Peroxy Radical Kinetics Resulting from the OH-Initiated Oxidation of 1,3-Butadiene, 2,3-Dimethyl-1,3-Butadiene and Isoprene

The laser flash photolysis/UV absorption spectrometry technique has been used to investigate the kinetics of the peroxy radical permutation reactions (i.e. self and cross reactions) arising from the OH-initiated oxidation of isoprene (2-methyl-1,3-butadiene), and of the simpler, but related conjugated dienes, 1,3-butadiene and 2,3-dimethyl-1,3-butadiene. The results of the two simpler systems are analysed to provide values of the rate coefficients for the 6 peroxy radical permutation reactions of the three types of isomeric peroxy radical produced in each system (T = 298 K, P = 760 Torr). The rate coefficients are all significantly larger than values estimated previously by extrapolation of structure-reactivity relationships based on the kinetics of a limited dataset of simpler radicals containing similar structural features. The results are discussed in terms of trends in self and cross reaction reactivity of primary, secondary and tertiary peroxy radicals containing combinations of allyl, -hydroxy and -hydroxy functionalities. Since the peroxy radicals formed in these systems are structurally very similar to those formed in the isoprene system, the kinetic parameters derived from the results of the simpler systems are used to assist the assignment of kinetic parameters to the 21 permutation reactions of the six types of isomeric peroxy radical generated in the isoprene system. Kinetic models describing the OH-initiated degradation of all three conjugated dienes to first generation products in the absence of NO_x are recommended, which are also consistent with available end product studies. The model for isoprene is considered to be a further improvement on that suggested previously for its OH-initiated oxidation in the absence of NO_x. The mechanism is further extended to include chemistry applicable to NO_x-present conditions, and calculated product yields are compared with those reported in the literature.     

Tropospheric Box-Modelling and Analytical Studies of the Hydroxyl (OH) Radical and Related Species: Comparison with Observations

Calculated and observed hydroxyl (OH) fields are presented. Calculated OH was obtained in three ways using (1) a photochemical box-model (2) a simple OH steady state approach and (3) a variant on (2) – the multiple equation steady state approach which assumes steady state for OH, HO2 and RO2 and hence obtains three simultaneous, non linear, equations. All three methods used data collected in June 1995 during the Weybourne Atmospheric Observatory Summer Experiment (WAOSE'95). Julian Days 169, 178, 179 and 180 displayed especially good data capture and were consequently chosen for study. The two steady state methods are essentially driven purely by observations and derive OH from the ratio of the relevant source and sink terms. The box-model was constrained where possible to observations; remaining unmeasured volatile organic compounds (VOCs) were initialised to an arbitrary low value of 10 ppt. Agreement between theory and experiment was usually around 50% and often better than this value, especially on J169, though discrepancies of up to a factor of 3 were occasionally apparent. Despite the inherent scatter, neither the box-model nor the simple steady state method were found to consistently over-estimate OH (a common feature of many numerical approaches) although this did occur to a certain extent using the multiple equation steady state approach, probably due to breakdowns in the steady state approximation. More data spread was evident in the box-model approach compared with the other methods. An analysis of the major sources and sinks of OH is presented for the three methods of calculation. Calculated and observed peroxy radicals are also presented. Calculated peroxy radicals were generally lower than that observed at night yet higher, sometimes by up to a factor of 7, during the day. Possible explanations for this result are explored.     

Kinetics of the Reaction of Hydroxyl Radicals with CH2Br2 and its Implications in the Atmosphere

Rate constants for the reaction of hydroxyl radicals with dibromomethanehave been measured by discharge flow-resonance fluorescence technique(DF-RF) over the temperature range 288–368 K. The derived Arrheniusequation is k₁=(1.51 ± 0.37)× 10^-12 exp(-(720 ±60)/T) cm³ molec.^-1 s^-1.The tropospheric lifetime of dibromomethane has been estimated to be 0.29years. An ozone depletion potential (ODP) value of 0.10 for dibromomethanehas been obtained.     

全球二维大气化学模式和大气化学成分的数值模拟

建立了一个全球二维纬向平均化学模式,模式包括了从90°S到90°N,从地面到20 km高度的大气.模式中应用的流场来自根据加热率计算得到的剩余环流.模式化学部分包括34种大气成分、104个化学反应和光化学反应.其中,甲烷、一氧化碳和氮氧化物排放分为季节性和非季节性排放源,并将其参数化为时间和纬度的函数再应用到模式中去.按1990年的甲烷、一氧化碳和氮氧化物的的排放水平模拟得到了多种大气组成的分布,模拟结果与观测有较好的一致性.由于模式考虑了一氧化碳的季节变化,模拟得到的OH自由基分布更为合理.模式的建立为今后进一步研究大气微量成分的全球循环过程及其长期变化提供了有效的手段.     

Photochemistry in the Arctic Free Troposphere: Ozone Budget and Its Dependence on Nitrogen Oxides and the Production Rate of Free Radicals

Local ozone production and loss rates for the arctic free troposphere (58–85° N, 1–6 km, February–May) during the TroposphericOzone Production about the Spring Equinox (TOPSE) campaign were calculated using a constrained photochemical box model. Estimates were made to assess the importance of local photochemical ozone production relative to transport in accounting for the springtime maximum in arctic free tropospheric ozone. Ozone production and loss rates from our diel steady-state box model constrained by median observations were first compared to two point box models, one run to instantaneous steady-state and the other run to diel steady-state. A consistent picture of local ozone photochemistry was derived by all three box models suggesting that differences between the approaches were not critical. Our model-derived ozone production rates increased by a factor of 28 in the 1–3 km layer and a factor of 7 in the 3–6 kmlayer between February and May. The arctic ozone budget required net import of ozone into the arctic free troposphere throughout the campaign; however, the transport term exceeded the photochemical production only in the lower free troposphere (1–3 km) between February and March. Gross ozone production rates were calculated to increase linearly with NO_x mixing ratiosup to 300 pptv in February and for NO_x mixing ratios up to 500 pptv in May. These NO_x limits are an order of magnitude higher thanmedian NO_x levels observed, illustrating the strong dependence ofgross ozone production rates on NO_x mixing ratios for the majority of theobservations. The threshold NO_x mixing ratio needed for netpositive ozone production was also calculated to increase from NO_x 10pptv in February to 25 pptv in May, suggesting that the NO_x levels needed to sustain net ozone production are lower in winter than spring. This lower NO_x threshold explains how wintertime photochemical ozone production can impact the build-up of ozone over winter and early spring. There is also an altitude dependence as the threshold NO_x neededto produce net ozone shifts to higher values at lower altitudes. This partly explains the calculation of net ozone destruction for the 1–3 km layerand net ozone production for the 3–6 km layer throughout the campaign.     

An Assessment of HOx Chemistry in the Tropical Pacific Boundary Layer: Comparison of Model Simulations with Observations Recorded during PEM Tropics A

Reported are the results from a comparison of OH,H₂O₂CH₃OOH, and O₃ observationswithmodel predictions based on current HO_x–CH₄reaction mechanisms. The field observations are thoserecorded during the NASA GTE field program, PEM-Tropics A. The major focus ofthis paper is on thosedata generated on the NASA P-3B aircraft during a mission flown in the marineboundary layer (MBL) nearChristmas Island, a site located in the central equatorial Pacific (i.e.,2° N, 157° W). Taking advantage of thestability of the southeastern trade-winds, an air parcel was sampled in aLagrangian mode over a significantfraction of a solar day. Analyses of these data revealed excellent agreementbetween model simulated andobserved OH. In addition, the model simulations reproduced the major featuresin the observed diurnalprofiles of H₂O₂ and CH₃OOH. In the case ofO₃, the model captured the key observational feature whichinvolved an early morning maximum. An examination of the MBL HO_xbudget indicated that the O(¹D) + H₂Oreaction is the major source of HO_x while the major sinks involveboth physical and chemical processes involving the peroxide species,H₂O₂ and CH₃OOH. Overall, the generally goodagreement between modeland observations suggests that our current understanding ofHO_x–CH₄ chemistry in the tropical MBL isquite good; however, there remains a need to critically examine this chemistrywhen both CH₂O and HO₂are added to the species measured.     

Chemistry of Halogen Oxides in the Troposphere: Comparison of Model Calculations with Recent Field Data

Reactive halogen species (RHS = X, XO, HOX, OXO; X = Cl, Br, I) are known to have an important influence on the chemistry in the polar boundary layer (BL), where they are responsible for ozone depletion events in spring. Recent field campaigns at Mace Head, Ireland, and the Dead Sea, Israel, identified for the first time iodine oxide (IO) at mixing ratios of up to 6.6 ppt and 90 ppt bromine oxide (BrO), respectively, by DOAS also at lower latitudes. These results intensified the discussion about the role of the RHS in the mid-latitude BL.Photochemical box model calculations show that the observed IO mixing ratios can destroy ~0.45 ppb ozone per hour. This is comparable to the rates of the known O₃-loss processes in the boundary layer. The model studies also reveal that IO, at these levels, has a strong influence on the BL photochemistry, increasing the OH/HO₂- and the NO₂/NO - ratios. In combination these changes lead to a reduction of the photochemical ozone formation, which - in addition - reduces ozone mixing ratios by up to 0.15 ppb/h.The studies for the Dead Sea case give no information on the heterogeneous process responsible for the bromine release, but they show that a total of 2 – 4 ppb of total bromine have to be released to explain the observed complete depletion of 60 ppb ozone in 2 – 3 hours.     

Determination of the Tropospheric Hydroxyl Radical by Liquidphase Scrubbing and HPLC: Preliminary Results

A preliminary study was carried out toexamine the feasibility of measuring tropospherichydroxyl radicals (OH) by liquidphase scrubbing andhigh performance liquid chromatography (HPLC). Thepotential advantages of this approach are itssimplicity, portability, and low expense. Thesampling system employs glass bubblers to trapatmospheric OH into a buffered solution of salicylicacid (o-hydroxybenzoic acid, OHBA). Rapidreaction of OH with OHBA produces a stable fluorescentproduct, 2,5-dihydroxybenzoic acid (2,5-DHBA), whichis determined by reverse-phase HPLC and fluorescencedetection. Our preliminary field results indicatethat this method is most suitable for OH measurementsin clean tropospheric air, where interferences fromother atmospheric species appear to be negligible orminor relative to polluted air. In clean air, thesampling period is about 45–90 minutes, which yieldsa detection limit of approximately 3–6 ×10⁵ radicalscm^-3. During an OHintercomparison experiment at the Caribou samplingsite in Colorado, our liquidphase scrubber method wascompared with the ion-assisted mass spectrometry (MS)method. Our results were within the same range asthose of the ion-assisted MS method (1–5 ×10⁶ radicals cm^-3) within our precision atthat time (about ±30–50%). Preliminary testsin Pullman, WA indicated that the method might alsofunction in moderately polluted air by acidifying thescrubbing solution or by adding a scavenger tosuppress interferences. In Pullman, mid-day OHconcentrations were usually in the range of 2–20 ×10⁶ radicals cm^-3. Nighttime OHconcentrations were always low, either at or slightlyabove the detection limit.     

Formulation and Evaluation of IMS, an Interactive Three-Dimensional Tropospheric Chemical Transport Model 2. Model Chemistry and Comparison of Modelled CH4, CO, and O3 with Surface Measurements

In part two of this series of papers on the IMS model, we present the chemistry reaction mechanism usedand compare modelled CH₄, CO, and O₃ witha dataset of annual surface measurements. The modelled monthly and 24-hour mean tropospheric OH concentrationsrange between 5–22 × 10⁵ moleculescm^–3, indicating an annualaveraged OH concentration of about 10 × 10⁵ moleculescm^–3. This valueis close to the estimated 9.7 ± 0.6 × 10⁵ moleculescm^–3 calculated fromthe reaction of CH₃CCl₃ with OH radicals.Comparison with CH₄ generally shows good agreementbetween model and measurements, except for the site at Barrow where modelledwetland emission in the summer could be a factor 3 too high.For CO, the pronounced seasonality shown in the measurements is generally reproduced by the model; however, the modelled concentrations are lower thanthe measurements. This discrepancy may due to lower the CO emission,especially from biomass burning,used in the model compared with other studies.For O₃, good agreement between the model and measurements is seenat locations which are away from industrial regions. The maximum discrepancies between modelled results and measurementsat tropical and remote marine sites is about 5–10 ppbv,while the discrepancies canexceed 30 ppbv in the industrial regions.Comparisons in rural areas at European and American continental sites arehighly influenced by the local photochemicalproduction, which is difficult to model with a coarse global CTM.The very large variations of O₃ at these locations vary from about15–25 ppbv in Januaryto 55–65 ppbv in July–August. The observed annual O₃amplitude isabout 40 ppbv compared with about 20 ppbv in the model. An overall comparison of modelled O₃ with measurements shows thatthe O₃seasonal surface cycle is generally governed bythe relative importance of two key mechanisms that drivea springtime ozone maximum and asummertime ozone maximum.     

Changing trends in tropospheric methane and carbon monoxide: A sensitivity analysis of the OH-radical

A sensitivity analysis is performed in order to study recently observed changes in atmospheric methane and carbon monoxide trends. For the analysis we have adapted a one-dimensional transport/chemistry model in order to comply with changes in vertical transport, stratosphere-troposphere flux of ozone, the water vapour cycle and the short-wave radiative transfer. In addition we have formulated an improved relationship which expresses the steady state OH concentration in terms of longer lived compounds which has a fair agreement with the one-dimensional model results. An analysis of the observed changes and trends in methane and carbon monoxide shows that both emissions and changes in global OH concentrations can be main causes for the observed changes. Average methane emissions have slowed down, particularly in the NH, in the last five years, though perhaps not very significantly. Carbon monoxide emissions are decreasing faster in the last couple of years than in the period 1983–1990. The study suggests that climate fluctuations (tropospheric water vapour, temperature and convective activity) and the stratospheric ozone depletion (tropospheric UV radiation) have a significant influence on tropospheric composition and thus on trends in methane and carbon monoxide concentrations.The IMAU is partner in the Netherlands Centre for Climate Research (CCR).     

Doppler Detection of Hydroxyl Column Abundance in the Middle Atmosphere: 2. Measurement for Three Years and Comparison with a 1D Model

The hydroxyl column abundance in the middle atmosphere is measured for 3 years in Tokyo by ground-based uv absorption spectroscopy utilizing the Doppler shift of the solar spectrum due to the solar rotation. The diurnal and seasonal variations of the measured column abundance are found to be reproduced qualitatively by a one-dimensional photochemical-diffusive model; however, the measured column abundances are smaller than the calculated by 20 to 30% systematically.     

Urban Atmospheric Chemistry During the PUMA Campaign 2: Radical Budgets for OH,HO<Subscript>2</Subscript> and RO<Subscript>2</Subscript>

A detailed photochemical box model was used to investigate the key reaction pathways between OH, HO₂ and RO₂ radicals during the summer and winter PUMA field campaigns in the urban city-centre of Birmingham in the UK. The model employed the most recent version of the Master Chemical Mechanism and was constrained to 15-minute average measurements of long-lived species determined in situ at the site. The results showed that in the summer, OH initiation was dominated by the reactions of ozone with alkenes, nitrous acid (HONO) photolysis and the reaction of excited oxygen atoms atoms with water. In the winter, ozone+alkene reactions were the primary initiation route, with a minor contribution from HONO photolysis. Photolysis of aldehydes was the main initiation route for HO₂, in both summer and winter. RO₂ initiation was dominated by the photolysis of aldehydes in the summer with a smaller contribution from ozone+alkenes, a situation that was reversed in the winter. At night, ozone+alkene reactions were the main radical source. Termination, under all conditions, primarily involved reactions with NO (OH) and NO₂ (OH and RCO₃). These results demonstrate the importance of ozone+alkene reactions in urban atmospheres, particularly when photolysis reactions were less important during winter and at nighttime. The implications for urban atmospheric chemistry are discussed.     

中国及邻近地区污染排放对对流层臭氧变化与辐射影响的研究

利用双向耦合的区域气候模式和大气化学模式系统, 研究了中国与邻近地区人为污染排放引起对流层臭氧变化和产生的辐射强迫.结果表明, 污染排放对对流层臭氧含量的影响有明显的季节变化, 对北方的影响不如南方显著, 西部的季节变化稳定且小于东部, 内陆污染地区各季节臭氧柱含量的变化量均较高.对整个模拟区域而言, 臭氧变化量的年平均值为30.928 DU, 春季最大为32.168 DU, 而空间分布变化在12～38 DU之间.臭氧变化量对北方地区辐射的影响较小, 而对低纬和华东地区影响较大, 臭氧变化量引起的晴空地气系统短波辐射强迫、长波辐射强迫的平均值分别是0.185 W·m-2和0.464 W·m-2, 标准化短波辐射强迫与净辐射强迫值为0.006 W·m-2·DU-1和0.021 W·m-2·DU-1.气候反馈过程对对流层臭氧含量的影响范围在-0.470～0.752 DU之间, 包含气候反馈过程的区域年平均臭氧变化量是30.942 DU.在气候反馈条件下, 臭氧变化量的短波和长波辐射强迫分别是0.249 W·m-2及 0.482 W·m-2, 标准化的短波与净辐射强迫值为0.008 W·m-2·DU-1和0.024 W·m-2·DU-1.臭氧变化量导致地表温度的变化范围在±0.80 K之间.     

Reduction of a Model Describing Ozone Oscillations in the Troposphere: Example of an Algorithmic Approach to Model Reduction in Atmospheric Chemistry

We consider a simplified reaction mechanism from tropospheric chemistryconsisting ofsix chemical species involved in ten dynamic processes. The concentrations of all six speciesundergo temporal oscillation for some parameter values. An asymptoticapproach to reduction of the original six-variable model to a four-variable version is considered together with detailed explanation of the procedure as well as discussion of conditions under whichthe reduction is possible. The reduced system nearly quantitatively represents the oscillating behavior of thefull model and allows elucidation of its basic dynamical features. This approach also can be used to compare various small models of atmospheric chemistry, and to determine their underlying dynamic structure. The method can be systematically appliedto larger atmospheric models.     

Lagrangian Estimations of Ozone Loss in the Core and Edge Region of the Arctic Polar Vortex 1995/1996: Model Results and Observations

Ozone evolution and diabatic descent in the Arctic polar vortex in winter 1995/1996 was studied with a newly developed diabatic trajectory–chemistry model (DTCM). To study the chemical and dynamic evolution of the species in the polar vortex, 400 diabatic trajectories were calculated in the vortex core and edge region by using three-dimensional (3-D) wind data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). The averaged diabatic descending motion and ozone behavior were obtained for particles started from the core and from the edge region of the vortex. The difference in ozone-loss rates as well as the difference in descending rates between the vortex core and the vortex-edge region was not statistically significant. The average cumulative ozone loss of 65 ± 16% in the vortex core obtained from the model calculations was consistent with the estimates obtained with a different method (Match experiment). The model results for the vortex core were compared with those obtained using trajectories with the vertical winds calculated on the basis of radiative cooling rates as used by the SLIMCAT 3-D chemical transport model. Although the trajectories based on cooling rates exhibited lower descending rates than those based on 3-D analyzed wind data, the ozone behavior was similar for both types of trajectory. Ozonesonde data from two stations (Ny-Alesund in the vortex core and Yakutsk in the vortex edge) were compared with the model results. For Lagrangian estimation of the ozone loss at these stations, the descending rates obtained by the diabatic trajectory calculations were used. Good agreements were obtained between the model results and observations for both the vortex core and edge region. These results suggest that strong ozone depletion occurred not only in the core, but also in the edge region of the vortex, and that air masses from the mid-latitudes did not appreciably affect the degree of ozone depletion in this winter–spring period. The sensitivity of the model to different descending rates and to the presence of large nitric acid trihydrate (NAT) particles was also examined.     

Iodine Chemistry and its Role in Halogen Activation and Ozone Loss in the Marine Boundary Layer: A Model Study

A detailed set of reactions treating the gas and aqueous phase chemistry of the most important iodine species in the marine boundary layer (MBL) has been added to a box model which describes Br and Cl chemistry in the MBL. While Br and Cl originate from seasalt, the I compounds are largely derived photochemically from several biogenic alkyl iodides, in particular CH2I2, CH2ClI, C2H5I, C3H7I, or CH3I which are released from the sea. Their photodissociation produces some inorganic iodine gases which can rapidly react in the gas and aqueous phase with other halogen compounds. Scavenging of the iodine species HI, HOI, INO2, and IONO2 by aerosol particles is not a permanent sink as assumed in previous modeling studies. Aqueous-phase chemical reactions can produce the compounds IBr, ICl, and I2, which will be released back into the gas phase due to their low solubility. Our study, although highly theoretical, suggests that almost all particulate iodine is in the chemical form of IO-3. Other aqueous-phase species are only temporary reservoirs and can be re-activated to yield gas phase iodine. Assuming release rates of the organic iodine compounds which yield atmospheric concentrations similar to some measurements, we calculate significant concentrations of reactive halogen gases. The addition of iodine chemistry to our reaction scheme has the effect of accelerating photochemical Br and Cl release from the seasalt. This causes an enhancement in ozone destruction rates in the MBL over that arising from the well established reactions O(1D) + H2O 2OH, HO2 + O3 OH + 2O2, and OH + O3 HO2 + O2. The given reaction scheme accounts for the formation of particulate iodine which is preferably accumulated in the smaller sulfate aerosol particles.     

三维对流云对大气光化学组分的再分布作用及其化学效应

利用一个三维的冰雹云模式与化学组分输送模块耦合，得到云输送引起大气光化学组分的再分布，然后用一个包含详细气相化学反应机制的箱模式研究了云输送引起的气相体积分数的变化及其对大气化学系统产生的影响。结果表明，云输送后O3体积分数大于无云个例，但其后两天内两者的变化趋势相差不大；HNO3、NO2、NO3、PAN等的体积分数均明显高于无云个例，分别增长了87％、70％、62％和49％，其中NO2体积分数的增加主要由于云输送造成，而NO3、HNO3、PAN主要是输送对化学扰动的结果。两天内OH和HO2自由基体积分数比无云个例平均增长了13％和11％。     

Simulation Chamber Studies of the Atmospheric Oxidation of 2-Methyl-3-Buten-2-ol: Reaction with Hydroxyl Radicals and Ozone Under a Variety of Conditions

This article presents a complete study of the diurnal chemical reactivity of the biogenic volatile organic compound (BVOC), 2-methyl-3-buten-2-ol (MBO) in the troposphere. Reactions of MBO with OH and with ozone were studied to analyse the respective parts of both processes in the global budget of MBO atmospheric reactivity. They were investigated under controlled conditions for pressure (atmospheric pressure) and temperature (298 ± 2 K) using three complementary European simulation chambers. Reaction with OH radicals was studied in the presence of and in the absence of NO _x . The kinetic study was carried out by relative rate study using isoprene as a reference. The rate constant found for this reaction was molecule⁻¹ cm³ s⁻¹. FTIR spectroscopy, DNPH- and PFBHA-derivatisation analyses were performed for reactions with both OH radicals and ozone. In both reactions, the hydroxycarbonyl compound, 2-hydroxy-2-methylpropanal (HMPr) was positively identified and quantified, with a yield of in the reaction with OH, and a yield of and 0.84 ± 0.08 in the reaction with ozone under dry (HR < 1%) and humid conditions (HR = 20%–30%). A primary production of two other carbonyl compounds, acetone , and formaldehyde was found in the case of the dry ozonolysis experiments. Under humid conditions, only formaldehyde was co-produced with HMPr as a primary carbonyl compound, with a yield of . For the reaction with OH, three other carbonyl compounds were detected, acetone , formaldehyde and glycolaldehyde . In addition some realistic photo-oxidation experiments were performed to understand in an overall way the transformations of MBO in the atmosphere. The realistic photo-oxidation experiments were conducted in the EUPHORE outdoor simulation chamber. It was found that this compound is a weak secondary aerosol producer (less than 1% of the carbon balance). But it was confirmed that it is a potentially significant source of acetone, Δ[Acetone]/Δ[MBO] = 0.45. With our experimental conditions ([MBO]₀ = 200 ppb, [NO]o = 50 ppb), an ozone yield of Δ[O₃]/Δ[MBO] = 1.05 was found.     

Urban Atmospheric Chemistry During the PUMA Campaign 1: Comparison of Modelled OH and HO<Subscript>2</Subscript> Concentrations with Measurements

Photochemical box modelling was undertaken to investigate OH and HO₂ radical chemistry during summer and winter field campaigns in the urban city-centre of Birmingham in the UK. The model employed the most recent version of the Master Chemical Mechanism (v3.1) and was constrained to 15-minute average measurements of long-lived species determined in situ at the site. The model was used to predict OH and HO₂ concentrations for comparison with measurements made by the fluorescence assay by gas expansion technique. Whilst there was generally good agreement between the modelled and measured OH concentrations, particularly during summer, there was sometimes a significant model under-prediction during daylight hours, which significantly skews the overall model: measured agreement. There were less measured data available for HO₂, but the agreement between model and measurement for the days where measurements existed were less good than for OH, with one or two exceptions. The modelled:measured ratios between the hours of 11:00–15:00 h for OH were 0.58 and 0.50 for summer and winter respectively. For HO₂, the same ratios were 0.56 in the summer and 0.49 in the winter. Sensitivity studies were conducted to attempt to understand the model-measurement discrepancy. The predicted radical concentrations were particularly sensitive to changes in NO_X concentrations. Constraining the model to the observed HO₂ concentrations made the OH predictions worse. These results highlight the fact that there are many complexities in urban areas and that more highly-instrumented campaigns are required in the future to further our understanding.