NOTE: The nws-changes mailing list is no longer active. The list archives are made available for historical reasons.
Notice of Intent to Change the NCEP Rapid Update Cycle (ASCII below, Word document also attached) Brief description of change(s): The Rapid Update Cycle will increase its horizontal and vertical resolution, and also improve its data assimilation and numerical prediction model. The horizontal resolution will change from 40 km to 20 km (covering the same domain), and the number of vertical levels will be increased from 40 to 50. A 3-dimensional variational analysis in the RUC?s isentropic/sigma hybrid coordinate system will replace the current optimal interpolation analysis. The analysis will also now include an initial cloud analysis, with assimilation of GOES cloud-top pressure, to modify RUC 1-h forecast hydrometeor fields. The 20km RUC model includes upgrades in explicit cloud microphysics, convective parameterization, and land-surface scheme. The 20km RUC will run out to at least 6h every hour, as opposed to the current 3h forecast at initial times not divisible by 3. It will also begin to use the 06z and 18z Eta runs, along with the 00z and 12z runs, for its boundary conditions. ?Look-alike? 40km grids will also be produced from the 20km RUC, allowing an easy transition by users. More information about the 20km RUC is available at http://ruc.fsl.noaa.gov/20kmRUC-apr01.htm (a slightly extended version of a CAFTI presentation from 22 March 2001). Reason for Change(s): To improve RUC forecasts, especially for precipitation, clouds, and surface variables. The 20km RUC has been in real-time testing at FSL since July 2000, and with all of the components listed above since December 2000. Schedule for change: Final testing completed: 5 June 2001 CAFTI Approval: Expected implementation: 30 June 2001 Description of testing: The 20km RUC model has been in real-time testing at FSL since July 2000, and with the assimilation system since December 2000. Both case studies and statistical verification show considerable improvement for the RUC. Statistical comparisons with the 40km RUC have been made, including verification against rawinsonde and METAR observations, and against NCEP precipitation analyses. Statistical summaries: Initial summaries from FSL 20km RUC testing are available inside the RUC presentation at http://ruc.fsl.noaa.gov/20kmRUC-apr01.htm . These will be updated for longer periods in May, and more detail will be provided. Rawinsonde verification is made every 12h for wind, temperature, height, and relative humidity at mandatory levels. The raob verification generally shows similar performance to the 40km RUC, but with slightly reduced standard deviations at 250-150 hPa, and with reduced RH and temperature bias. Surface analyses and forecasts from the 20km RUC are verified every 3h against METAR data, with standard deviation and bias of 2m temperature and dewpoint, and 10m vector wind and wind speed. The surface verification shows a strong reduction in temperature, dewpoint, and wind speed biases, and an increased fit to METAR data. All of this verification has been done using the 40km ?look-alike? output grids from the 20km test RUC. Cloud-top forecasts are evaluated against NESDIS cloud-top products. Finally, 24h precipitation totals (from summing two 12-h RUC forecasts) have been verified against the NCEP 24h precipitation analysis for equitable threat score, bias, and categorical mean absolute error. These indicate that the 20km RUC significantly improves forecasts for 24h totals over 1?, both in winter and spring. Forecast examples: The performance of the 20km RUC running in real-time at FSL can be monitored and compared with the 40km RUC through the RUC web site (http://ruc.fsl.noaa.gov ) Also, examples from FSL 20km RUC testing are available inside the RUC presentation at http://ruc.fsl.noaa.gov/20kmRUC-apr01.htm. Anticipated impact on forecasts: Improved precipitation forecasts, especially in the warm season, improved cloud forecasts, improved wind and temperature forecasts near the tropopause. Also, the 20km RUC provides improved accuracy in surface forecasts with elimination of an insufficient diurnal cycle (40km RUC is too cool in the day and too warm at night). Human point of contact for further information: Stan Benjamin, 303-497-6387; e-mail: Stan.Benjamin@xxxxxxxx Field Evaluation: Aviation Weather Center and Storm Prediction Center and NWS regions will all participate in an evaluation during May. Storm Prediction Center will include a focus on the 20km RUC during its spring verification program. Grids from the 20km RUC runs at FSL are available via ftp already, and grids from the NCEP test 20km RUC will be available by 10 May. A web forum on the 20km RUC evaluation has been set up at http://ruc.fsl.noaa.gov/forum/eval20. A similar forum for the 40km RUC was very successful in disseminating information to RUC users during its evaluation, and has been continued since that time. Approvals: EMC Director: Stephen Lord Date: NCEP Director: Louis Uccellini Date: Future changes: A further upgrade to the RUC will be proposed in 2002, including improvements in the cloud/hydrometeor analysis using radar and surface observations, improvements in the microphysics scheme (developed primarily by NCAR/RAP) and convection scheme, and an increase in horizontal resolution to approximately 12 km. ____________________________________________ Stan Benjamin benjamin@xxxxxxxxxxxx NOAA Forecast Systems Lab 303-497-6387 - phone Chief, Regional Analysis and 303-497-4176 - fax Prediction Branch http://ruc.fsl.noaa.gov 325 Broadway Boulder, CO 80305-3328 USA
nws-changes archives: