ONGOING GEOS-CHEM ACTIVITIES IN JACOB GROUP

• Tropospheric ozone-NO_x-VOC chemistry (Mat Evans, Arlene Fiore, Qinbin Li, Rynda Hudman)
• Aerosol chemistry (Rokjin Park, Becky Alexander, Duncan Fairlie, Yang Liu)
• Oxygenated organics (Brendan Field)
• Biogenic VOC emissions (Dorian Abbot, May Fu, with Randall Martin and Kelly Chance)
• Methane (Yaping Xiao, with James Wang)
• CO₂ (Parvadha Suntharaligam, Qinbin Li)
• Methyl halides (Paul Palmer)
• Mercury (Noelle Eckley, Rokjin Park)
• Inverse modeling of CO and CO₂ (Colette Heald, Paul Palmer, Dylan Jones, Parvadha Suntharalingam, with Yuxuan Wang)
• CO/CO₂ satellite OSSEs and chemical data assimilation (Dylan Jones)
• Interface with GISS GCM (Loretta Mickley, Shiliang Wu)
• fvDAS simulation capability (Brendan Field, Bob Yantosca)
• MPI parallelization (Jack Yatteau, Bob Yantosca, with NCCS)

​

​

TROPOSPHERIC OZONE-NO_x-VOC SIMULATION

• ~80 species, 300 rxns–detailed oxidation pathways for ethane, propane, C_4-5 alkanes, propene, isoprene
• SMVGear chemical solver
• Fast-J radiative transfer code including 1-D cloud, aerosol effects
• Stratosphere: simple chemical processing, Synoz ozone (x-tropopause flux of 475 Tg yr^-1).
• Anthropogenic emissions from GEIA (NO_x), Logan (CO), various sources (VOCs); biofuels/biomass burning from Logan; biogenic from GEIA (modified); lightning from Price/Pickering.
• Yearly scaling of anthropogenic emissions using inventory/economic data, of biomass burning emissions using satellite firecounts.
• Dry deposition from big-leaf resistance-in-series scheme (Wesely with extensions/modifications)
• Wet deposition from convective updrafts, rainout/washout

​

Liu et al., JGR 2001: Constraints from ²¹⁰Pb and ⁷Be on wet deposition and transport in a global three-dimensional model driven by assimilated meteorological fields [v2.2, 1991-1996]

Development/evaluation of wet deposition algorithm for GEOS-CHEM

• No global bias in simulation of surface ²¹⁰Pb and ⁷Be data
• Aircraft data are ambiguous viz. cirrus sink (not included in std code)
• Algorithm extended to gases on basis of Henry’s law partitioning, retention efficiency upon freezing (Jacob document for GMI, to be published somewhere…)

Bey et al. JGR 2001a:Global modeling of tropospheric chemistry with assimilated meteorology: model description and evaluation [v3.2, 1994]�General description of tropospheric ozone-NO_x-VOC simulation

• Ozone (O_x) production on high side of literature range; fast-J treatment of clouds appears to be a factor. Rate went down in v4.27 due to revised chemistry (esp. O¹D + N₂) and has crawled down since

• Ozone STE ~3x too high: problem still there in GEOS-3, circumvented with Synoz

• Simulation of ozonesonde data within 5-10 ppbv, no systematic bias

Global mean tropospheric OH concentration (methylchloroform lifetime)

• observed MeCl lifetime: 5.7+/-0.7 yrs
• model lifetime keeps going up! 5.1 yrs (Bey2001, v3.2), 5.6 yrs (Martin2003, v4.26), 6.4 yrs (Fiore, v4.33), 6.8 yrs (Park, v5.3)

addn’l VOC sources of CO

O(¹D)+N₂

GEOS-3

(change in

benchmark)

Benchmark 1-month run

Additional VOC emissions, decline in OH have fixed 10-30 ppbv CO underestimate of Bey et al. [2001]

aerosols

(hv)

Li et al., GRL 2001: A tropospheric ozone maximum over the Middle East (v4.6, 1993-1997)�

• Middle East summer max in the model is due to outflow from S and E Asia in UT, and from Europe in LT
• Observational evidence for this maximum in MOZAIC and SAGE-II observations, but not in TOMS residuals. More data are needed.

Circles are sonde and MOZAIC obs

Liu et al., JGR 2002: Sources of tropospheric ozone along the Asian Pacific rim: an analysis of ozonesonde observations [v4.6, 1995-1997]

Observed

GEOS-CHEM

Hong Kong sonde and model profiles,

12/24.96 and 1/8/97

Model ozone concentrations and fluxes, 200 hPa

Stratospheric ozone tracer at longitude of Hong Kong

• Good unbiased simulation of climatologies at HK and Japanese stations except for summer monsoon

​

• Success in simulating day-to-day variability over Hong Kong

300-120 hPa

700-300 hPa

850-700 hPa

Time series of ozone

at Hong Kong, 1996

Martin et al., JGR 2002: Interpretation of TOMS observations of tropical tropospheric ozone with a global model and in situ observations [v4.11, 1996-1997]�

• Include optical and chemical effects of dust 🢫OH decreases by 9%

​

• Overestimate of ozone (5-10 ppbv) and NO_x (x2) over tropical Pacific

Martin et al., JGR 2003: Global and Regional Decreases in Tropospheric Oxidants from Photochemical Effects of Aerosols [ v4.26, 1996-1997]

Δ OH (%)

Δ O₃ (ppbv)

Difference in OH and ozone mixed layer concentrations in simulations with vs. without aerosol effects on photolysis rates and on reactive uptake of HO₂, NO₂, NO₃

August

Inclusion of aerosol effects on photolysis frequencies and heterogeneous chemistry using off-line monthly mean aerosol fields from GOCART

Martin et al., JGR 2003 (cont.)��LARGE MODEL OVERESTIMATE OF O₃ OVER S. ASIA:�aerosol effects are not enough to fix it

GEOS-CHEM with full aerosol photochemistry

GEOS-CHEM w/o radiative effects or uptake of HO₂, NO₂, or NO₃ by aerosols

MOZAIC aircraft observations (1995-99)

…this is a very puzzling problem!

Fiore et al., JGR 2002: Background ozone over the United States in summer: origin, trend, and contribution to pollution episodes [v3.3, 1995]

• mean bias = +3 ppbv; good simulation of pdfs up to 70 ppbv, trends, precursors, correlations.

​

• Background over Gulf of Mexico too high; excessive convection

​

• Positive bias in urban coastal areas: BL horizontal resolution problem

​

• uses SAMI inventory for eastern U.S. (never put in standard code; little difference with GEIA)

GEOS-CHEM

Fiore et al., JGR 2003a: Application of empirical orthogonal functions to evaluate ozone simulations with regional and global models [v3.3, 1995]

EOF 1:

East-west

EOF 2:

Midwest-

Northeast

EOF 3:

Southeast

OBS (AIRS)

GEOS-CHEM 2°x2.5°

r² = 0.74

Slope = 1.2

r² = 0.27

Slope = 1.0

r² = 0.90

Slope = 1.0

r² = 0.68

Slope = 1.0

r² = 0.54

Slope = 0.8

r² = 0.78

Slope = 1.0

• good GEOS-CHEM simulation when projected on observed EOFs 🢫successful simulation of synoptic processes driving regional ozone episodes

Fiore et al., ready to go to JGR: Variability in surface ozone background over the United States: implications for air quality policy [v4.33, 2000-2001]

Monthly mean pm conc.

Time series

Good simulation of temporal variability; main problem is background overestimate in southeastern U.S. in summer (GEOS-3 did not fix problem)

Li et al., JGR 2002a: Transatlantic transport of pollution and its effects on surface ozone in Europe and North America (v4.16, 1993-1997)

Mar-Aug 1997 time series

1993-1997 stats

• Excellent simulation of ozone and CO at Sable I., Mace Head, Iceland (means, pdfs, time series, correlations)

Li et al., JGR 2002b: Stratospheric versus pollution influences on ozone at Bermuda: Reconciling past analyses (v4.16, 1996)

r = 0.82, bias –1.8 ppbv

model ozone

source attribution

3-d back-trajectory facility in GEOS-CHEM (T.D. Fairlie)

• Tagging of ozone by region of origin identifies U.S. pollution as dominant contributor to high-ozone events in Bermuda

Li et al., JGR2002b (cont.)

• Successful simulation of April ozonesonde data over N America strengthens case against stratospheric influence at Bermuda

Bey et al., JGR2001b: Asian chemical outflow to the Pacific in spring: origins, pathways, and budgets [V3.02, 1994]

• Successful simulation of Asian outflow pathways (WCBs, mixing of fuel and biomass burning effluents) – verified in TRACE-P
• Vertical profiles of NO and PAN, here and elsewhere, are usually within factor of 2, while HNO₃ is biased high in remote FT by factor of 2-3 – HNO₃ simulation is improved in GEOS-3 due to more frequent precip

Simulation of PEM-West B data

Triangles: obs

Circles: model

Li et al., ready to go to JGR: Export of NO_y from the North American boundary layer: a global model analysis of aircraft observations [v4.26, 1997]

Simulation of NARE 1997 aircraft observations of N American outflow off Nova Scotia

• Good unbiased agreement for CO, O₃, NO;
• Overestimates in the free troposphere of NO_y (35%) and PAN (50%) reflect a northern midlatitudes problem that is most severe in GEOS-STRAT

Aircraft HNO₃ evaluation: comparison of different model versions (A. Fiore)

PAN

(GEOS-3; v. 4.33)

​

(GEOS-1; v. 3.2)

Aircraft PAN evaluation: comparison of different model versions (A. Fiore)

Mat Evans: γ_N2O5 = 0.1 in standard model is too high: implications for NO_x, O₃

Improved representation:

• = f(T, RH) for sulfate, sea salt
• = 0.01 for dust
• = 0.005 for carbonaceous

Snapshot for January 1: global mean γ_N2O5= 0.025

Ozone increase for April 2001 with γ_N2O5 =0.01 vs. 0.1

…effect with best estimate of γ still TBD

Mat Evans: Simulation of TRACE-P Asian outflow of NO_y (γ_N2O5=0.01)

NO_y

PAN

HNO₃ + NO₃^-

NO

PAN/NO_y

NO_y

PAN

HNO₃ + NO₃^-

NO

(HNO₃+NO₃^-))

/NO_y

PAN/NO_y

(HNO₃+

NO₃^-))/NO_y

• Tropical overestimate probably due to biomass burning source
• γ =0.01 (vs. 0.1) helps simulation of NO in free troposphere

Palmer et al., 2001:Air mass factor formulation for spectroscopic measurements from satellites: application to formaldehyde retrievals from GOME ��Palmer et al., 2003b: Mapping isoprene emissions over North America using formaldehyde column observations from space, JGR [v4.4, 1996]�

Comparisons to surface HCHO data using different isoprene emission inventories

• GEIA isoprene emission inventory as used in GEOS-CHEM results in 20% high bias in HCHO simulation
• “GOME isoprene inventory” derived from top-down constraints and isoprene-HCHO relationship from GEOS-CHEM gives better simulation (but has not been implemented in the standard GEOS-CHEM)

GEIA

BEIS2

GOME

Abbot et al., GRL 2003: Seasonal and interannual variability of isoprene emissions � as determined by formaldehyde column measurements from space [v4.26, 1997]

• Regional discrepancies to be investigated after updating GEOS-CHEM isoprene emissions to GBEIS (in progress)
• Should we update our land surface data base? To MODIS?

​

INTEGRATION OF TRACE-P, MOPITT, AND GEOS-CHEM �TO QUANTIFY CARBON MONOXIDE SOURCES FROM ASIA

TRACE-P CO DATA

(G.W. Sachse)

CONCLUSIONS:

• A priori Chinese emissions too low by 50%

(domestic burning)

• A priori SE Asian biomass burning emissions

too high by 60%

• Japan, Korean emissions correct within 20%

A PRIORI

EMISSIONS

(customized for TRACE-P)

Fossil and biofuel

[D.R. Streets, ANL]

Daily biomass burning

(satellite fire counts)

GEOS-CHEM

CTM

MOPITT CO

March-April 2001

INVERSE

ANALYSIS

validation

chemical

forecasts

top-down

constraints

Liu et al., JGR2003: Transport pathways for Asian combustion outflow over the Pacific: Interannual and seasonal variations [v4.13; 1994, 1996, 1998, 2000-2001]

• Successful simulation of WCB motions; altitude of outflow is often a few km off but unbiased;

​

• excellent simulation of post-frontal boundary layer advection;

​

• difficulty with timing of convection.

Simulation of TRACE-P outflow pathways using CO as tracer

Heald et al., , JGR 2003a: Biomass burning emission inventory with daily resolution: application to aircraft observations of Asian outflow [v4.20, 2001]

(AVHRR)

Climatology

2001 monthly

2001 daily

Biomass

burning

emissions

• Using 2001 vs. climatological emissions improves simulation
• No further improvement by using daily vs. monthly emissions

Palmer et al., JGR 2003a: Inverting for emissions of carbon monoxide from Asia using aircraft observations over the western Pacific [v4.33, 2001] �

Simulation of CO aircraft observations in TRACE-P shows that

• Model transport error is 20-30% at all altitudes
• Streets inventory of Chinese anthrop emissions is 50% too low (Logan inventory used in standard GEOS-CHEM is OK)
• GEOS-CHEM biomass burning in SE Asia is 3x too high

​

Relative error in model simulation of TRACE-P CO observations

Heald et al., JGR 2003b: Transpacific satellite and aircraft observations of Asian pollution [v4.33, 2001]�

MOPITT

GEOS-CHEM

Difference

• GEOS-CHEM sampled with MOPITT averaging kernels and along MOPITT orbit track

• R² = 0.87, bias -4.6 ppbv

​

• Regional underestimate in SE Asia (need to reduce biomass burning by 50-60%); consistent with Palmer et al. TRACE-P inversion

​

• (not shown here) Succesful model simulation of events of Asian outflow, transpacific pollution

Mar-Apr

2001 mean

Colette Heald: Inverse modeling of Asian CO sources �using MOPITT data [v4.33, 2001]

Objective: Develop top-down constraints on Asian sources of CO based on synthesis of MOPITT and TRACE-P aircraft observations

Preliminary Results

(MOPITT only)

a priori

a posteriori

A priori: Streets (FF, BF),

Logan (BB)

A posteriori

FFCHKJ FFSEA FFIN BBCH BBSE BBIN ROW/10

Dylan Jones: constructing the covariance matrix for the model transport error

CO Mixing Ratio (ppb)

Use GEOS-CHEM chemical forecasts of CO for TRACE-P, assume that differences between successive (48-hr and 24-hr) forecasts are representative of the covariant error structure (NMC method)

Square root of variance�based on 49 pairs of �forecasts for �Feb-April, 2001

8 km

1.5 km

Qinbin Li, Rynda Hudman, Yuxuan Wang: �hindcast simulations for summer 2004 field studies �(v. 5.04) Nested 1°x1° Grid Over North America

• Conduct multi-year (1997, 2000-2002) simulations (CO, O₃, aerosols) to examine interannual variability in export pathways.

​

1^ox1^o CO at 0.5 km altitude, July 1 2001

Parvadha Suntharalingam: CO₂ Simulation Capability in GEOS-CHEM

FOSSIL FUEL

BIOSPHERIC

EXCHANGE

BIOMASS BURNING

BIOFUELS

OCEAN EXCHANGE

NWR

MID

MLO

BME

Model evaluated against measurements from NOAA-CMDL sites

Includes diurnal cycle

Parvadha Suntharalingam: constraints on Asian CO₂ fluxes from CO/CO₂ correlations in Asian outflow

• Modeled CO₂/CO ratios higher than observations
• Modeled boundary layer CO₂ is higher than observations
• Reconciliation of modeled CO₂ with observed CO₂ and CO₂/CO ratios requires a reduction in a source with a high CO₂/CO emissions ratio
• Better agreement between model and observations achieved with a 40% reduction in Chinese biospheric emissions

MODEL

OBS

REGION

Offshore China

Yaping Xiao: Global ethane simulation (v4.33, 1994)

Global evaluation:

Columns, aircraft profiles, surface sites

New info from TRACE-P:

• Streets’ Asian emission
• Russian ind. emission * 2.5
• Biomass burning * 0.3

Yaping Xiao: Improve understanding of CH₄ sources with TRACE-P aircraft observations (v4.33, 2001)

• Asian anth. total: Streets’ 95Tg/yr (as compared to Wang et al. 135 Tg/yr)
• Constraint from CH₄-C₂H₆-CO correlations: scale down Eurasian anth. by 30%
• With the optimized emissions, no distinct bias in comparing with TRACE-P or CMDL

Superimpose Streets’ Asian emissions

European anth. * 0.7

Preliminary inventory from James Wang

COUPLED AEROSOL-CHEMISTRY SIMULATION CAPABILITY IN GEOS-CHEM

• H₂SO₄-HNO₃-NH₃- H₂O aerosol thermodynamics
• GEIA sulfur emissions (scaled)
• GEIA ammonia emissions w/ T-dep seasonal variation
• ISORROPIA (slow) or RPMARES (fast) thermo module
• Sulfur oxidant chemistry: OH, H₂O₂, O₃, NO₃
• OC and EC (hydrophillic and hydrophobic)
• Soil dust (four size classes)
• Sea salt (two size classes)

​

• Coupling of aerosol with ozone chemistry through
• Aerosol effects on photolysis rates
• Sulfur oxidants
• HNO₃(g)/NO₃^- partitioning
• Heterogeneous chemistry

​

• Aerosol simulation can be either coupled (“on-line”) or uncoupled (“off-line”)

​

​

Park et al., JGR 2003: Sources of carbonaceous aerosols over the United States and implications for natural visibility (v4.23, 1995, 1998)

OC

EC

• Fuel combustion sources from Cooke et al., with biofuels and seasonal variation (N America) added; GEOS-CHEM biomass burning; secondary OC from terpenes
• Top-down constraints applied to improve U.S. emission estimates; biofuel source increased by 65%, other changes smaller

Model vs. observed (IMPROVE) 1998 annual concentrations

Rokjin Park: background SO₄^2—NO₃^—NH₄⁺ aerosols in U.S. and implications for AQ standards (v5.03, 2001)

• SULFATE 2001 comparison for non-urban U.S. sites: high correlation, 25% low bias in summer (excessive scavenging?), other seasons better

Rokjin Park: comparisons to U.S. NH₄⁺ and NO₃^- observations

• NH₄⁺: 2x high bias in fall, 5-25% bias in other seasons

• NO₃^-: 3x high bias in summer-fall, better but still high in other seasons.

…appears to be driven by NH₄⁺ overestimate

GEOS-CHEM vs. Gilliland seasonal variation of NH₃ emissions:

Rokjin Park: evaluation with EMEP 2001 aerosol observations in Europe

• Good simulation for sulfate, no apparent bias

​

• 40-60% overestimate of ammonium in summer-fall (25% annual)

​

• confusing picture for nitrate; high bias in summer-fall

Yang Liu and Rokjin Park: simulation of surface PM2.5 and MISR satellite AOT over U.S. [v5.3, 2001]

Note: this comparison does not include model dust or sea salt yet!

Yang Liu and Rokjin Park, cont.

1. General model overestimate in midwest (ammonia)
2. General model underestimate in west (don’t include dust – also urban bias in obs?), but overestimate in NW in summer (probably OC)

Rokjin Park and Yang Liu, very preliminary: �GEOS-CHEM AOTs

Becky Alexander, Rokjin Park: oxygen isotope tracers of sulfate chemistry [v5.03, 2001]

Δ¹⁷O sulfate (tracer of oxidation by O₃) in standard GEOS-CHEM simulation is only significant during winter at high northern latitudes (when H₂O₂ is titrated)

Δ¹⁷O is too low compared with measurements from various locations in California, Antarctica, and the 1997 pre-INDOEX cruise

Δ¹⁷O sulfate simulation

January 2001

July 2001

…fix by adding aqueous oxidation of S(IV) in sea-salt aerosols

Estimate that 44 -74% of marine SO₂ originating from DMS is oxidized to sulfate by O₃ on sea-salt aerosols

Becky Alexander (cont.)

GEOS-CHEM 2001 sea-salt emissions:

5700 Tg/year

95% supermicron

35% northern hemisphere

May 2001

Li et al., GRL2000: Atmospheric hydrogen cyanide (HCN): biomass burning source, ocean sink? (v3.2, 1993-1994)��Li et al., JGR2003: Model Evaluation of the Atmospheric Budgets of HCN and CH₃CN: Constraints From Aircraft and Ground Measurements [v4.33, 2001] �

• Simulation of HCN and CH₃CN includes two-film model for ocean uptake (applied since to acetone, methanol, DMS)

TRACE-P data

Bell et al., JGR 2002: Methyl iodide: atmospheric budget and use as a tracer of convection in global models [v4.3, 1993-1994]

Simple model

for ocean source

Observations

​

Model

(GEOS-CHEM)

Define Marine Convection Index (MCI) as ratio of upper tropospheric (8-12 km)

to lower tropospheric (0-2.5 km) CH₃I concentrations

• MCI over Pacific ranges from 0.11 (Easter Island dry season) to 0.40 (observations over tropical Pacific
• GEOS-CHEM reproduces observed MCI with little global bias (+11%) but poor correlation (r² = 0.15, n=11)

MCI: 0.40 (obs)

0.22 (mod)

MCI: 0.16 (obs)

0.14 (mod)

Paul Palmer: Inverse modeling of CH₃Br and CH₃Cl sources using constraints from aircraft data

CH₃Br a priori budget terms

CH₃Br has been declining by 5% yr^-1 in 1990s – need slab ocean model to represent?

Jacob et al., JGR 2002: Atmospheric budget of acetone [v3.03, 1994]

observations

• Simulation of propane, i-butane (acetone precursors) using Piccot 1992 inventory requires doubling of emissions outside Europe and N. America
• Fit to acetone observations achieved by invoking a net ocean source, but that’s probably not right (new info from TRACE-P)

Jacob, Field: Global budget of methanol

Objective: Evaluate understanding of current global methanol budgets through simulation of aircraft observations

obs

model

TRACE-P aircraft data (Mar-Apr 2001)

• Compairsons with data from SONEX, PEM-West B, PEM-Tropics B, TRACE-P, ITCT-2K2 are (so far) difficult to interpret in terms of biases in sources

Noelle Eckley, Rokjin Park: Global Transport of Mercury Compounds

• Ongoing work: incorporate real chemistry (coupled with oxidant model), improved sources
• Collaboration with U. Washington (ocean slab model)

Preliminary GEOS-CHEM results with fixed (mean) Hg species lifetimes:

model values are in ball park of observations

​

Hg(0)

Hg(II)

Loretta Mickley: Use GEOS-CHEM for simulation of past and future atmospheres through linkage with GISS GCM

GISS GCM II’

Δclimate

GEOS-CHEM

• GGs, aerosols, land surface,

solar flux

First application: Investigation of the effect of future climate change on US air quality (Mickley, Wu); collaboration with EPA/ORD

Air mass fluxes, other met. variables

Δ Emissions

Δchemistry,

aerosols

Feedback on climate forcing

Dylan Jones: chemical data assimilation in collaboration with GMAO

• Current ozone assimilation at GMAO (Steven Pawson, Ivanka Stajner) uses off-line CTM and observations from TOMS and SBUV/2
• Linear GEOS-CHEM tropospheric ozone chemistry (production rates and loss frequencies) is included but the CTM does not include convection
• Implementation of this linear chemistry in on-line fvGCM (to allow in particular for convection) is ongoing
• Next step is implementation of on-line linear CO
• Long-term goal is the implementation of full tropospheric chemistry in the assimilation system

• Applications:
• chemical data assimilation for MOPITT, SCIAMACHY, TES,…
• chemical forecasting