Industrial emissions: Measurements. NMVOCs, methane, SO 2, NO x, NH 3. Areas: refineries, petrochemistry natural gas production, upstream/downstream

Transcription

1 Solar Occultation Flux measurements of NMVOCs Johan Mellqvist Dep of Earth and Space Sciences Chalmers University i of Technology, Göteborg,Sweden, johan.mellqvist@chalmers.se

2 Optical remote sensing group Industrial emissions: Measurements with SOF and tracer correlation of NMVOCs, methane, SO 2, NO x, NH 3. Areas: refineries, petrochemistry natural gas production, upstream/downstream t Ship emissions: Airborne measurements of SO 2, NO x, CO2, particle number and size. Emission factors (g/fuel) and total emissions (g/s) Volcanic emissions: DOAS-network at 21 volcanoes for SO 2 (NOVAC), Bo Galle 2

3 Comparison between NOVAC and GEIA, SO 2 emissions from volcanoes (Arellano Chalmers) NOVAC GEIA kg/s

4 Typical areas in a refinery from which diffuse leakages of volatile organic compounds (VOC) occur : Crude oil 2: Water treatment 3: Process area 4: Product tanks 5: Product loading 2 6. Flaring Alkanes, 70-80% Aromatic VOC, 10-20% Alkenes 1-5%

5 Standard method for the determination of fugitive emissions i of VOCs from industry Storage tanks: Empirical calculation based on formulas derived by the US EPA and API(AP 42) Process industry: Leak kdetection by sniffer+api correlation Water treatmentt t areas, flaring, ship and truck loading: emission factors by US EPA 5

6 Why measure diffuse VOCs The standard approach with calculations means that the industrial site is treated as an ideal industry. Leaks are hence not accounted for. Measurements with different techniques (Solar Occultation Flux (SOF), Differential Absorption LIDAR (DIAL), Airborne, Tracer correlation) often shows large discrepancies compared to estimates.

7 The Solar Occultation Flux method (SOF) and mobile DOAS Solar tracker 1) The number of various molecules above the vehicle are estimated from spectroscopic analysis of direct solar IR light (SOF) and zenith scattered UV light (mobile DOAS). 2) The measurements are conducted while driving and hence is it possible to measure the total mass of molecules along the roads traveled. 3) The total mass is multiplied by the wind which yields the flux in 7 kg/s.

8 Experience with SOF Regular surveillance of Swedish refineries since 7 years (60 surveys). It provides the basis for national VOC reporting. Total refinery measurements down to the level of individual leaks are conducted. Participation in several international campaigns, such as Mexico city 2004, Texas 2006/2009/2011/2012, France 2008 the Netherlands 2008/2010, Antwerpen 2010 In general the measurements show % larger VOC emissions than reported. The uncertainty t of the SOF measurements is 20-30%. This uncertainty is derived from error estimations, validation exercises and instrument comparisons. (The wind uncertainty is the largest error source) 8

9 9

10 10

11 JM23 Fenceline measurements are typically conducted at 0.5 to 3 km distance from the sources s. The assumption is then that the plume is distributed from the ground up to several hundred meters height and that the wind varies little with height. (This has been validated by airborne measurements during TexAQS [Mellqvist, JGR 2009]).

12 Slide 11 JM23 main assumptaion during the campaign Johan Mellqvist,

13 SOF measurements of olefines and alkanes from petrochemical conglomerate, Mt Belvieu Texas 2009 C2H4 C2H4:453 kg/h C3H6: 229 kg/h C3H8: 1615 kg/h 12 Fi 1 SOF lk h l t fm tb l i f I 10 S i Rd 5 J

14 Ethene emissions from Mt Belvieu, Houston, measured during Texas 2006 by SOF and airborne measurements by NOAA (Joost de Gouw, EST 2009)

15 Ethene Houston ship channel 2006 Formaldehyde d Texas city 2009 Alkanes Rotterdam

16 SOF versus inventory for recent studies Source region Species SOF (kg/h) 2006/2009/2011 Inventory (kg/h) Houston ship channel 2006/2009/2011 (Texas, USA) Total Rotterdam 2008/2010 (The Netherlands) Ethene Propene 804/ 580/ / 1653/ Alkanes 11528/ 10134/ SO2 NA/ 3364/ NO2 NA/ NA/ Alkanes Total Le Havre 2008 Alkanes (France) Antwerpen Alkanes (Belgium) SO2 NO

17 Some responses by modellers and authorities TCEQ (Texas Commission on Environmental Quality) increases industrial VOCs by ~6X in ozone modelling (SIP model) Scientists at NOAA (Kim ACP, 2011) have upscaled VOC emissions based on SOF 2006 data in Texas. This improved the agreement with airborne measurements. EPA has admitted that the API model is underpredicting VOC emissions. They claim to be working to improve the emission factors 16

18 Ozone modelling for the TexAQS 2006 campaign by using upscaled VOC and NOx emissions based on SOF and comparison to airborne measurements (Kim et al, ACP 2011) Air SOF and S OF upda rborne o d airborn Orig ated inv observa ne upda ventory ginal inv tions ted inv ventory 17