Assessing the Lightning NOx Parameterization

in GEOS-Chem with HNO3 Columns from IASI

Matthew Cooper1 Randall Martin1,2, Catherine Wespes3,

Pierre-Francois Coheur3, Cathy Clerbaux3,4, Lee Murray5

1. Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia, Canada. 2. Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA 3. Spectroscopie de l’Atmosphère, Chimie Quantique et Photophysique, Université Libre de Bruxelles, Brussels, Belgium 4. Sorbonne Universités, UPMC Univ. Paris 06; Université Versailles St-Quentin; CNRS/INSU, LATMOS-IPSL, Paris, France 5. NASA Goddard Institute for Space Studies and Lamont-Doherty Earth Observatory of Columbia University, NY, USA

𝑁𝑁𝑁𝑁 𝑁𝑁𝑁𝑁2

𝐻𝐻𝑁𝑁2

𝑁𝑁3

𝐶𝐶𝑁𝑁

𝐻𝐻𝑁𝑁𝑁𝑁3

20-24 Tg

3-13 Tg

4-15 Tg

4-8 Tg

HNO3 Observations Can Be Used To Constrain NOx Emission Inventories

𝑁𝑁𝐻𝐻

GEOS-Chem HNO3 Biased Low Over SE Asia IASI Annual Mean Tropospheric HNO3 Column

mol

ec/c

m2

IASI – GEOS-Chem Percent Difference

mol

ec/c

m2

Perc

ent

GEOS-Chem Annual Mean Tropospheric HNO3 Column

CO Observations Can Identify Transport Of Combustion Sources

Anthropogenic, Biomass burning sources emits NOx, CO Long lifetime of CO allows for evaluation of transport Lightning NOx source is independent CO

GEOS-Chem has 4 lightning regimes: Land – Ocean – Tropics – Midlatitudes SE Asia lightning properties are different, which may lead to biases

Sim

ula

ted

IC

/CG

Fla

sh R

atio

Regional Variations In Lightning Properties Not Captured By GEOS-Chem

Simulated IC/GC Flash Ratio

Fla

sh F

ootp

rint [

km2 ]

600 400 200 0

Lightning Properties Observed from Space [Beirle et al., 2014]

The GEOS-Chem Lightning NOX Parameterization

Prescribed NOx/flash (260 mol/flash)

Flash rates match satellite observations

Prescribed vertical profiles

Emissions added

to GEOS-Chem

4°x5° grid boxes

LNOx Plumes have Sub-Grid Spatial Scale. NOx Chemistry is Non-linear.

Early subgrid plume has high NOx concentration Leads to rapid HNO3 production, lower OPE This is not represented in the model

Actual Plume

PHN

O3 (

mol

ec c

m-3

s-1 )

/ O

PE (m

ol/m

ol)

4°x5° GC grid box

NOx concentration (ppbv)

A Simple LNOx Plume Pre-processor

GC Grid Box NOx

HNO3 + 15O3

Current Mechanism With Pre-Processing

• Adding 0.5 Tg N to Southeast Asia gives better agreement

GC Grid Box

NOx

-1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0

HNO3 Tropospheric Column (IASI – Standard GEOS-Chem)/IASI (IASI – GC w/ PreProcessor)/IASI

92% Bias 6% Bias

Conclusions

GEOS-Chem underestimates HNO3 in Southeast Asia

Lightning NOx is most likely contributor to bias

A preprocessor representing early stages of lightning plume improves the simulation

Bonus Slides

Inferring Tropospheric Contribution to Total Column Apply IASI averaging kernel to GC output = “total column” Set GC troposphere to zero then apply kernel = “strat column” Infer tropospheric column = “total” – “strat”

Sensitivity

GC Total HNO3 Column Total Column Smoothed w/ IASI Kernel

GC Tropospheric HNO3 Column

Alti

tud

e (h

Pa)

IASI Vertical Sensitivity

Tropospheric Column Smoothed w/ IASI Kernel

Convective Effects

Scaled Mass Flux by 25% Scaled scavenging efficiency by 50%

Southeast Asia/ West Pacific is highly convective region HNO3 highly soluable – easily washed out Can have an impact, but its small 20% change vs 200% bias

GEOS-Chem lightning scaled by observed flash density

SE Asia lightning flashes are longer, more radiant than elsewhere in tropics

Intra-cloud lightning more common in SE Asia

[Beirle et al., Nat. Hazards Earth. System. Sci. Discuss, 2014]

Sim

ula

ted

IC

/CG

Fla

sh

Ratio

Some Regionally Varying Lightning Properties that could Affect SE Asia

Lightning Properties Observed by OTD/LIS

Simulated IC/GC Flash Ratio

Vertical Distribution of NOx Emissions

NOx lifetime increases with altitude Not much change in column near emissions region Change (~30%) is small compared to the bias vs IASI

(~200%)

Tropospheric HNO3 Column Ratio: Adjusted/Standard 1.3

1.2 1.1 1 .9 .8 .7

Vertical Distribution

LNOx Plume Pre-processor VS

Prescribed Flash Yield • Adding 0.5 Tg N to Southeast Asia gives better

agreement

Ozonesonde Standard 3.5x Nox

Pre-Process

HNO3 Tropospheric Column (IASI – Standard GEOS-Chem)/IASI

(IASI –GEOS-Chem w/ Pre-Processor)/IASI

Assessing the �Lightning NOx Parameterization�in GEOS-Chem with �HNO3 Columns from IASISlide Number 2GEOS-Chem HNO3 Biased Low Over SE AsiaCO Observations Can Identify Transport �Of Combustion SourcesSlide Number 5The GEOS-Chem Lightning NOX ParameterizationLNOx Plumes have Sub-Grid Spatial Scale.�NOx Chemistry is Non-linear.�A Simple LNOx Plume Pre-processor ConclusionsBonus SlidesInferring Tropospheric Contribution to Total ColumnConvective EffectsSlide Number 13Vertical Distribution of NOx EmissionsLNOx Plume Pre-processor �VS �Prescribed Flash Yield