A comparison of very short lived halocarbon (VSLS) and DMS aircraft measurements in the tropical west Pacific from CAST, ATTREX and CONTRAST

Abstract. We present a comparison of aircraft measurements of halogenated very short lived substances (VSLSs) and dimethyl sulphide (DMS, C 2 H 6 S) from a co-ordinated campaign in January–February 2014 in the tropical west Pacific. Measurements were made on the NASA Global Hawk, NCAR Gulfstream-V High-performance Instrumented Airborne Platform for Environmental Research (GV HIAPER) and UK Facility for Airborne Atmospheric Measurements (FAAM) BAe-146 (see Sect. 2.2) using four separate gas chromatography–mass spectrometry (GC-MS) instruments: one operated by the University of Miami (UoM), one from the National Center for Atmospheric Research (NCAR) and two from the University of York (UoY). DMS was measured on the BAe-146 and GV. The instruments were inter-calibrated for halocarbonsduring the campaign period using two gas standards on separate scales: a National Oceanic and Atmospheric Administration (NOAA) SX-3581 standard representative of clean low-hydrocarbon air, and an Essex canister prepared by UoM, representative of coastal air, which was higher in VSLS and hydrocarbon content. UoY and NCAR use the NOAA scale/standard for VSLS calibration, and UoM uses a scale based on dilutions of primary standardscalibrated by GC with FID (flame ionisation detector) and AED (atomic emission detector). Analysis of the NOAA SX-3581 standard resulted in good agreement for CH 2 Cl 2 , CHCl 3 , CHBr 3 , CH 2 Br 2 , CH 2 BrCl, CHBrCl 2 , CHBr 2 Cl, CH 3 I, CH 2 ICl and CH 2 I 2 (average relative standard deviation (RSD) 10 %). Agreement was in general slightly poorer for the UoM Essex canister with an RSD of 13 %. Analyses of CHBrCl 2 and CHBr 3 in this standard however showed significant variability, most likely due to co-eluting contaminant peaks, and a high concentration of CHBr 3 , respectively. These issues highlight the importance of calibration at atmospherically relevant concentrations ( ∼  0.5–5 ppt for VSLSs; see Fig. 5 for individual ranges). The UoY in situ GC-MS measurements on board the BAe-146 compare favourably with ambient data from NCAR and UoM; however the UoY whole-air samples showed a negative biasfor some lower-volatility compounds. This systematic bias could be attributed to sample line losses. Considering their large spatial variability, DMS and CH 3 I displayed good cross-platformagreement without any sampling bias, likely due to their higher volatility. After a correction was performed based upon the UoY in situ vs. whole-air data, all four instrument datasets show good agreement across a range of VSLSs, with combined mean absolute percentage errors(MAPEs) of the four platforms throughout the vertical profiles ranging between 2.2 (CH 2 Br 2 ) and 15 (CH 3 I) % across a large geographic area of the tropical west Pacific. This study shows that the international VSLS calibration scales and instrumental techniques discussed here are in generally good agreement (within ∼  10 % across a range of VSLSs), but that losses in aircraft sampling lines can add a major source of uncertainty. Overall, the measurement uncertainty of bromocarbons during these campaigns is much less than the uncertainty in the quantity of VSLS bromine estimated to reach the stratosphere of between 2 and 8 pptv.