Mesoscale Modeling2014, 2015 and beyond …

Geoff DiMegoDecember 2, 2014

geoff.dimego@noaa.gov 301-683-3764

TOPICS• Implementations in 2014• RTMA-URMA upgrades in 2015• HiResWindow upgrade in 2015• North American Mesoscale upgrade in

2015• NARRE/SREF and HRRRE/CASE in 2016-

20182

Implementations in 2014• RTMA v2.2 with URMA 1/28/2014

– Drop legacy reject list 5/13/2014 OCONUS– Drop legacy reject list 8/12/2014 CONUS

• RAP v2 2/25/2014• SREF v6.1 4/2/2014• HiResWindow v6.0 6/11/2014• NAM v3.1,nests, DGEX 8/12/2014

– DNG winds corrected 11/13/2014

• HRRR v1 9/30/2014• HYSPLIT v7.2.2 [see BACKUP] 9/30/2014 3

RTMA/URMA UpgradesTo V2.3 in Q2 & To V2.4 in Q4

Manuel Pondeca, Steve Levine, Jacob Carley, Geoff Manikin, Jeff McQueen, , Yanqiu Zhu, Ying Lin, Dave Parrish, Jim Purser

Manuel.Pondeca@noaa.gov 301-683-3656

RTMA/URMA v2.3 Upgrade1. Replace RAP first guess with HRRR or HRRR+NAMnest blend for

RTMA-CONUS and URMA [gathering stats & feedback, discussion & decision next week 9th & 10th December 2014]

2. Apply new “buddy check” to make rejection less likely 3. Modify gross-error check to account for terrain variability4. Analyze Total Cloud Amount (a.k.a. Sky Cover) 5. Extend Stage IV gridded precipitation look-back period used by

URMA to 7 days [see BACKUP SLIDES]6. Synchronize all RTMA/URMA applications to use the same code

5

RTMA/URMA v2.4 Upgrade1. Analyze Min/Max 2m Temperature, cloud ceiling height,

significant wave height, and 10m wind speed as a scalar variable2. Use EMC/GFE common topography and land/sea mask fields3. Use enhanced variational obs quality control within the GSI 4. Add automatic updating of static accept- and reject-lists5. Extend CONUS domain westward to support OPC6. Replace GFS first guess with GFS+HiResWindow blend for GUAM7. Implement URMA for Alaska (if late arriving obs warrant)8. Add GLERL-type coastal-ob-adjustment to improve wind analysis

over Great Lakes9. Improve first guess 10-m winds via downscaling [see BACKUP]10.Prepare for first 10-year reanalysis [SS-BLENDER & USFS] 6

URMA (UnRestricted Mesoscale Analysis)

• Identical to RTMA except with a 6 hr delay • Hence, unrestricted by real-time requirement• Allows use of late arriving observations• Allows potential use of temporal consistency

checking ala CLARIS – MODIS • Allows use of shortest (1 hr) HRRR & NAM-

nest forecasts as first guess for analysis7

HiResWindowUpgrade to V6.1

Q2FY2015Matt Pyle, Hui-ya Chuang, Zavisa Janjic, Ying Lin, Eric RogersEric Aligo, Jacob Carley, Brad Ferrier, Shun Liu, Perry Shafran

Matthew.Pyle@noaa.gov 301-683-3687

HiResWindow Upgrade to V6.1

1. Increase vertical resolution from 40 to 50 levels2. Add requested products such as ceiling height and

lightning3. Incorporate a post-processed ensemble product

generator [CASE=Convection-Allowing Scale Ensemble]

4. Eliminate production of intermediate GRIB1 files5. Improve forecasts of echo top height [AWC]6. On Phase 2, pre-emption should not be necessary

9

10

Feature v6.1 ConfigurationModel code version WRF v3.6 (ARW)

September 2014 version (NMMB)

Horizontal grid spacing 3.0 - 3.6 km NMMB3.5 - 4.2 km WRF-ARW

Vertical levels 50Microphysics (ARW) WSM6 (looking to improve quality of echo top

forecast for AWC)

Microphysics (NMMB) Ferrier/Aligo – adopt NAM-nest version (to enhance convective storm structure & performance)

Radiation (NMMB) RRTMv3 (updated code for improved efficiency & interaction/integration with other physics)

Initial conditions RAP & GFS

HiResWindow v6.1 Overview

North American Mesoscale (NAM) Upgrade to V3.2Q4FY2015 or Q1FY2016

Eric Rogers, Tom Black, Hui-ya Chuang, Mike Ek, Zavisa Janjic, Dennis Keyser, Ying Lin, Jeff McQueen, David Parrish, Matt Pyle, Wan-Shu Wu

Jacob Carley, Brad Ferrier, Jim Abeles, Eric Aligo, Ed Colon,, George Gayno, Dusan Jovic, Hsin-Mu Lin, Yangrong Ling, Shun Liu, John Michalakes, Perry

Shafran, Ratko Vasic, Jeff Whiting, Yihua Wu, Yanqiu Zhu

Eric.Rogers@noaa.gov 301-683-3682

NAM v4.0 in Late 2015-Early 2016Partial & Tentative List of Changes – 1.

• Upgrade NMMB prediction model• Increase resolution of CONUS and Alaska nests to 3 km• Physics upgrades to reduce warm season high QPF bias in CONUS nest• LSM upgrades to reduce winter afternoon cool bias over the CONUS

• Upgrades to GSI analysis and data assimilation• Replace NDAS’ 3 hr update frequency with RAP-like hourly cycle of NAM Rapid

Refresh (NAMRR) • Continue NDAS-like reach-back to GDAS with catch-up cycle every 6-h or 12-h• 4-d version of hybrid ensemble variational analysis• Diabatic digital filter radar-derived temp. tend. & cloud analysis [in BACKUP SLIDES]• Tropical cyclone relocation [Sandy Supplemental]

• Add hourly RAP-like18 hr & HRRR-like 15 hr forecasts via NAMRR12

NAM v4.0 in Late 2015-Early 2016Partial & Tentative List of Changes – 2.

• Upgrade observation processing• Accommodate TAC –to–BUFR conversion of international transmissions including high density

radiosondes with thousands of levels in vertical• Use Doppler obs from MRMS ingest on IDP with 3 new Caribbean sites [SS]• Improve radial wind qc and start moving MRMS into GSI [NextGen]

• New observations in analysis and data assimilation• SEVIRI, NOAA17-18 SSMIS, Metop_B (IASI, HIRS4, AMSUA, MHS), Himawari, NPP (ATMS, CRIS), … just to

name a few ^_^• New sources of aircraft observations [Sandy Supplemental]• Tower & nacelle obs from wind power producers

• Improve / enhance output products• Convert all NAM gridded output from WCOSS post-processing to GRIB2• Remove legacy output grids/products (will discuss this Thursday AM)

13

NARRE + SREF HRRRE + CASE

• 9-12 km continental• NARRE [North America Rapid Refresh Ensemble]

– Hourly– 18-24 hr forecast range– Subsumes RAP + NAMRR

• SREF [Short Range Ensemble Forecast]

– 6 hourly – 84-96 hr forecast range with some

extensions of NARRE members– Subsumes NAM

• 3 km CONUS, AK, HI, PR• HRRRE [High resolution Rapid Refresh Ensemble]

– Hourly– 15-18 hr forecast range– Subsumes HRRR + NAMRRnests

• NCASE [Convection Allowing Scale Ensemble]

– 6 hourly– 48-60 hr forecast range with some

extensions of HRRRE members– Subsumes HiResWin + NAMnests

14

EMC Model Evaluation Group

Weekly MeetingThursday 11:30 AM

Room 2890

B A C K U P S L I D E S

16

RTMA/URMA: What’s gone wrong●13 km resolution RAP did not properly resolve complex terrain features, sometimes produces inaccurate forecast●Many mesonet obs were on a WFO/region provided reject list and withheld from the analysis●Obs not on reject list fail gross error check due to large (>10 F) O-B innovation

What we’ve done about it●Relaxed gross error check over complex terrain, added buddy check to “save” obs previously thrown out●Removal of obsolete WFO-provided reject lists (ops and parallel)●Background now blend of HRRR (3 km) and CONUS NAM nest (4 km)

Courtesy Steve Levine

OPER RTMA ANALYSIS (2m T, ℉)

Ob Key:

Accepted

Rejected

Partial (Crossval)

10/3/1410Z

Courtesy Steve Levine

PAR RTMA ANALYSIS (2m T, ℉)

10/3/1410Z

Ob Key:

Accepted

Rejected

Partial (Crossval)

Courtesy Steve Levine

Differences at key obs (2m T, )℉Site Ob Val Oper BG Oper Anl Para BG Para Anl

KS03 46.31 57.83 55.13 57.83 50.99

ODT92 44.69 44.69 41.45 54.77 47.75

BUCO3 56.93 48.83 48.29 55.13 54.95

EMIO3 55.49 53.87 52.61 55.67 53.69

ODT59 37.85 43.25 41.81 44.15 39.65

CTHLT 50.81 45.59 47.03 54.05 51.71

ODT26 51.53 48.11 48.11 54.23 50.45

Courtesy Steve Levine

Changes to Downscaled Grids

• 5 km CONUS / 6 km Alaska DNG grids extended to 192-h via DGEX

• Addition of Haines Index for Fire weather

• Improved 10-m wind treatment• Use mass-consistent wind field

model• Based on velocity potential,

incorporating local terrain gradients

• DGEX extension will not be on AWIPS simultaneously with NAM implementation, this will probably occur sometime during fall 2014.

21

Ops downscaled NAM AKNEST

(3 km)

Parallel downscaled NAM AKNEST (3 km)

Improved representation of the effects of local

terrain on winds Courtesy Jeff McQueen

Diagnostic Wind Downscaling• Compute the velocity potential

– solve the Poisson Eqtn using Gauss-Seidel Method from the terrain gradients

U=U + S*(PHI i+1 - PHI i-1)/ΔXV= V + S*(PHI i+1 - PHI i-1)/ΔX

where PHI=U * ΔZ/ΔX + V * ΔZ/ΔY = velocity potential ΔZm = parent model topographyΔZd = 2.5 km NDFD grid topographyS=Abs (ΔZm – ΔZd)/ ΔZd (implemented 11/13/14)

22Courtesy Jeff McQueen

Unified Smartinit Downscaling UpgradeQ3FY2015

Significant upgrade that introduces: - Expanded CONUS 2.5 km domain;- GRIB2 input/output;- Physics upgrade (adiabatic wind adjustments, improved coastline adjustments for lakes, temperatures in valleys);

- NLDAS 2 m temperature, spec. humidity option; - New products (80 m winds for energy sector, wx type)- Code optimization with IBM.

23Courtesy Jeff McQueen

Precipitation URMA, implemented on 28 Jan 20146-.

6-hourly multi-sensor precipitation estimates from the 12 ConUS River Forecast Centers (RFCs) are mosaicked into a national product (the NCEP Stage IV) and remapped to the ConUS and Northwest NDFD grids.

Courtesy Ying Lin

6h pcp accum ending 00Z 5 Nov 2013

at 14:35Zat 13:35Z 5 Nov

at 17:35Z

The RFCs generally transmit their 6-hourly analysis files covering the 12Z-12Z 24h period in the several hours after 12Z. Complete ConUS coverage is usually achieved by 18:35Z.

25Courtesy Ying Lin

Upcoming URMA upgrade (Q2FY15)

06h accum ending 18Z 20140928.

Nov 2014: NCO implemented an upgraded database for incoming QPE data from the RFCs.In the upcoming RTMA/URMA upgrade package, additional re-mosaics for 6-hourly Stage IV/precip URMA will be made at 1/3/5/7 days after ending of the accumulation time.In addition, hourly QPEs from the 8 Eastern/Central RFCs are first summed into 6-hourly totalsthen combined with 6-hourly QPEs from the four Western RFCs, to take into account of regionaldifferences in base (primary) analysis.

At 23:35Z 28 Sept – data from MBRFCdid not arrive until 15:17Z 30 Sept

Re-run made 3-days later Courtesy Ying Lin

27

0-3 h PCP verification

From 20130223 to 20130712

Red – old opnl NAM before 8/12/2014Blue – new opnl NAM since 8/12/2014Green – new with new cloud analysis

Courtesy Shun Liu

28

3-6 h PCP verification

From 20130223 to 20130712

Red – old opnl NAM before 8/12/2014Blue – new opnl NAM since 8/12/2014Green – new with new cloud analysis

Courtesy Shun Liu

29

24 h PCP verification

From 20130223 to 20130712

Red – old opnl NAM before 8/12/2014Blue – new opnl NAM since 8/12/2014Green – new with new cloud analysis

Courtesy Shun Liu

North American Rapid Refresh ENSEMBLE (NARRE) 2016-2018

● NMMB (from NCEP) & ARW (from ESRL) dynamic cores● Common use of NEMS infrastructure and GSI analysis● Common NAM parent domain at 9-12 km● Ensemble made up of equal numbers of NMMB- & ARW-based

configurations (number of members T.B.D.)● Hourly updated with forecasts to 18-24 hours● NMMB & ARW control data assimilation cycles w/ hourly updating

and periodic 3-6 hr catch-up periods (to reach-back to GDAS)● SREF 84 hr forecasts are extensions of the 00z, 06z, 12z, & 18z runs

– a NAM could be NMMB control run for continuity sake.● SREF will be at same 9-12 km resolution as NARRE by then● SREF will have 20-30 members with some coming from extensions of

NARRE which for a while will have fewer members than SREF30Courtesy Eric Rogers

High Resolution Rapid Refresh ENSEMBLE (HRRRE) 2017-2019

● Each member of NARRE contains 3 km nests ● CONUS, Alaska, Hawaii & Puerto Rico/Hispaniola nests● The two control runs initialized with radar data & other hi-res obs

● This capability puts NWS/NCEP[+OAR/ESRL] in a position to ● Provide NextGen Enroute AND Terminal guidance (FWIS-like)● Provide PROBABILITY guidance with full Probability Density Function

specified, hence uncertainty information too● Provide a vehicle to improve assimilation capabilities using hybrid

(EnKF+4DEnVar) technique with current & future radar & satellite● Address Warn-on-Forecast as resolutions evolve towards ~1 km

31Courtesy Eric Rogers

HYSPLIT Dispersion Model

Overview of Changes in Version 7.2•ALL – Update to unified HYSPLIT code and libraries•CTBTO – New application & NCO web GUI for model configuration

– Extend retention of GDAS/GFS ARL packed format files

/com/hysplit to 30 days•Volcanic Ash – Add ½ ˚ GFS hysplit format files •RSMC – radiological text product to MWOs (AWC, Anchorage and Honolulu)

– Add ½ ˚ GFS hysplit format files•WFO Hazmat – Improved Google Earth Graphics (NCO is backup)•Smoke – Include Canadian & Mexican emissions • Dust – No specific changes

Courtesy Jeff McQueen

Smoke verificationagainst NESDIS GASP smoke mask 9/11/14 18z

33Courtesy Jeff McQueen

CMAQ V4.6.3Inclusions of real-time smoke emissions

• NESDIS Hazard Mapping System (HMS) provides fire points• USFS BlueSky system run within HYSPLIT dispersion model provides fire

emissions (organic carbon, black carbon…)

34Courtesy Jeff McQueen

1h Avg Diurnal Ozone performanceCMAQ Prod vs V4.6.3

35

Diurnal forecast for West- Underprediction during day

Diurnal forecast for East- Overprediction at all times

Courtesy Jeff McQueen

Q4FY2015 Upgrades• AQ Upgrades to Emissions/Chemistry (CMAQ)

1. Upgrade CMAQ as needed with improved chemistry mechanism & vertical transport.

2. Upgrade emissions with 2008 base year estimates and projections.

3. Implement BlueSky V4.0 smoke emission software.

4. Inclusion of additional vertical levels5. Use NGAC aerosol lateral boundary

conditions6. Implement PM bias correction7. GRIB2 I/O coupling

• NAQFC : HYSPLIT v7.4 Dispersion Model1. Meteo Prep

– Grib2 input for NAM/GFS– Option for HRRR GRIB2 input

2. On-demand SDM Hysplit runs (volcanoes, RSMC, DHS, HAZMAT)

- Improved 4 km NAM coupling - Transition static RSMC graphics web page to Operations - Transition volcanic ash graphics to MAG - Add Transfer Coefficients Matrix (TCM) outputs for flexible

source specification3. Canned pre-configured runs

– Add trajectories for lake effect snow forecasts (WFO request)

4. Smoke/dust– Upgrade to BlueSky v4.0 smoke emissions– 3/4 km NAM nest coupling for short term smoke

predictions– Grib2 output5. All: Update to unified HYSPLIT code and libraries to V7.4

36