1 April 1998 The model spectral resolution was increased from T213 to TL319 (improved orography, reduced diffusion, but Gaussian grid unchanged). The other modifications introduced on this occasion (Cy18r5) were:
1. A new 2-time level semi-Lagrangian numerical scheme: this made possible the use of a Linear Gaussian Grid reflected by the TL notation already introduced for the EPS configuration in Dec. 1996;
2. Use of a new 2'30" orography reference dataset; as a result, some large errors over Antartica have been corrected and spurious noise in other areas suppressed;
3. Improved sea-ice limits determination.
Other modifications introduced at the same time (Cy18r6) were:
1. the use of both significant and standard wind level winds, temperatures and humidities from radiosondes (geopotential is no longer used);
2. the use of extra off-time data, mostly SYNOPs and DRIBUs;
3. the use of 1D-Var SSM/I total column water vapour;
4. extension of the use of GOES high-resolution winds to the northern extratropics;
5. the use of more TOVS channels over land;
6. the observation operator for 10m winds is now unified for scatterometer, DRIBU and SYNOP observations.
Two articles from the ECMWF Newsletter describing the changes are available:
- "Progress with wind-wave interaction" from Newsletter No.80
- "Recent improvements to the ECMWF 4D-Var data assimilation" from Newsletter No.81
12 August 1998 METEOSAT5/INDOEX SATOB products have been introduced in the data assimilation.
21 October 1998 Cycles 19r1 A formulation of uncertainties associated to physical processes (stochastic physics) was introduced in the ECMWF Ensemble Prediction System.