[an error occurred while processing this directive]
Operational implementation of 20km RUC. 40km RUC-2 stops running at this point.
These changes are primarily for BUFR output, including rotating winds to earth-relative (10m, profiler, storm-relative), setting 1h precip properly (instead of 3h precip on every 3rd hour), and using 2m temps, pressure, and elevation from the RUC 'topomini' file (see http://ruc.fsl.noaa.gov/forum/eval20/?read=1053). The ceiling (cloud base) calculation now properly uses the snow mixing ratio, as well as those of cloud water ice. Also, the precipitation type now properly includes sub-grid-scale precipitation (it did not before).
Land-surface model fixes for thin snow cover. Fix to shortwave radiation cooling/heating in atmosphere.
Fix so that precip type arrays output on 40km grid are not smoothed, thus avoiding substantial loss to areal coverage.
A fix was made today (by Geoff Manikin and NCEP/NCO) to correct the date stamp in the RUC BUFR files. The stamp had been running 1 day ahead since January 1. The problem was in a data statement for January, and since the new format for these files began last April, it wasn't until our first January with the RUC20 that the problem surfaced.
HYBFRONT - 3-d variational analysis
A 3-d variational (3DVAR) analysis was implemented replacing the previous optimal interpolation (OI) analysis. Details in http://ruc.fsl.noaa.gov/forum/eval20/?read=1121 and in a TPB at http://ruc.fsl.noaa.gov/ppt_pres/RUC-3dvar-tpb-May03.htm.
Added 0-1km helicity and fixed some other products in post processing -- described in http://ruc.fsl.noaa.gov/ppt_pres/RUC-3dvar-tpb-May03.htm.
Script producing RUC BUFR files
A fix was made today (by Geoff Manikin and NCEP/NCO) to correct a bug in this script which had resulted in the use of the early 'A' RUC analysis run at 0000 and 1200 UTC before most of the rawinsonde is available. Now the BUFR data uses the correct RUC 'B" analysis including the rawinsonde data. The 'B" analysis is the initial condition for RUC model runs at 0000/1200.
BUFR station list
A few stations were added to the BUFR station list, also added at the same time to the Eta BUFR station list.
HYBFRONT - gtresids_sfc_tquv.f, hybhqc.f
A couple of modifications to the 3DVAR initialization were implemented at 15Z today (24 Sep):
HYBPOST - hb2mw.f
Fix max wind level diagnostic to prevent intermittent spurious wind max values. The spline interpolation used sometimes can give wild extrapolation from RUC native levels that are very close together. This code fix constrains the spline value to be within 3% of the native level wind max, and adds some other constraints.
HYBETABC (HYBEXT) - (grdtr3d.f)
Fix interpolation routine for interpolating Eta grids to RUC vertical/horizontal grid for boundary conditions. Under a certain rare condition, a divide by zero could occur, so this possibility was eliminated by adding a line of code. A crash that occurred in the 12z 10 Oct run was fixed by this change.
HYBFRONT - (rd_cld.f,hybctp.f)
A change was made to the operational RUC at NCEP today to fix a cloud assimilation error. Cloud was inadvertently being cleared in situations with partial cloud cover in the GOES data in the proximity of a grid point. The overall impact is fairly minor.
Major change package, including changes to HYBCST ,
HYBPOST, HYBCST_PRE, and HYBFRONT.
Changes in package described here
RUC analysis changes made on 14 April as part of change package were pulled today due to intermittent moisture/CAPE noise in analysis fields.
Revised version of RUC analysis implemented, including a revised and more robust version of PBL-based assimilation of surface observations. Also includes constraints for moisture and temperature increments (correction to background) and innovations (obs-minus-background) to improve coherence in analysis fields. More information available here
Fix to treatment of Eta/NAM data in RUC boundary condition code to avoid problem with non-physical lapse rate near surface. See RUC forum for more.
At this time, recovered RUC analysis version with 28 Sept 04 changes. Due to an inadvertent problem with the switch of NCEP Central Operations to a new computer on 25 Jan 05, an older version had been implemented. Today, the 28 Sept 04 version was reinstated. condition code to avoid problem with non-physical
Soil moisture updates. Background 1h forecast copied from FSL to NCEP to update soil moisture, which has become too dry at NCEP. The FSL RUC runs includes a soil temp/moisture nudging procedure, to be included in the RUC13, which greatly reduces this problem.
GETBUFR - getbufr.f - Fix to observation handling problem in case with raob with wind but not mass data data. In this case, winds were being assigned to incorrect station. This was a rare event -- see RUC forum.
HYBFRONT - hybanl_qc.f - Fix to avoid multivariate height solution when no wind or height available. [an error occurred while processing this directive]