Operational implementation. RUC-1 stops running at this point. End of internal implementation status for RUC-2.
HYBCSTSP - (hybcst.f)
The RUC2 crashed late in the fcst runs last Fri at 12z, 15z, and 18z. The problem was a very low water vapor mixing ratio value at the top level in the model, which eventually became negative. A piece of code designed to prevent this in the model had left the top level unprotected, which is where the problem occurred. The actual crash was in long-wave radiation, which did not appreciate being given a negative water vapor path length. At 21Z on Friday, the RUC2 was forced to use the 9h Eta as a first guess, which temporarily got rid of the dry spot at the top, and the RUC2 ran reliably for the weekend. This morning at 14Z, a fix was put in to extend the check to the top level so that this problem will not occur again.
HYBFRONT, HYBPOSTL (capcin.f)
New version of CAPE/CIN diagnostic which gets rid of "holes" in CAPE field caused by cloud top being calculated from the bottom up. Sometimes, weak stable layers caused the calculation to miss most of the unstable layers.
See RUC2 forum message for more details.
HYBFRONT, HYBPOSTL (capcin.f)
New version of capcin.f pulled out due to GRIBbing problems. Apparently, the new routine works OK, but something was amiss in its implementation, so this will get fixed in the future.
HYBFRONT, HYBPOSTL (capcin.f)
New version of capcin.f with many more fixes in than 17 April version was implemented. This new version was developed by John Brown of FSL in conjunction with Storm Prediction Center meteorologists, especially Phillip Bothwell. See the forum message for more details.
Effect on output - Much improved CIN, no holes in middle of CAPE and more accurate CAPE calculation.
Y2K fixes Analysis changes - smaller horizontal error correlation near sfc for T/Td, slightly less dependence on stability => improved sfc T/Td/wind fit in mtns, better near sfc and lower tropospheric structure overall - fix to use of cloud drift winds => will have much more effect (over water only). Weight had been getting set to very low values, and this is being fixed. West Coast RUC users should start seeing some differences. Model changes - fixes to radiation - more cooling at night, more heating in daytime -- A big improvement in the RUC2 diurnal cycle. This fix was removal of the application of the water vapor mixing ratio from the model top level (~60 mb) up to 0 mb. This extra water vapor was acting a bit like a cloud at the top of the model, and damping both longwave and shortwave temperature changes beneath. - fixes to evaporation physics - further reduction of cool bias in daytime - fix to convective precipitation to average over deeper layer near surface within the Grell scheme - reduces excessive precipitation over warm water and improves convective precipitation over Florida and along Gulf Coast and along south Atlantic Coast. Diagnostic fixes - CAPE/CIN - mix lowest 30-40 mb instead of only 20-25 mb - less jumpiness from analysis to 1h fcst. CAPE users will like this improvement (developed in collaboration with SPC forecasters), - tropopause level fix - smoother trop fields due to a fix in the algorithm. See http://maps.fsl.noaa.gov/vartxt.cgi for more details on RUC tropopause levels fields and other diagnosed RUC fields (e.g., SLP, precip type, snow accumulation...) - max wind level fix in case wind is zero at all levels ( not likely, but...) GRIB table fixes - improved precision in several variables, including soil moisture. This fix is very important since the RUC soil moisture has been unable to dry out properly since the GRIB precision was not sufficient to resolve 1-h changes in this field. The RUC soil moisture is too high overall now due to this problem, and will finally start to adjust with this change, but it will take a few weeks or longer. Eta boundary conditions - The Eta model changed its definition of relative humidity on 11/3/98 from with respect to water to a water/ice combination. Although this change gave more realistic RHs for forecasters thinking about ice clouds, it was unannounced and started causing problems for the RUC and local models using Eta for initial conditions and still assuming that the Eta RH was with respect to water. There is far too much ice cloud in the RUC since 11/3 because of this problem, and especially near boundaries. This change accounts for the change to Eta RH definition.
VAD winds turned off in RUC. Brad Ballish at NCEP noticed a large number of zero VAD winds from different stations getting into the global model and after some further review by NCEP and us, the VAD winds were turned off temporarily from all NCEP models today, the RUC as of 1900 UTC. These bad winds often had the highest VAD quality control flag.
Effect on output - There appears to be some improvement in overall lower-tropospheric wind forecasts from the RUC from turning off the VADs. Obviously, we are losing the benefit of the good VAD winds, but also avoiding contamination from the zero winds that were getting thru the QC.
GETBUFR (GETBUFR.f)
This program reads in preprocessed BUFR data and writes out files to be used by the RUC-II analysis system. It was changed to correct the problem that the model has experienced since the new decoding system made more Canadian surface data available to the RUC-II. Specifically, there was no accounting for missing altimeter setting values in the GETBUFR code, and those values actually missing were given an initialized value of 99999, which creates 999mb surface lows scattered across Canada.
Effect on output - Got rid of anomalous 999 lows around Canada.
RUC2 run script
This script was modified to correct the FAA's problem in obtaining the RUC-II grid #236 output files. The files must be routed with a KWBG header for OSO to transmit them to the FAA.
Effect on output - None, except header identifier.
HYBFRONT (gtraob.f)
Significant error about RUC-2 use of significant level temperature and dewpoint temperature data fixed. Previously, the RUC was not using significant level temp/dewpoint data, and worse, was forcing in linearly interpolated values between mandatory levels at raob sites. See the forum message for more details.
Effect on output - Much improved sounding structure, CAPE/CIN, improved precipitation, and probably many other things.
hybint,hybpostl (hybint.f,hb2p.f), hybcstpr, hybfront (hybanl.f)
The code changes include
Effects:
Speed of processing -- The processing of the RUC at NCEP has speeded up fairly dramatically since late last week, as some of you may have noticed. NCEP allocated 10 processors on the Cray for all runs for the RUC as part of improving the use of the Cray for the overall NCEP model suite.
The effects of this change are that RUC forecast grids at 0300, 0600, 0900, 1500, 1800, and 2100 UTC are available about 25-30 min earlier than previously. Analyses at all times are available about 5-8 min earlier, as are the 3-h forecasts at non-3-hourly times.
HYBCSTPR (getveg.f), soilsave.dat
Error causing vegetation fraction to be set to zero is fixed. Soil moisture and temperature fields at all levels are re-initialized from field from FSL. This field started on Thurs 3 June as half from the control 40km MAPS run and half from the RUC to eliminate the obvious hot/dry bias of the RUC but also to improve on the slight cool/moist bias during the day of MAPS. This was put into the parallel MAPS cycle on this day, and then a new soilsave.dat from this parallel run was sent to NCEP to restart the soil moisture and temperature for this time. See the forum message and other associated forum messages for more details.
Effect on output - Substantially reduces hot/dry surface bias in 2100 - 0900 UTC range. Slightly increases cool bias at 1200-1800 UTC. Improves precipitation forecasts, 850/700 temp forecasts.
HYBNCEP (hybncep.f)
Fixes were made to fix errors in the moisture fields used in the Rapid Update Cycle (RUC) boundaries. The primary error was that saturation vapor pressure is calculated using virtual temperature rather than temperature. The second error is that the saturation vapor pressure at temperatures below -15 deg C was calculated with respect to ice instead of water. The impact of the fixes to these errors is to eliminate much of the erroneous precipitation along the RUC boundaries over warm oceans and slightly reduce some of the excessive areal coverage of RUC precipitation over warm oceans.
Examples of 12-h RUC precipitation before and after the change may be viewed at http://maps.fsl.noaa.gov/ruc_bc_fix.html.
Backup RUC from FSL
Due to the NCEP fire and need to move the NCEP IBM SP to a better location, a version of the RUC was run at FSL until 15 November 1999. See http://maps.fsl.noaa.gov/rucbackup_diff.html for a log of differences in the backup RUC from the NCEP RUC. See also the RUC forum http://maps.fsl.noaa.gov/forum/eval for information on the backup RUC.
NCEP RUC on IBM restarted
Faster post-processing with this version. Analysis and each output time post-processed as they become available, rather than waiting until end of forecast model.
HYBFRONT (capcin.f, hybanl.f), HYBPOST (capcin.f, mapsout_post.f), HYBINT (hybint.f)
Upgrade of post-processing. 8 new variables added (sfc wind gust, visibility, cloud base, stable cloud top, PBL height, convective cloud top, equilibrium level, pressure of max theta-e), and fix made so that 3h sfc pressure change works again.
HYBINT (write_grib.f)
Fix made so that GRIB header has proper month and date of month (fix to leap year problem).
BUFR dump for RUC
VAD wind profiles now getting into RUC again every third hour (but not at 01,02,04,05 ... UTC)
HYBINT (HYBINT.f, grib.table)
Tropopause pressure added as output variable. Output for FAA added at 4,5,7,8 hours output.
HYBFRONT (PIREPIN.f, SFCIN.f, RAOBIN.f, MAPSCON)
Changes to prevent large numbers of obs from exceeding array sizes. New checks introduced, and limit increased by factor of two. This all resulted from a change to MDCRS to allow Federal Express observations in at NCEP.
SST analysis
Changed to use coarser resolution SST analysis at NCEP; similar change for Eta made previously. Based on NCEP/EMC investigation of SSTs on Eta model forecasts. Will be changed back to improved high resolution analysis later.
BUFR script providing data to RUC
Change in surface data provided to RUC, resulted in no surface wind data being available.
GETBUFR (RUC program)
Surface winds now getting in (but as we later discovered with wind speed off by 50% factor).
GETBUFR (RUC program)
Surface winds now getting in with proper wind speed.
SST used
Changed to new 2dVAR SST analysis from NCEP. More information here from Geoff Manikin (NCEP/EMC) on RUC Forum
HYBFRONT (exsfcobs.f)
Fix made for extrapolation of METAR temp/dewpoint to
RUC surface elevation in RUC analysis. Background lapse
rate used in this calculation, and was allowed to be
as stable as isothermal, resulting in possibility of divide by zero
condition. Changed to be slightly less stable than
divide by zero (0.0005 K/m) to avoid this condition.
This fix solved at least one RUC crash that occurred
in February and may explain other occasional RUC crashes
that have occurred.
HYBINT (helic.f)
Fixes/improvements to helicity and storm motion calculations. Lower (and more accurate) helicity values. Now uses Bunkers et al. storm motion. Same as version of helic.f developed for 20km RUC.