5 seasonal and extended ranges (day 16-46the extended range and seasonal (SEAS5) ensembles), 18km, 9km (used by HRES and medium range ensemble) , 5km (a possible future resolution).

Fig2.1.61.2-2: Magnification of rainfall forecasts as in Fig2.1.51.2-1. The bottom right plot shows the observed rainfall for comparison.  Higher resolution gives better representation of the rain shadow effect but it still is underestimated (e.g. northeast Northeast England).

Fig2.1.1.5 2-1 and Fig2.1.6 1.2-2 illustrate the impact of resolution on rainfall, particularly orographic enhancement and the rain shadow effect.  The effect of increased model resolution on:

...

Note:   An estimate of the range of totals likely within each gridbox is made by the Point Rainfall products.