2 January 1985 Indian SATOB excluded from analysis if data mislocated.
15 January 1985 Forecast model cycle 20. Surface pressure is computed from analysed field instead of first-guess in post-processing.
19 February 1985 SATOB winds over land between 20°N and 20°S were included in the operational analysis. Previously only SATOB winds over the sea had been used.
26 February 1985 High resolution satellite soundings data (TOVS) were included in the operational analysis. TOVS data are similar to SATEMs but have 250 km horizontal resolution compared to 500 km for SATEMs.
In the analysis the matrix size was increased, the large scale wind correlation was introduced and the height overlap was changed.
7 March 1985 NOAA-9 TOVS and SATEMs are included in analysis.
1 May 1985 Forecast model cycle 22. The T106 model became operational. The spectral representation in the horizontal is truncated at wavenumber 106. The vertical resolution was not changed.
The vertical diffusion scheme was modified which allows larger time-steps to be used
(t = 900s). A parametrisation scheme of shallow cumulus convection has been introduced.
The Kuo cumulus convection scheme has been replaced by a modified Kuo scheme.
Evaporation of large-scale rain and melting of snow have been reformulated.
A new cloud prediction scheme was formulated to overcome the deficiencies of the previous scheme (too many deep clouds, too little tropical cirrus, too little subtropical cloudiness, poor representation of the diurnal variation in cloudiness).
Sea ice has been re-specified and is now assumed to exist whenever sea temperatures are below -2°C (previously 0°C).
9 May 1985 Satellite thicknesses south of 60°S were excluded from analysis, but was retained in statistics.
21 May 1985 A more stringent first-guess check for high resolution satellite soundings was introduced.
11 June 1985 Include all NOAA-6 data in analysis (still excluded south of 60°S).
18 July 1985 Forecast model cycle 24. The production of rain in convective parametrisation scheme and the evaporation efficiency over land was modified due to the over-forecasting of precipitation (particularly over Europe) and the marked negative bias in the 2m temperature in daytime (in particular when dry, sunny conditions were observed). After the correction was introduced, the over-forecasting of precipitation was eliminated and the 2m temperature bias was reduced.
31 July 1985 Surface temperature set to climatology over sea if SST analysis not received or rejected.
1 August 1985 Remove change of 31 July 1985. Surface temperature set to first-guess over sea if sea surface temperature analysis (SST) not received or rejected.
8 October 1985 11 thickness layers used in analysis for TOVS and SATEMs instead of 14.
15 October 1985 Modification of rejection criteria for NMC SST analysis (not used on lakes or near coast).
17 October 1985 Modification of the stratospheric Hough smoothing functions in analysis, giving a smoother height and wind analysis.
4 October 1985 Removing SATEMs and TOVS north of 70°N in analysis.
3 December 1985 Modification of the horizontal interpolation in sea surface temperature analysis.
9 December 1985 Forecast model cycle 25. The solar heating rates were modified and the subterranean extrapolation was changed. An error in the vertical diffusion was corrected.