Atmospheric mixing ratios of C2H5I, CH2ClI, CH2Br2, and CHBr3 were quantified from whole air samples collected aboard the FAAM BAe 146-301 atmospheric research aircraft during flights over the tropical Atlantic Ocean (30°N to 10°S) as part of the Trade Wind Ozone Photochemistry Experiment (TROMPEX) in September 2009. Mixing ratios of C2H5I and CH2ClI increased with decreasing latitude, and concentrations of all measured halocarbons were elevated with increasing proximity to the oceanic upwelling along the coast of Mauritania. Vertical profiles highlighted the rapid photolytic degradation of these compounds within the tropical troposphere, with CH2ClI concentrations reduced by 50% at the top of the boundary layer. The spatial distributions of CH2ClI indicated an open ocean source.
 Very short-lived halocarbons (VSLH, lifetimes <6 months) are significant sources of reactive halogens to the troposphere and lower stratosphere and make an important contribution to photochemical ozone depletion [von Glasow et al., 2004; Yang et al., 2005; Montzka et al., 2011; Saiz-Lopez et al., 2012]. In the lower stratosphere, 5–40% of the bromine present is thought to derive from VSLH entering via rapid vertical transport from the marine boundary layer (MBL) [Kritz et al., 1993; Montzka et al., 2011]. Due to greater convection, emissions at tropical latitudes are more likely to reach the stratosphere; hence, identification of major bromocarbon sources in this region is of significant global atmospheric importance. In the free troposphere, biogenic organohalogen decomposition is predicted to reduce ozone levels by 5–30% via catalytic ozone destruction by bromine oxide radicals (BrO) and reduced ozone production due to BrO reactions with nitrogen oxides [von Glasow et al., 2004; Yang et al., 2005]. In the tropical MBL, reactive iodine has been shown to be an important control in the photochemical destruction of ozone [Read et al., 2008; Mahajan et al., 2009; Jones et al., 2010]. CH3I, C2H5I, CH2I2, and CH2ClI are the primary organic sources of iodine atoms, although these precursors do not explain the total reactive iodine abundance in the tropical Atlantic MBL [Mahajan et al., 2009; Jones et al., 2010]. Observations from the Cape Verde Atmospheric Observatory (16°51′N, 24°52′W) in the tropical east Atlantic Ocean indicate that reactive bromine and iodine chemistry increases surface photochemical ozone destruction by ~50% and is widespread in this region [Read et al., 2008].
 Organic halogens are released from marine macro algae [Goodwin et al., 1997; Carpenter et al., 2000], from phytoplankton [Tokarczyk and Moore, 1994], from marine bacteria and detritus [Hughes et al., 2008], and via photochemical breakdown of dissolved organic material in the ocean [Happell and Wallace, 1996; Richter and Wallace, 2004]. However, there are major uncertainties regarding the contribution from these different sources. The global distributions and controls of VSLH emissions are also not well known, exacerbated by large spatial variability in sea-air fluxes. It has been suggested that biologically active oceanic upwelling regions are a major localized source of VSLH emissions [Quack and Wallace, 2003], contributing a substantial fraction of CH2Br2 and CHBr3 to the upper troposphere/lower stratosphere [Quack et al., 2004]. More recent analysis has shown that oceanic emissions of these species from the Mauritanian upwelling region cannot explain the high atmospheric concentrations observed in this region [Quack et al., 2007; Carpenter et al., 2009].
 Models assume a variety of scenarios [Bell et al., 2002; von Glasow et al., 2004; Yang et al., 2005; Warwick et al., 2006] for oceanic VSLH emissions, characterized according to geographical region. These require better validation by global oceanic and atmospheric measurements, as well as a better understanding of the underlying processes in order to predict the impact of environmental change on emissions. This study aims to quantify horizontal and vertical chemical gradients in VSLH concentrations over the tropical East Atlantic in order to better understand the extent of active halogen chemistry in an area where the impacts have been identified to be substantial [Read et al., 2008].
 Sampling was carried out onboard the FAAM BAe 146-301 research aircraft in the boundary layer and lower free troposphere between the Canary Islands, the island of São Vicente in the Cape Verde archipelago, and the Mauritanian upwelling from 7 to 9 September 2009 (flights B475-B479). The position of the upwelling was identified during the flight as corresponding to a 1–2°C drop in sea surface temperature (SST) measured from a downward facing (Heimann KT 19.82) radiometer mounted to the exterior of the aircraft, later confirmed from MODIS chlorophyll-a averaged data for the week of the flight (7–14 September 2009) (Figure 1). The algal bloom size appears slightly smaller relative to previous campaigns in this area [Carpenter et al., 2009; Quack et al., 2007].
 Air samples were collected at an average frequency of 1 min−1, pressurized into evacuated 3 L Silco Steel® (Restek Corporation) cylinders at 2–3.25 bar using a metal bellows pump (Senior Aerospace). Samples were analyzed at the University of York within 2 weeks of being collected. Analysis was carried out using a Perkin Elmer AutosystemXL gas chromatograph coupled to a Perkin Elmer Turbomass mass selective detector with an electron ionization source and operating in single ion mode. Three liters of each sample was passed through an ME™-series Nafion® moisture exchanger and a counterflow MD™-series Nafion® gas drier (Permapure™) prior to preconcentration of volatile components on a Peltier cooled (−30 °C) adsorbent trap (Perkin Elmer TurboMatrix300). Analytes were thermally desorbed at 300°C and separated using a CP-PoraBOND Q column (0.32 mm ID, 25 m, 5 µm, Varian Chrompack). CH3I was quantified using the CH3I+ ion (m/z 142) with I+ (m/z 127) as a qualifier ion. C2H5I, CH2ClI, CHBr2Cl, CH2Br2, and CHBr3 responses were measured using m/z 156/127, 176/178, 129/127, 174/176, and 173/171 as quantification/qualification ions, respectively.
 VSLH were calibrated according to the NOAA (National Oceanic and Atmospheric Administration) 2003/2004 scale (excluding C2H5I) using gas standard SX-3570 (created 2009). Because this standard was not available during the Trade Wind Ozone Photochemistry Experiment (TROMPEX) analysis period, we calibrated using a permeation oven system [Wevill and Carpenter, 2004] used routinely in our laboratory to provide parts per trillion (ppt) gas phase halocarbon mixtures from dynamic dilution of each pure compound with a N2 flow over the permeation tubes (Eco Scientific). Comparison of the permeation oven system to the NOAA standard in November 2009 provided the necessary data to adopt the NOAA scale for this data set (excluding C2H5I). The agreement between these scales ranged from 2% to 22% [Jones et al., 2011].
 The SX-3570 standard consists of 18 halogenated species at parts per trillion concentrations in a 35 L electropolished air sampling cylinder (Essex Cryogenics). The concentration of CH3I, CH2Br2, and CHBr3 in the gas standard is known to be stable, and scale relationships are available publicly (http://www.esrl.noaa.gov/gmd/hats/standard/scales.html). CH2ClI and CHBr2Cl concentrations are based on a preliminary scale with the stability in the Essex cylinder unknown. The stability of the standard SX-3570 was tested by intracomparison with an equivalent NOAA standard SX-3576 created in 2011. The cylinder SX-3576 was analyzed blind and quantified using SX-3570. The concentrations measured in SX-3576 matched the concentrations quoted by NOAA to within 8% for the maximum deviation (CH3I), showing good stability for all compounds.
3 Results and Discussion
3.1 Air Mass Origins
 NOAA Hysplit back trajectories were calculated for the time and location of each data point in order to establish the origin of the air parcel at each specific latitude, longitude, and altitude where samples were collected. The data were then categorized according to five distinct air mass origins (Figure 1). The majority of the air masses followed the expected northeasterly trade winds and traveled parallel to the African coastline, without any contribution from continental air. A few of the air masses arriving at high altitude originated from over Africa; moderate levels of CO (77–83 ppb) and O3 (23–25 ppb) in these air masses indicated that they were not strongly influenced by contaminated or biomass burning sources. VSLH levels in these air masses were not depleted and in some cases were enhanced relative to those in open ocean trajectories, suggesting either a terrestrial source or strong coastal input (Figure 2). High O3 (blue +, Figure 2e) was observed in open ocean trajectories over the Canaries which will have significant European influence (Figure 1). The presence of hurricane Fred south of flight B477 (09 September 2009), which traveled west from Cape Verde islands at 17°N, produced unusual, southerly trajectories which looped around and were possibly influenced by the Southern Hemisphere African continent. This is supported by the unusual CO and O3 concentrations for these trajectories (Figures 2e and 2f), which displayed relatively low O3 characteristic of Southern Hemisphere air masses but also enhanced CO which may be attributable to African biomass burning. O3 concentrations are noticeably higher over the upwelling region; this could be attributed to the emission of NOx from shipping close to the coast of Mauritania.
 Average VSLH mixing ratios measured during TROMPEX are shown in Table 1. CH2I2 was not detected in any of the canister samples, likely due to its rapid photolysis (<5 min [Montzka et al., 2011]) resulting in levels below the limit of detection at the altitudes sampled. The mean MBL mixing ratio of CH2ClI of 0.27 pptv agrees well with springtime measurements in the western north Pacific of 0.27 pptv [Kurihara et al., 2010] and previous measurements made in the Mauritanian upwelling region of 0.24 pptv in the summer of 2007 [Jones et al., 2010]. The mean mixing ratio, however, is skewed by the presence of the upwelling and by changes in altitude. Outside of the upwelling region, the average MBL background level was between 0.1–0.2 pptv (Figure 3), which is more consistent with fieldwork measurements of 0.11 pptv at Mace Head, Ireland [Carpenter et al., 1999; Carpenter et al., 2000], 0.1 pptv at Christmas Island [Varner et al., 2008], and with long-term measurements at Hateruma Island (east of Taiwan, 24°3′N 123°48′E) and Cape Ochiishi (Oshiishinishi, Nemino, Hokkaido 43°9′N 145°30′E) of 0.12 and 0.18 pptv, respectively [Yokouchi et al., 2011]. To our knowledge, these are the first airborne measurements of CH2ClI reported to date. Due to its short atmospheric photodissociation lifetime of ≈2 h [Rattigan et al., 1997], mixing ratios of CH2ClI at altitudes >MBL (≈500 m) were very low (<0.02 pptv). This suggests that the majority of CH2ClI is photolyzed below 500 m, releasing iodine atoms into the boundary layer.
 Figure 3 shows that mixing ratios of CH2ClI in the MBL (defined as 500 m above mean sea level in this region shown by NOAA Hysplit analysis) were elevated above background over the open ocean as well as over the Mauritanian upwelling, a known source region [Jones et al., 2010]. Concentrations were higher in the morning than in the afternoon presumably due to rapid photolysis of CH2ClI; this was also apparent in the two flights between the Canaries and Cape Verde Islands (Figure 4, top). Back trajectory analyses (Figure 1) showed that the open ocean regions with elevated CH2ClI corresponded to trade wind- and hurricane-influenced air masses of marine origin. These air masses had not been influenced by any coastal or upwelling regions for at least 24 h, which is more than 10 times the photolytic lifetime of CH2ClI. CH2ClI can be produced directly from phytoplankton or indirectly from the photolysis of CH2I2 in surface seawater [Martino et al., 2005; Jones and Carpenter, 2005]. The observations suggest an open ocean, possibly phytoplankton-derived source of either CH2ClI, its precursor CH2I2, or both.
 During the open ocean flights between the Canaries (25°N) and Cape Verde Islands (17°N), a latitudinal dependence of CH2ClI and C2H5I was observed with their concentrations doubling and quadrupling, respectively. The mixing ratio increased with decreasing latitude between these latitudes and broadly corresponded to changes in sea surface temperature (Figure 4). CH2ClI fluxes measured during the RHaMBLe (Reactive Halogens in the Marine Boundary Layer) campaign conducted in the same geographical region between the Canaries and Cape Verde Islands in summer 2007 (Figure 4, bottom, Jones et al. ) showed a similar pattern; thus we suggest that the latitudinal dependence of MBL-iodinated VSLH may be driven by temperature or photochemically driven processes in the surface ocean.
 CH3I was measured in all of the samples, but due to problems with the calibration at the time of analysis, the concentrations could not be quantified. The trend in the CH3I data was very similar to CH2Br2 and CHBr3 (section 3.3) with concentrations increasing over the upwelling region. CH3I concentrations also increased with decreasing latitude, similarly to CH2ClI, although these did not show elevated amounts over the open ocean.
 The mean C2H5I mixing ratio was 0.13 pptv, in agreement with measurements from Hateruma Island (0.15 pptv) and Cape Ochiishi (0.08 pptv, Yokouchi et al. ), Asian seas (0.09 pptv, [Yokouchi et al., 1997]), and at Mace Head (0.06 pptv, [Carpenter et al., 1999]).
 CHBr3 and CH2Br2 both showed distinct upwelling influence (Figures 2a and 2b) with CHBr3 concentrations increasing to 3–5 pptv, more than threefold higher than the mean background MBL CHBr3 concentration in that area. Bromocarbon levels in the southerly air masses were similar to, although at the low end, of those in Northern Hemisphere open ocean air. CH2Br2 concentrations were strongly correlated to those of CHBr3 (R2 = 0.89) throughout the flights (Figure 2h (inset)). The injection of fresh emissions from the upwelling into the free troposphere is indicated in Figure 2h, which shows CH2Br2/CHBr3 ratios grouped by air mass origin. CH2Br2 has an atmospheric lifetime of 124 days and is primarily removed via reaction with OH [Montzka et al., 2011]. CHBr3 is photolyzed within 24 days; therefore, an increasing CH2Br2/CHBr3 ratio suggests aging of the air mass, if the emission ratio remains constant. The southerly and African air masses, plus some trade wind trajectories from Europe, contain relatively aged emissions.
 The mean mixing ratios of CH2Br2 and CHBr3 of 1.01 and 1.08 pptv are within the range of previous, ship-based measurements in the area. Higher mean mixing ratios of 2.4 pptv for CH2Br2 and 6.2 pptv for CHBr3 [Quack et al., 2007] were observed in spring 2005 and lower mean mixing ratios of 0.4 pptv for CH2Br2 and 1.1 pptv for CHBr3 [Carpenter et al., 2009] during summer 2007. Over the open ocean, CHBr3 surface concentrations are generally higher than those of CH2Br2 [Butler et al., 2007; Kurihara et al., 2010]. However, throughout the flights, background levels of CH2Br2 were generally higher than CHBr3. This can be explained by considering the altitude at which the TROMPEX samples were taken (500–3000 m) and from the shorter lifetime of CHBr3 compared to that of CH2Br2, indicating an aged background air mass.
 Unlike the iodocarbons, the brominated species did not show a clear latitudinal dependence.
 The Mauritanian upwelling was confirmed as a significant regional source of CHBr3, CH2Br2, CHCl3, C2H5I, and CH2ClI to the MBL. However, mixing ratios over the upwelling were similar to if not slightly lower than observed in coastal regions [e.g., Carpenter et al., 1999; Quack and Wallace, 2003], which cover a much larger total surface area, implying that coastal upwelling systems are not strong global sources of VSLH [Carpenter et al., 2009]. Ratios of CH2Br2/CHBr3 revealed distinct clusters of data which correspond to different air mass origins, with the lowest ratios corresponding to fresh emission from the upwelling region. The ubiquitous presence of CH2ClI over the open ocean shows that it must have an open ocean source. Although concentrations measured were very low, CH2ClI has an extremely short lifetime and is an important source of reactive iodine to the MBL [Jones et al., 2010]. Iodocarbons showed a latitudinal dependence with concentrations increasing with decreasing latitude similarly to SST, in line with previous flux measurements of CH2ClI in the same region. Greater iodocarbon emissions would result in an overall increase in tropospheric ozone destruction towards the equator.
 We would like to thank FAAM (Facility for Airborne Atmospheric Measurements) for the aircraft facilities, the Natural Environment Research Council (NERC) for funding the TROMPEX flying hours, and NEODAAS (NERC Earth Observation Data Acquisition and Analysis Service) for providing MODIS Chl-a images. S.J.A. thanks NCAS for funding his studentship.