ncep chemistry modeling overview and status (with a focus on nems aq development) sarah lu...

NCEP Chemistry Modeling NCEP Chemistry Modeling Overview and StatusOverview and Status

(With a focus on NEMS AQ development)(With a focus on NEMS AQ development)

Sarah LuNOAA/NWS/NCEP

Environmental Modeling Center

with acknowledgments to many colleagues and collaborators

National Central University, Chung-Li, Aug 26th, 2009

Acknowledgments:Acknowledgments:

EMC AQ group Jeff McQueen, Ho-Chun Huang, Youhua Tang, Dongchul Kim, Marina Tsidulko, Caterina Tassone

EMC UMIG group Mark Iredell, Henry Juang, Shrinivas Moorthi, Tom Black, Jun Wang, Weiyu Yang, Ratko Vasic, Ed Colon

EMC GMB Yu-Tai Hou, Suranjana Saha, Fanglin Yang, Xu Li, Jesse Meng, Yuejian Zhu, Jongil Han, John Ward

EMC GSI group John Derber, Russ Treadon, Daryl Kleist, Haixia Liu

CPC Craig Long, Shuntai Zhou

NWS OST Paula Davidson, Ivanka Stajner

OAR ARL Daewon Byun, Pius Lee, Roland Draxler, Ariel Stein, Hsin-Mu Lin, Daiwen Kang, Daniel Tong, Shao-cai Yu

GSFC Arlindo da Silva, Mian Chin, Thomas Diehl

EPA Ken Schere, Rohit Mathur, Jon Pleim

Howard University Everette Joseph, William Stockwell

NESDIS Shobha Kondragunta, Quanhua Liu, Yong Han, Brad Pierce


Comparison of RAQMS OMI+TES reanalysis with IONS ozonesondes

(373 sondes, August, 2006)

PI: ANNE M. THOMPSON Penn State

Fishman, J et al., “Remote Sensing of Tropospheric Pollution from Space”, BAMS June 2008Pierce et al. “Impacts of background ozone production on Houston and Dallas, TX Air Quality during the TexAQS field mission”, Accepted JGR-Atmospheres, February, 2009

Tropospheric biases: +/- 20%

The TES+OMI assimilation results in significant reductions in column, tropospheric (>100mb), and stratospheric (<100mb) biases (all less then 1%)

However, the low tropospheric biases are the result of compensating errors in the upper and lower troposphere.

Extensive chemistry modeling efforts within NOAA Research Laboratories (e.g., ESRL, ARL, GFDL) and NESDIS.

Brad Pierce (NESDIS/STAR)


OutlineOutline

NCEP current weather-air quality capabilitiesNational AQ Forecast CapabilityGlobal ozone assimilation

NCEP R&D activities National Environmental Modeling SystemNEMS Interactive atmosphere-chemistry modeling

Proposed enhancementsImpact of dynamic lateral BCs on AQ forecastsImpact of aerosols on weather forecasts

Conclusions


NCEP Current Weather-AQ CapabilitiesNCEP Current Weather-AQ Capabilities

6

Model Components: Linked numerical prediction systemOperationally integrated on NCEP’s supercomputer

NCEP mesoscale NWP: WRF-NMMNOAA/EPA community model for AQ: CMAQ

Observational Input: NWS weather observations; NESDIS fire locationsEPA emissions inventory

National Air Quality Forecast National Air Quality Forecast CapabilityCapability

End-to-End Operational CapabilityEnd-to-End Operational Capability

Gridded forecast guidance productsOn NWS servers: www.weather.gov/aq and ftp-serversOn EPA serversUpdated 2x daily

Verification basis, near-real time: Ground-level AIRNow observations Satellite smoke observations

Customer outreach/feedbackState & Local AQ forecasters coordinated with EPAPublic and Private Sector AQ constituentsWebsite monitoring

AQI: Peak Oct AQI: Peak Oct 44

EPA Monitoring Network

Paula Davidson (NWS OST)


442 grid cells

265gridcells

268 grid cells

259gridcells166

Grid cells

142142gridcells

CONUS “5x” Domain1. OPS: AQFC Sept. 07 2. EXP: AQFC/CB05 June. 083. DEV: AQFC/CB05-AERO-4

Eastern “3x” DomainSept 05

Northeast US “1x” Domain

Sept 04

Expansion of coverageExpansion of coverage

Jeff McQueen (EMC)


NCEP Air Quality Forecast VerificationNCEP Air Quality Forecast Verification

Bia

s (p

pb)

Bia

s (p

pb)

-10

20

Almost the same for NW and Mid WestHigher for NE, SE and Low Miss Valley (increase positive bias)Higher for SW (improve negative bias)

Production Experimental

Jeff McQueen (EMC)

http://www.emc.ncep.noaa.gov/mmb/aq

8 h Avg Ozone Obs vs Fcst8 h Avg Ozone Obs vs Fcst


Global Ozone Assimilation in GSIGlobal Ozone Assimilation in GSIWhy assimilate ozone

Ozone forecastsUV Index Forecasts Air Quality Forecasts

Needed for assimilating radiances from IR instruments (e.g. HIRS, AIRS) where ozone influences the accuracy of determining temperatures.

Parameterized ozone physics in GFSProduction and destruction are parameterized from monthly and zonal mean dataset derived from NRL 2D ozone chemistry model

Current and future ozone products to be assimilated at NCEPGFS currently assimilating only NOAA-17 SBUV/2 (nadir obs)Probable data update to NOAA-18 and possible for NOAA-19OMI and GOME-2 total ozone being tested in parallel

offers greater horizontal and latitudinal coverageNRT MLS ozone profile product is being evaluated.OMPS (NPP and NPOESS)

Craig Long (CPC)


Total Ozone Analysis Improvements byTotal Ozone Analysis Improvements byAssimilating OMI TOz in addition to SBUV/2Assimilating OMI TOz in addition to SBUV/2

More Structure

Tighter Gradients

Craig Long and Shuntai Zhou (CPC)


An Overview of National Environmental An Overview of National Environmental Modeling System (NEMS) Modeling System (NEMS)


Earth System Modeling FrameworkEarth System Modeling Framework Modeling framework for the geo-science

community

A software infrastructure that enables different weather, climate, and data assimilation components to operate together on a variety of platformsEarth system models that can be built, assembled and reconfigured easily, using shared toolkits (e.g., data communications, time management, message logging, re-gridding, and error handling) and standard interfaces A growing pool of Earth system modeling components that, through their broad distribution and ability to interoperate, promotes the rapid transfer of knowledge. Community effort, partially supported by NOAAESMF superstructure (grid component, state, and coupler) required for all NEMS componentsESMF infrastructure optional


National Environmental Modeling System National Environmental Modeling System (NEMS)(NEMS)

Earth Science Modeling Framework (ESMF)http://www.esmf.ucar.edu

NEMS atmosphereWrite history and Post processorNestingAerosols and ChemistryLandOcean, waves and sea iceIonosphereEnsembleData assimilation

Unified Modeling Infrastructure Group, led by Mark Iredell

NCEP UMIG group routinely meets with GSD and GFDL groups

http://www.esmf.ucar.edu/


NEMS AtmosphereNEMS Atmosphere

Atmosphere

Dynamics PhysicsDyn-PhyCoupler

NMM-B

Spectral

FIM

Color Key

Generic Component

Generic Coupler

Completed Instance

Under Development

NAM Phy

GFS Phydo nada

unified atmosphereIncluding digital filter

Future Development

ARW

FVCORE

FISL

Navy

Navy

adjoints

Mark Iredell (EMC)

The goal is one unified atmospheric component that can invoke multiple dynamics and physics.

At this time, dynamics and physics run on the same grid in the same decomposition, so the coupler literally does nothing.

FY2010 operational implementation for NEMS NMM-B

Chemistry

GOCART

AQF chem

reduced chemistry


Developing an interactive atmosphere-Developing an interactive atmosphere-chemistry forecast systemchemistry forecast system

In-line chemistry advantageConsistent: no spatial-temporal interpolation, same physics parameterizationEfficient: lower overall CPU costsEasy data managementAllows for feedback to meteorology

Requirements:Meteorology and chemistry should be initialized with GSIConform to NCO CCS computer architecture Conform to NCO software & I/O standards (GRIB/BUFR)

NEMS AQ development:NMM-B Chem

In support of regional AQF systemGFS coupled with GOCART

Potential for improving weather forecasts (by improving aerosol-radiation feedback in GFS and atmospheric correction in GSI)Providing LBCs for regional AQF aerosol predictions


NEMS Tracer Experiments: NEMS Tracer Experiments:

NMM-B and GFSNMM-B and GFS


NEMS NMM-B tracer experimentNEMS NMM-B tracer experiment

Youhua Tang (EMC)


GB EAS WAF SAM NAM

T62 L64 30-day experiments: CTR, CLD (Ferrier cloud microphysics), DYN (Adiabatic), SAS (Simplified Arakawa-Schubert convection), TVD (Flux-limited vertical advection)

GLB_SFC GLB_UTLS GLB_ALL

Change in global sum

-2.5

-2

-1.5

-1

-0.5

0

0.5

1 2 3CTRL

CLD

DYN

SAS

TVD-1.37% 0.03% (diffusion off)

NEMS GFS tracer experimentNEMS GFS tracer experiment

IC = 2009/01/01 00Z

Change in total mass loading (scaled by initial values)


Zonal mean cross section for SAM_SFC & SAM_UTLS (IC=20090101)Flux-limited vertical advection reduces (but does not eliminate) negative tracer values

NEMS GFS tracer experimentNEMS GFS tracer experiment


Global aerosol forecast and Global aerosol forecast and analysis system (GFS-GOCART)analysis system (GFS-GOCART)


Goddard Chemistry Aerosol Radiation and Transport Model Goddard Chemistry Aerosol Radiation and Transport Model (GOCART)(GOCART)


Global Forecast System (GFS)Global Forecast System (GFS)

Global spectrum model for NCEP operational medium range forecasts

RESOLUTION T382 horizontal resolution (~ 37 km)64 vertical levels (from surface to 0.2 mb)

MODEL PHYSICS AND DYNAMICSVertical coordinate changed from sigma to hybrid sigma-pressureNon-local vertical diffusionSimplified Arakawa-Schubert convection schemeRRTM LW radiation schemeMD Chou SW radiation schemeExplicit cloud microphysicsNoah LSM (4 soil layers: 10, 40, 100, 200 cm depth)

INITIAL CONDITIONS (both atmosphere and land states)NCEP Global Data Assimilation System (GDAS)


Gridpoint Statistical Interpolation (GSI)Gridpoint Statistical Interpolation (GSI)

Global/regional analysis system for operational weather forecasts

NCEP 3DVAR ASSIMILATION SYSTEMImplemented with WRF-NMM into the NAM system in June, 2006Implemented for replacement of SSI in the GFS system in May, 2007

SCIENTIFIC ADVANCESGrid point definition of background errors Inclusion of new types of data (e.g., AIRS radiance, COSMIC GPS)Advanced data assimilation techniques (e.g., improved balance constraints)New analysis variables (e.g., SST)

CODE DEVELOPMENTGMAO collaboration through NASA-NOAA-DOD Joint Center for Satellite Data Assimilation (JCSDA)Evolution to Earth System Modeling Framework (ESMF)


Nick Nalli (NESDIS)

Impact of aerosols on AVHRR Pathfinder Atmospheres (PATMOS) OI multichannel SST (MCSST) retrievals


Aerosol effect on HIRS brightness

temperature retrieval

Aerosol Effect on hirs3_n17

-5

-4

-3

-2

-1

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

BT

dif

fere

nc

e (

K)

.

Quanhua Liu (NESDIS)


Global aerosol forecast and analysis systemGlobal aerosol forecast and analysis system

Global forecast and analysis systemModeling

Emissions

Data Assimilation

Regional AQF

Dynamic LBCs

SST Analysis

Atmos. Correction

Validation

Various datasetsAERONET, OMI, CALIPSO

Satellite data

Algorithm

GOCART

MODIS fire emissions

NASA obs and tech

ROSE project

NCEP DSSs

Color key

Goal: Improving weather and air quality forecasts by incorporating prognostic aerosols in GFS and assimilating global aerosol information in GSI via NCEP-

NASA/GSFC-Howard University collaborations


Multiple, complementary approaches:On-line systems including GOCART:

GFS/GOCART: new capability being developed GEOS-5/GOCART: NASA/GMAO real-time system GFS~GEOS-5/GOCART: Hybrid model (GEOS-5 dynamics + GFS physics)

Off-line GOCART CTM Driven by GFS meteorology

Phased development:Development of prototype systemTransition to real time systemTransition to operational productionPrototype system extended to include ozone chemistry (if resources available)Transition to NCEP’s climate system (if resources available)

NEMS/GFS-GOCART

Dust-only offline GFS-GOCART

Global aerosol forecast and analysis systemGlobal aerosol forecast and analysis system(-cont’d)(-cont’d)


http://www.emc.ncep.noaa.gov/gc_wmb/dkim/web/html/dust_day.html


-14.6E 23.5N, 2006072214

0

2

4

6

8

0 0.1 0.2 0.3 0.4 0.5

Aerosol Extinction (1/km)

Hig

ht

(km

)

CALIPSO

MODEL

Comparisons between Model and CALIPSO

(2006072214)

-12.6E 14.9N, 2006072214

0

2

4

6

8

0 0.05 0.1 0.15 0.2 0.25 0.3


Hig

ht

(km

)

CALIPSO

MODEL

Dongchul Kim (EMC)


Comparisons between Model and CALIPSO

(2006072705)

-53.3E 21.4N, 2006072705

0

2

4

6

8

0 0.1 0.2 0.3 0.4 0.5


Hig

ht

(km

)

CALIPSO

MODEL

-52.5E 25.0N, 2006072705

0

2

4

6

8

0 0.1 0.2 0.3 0.4 0.5


Hig

ht

(km

)

CALIPSO

MODEL

Dongchul Kim (EMC)


Resources !! Code optimization needed

The inclusion of 15 passive tracers leads to ~45% increase in wall time

The 3d atmosphere file sizes increased by the factor of 2.4-2.7

Needed capabilities

Convective transport (under testing for RAS)

Tracer scavenging

Positive definite advection with mass conserving

Challenges for incorporating chemistry Challenges for incorporating chemistry component into NEMS GFS:component into NEMS GFS:

The chemistry modeling efforts will lead to scientific advances and technical upgrades in the NEMS


NOAA medium range weather forecasts

Climatology-based aerosol distributions are used in the GFS and background aerosol conditions are assumed in the GSI Community Radiative Transfer Model (CRTM)

Global aerosol products will improve the representation of aerosol distributions and variations within the GFS/GSI system

NOAA air quality forecasts

Default static boundary conditions are used for the developmental aerosol air quality predictions

Global aerosol products will provide improved aerosol lateral boundary conditions for the AQF system and, consequently, improve AQF aerosol forecasts

Proposed EnhancementsProposed Enhancements


The impact of aerosols on medium The impact of aerosols on medium range weather forecastsrange weather forecasts


U-wind Cross Section at 10W

The intensity and location of African Easterly Jet are affected by background aerosol loading (via direct radiative effect)

Climate Forecast System (CFS):Climate Forecast System (CFS):GFS coupled with GFDL MOM3GFS coupled with GFDL MOM3

OPAC climo. GOCART climo.


RMS errors of NH temp for 00Z forecasts

Pressure

Forecast hours

GDAS experiments with different aerosol representations:T126 L64; PRC (climatology) vs PRG (time varying)

RMSE reduced

RMSE increased


North America temperature verification

Temperature biases reduced by ~ 10% in lower atmosphere

Climo.

Time-varying


The impact of lateral boundary The impact of lateral boundary conditions on air quality forecastsconditions on air quality forecasts


4 0 6 0 8 0 1 0 0 1 2 0 1 4 0O 3 (p p b v)

0

4000

8000

12000

16000

Alt

itud

e ab

ove

Sea

Lev

el (

m)

ObservedFixed LBCRAQMS LBCMOZART LBCGFS-O3 LBCIONS LBC1IONS LBC2

Beltsville 20060803 17.92 UTC

4 0 6 0 8 0 1 0 0 1 2 0 1 4 0O 3 (p p b v)

0

4000

8000

12000

16000

Alt

itud

e ab

ove

Sea

Lev

el (

m)


Boulder 20060803 19.33 UTC

0 100 200 300 400O 3 (p p b v)

0

4000

8000

12000

16000

Alt

itud

e ab

ove

Sea

Lev

el (

m)


Trinidad Head 20060803 21.02 UTC

4 0 6 0 8 0 1 0 0 1 2 0 1 4 0O 3 (p p b v)

0

4000

8000

12000

16000

Alt

itud

e ab

ove

Sea

Lev

el (

m)


Huntsville 20060803 17.53 UTC

0 100 200 300 400 500O 3 (p p b v)

0

4000

8000

12000

16000A

ltit

ude

abov

e Se

a L

evel

(m

)


Bratt's Lake 20060803 21 UTC

Youhua Tang (EMC)

Ozone Lateral Boundary Conditions Tests Ozone Lateral Boundary Conditions Tests

Tang et al., The impact of chemical lateral boundary conditions on CMAQ predictions of tropospheric ozone over the continental United States, Environmental Fluid Mechanics, 2008

Obs (IONS), Obs (IONS), FixedFixed, , RAQMSRAQMS, , MOZARTMOZART, , GFS-O3GFS-O3


Aerosol Lateral Boundary Conditions Tests: Aerosol Lateral Boundary Conditions Tests: Trans-Atlantic dust TransportTrans-Atlantic dust Transport

During Texas Air Quality Study 2006, the model inter-comparison team found all 7 regional air quality models missed some high-PM events, due to trans-Atlantic Saharan dust storms.

These events are re-visited here, using dynamic lateral aerosol boundary conditions provided from dust-only off-line GFS-GOCART.

Corpus Christi - Nat, TX 2006 Observed

CO

NC

(ug/

m3)

60 50 40 30 20 10 0

Corpus Christi - Nat, TX 2006CMAQ base run

CO

NC

(ug/

m3)

60 50 40 30 20 10 0

Corpus Christi - Nat, TX 2006

CMAQ+GFS-GOCART LBC

CO

NC

(ug/

m3)

60 50 40 30 20 10 0

Thomas Jefferson Sch, TX 2006 Observed

CO

NC

(ug/

m3)

80 70 60 50 40 30 20 10 0

Thomas Jefferson Sch, TX 2006 Observed

CO

NC

(ug/

m3)

80 70 60 50 40 30 20 10 0

Thomas Jefferson Sch, TX 2006CMAQ base run

CO

NC

(ug/

m3)

80 70 60 50 40 30 20 10 0

Thomas Jefferson Sch, TX 2006

CMAQ+GFS-GOCART LBC

CO

NC

(ug/

m3)

80 70 60 50 40 30 20 10 0

Karnack C85, TX 2006 Observed

CO

NC

(ug/

m3)

60 50 40 30 20 10 0

29JUL 31JUL 02AUG 04AUG 06AUG 08AUG 10AUG

Karnack C85, TX 2006 Observed

CO

NC

(ug/

m3)

60 50 40 30 20 10 0


Karnack C85, TX 2006CMAQ base run

CO

NC

(ug/

m3)

60 50 40 30 20 10 0


Karnack C85, TX 2006

CMAQ+GFS-GOCART LBCC

ON

C (u

g/m

3)

60 50 40 30 20 10 0


Youhua Tang and Ho-Chun Huang (EMC)


NCEP is developing NEMS as next-generation weather forecast system

NEMS R & D efforts continue in interactive atmosphere-chemistry modeling system

NMM-B + Chem

GFS-GOCART

NCEP modeling efforts leverage common modeling framework (ESMF), shared software development (via NOAA-NASA-DOD JCSDA), and research collaborations, such as

GSI ozone and aerosol data assimilation working group (EMC AQ group)

Co-Ops Biomass Burning Emission Committee (Jeff Reid and Shobha Kondragunta)

AeroCOM (Michael Shulz, Stefan Kinne, and Mian Chin)

GEMS/MACC community

In ConclusionIn Conclusion


THANK YOU

ncep chemistry modeling overview and status (with a focus on nems aq development) sarah lu...

Documents

tx air quality

aq forecastsimpact of

emc aq groupjeff mcqueen

x domainsept

x domainops

tropospheric biases

tes omi assimilation

jongil han