NOTE: The nws-changes mailing list is no longer active. The list archives are made available for historical reasons.
On 17 July 2002, several upgrades were made to NCEP's HiResWindow (used to be known as Threats) runs. These are made daily on an experimental basis in the run-slot of NCEP's Operational Production suite reserved for the GFDL Hurricane model runs. Any need to run the hurricane model ALWAYS takes precedence over a HiResWindow run. Four large domain and two small domain nests are run on a rotating schedule through the day: Cycle Time Large Domain Small Domain 00z Alaska Hawaii 06z Western Puerto Rico 12z Central Hawaii 18z Eastern Puerto Rico Several of my CAFTI briefings in the recent past have referred to these runs and this upgrade. The most recent dealt with the implementation of a special selectable 4 km configuration in an On-Call Emergency Response capability for tracking hazardous releases which was declared Operational on 25 June 2002. This briefing can still be viewed at http://www.emc.ncep.noaa.gov/mmb/mmbpll/hiresw.cafti/ Slides 12, 13, 14, 21, 22, 23, 24, 38 and 39 are most relevant to the upgrades listed below: 1) Model: We are replacing the hydrostatic Meso Eta Model with the Nonhydrostatic Meso Model (NMM - see the initial paper by Janjic, Gerrity, and Nickovic [MWR,2001, Vol. 29, No. 5, 1164-1178] for model equations, solution techniques etc.). This is the first nonhydrostatic model to run routinely at NCEP. While much of the original Meso Eta code remains - especially for the physics, the NMM has been tailored to deal especially with high resolutions. First and foremost, the NMM uses nonhydrostatic dynamics which are required to account for the vertical accelerations that become significant at horizontal grid resolutions less than 10 km. Among numerous other refinements, the NMM replaces use of the step-mountain eta coordinate with a hybrid one where the lowest layers are terrain-following sigma topped by constant pressure surfaces. This choice will avoid the problems encountered at high resolution (10km or finer) with the step-mountain coordinate with strong downslope winds and will improve placement of precipitation in mountainous terrain. 2) Horizontal Resolution: The horizontal resolution will increase from 10 km to 8 km for all nests except Alaska which goes from 12 km to 10 km. We run the identical model code (array sizes) for the four large domain nested runs and the need to adequately include Juneau in the Alaska nest requires us to cover a larger area and, therefore, use a slightly coarser resolution than for the three CONUS domains. 3) Vertical Resolution: Despite the statement in slide 38 that we are increasing the number of levels from 50 to 60, the hydrostatic nests were already running with 60 levels and we are keeping the same number of vertical levels in the NMM. There is some redistribution of levels as a consequence of using terrain-following sigma instead of step- mountain eta. The first model layer above ground in the step-mountain coordinate was thinnest over the ocean and became thicker as you moved to higher and higher terrain. The opposite is true in sigma where the thinnest surface layers are over the highest mountain peaks. To keep these mountain top layers from getting too thin (which can lead to vertical stability problems), the depth of the first sigma layer has been set to 40 m over the oceans compared to the 20 m used there in the Meso Eta. So, except near the ocean's surface, the boundary layer in the NMM will have more vertical resolution than in the Meso Eta. Product availability: These runs are viewable at Eric Rogers' web page http://www.emc.ncep.noaa.gov/mmb/mmbpll/nestpage/ Output from the runs are in the form of GRiB grids and BUFR soundings can be ftp'ed from either the NCEP or NWS/OPS sites. The NCEP site is at ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/eta/para/ and will contain directories named hiresw.yyyymmdd where yyyy is the year, mm is the month and dd is the day. The NWS/OPS site is at ftp://tgftp.nws.noaa.gov/SL.us008001/ST.opnt/ and will contain directories named MT.meso_CY.cc/RD.yyyymmdd/ where cc is cycle time either 00, 06, 12 or 18 and yyyy is the year, mm is the month and dd is the day. Each nested domain has its own output grid at full model resolution. Eric Rogers has been able to nearly duplicate the computational domain in the coverage of these grids as follows: Domain Res. Grid Type GRiB Grid # Eastern US 8 km Lambert Conformal 245 Western US 8 km Lambert Conformal 246 Central US 8 km Lambert Conformal 247 Alaska 10 km Polar Stereographic 249 Hawaii 0.075 degree Lat/lon 250 Puerto Rico 0.075 degree Lat/lon 248 Eric's model and product grid page has additional information: http://www.emc.ncep.noaa.gov/mmb/etagrids/ REMEMBER, the large domain HiResWindow runs are only made when there are NO GFDL Hurricane Model runs being requested by NCEP's Tropical Prediction Center (aka National Hurricane Center). At this time of year, running the GFDL model will be a fairly frequent occurrence. Eric Rogers' website has a link to a log file http://www.emc.ncep.noaa.gov/mmb/mmbpll/nestpage/hiresw_log which is updated at the completion of the generation of his web graphics. This file is updated about an hour after the HiResWindow runs finish and it only reflects successful runs. There is no single place that tells you whether a GFDL run(s) is being made. It is possible to determine what has run after the fact by checking the contents of sites where the hurricane runs are made available. The NCEP site to check is ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/hur/prod/ while the NWS/OPS site is at ftp://tgftp.nws.noaa.gov/SL.us008001/ST.opnl/ with directories named MT.ghm_CY.cc/RD.yyyymmdd/ where cc is cycle time either 00, 06, 12 or 18 and yyyy is the year, mm is the month and dd is the day. If the directories don't exist or if there are no files in them, then you should expect large domain HiResWindow runs to have been made. Please direct all questions to: Geoff.DiMego@xxxxxxxx or Eric.Rogers@xxxxxxxx
nws-changes archives: