Testing and Upcoming Implementation of the 20km RUC Stan Benjamin NOAA / FSL [email protected]303-497-6387 http://ruc.fsl.noaa.gov Georg Grell Steve Weygandt Tracy Lorraine Smith Tanya Smirnova Barry Schwartz Dongsoo Kim Dezso Devenyi Kevin Brundage John M. Brown Geoff Manikin – NCEP/EM Ralph Petersen – NCEP/E NCEP Production Suite Review 11 December 01 – updated 15 Feb 0
34
Embed
Testing and Upcoming Implementation of the 20km RUC Stan Benjamin NOAA / FSL [email protected] 303-497-6387 Georg Grell Steve.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Testing and Upcoming Implementation of the 20km RUC
Stan BenjaminNOAA / [email protected] 303-497-6387http://ruc.fsl.noaa.gov
Georg GrellSteve WeygandtTracy Lorraine SmithTanya SmirnovaBarry Schwartz Dongsoo KimDezso DevenyiKevin BrundageJohn M. BrownGeoff Manikin – NCEP/EMCRalph Petersen – NCEP/EMC
NCEP Production Suite Review11 December 01 – updated 15 Feb 02
2
Key goals for 20km RUC• Focus on improving RUC performance for
• QPF, especially for convective precipitation• Surface forecasts• Cloud/icing 3-d fields
• Take advantage of increased computer power40 km / 40 levels 20 km / 50 levels
• Improved analysis in native coordinate, cloud analysis
3
RUC20 Testing Since May 2001
Ongoing real-time 1-h cycle at FSL and NCEP Retrospective period tests – full 1-h cycle
Feb 2001 at NCEP and FSL July 2001 at NCEP May 2000 at FSL
Examination of verification against precipitation, radar surface observations rawinsonde observations
Web page products – daily look at maps For both RUC20/40 analyses and forecasts
4
RUC20 Implementation schedule
CAFTI information items- March, Sept, Nov 2001, Feb 2002
Complete retrospective tests Feb 2001 – cold season - complete July 2001 – warm season – almost complete
Real-time tests – late Feb - March 2002 CAFTI approval – late March 2002 Operational implementation – 9 April 2002
5
Subset of full domain
20km RUC 40km RUC
Model terrain elevation
6
Subset of full domain
20km RUC 40km RUC
Model terrain elevation
7
40km RUC40 levels
20km RUC50 levels
10 new levels7 – upper levels – 330 – 500 K3 – lower levels – 270 – 290 K
RUC native coordinate levels
W-E vertical cross-section - Pacific – w. coast6h fcsts valid 0000 UTC 24 Feb 2001
800 1600 2400 3200
8
Current Operations – RUC40 RUC20
Resolution 40 km, 40levels 20 km, 50 levelsAnalysis Optimal interpolation on Modified OI technique on generalized
on generalized surfaces surfaces, hydrometeor analysis w/ GOES…, use raw instead of interp. obs
Assimilation Intermittent 1-h cycle Intermittent 1-h cycleStable clouds Mixed-phase cloud microphysics MM5), New version of MM5/RUC / precipitation explicit fcst of cloud water, rain water, microphysics (FSL and NCAR),
snow, ice, graupel, much shorter microphysics time step, no. concentration of ice particles less graupel, more supercooled liq water
Sub-grid-scale Grell (1993) New Grell scheme w/ ensemble cloud, precipitation shallow convection, detrainment of
cloud water to microphysics Turbulence Burk-Thompson explicit TKE scheme Modified Burk-ThompsonRadiation MM5 LW/SW scheme, f(hydrometeors) MM5 scheme with fix to SW lag error
daily NESDIS snow/ice (same as Eta), albedo0.14 monthly NDVI veg frac (same as Eta), cycled soil moisture/temp, snow depth/temp
Lat boundary conds Eta model initialized every 12h Eta model initialized every 6h
Rapid Update Cycle – Present and Next Version
9
Key RUC20 Modifications Since May 2001
- Forecast model- revisions to Grell ensemble convective parameterization- use of GOES cloud-data to prompt convective initiation- revised vertical advection / hybrid isentropic-sigma coordinate movement
Prompted by PC Spring Program evaluation, model stability- Analysis
- 3dVAR found to give slightly poorer wind forecast for 3h period (not at 12h)- Decision made to introduce OI analysis within 3dVAR framework, defer 3dVAR analysis implementation for ~3 months after RUC20 impl.Prompted by FAA use of RUC for Free Flight
Data Type ~Number Freq.Rawinsonde (inc. special obs) 80 /12hNOAA 405 MHz profilers 31 / 1hBoundary-layer (915 MHz) profilers ~24 / 1h * VAD winds (WSR-88D radars) 110-130 / 1h Aircraft (ACARS) (V,temp) 1400-4500 / 1hSurface/METAR - land (V,psfc,T,Td) 1500-1700 / 1hBuoy 100-150 / 1hGOES precipitable water 1500-3000 / 1hGOES cloud drift winds 1000-2500 / 1hGOES cloud-top pressure ~40km res / 1hSSM/I precipitable water 1000-4000 / 6hGPS precipitable water ~80 / 1h * Ship reports 10s / 3hReconnaissance dropwinsonde a few / variable
Data for 20-km RUC at NCEP New for 20km RUC at NCEP
Availability pending
11
Improved use of obs in RUC20 analysis
• More accurate forward models-- simulate observations from background forecast as closely as possible• 2m T/q, 10m winds from METARs, buoys• Use of background grid-point with matching land-use for coastal stations instead of bilinear interpolation to ob location• Account for elevation differences (model vs. ob) for precipitable water• Better handling for near-surface raob structure
VerificationCloud-top pressurebased on NESDIS product
Cloud-top forecast verification - correlation coefficient between forecast and NESDIS cloud-top product- much improved cloud forecasts even at 12h28 Sept – 2 Oct 2001
40km RUC 20km RUC w/ cloud analysis
17
Key changes in 20km RUC forecast model for QPF
• 20km/50 level resolution – resolved precipitationimproved – grid volumes are now ~5 times smaller
• Cloud microphysics time step much smaller -- 10 2 min
• Revised hybrid coordinate definition, verticalregridding for all variables
• Bug fixes in interface between Grell scheme and rest of model
• New ensemble cumulus parameterization – Grell• Revised version of MM5/RUC microphysics
• BUFR sounding output identical to Eta– Current – separate monolithic files for each output time only– New – station files with all output times - ‘monolithic’ file w/ all stations, output times
• RUC20 forum for questions (under RUC web site)• Backup for RUC20 at FSL (similar to current RUC40)
RUC web site - http://ruc.fsl.noaa.gov - 20km test RUC products
34
Summary - the 20km RUC• Schedule for implementation
– Current – real-time, retro testing at NCEP and FSL– Mar 2002 – field test for RUC users– Implementation – 9 April 2002, 3dVAR follows by ~ 3 months
• 20 km/50 level 1 hr version– with model improvements including cloud microphysics, convection, land-
surface– Improved optimal interpolation analysis, cloud/hydrometeor analysis using
satellite combined with explicit cloud fcsts in RUC– Improvements in warm- and cold-season precipitation and cloud/icing
forecasts, surface forecasts, wind/RH/temp aloft
• 10 km RUC – testing at FSL in 2002 w/ current RUCand WRF (NCAR-mass coordinate) models
RUC web site - http://ruc.fsl.noaa.gov - 20km test RUC products