One of the unique aspects of COBRA is that it will bring together a very wide range of instruments for both in-situ measurements and fluxes of halogens and other trace gases and particles. Some of this instrumentation is newly developed and/or will be developed within the timeframe of COBRA (yet largely funded from elsewhere), namely cavity-enhanced DOAS for in-situ and sensitive measurements of halogen radicals and I2 and a denuder-based ICP-MS technique for in-situ measurements of I2.
Cavity-enhanced differential optical absorption spectroscopy (CE-DOAS) is a variant of the very wellestablished long-path DOAS (LP-DOAS) technique. CE-DOAS, recently demonstrated by A. Ruth and colleagues (Fiedler et al. 2003), uses a short-arc broad-band Xe lamp source which is fed into an optically stable cavity between two highly-reflecting spherical mirrors (similar to a cavity ring down experiment); light that emerges from the cavity is then spectrally dispersed using a conventional spectrometer with a CCD detector. We will collaborate closely with Dr Ruth (University College Cork, Ireland) to develop this technique into a field instrument for measuring BrO (315-360 nm), IO (413-453 nm), OIO and I2 (535 - 575 nm). Anticipated detection limits are a factor of ~ 5 worse than our LPDOAS (e.g. [BrO] > 4 ppt), but this should be acceptable for measuring the high levels of halogen oxides in polar springtime. For atmospheric measurements, we will determine the effective path length of the cavity by reference to the H2O overtone bands at 445 and 571 nm using a hygrometer to measure the atmospheric humidity, and independently measured O3 in the Huggins bands for the 330 nm region.
The CE-DOAS technique is being developed as part of other current projects. An additional in-situ technique for determination of molecular iodine will be developed specifically for COBRA, based on the methodology of Hoffman et al. (2005), using multiple denuder banks analysed by ICP/MS. This will allow collection of semi-continuous measurements of I2 over several diurnal periods. We estimate it will take 1 year after the start date of COBRA to build and test field versions of these systems before deployment (the development of the CE-DOAS is starting in October 2005).
To date there are, to our knowledge, no published reports of co-measured BrO, IO, Hg and O3 in the Arctic. COBRA will be the most comprehensively instrumented Polar halogen atmospheric chemistry campaign to date, and be used to generate a detailed inventory of halogen compounds in Arctic air, using for the first time a combination of cavity-enhanced (in-situ) DOAS, active long-path DOAS, MAXDOAS, off-line ICP-MS, on-line GC-MS, and on-line APCI-MS, in conjunction with a wide range of supporting data including mercury speciated in the gas, particle and snow phase, particle measurements including iodine enhancement in the ultrafine mode, and ancillary trace gas measurements. The 7 week long field experiment during spring at Kuujjuarapik will collect data on an extensive range of halogen species (I2/IO/BrO/OIO, Br2/BrCl, organohalogens including CH2I2, CH2IBr, CH2ICl, CH2Br2, CHBr2Cl, CHBr3), volatile organic compounds (non- methane hydrocarbons, C2-C5 carbonyl compounds), DMS and inorganic species (O3, Hg, NO, NO2, CO) in the gas phase, and photolysis frequencies, focused on measuring the key parameters needed to constrain or validate the developed models. Non-methane hydrocarbons will be used as tracers for active Br and Cl chemistry, e.g. by analysing for selective removal of acetylene compared to benzene, allowing comparison of in-situ measured reactive halogens and integrated, calculated Br and Cl concentrations. Boundary layer meteorology in the field will be characterised via tethered balloon and kite systems, giving boundary layer height up to about 350 m, and ozone concentrations to 250 m.
The Kuujjuarapik campaign will also produce data on aerosol composition and properties. Particle number concentrations greater than 3 nm and 10 nm will be measured using two counters, also providing a measure of the ultra-fine particle population continuously (1 Hz). Particle size distributions between 3 and 3000 nm diameter will be made using a Differential Mobility Particle Sizer (DMPS) and optical particle counters (OPC). Impactor measurements and off-line analysis using TEM/ESEM and EDX will provide atomic composition and the mixing state of iodine in bulk particles.
Flux measurements to be made on the sea-ice during the Kuujjuarapik campaign are described in detail in A4. Following each campaign/study, data will be validated by individual groups, and recorded in common NASA Ames format. Co-I Evans will act as data manager to oversee this activity, and to co-ordinate the submission of data to the BADC. Data would be made publicly available two years from date of submission.
Knowledge of the precise nature of the sea ice/snowpack/frost flower surfaces will be key to interpreting the field trace gas measurements and understanding (and thereby providing parameterizations of) the link between the halogen source and atmospheric concentrations. Surfaces will be characterized according to several variables. Samples will be collected in the field: both the immediate surface (such as frost flowers or upper snowpack) and the underlying ones (such as brine under frost flowers) will be collected separately at regular intervals during the measurement cycle. Part of each sample will be returned to the UK for off-line analysis of major ions using an ion chromatograph. Existing capability includes sodium, chloride, nitrate, sulphate, bromide. Iodide will be quantified by an external laboratory under contract. In addition, we will measure the pH of the various surfaces, given the pH-dependence of a number of the relevant reactions.
The specific surface area of different surfaces will be measured using the BET isotherm method previously used successfully on snow. If we are successful in a parallel bid, then we will have our own SSA equipment funded already, which will be made under contract by local suppliers following the method specified in Legagneux et al. (2002). In case the parallel bid fails, we will instead collect samples in the field and transport them under liquid nitrogen (using cryogenic travel-approved shipping Dewars) to Europe, where F. Domine is willing to work with us to analyse a reduced number of samples using his apparatus.
Samples of under-ice diatoms will be collected by scraping the underside of the sea ice with simple tools, both through holes drilled in the ice and at the edge of open water leads. At least 40 samples will be collected, and bagged via a simple protocol involving dark cold storage, for microscope examination and analysis back at BAS Cambridge. Drill, tools and operator training will be provided. The analysis will result in a comprehensive statement of numbers of each species of diatom in each sample, in order to relate to the potential for production of iodine compounds from literature values.
In activity 4, we aim to determine what type of surfaces are the most effective halogen atom producers; first year sea-ice on which frost flowers exist, open leads from which ice algal emissions can ventilate, surface diatoms on sea-ice, or ventilated snowpack enriched in surface deposited sea-salt? This wideranging activity can be broken down to smaller sub-objectives, i.e. to:
A combination of tried and tested micrometeorological measurement techniques, described below, at small ice-camps on the sea-ice will be used to assess the relative roles of different surfaces (characterised as described in activity A3), with different ionic compositions, in halogen, ozone and particle exchange. We will analyse I2 and halocarbons using off-line adsorbent tube/mass spectrometry methodologies, simply requiring pumps and an automated tube sampler. Measurements of Br2, BrCl (by APCI mass spectrometry- see Spicer et al. (2002)), O3, and particles will be made in situ, requiring an "Arctic proofed" enclosure mounted on sleds i.e. a movable ice camp, and powered from a generator located near the field station (using 3-phase extension cables of up to 1 km and a distribution box mounted on one of the sleds). Because the ice-camps will be within 1 km of the shore, the instrumentation/sleds can be pulled by skidoo back to the shore and stored in a shipping container when not in use, reducing any potential damage. The surfaces studied will include artificial leads created in low wind and low temperature (<- 20C) conditions by digging small holes (~ 1 m x 1 m) into the young sea ice within the first 2 weeks of the Kuujjuarapik campaign (mid February to early March), when temperatures are sufficiently low. Almost any newly-frozen lead under low wind and low temperature conditions will grow frost flowers (e.g. Martin et al., 1996 and references therein). Depending on the temperature, the frost flowers have different crystal natures and growth habits; laboratory studies show that at temperatures of about -20C, structures form after ~ 2.5 days (Martin et al., 1996). In the field, colleagues at DAMPT, Cambridge (Grae Worster, pers. comm., 2005) have found it straightforward to create leads at Spitzbergen and we aim to follow their approach. As these leads evolve from open water to frozen sea-ice, we will characterise both ice algae emissions from the open lead, and later frost flower emissions as a function of age. The Kuujjuarapik station director, Claude Tremblay, along with locally employed Inuit people will be responsible for the creation of the leads and will provide safety measures on the sea-ice.
Dynamic flux chambers will be deployed simultaneously over (i) open leads (to investigate frost flower sources), and (ii) snow pack surfaces. We will measure I2, Br2, BrCl, halocarbons, O3, and sizesegregated and ultrafine particles in the chambers. We estimate a total power consumption of 7 kW for these experiments. We will construct a ~ 1 m x 1 m chamber, coated with Mylar, with a minimum volume of ~ 80 L to fit over the leads. The chamber will be stirred to avoid vertical concentration gradients. In each case multiple air samples will be collected from the chambers over relatively short periods before re-exposing the underlying surfaces to the atmosphere. The procedure of alternate measurement and re-exposure will be repeated over a period of several days and at different locations to; minimise microclimate induced changes that will occur within the chambers and so influence possible emission processes; and to account for spatial heterogeneity in the sources. Chambers that are opaque or transparent to UV radiation will be deployed simultaneously to simulate arctic day-night changes in emission sources. Temperature, pH and light levels will be measured within the chambers. The flux gradient technique (using the same sample and analytical techniques as used in the chambers) will be used to (i) confirm open leads as sources (using continuous concentration profile measurements upwind and downwind of the lead), and (ii) to measure emissions from snowpack sources whose source fetch is unaffected by leads. The efficacy and quality of the flux gradient measurements to be used in the final analysis to assign source footprints and strengths will be controlled by measurements made using turbulence and sensible heat flux measurements with sonic anemometers.
The eddy correlation technique in conjunction with a fast response chemiluminescence ozone analyser will be used to determine fluxes of ozone deposition to snowpack and fetches influenced by open leads. Finally, the role of sea salt aerosol in air in liberating halogens will be investigated using models constrained by measured halogen oxide concentrations and size-resolved aerosol distributions. The model of McFiggans et al. (2000, 2002), extended to include full bromine chemistry and treatment of aerosol microphysics and aqueous chemistry will be initialised by measured gas phase concentrations and aerosol distribution measurements and driven by halogen precursor fluxes and measured radiometry to predict the concentrations of reactive halogen species (RHS). Whilst source heterogeneity may not allow direct comparison with measured RHS, a predicted climatology of RHS based on a range of emission scenarios may be constructed and used to estimate the impact on ozone concentrations and HOx species.
The goal of the laboratory experiments is to separate and adequately quantify the impact of relevant physical and chemical processes driving the exchange of trace gases and kinetics. These experiments constitute an interface between the field measurements and the models, and are designed to improve both the interpretation of the field data and model parameterizations. The experiments build on previous experience in growing artificial frost flowers [Martin et al., 1996], and in investigating O3 loss over frozen salt solutions in the presence and absence of UV-radiation [Hutterli et al., 2005; Oum et al., 1998; Rankin, 2004]. Recent results suggest a significant O3 loss over frozen NaBr and NaI solutions in the dark and the production of compounds that photolyse to destroy O3. The O3 losses increased in the presence of frost flowers.
A new, closed environmental chamber will be built to accommodate the connection of a subset of the gas-phase and aerosol monitoring instruments used in the field campaigns (iodine/bromine halocarbons, I2, IO, O3, NOx, total condensation particle counts with 3nm cut-off). Vertical temperature gradients in the chamber across the salt water and air in the headspace as well as relative humidity in the latter will be controllable over a large range covering typical values encountered in the Arctic and will allow growing various types and amounts of frost flowers. The composition of the inflowing air and the salt solution can be varied, and the chamber irradiated with UV. In addition to measuring the processed air, the concentration of major ions and the pH of the produced ice, brine and frost flowers will be measured. The specific surface area of the frost flowers will be determined and the evolution of the frost flowers recorded with a digital camera.
Experimental set-up will commence during autumn 2007, with initial lab experiments in late 2007. More specific experiments, based on the field work, will be conducted throughout 2008.
Here we will extend the work of Co-I Evans (2003) to include iodine and Hg chemistry and to describe the flux of material between the snow / ice / ocean surface and the boundary layer in a 1D framework. The parameterization of fluxes between the surface and the atmosphere will be based on the observations made in the other Activities, and will be used to investigate the linked flux of halogens, ozone and other species into or out from the surface. The methodology here will follow that of our previous work which used chemical coordinates (typically the O3 concentration but in this case the Hg concentration may prove useful) rather than temporal or spatial coordinates. This helps to remove many of the problems of interpreting observations collected at a fixed site and allows the emphasis of model / measurement failure to be towards the chemical / physical rather than dynamical processes.
The scientific goals are to:
A quantitative relationship between atomic iodine precursor strength and new particle nucleation and growth (Saiz-Lopez et al., 2005) has recently been reported in the coastal environment. There is no reason to suspect that such a physical relationship is not universal; therefore particle formation will be favoured wherever there is sufficient iodine atom flux and available ozone. Competition for the iodine- (and bromine-) containing species will be provided by condensation to available pre-existing aerosol particles. This measurement- and modelling-based activity seeks to investigate the condensed-phase fate of reactive halogen species (RHS) with particular emphasis on new particle formation.
There are a number of questions which will be addressed by this activity, including:
To address these, a range of instrumentation will be deployed to characterise the aerosol population as a function of wind sector (hence source region). These will include a differential mobility particle sizer (DMPS) system, condensation particle counters, an optical particle spectrometer and an impactor system loaded with TEM grids
A model of iodine cluster nucleation and growth, similar to that used in Saiz-Lopez et al (2005) will be coupled to the model of halogen cycling used in (A4) to investigate the impact of the iodine on aerosol composition and distribution. Fluxes derived in A4 will be used to drive the model and thus the potential for iodine species emitted from the sea, snow or ice surface to contribute to aerosol processes will be investigated, comparing model output with measured aerosol physical properties.