Revisions during 1993

1 Februuary 1993 Changes were introduced to the horizontal diffusion and to the cloud-radiation scheme (cycle 46), inclusion of shortwave optical properties for ice and mixed phase clouds, and revision to the clear-sky absorption coefficients.

21 April 1993 Cycle 47

4 August 1993 Changes were made in the model's physics to improve the representation of surface and planetary boundary layer processes (model cycle 48, cycle 47 was a technical change):

• introduction of entrainment at the top of the planetary boundary layer;



• increased entrainment in shallow clouds;



• modified roughness length and air-sea transfer coefficients;



• enhancements to the parametrisation of the soil processes (four layers with prognostic variables for soil moisture and temperature plus a skin layer temperature).

The meteorological impact of these changes are:

• an overall synoptic improvement of the predicted flow, particularly over Europe in summer;



• a significant reduction of the warm bias in continental boundary layer during daytime in summer;



• a large improvement of the humidity in the boundary layer, with a more realistic diurnal cycle.

The impact on the temperature in the daytime boundary layer is noticeable with a typical reduction in the warm bias at 2 metre of 2-3 degrees. However, an increase in the cold night-time bias has been noticed. Depending on the prevailing cloud cover, the 2 metre night temperatures will, in places, be too low by several degrees.

An article from the ECMWF Newsletter describing the changes is ginen in pdf (300 KB) file.

5 October 1993 A change to the boundary layer parametrisation was introduced, further to the change of 4 August 1993 (cycle 49). The profiles of 2 metre temperature in stable conditions were revised to reduce the coupling between the skin temperature and the temperature of the lowest model levels. The cold bias of the predicted night-time temperature was reduced by this change.

11 November 1993 The pre-processing was modified to correct radiosonde observations for systematic biases prior to their use in the data assimilation. The correction is applied to the radiosonde stations which present a significant bias in geopotential height due to radiative effects, on the sonde (long wave cooling and short wave heating, mainly in the stratosphere).