VOC emissions in the Middle East
from bottom-up inventories &
as seen by OMI
J.-F. Müller, T. Stavrakou, I. De Smedt, M. Van Roozendael
Belgian Institute for Space Aeronomy (BIRA-IASB)
GLOBEMISSION USER CONSULTATION MEETING Doha, Qatar, 24-25 November 2015
Outline
• The view from below: a first look at emission inventories over the Middle East
• First confrontation with satellite (OMI) HCHO
• Revising NMVOC emission inventories over the Middle East
• Apply inverse modelling of OMI data using a global CTM to quantify NMVOC emissions over Middle East
• Conclusions
A first look at inventories: natural vs anthropogenic
Impact of soil moisture stress on biogenic isoprene
• In Middle East: huge effect of soil moisture stress on emissions according to MEGAN
• However, very poorly constrained parameterization ! Does not account for the fact that plants are accustomed to lack of soil water
Anthropogenic emission inventories (here at 0.5°x0.5°)
• Large discrepancies between inventories • MACCity similar to RETRO, only higher • Poor representation of cities, oil/gas
related emissions • EDGAR4 : better but with strange features
Contributions of different emission sectors to NMVOC
emissions in the Middle East in EDGAR4.2
60%
26.5%
4.8%
6.5%
In summary : • Residential sector:
unimportant in Gulf area • Solvent industry: only at few
large cities • Road transport: many cities
missing! (Riyadh, Dubai, Basrah, etc.)
• Fossil fuel production: many
point sources, but distribution is questionable
Confronting EDGAR4.2 with OMI (at 0.5°x0.5°)
high oil-related EDGAR emissions in NE S.Arabia
low EDGAR emissions
in Southern Iraq not supported by OMI data!
Also: emission underestimations in Qatar, Emirates, Southern Iran coast, Northern Iraq…
Data provided by I. De Smedt @ BIRA-IASB Here only clear-sky pixels are used
x10-11
Can we improve the distribution of road emissions?
Obvious flaws in available population density map in the Middle East – e.g. from the SEDAC (http://sedac.ciesin.columbia.edu)
For highly urbanized countries*, use urban population distribution constructed from database of cities (Geonames, http://download.geonames.org/)
* : UAE, Qatar, S. Arabia, Bahrein, Iraq, Oman
Disparities in country-level
road emissions from EDGAR
Very low per capita emissions in e.g. Turkey & Egypt likely unrealistic
0 20 40 60
U.A.E.
Qatar
Oman
Kuwait
Iran
S. Arabia
Bahrein
Iraq
Lebanon
Jordan
Israel
Syria
Turkey
Egypt
Per capita NMVOC emissions (kg/year) in
2007
Adopt minimum of 10 kg NMVOC / year for all countries
Road emission adjustment
NMVOC sources due to oil/gas production :
• Evaporative losses (oil spills, etc.)
• Gas flaring and venting (mostly at oil wells)
• Gas leakages during transmission, storage, loading
• Transformation in refineries and petrochemical plants
Can we improve the distribution of emissions
related to oil/gas production?
Complex !
Need proxies for oil/gas extraction &
handling
Can we improve oil/gas emissions?
1. Refineries
Country Name (owner) Capacity
‘000 bpd
Latitude Longitude
EGYPT Alexandria El Mex (EGPC) 117 31.04 29.67
EGYPT Cairo Mostorod (EGPC) 142 27.02 31.00
EGYPT El Nasr (EGPC) 132 29.96 32.53
EGYPT El Sokhna (Bashandy Oil) 300 29.60 32.32
IRAN Arvand (Private) 115 30.44 48.18
IRAN Tabriz (NIOC) 112 38.08 46.29
IRAN Abadan (NIOC) 450 30.34 48.30
IRAN Arak (NIOC) 250 34.09 49.69
IRAN Tehran (NIOC) 225 35.69 51.42
IRAN Isfahan (NIOC) 265 32.65 51.67
IRAN Bandar Abbas (NIOC) 335 27.19 56.28
IRAQ Basrah (INOC) 210 30.53 47.80
IRAQ Daurah (INOC) 180 33.20 44.40
IRAQ Kirkuk (INOC) 170 35.47 44.39
IRAQ Baiji North (INOC) 150 34.94 43.49
ISRAEL Haifa (Oil Ref Ltd.) 196 32.82 34.99
KUWAIT Mina Al-Ahmadi (KNPC) 470 29.07 48.08
KUWAIT Mina Abdullah (KNPC) 270 29.32 47.49
KUWAIT Shuaiba (KNPC) 200 29.03 48.17
OMAN Sohar SRC (ORPIC) 116 24.35 56.73
QATAR Um Said (QP) 147 24.98 51.55
QATAR Lafan 146 25.91 51.55
S. ARABIA Rabigh (Aramco/Smitomo) 400 22.80 39.03
S. ARABIA Riyadh (Aramco) 120 24.46 46.82
S. ARABIA Ras Tanura (Aramco) 550 26.65 50.17
S. ARABIA Yanbu SAMREF (Aramco/Exxon) 400 24.08 38.07
S. ARABIA Jubail SATORP (Aramco/Total) 400 27.00 49.67
S. ARABIA Yanbu YASREF (Aramco/Sinopec) 400 24.08 38.07
S. ARABIA Jubail SASREF (Aramco/Shell) 305 27.00 49.67
UAE Takreer Ruwais 420 24.13 52.71
UAE Jebel Ali 120 25.00 55.05
Database of large refineries (here >100 000 bpd)
2. Terminals & other facilities
Country Name Capacity latitude longitude
IRAN Kharg Island 3 29.23 50.31
IRAN Lavan Isl. 0.2 26.81 53.26
IRAN Sirri Isl. 0.1 25.91 54.54
QATAR Halul Isl. 0.25 25.68 52.42
QATAR Ras Laffan 0.2 25.91 51.55
IRAQ Al Basrah 3 29.68 48.81
S. ARABIA Ras Tanura 3.4 26.65 50.17
S. ARABIA Ras al Ju’aymayah 3.0 26.95 50.05
S. ARABIA Yanbu 1.3 24.09 38.06
S. ARABIA Abqaiq 5.0 25.99 49.67
KUWAIT Mina al Ahmadi 2.0 29.13 48.13
UAE Fujairah 2.0 25.12 56.35 Database of terminals & processing sites
3. Gas flares Use flare CO2 emissions from nighttime VIIRS observations in the infrared
Rumailah oil fields, Iraq
Ghawar oil fields, Saudi Arabia
Can we improve oil/gas emissions?
Oil/gas-related emission adjustment
Methodology : • Keep EDGAR4 total emission over the area (5.4 Tg/year) • Re-distribute using the proxies : 3 Tg for flaring, 2 Tg refineries, 0.5 Tg/yr terminals
Are our proxies appropriate? Confront to OMI
Zooming into the Gulf area
Use oversampling to increase horizontal resolution of OMI data : map 7 years of OMI data to a 0.1°x0.1° grid
Zooming into the Gulf area
The impact of gas flares
Zooming into Egypt
Inverse modelling of emissions
• Use monthly averaged HCHO data averaged over 2009-2014
• Global CTM IMAGESv2 at 2°x2.5°
• Isoprene a priori emissions: neglect soil moisture stress
• Express e.g. anthropogenic emissions as
and optimize emission parameters fj
• Same for biogenic emissions
• Use adjoint model technique
m
j
jjj txfftxG1
),()exp(),,(
A priori emissions
Inverse modelling results: emission updates
• Moderate changes (mostly increases) in isoprene fluxes • Anthropogenic VOC emissions decreased over Gulf area, especially Southern Iraq • Anthropogenic VOC emissions increased over Turkey
Inverse modelling results: HCHO columns
• Nice posterior agreement over Gulf area and Turkey
• Underestimations persist e.g. over Egypt
Inverse modelling results: HCHO columns
Inverse modelling results: HCHO columns
Impact of anthropogenic emissions on ozone
• Ozone production is primarily NOx-limited in
the Middle East
• EDGAR4.2 requires substantial revision in the Middle East
• Isoprene emissions in MEGAN-MOHYCAN inventory is likely strongly
underestimated due to poor parameterization of soil moisture stress
• Inversion of emissions using OMI data indicate
overestimation of oil-related (flares) emissions
underestimation (factor 2) of anthropogenic emissions over Turkey
biogenic isoprene emissions are significant in Northern Iran, Iraq, Syria
• Ozone production in summer is primarily NOx-limited, although VOC
emission reduction would have a measurable effect (up to ~6 ppbv)
Conclusions