Observations of coincident high relative humidity and low surface ozone are common in air quality data sets, but models underpredict the strength of this correlation. We perform a statistical analysis of 28 years of ozone and meteorology observations taken as part of the Clean Air Status and Trends Network across the United States and find that vapor pressure deficit (VPD) is the strongest predictor of midday ozone in the spring, summer, and fall, and this correlation is strongest at sites with the largest leaf area index. We argue that stomatal regulation of dry deposition, which is known to have a VPD dependence that is not typically included in model parameterizations, can explain this relationship. Using a box model of ozone production and loss, we show that a negative ozone‐humidity slope is only achieved by the inclusion of VPD‐dependent dry deposition, suggesting that this mechanism may explain the observed ozone‐humidity correlation.

Abstract. The FORCAsT (FORest Canopy Atmosphere Transfer) model version 1.0 is updated to FORCAsT 2.0 by implementing five major changes, including (1) a change to the operator splitting, separating chemistry from emission and dry deposition, which reduces the run time of the gas-phase chemistry by 70 % and produces a more realistic in-canopy profile for isoprene; (2) a modification of the eddy diffusivity parameterization to produce greater and more realistic vertical mixing in the boundary layer, which ameliorates the unrealistic simulated end-of-day peaks in isoprene under well-mixed conditions and improves daytime air temperature; (3) updates to dry deposition velocities with available measurements; (4) implementation of the Reduced Caltech Isoprene Mechanism (RCIM) to reflect the current knowledge of isoprene oxidation; and (5) extension of the aerosol module to include isoprene-derived secondary organic aerosol (iSOA) formation. Along with the operator splitting, modified vertical mixing, and dry deposition, RCIM improves the estimation of first-generation isoprene oxidation products (methyl vinyl ketone and methacrolein) and some second-generation products (such as isoprene epoxydiols). Inclusion of isoprene in the aerosol module in FORCAsT 2.0 leads to a 7 % mass yield of iSOA. The most important iSOA precursors are IEPOX and tetrafunctionals, which together account for >86 % of total iSOA. The iSOA formed from organic nitrates is more important in the canopy, accounting for 11 % of the total iSOA. The tetrafunctionals compose up to 23 % of the total iSOA formation, highlighting the importance of the fate (i.e., dry deposition and gas-phase chemistry) of later-generation isoprene oxidation products in estimating iSOA formation.

At the University of Michigan Biological Station during the 2016 AMOS field campaign, isoprene concentrations typically peak in the early afternoon (around 15:00 local time, LT) under well-mixed conditions. However, an end-of-day peak (around 21:00 LT) occurs on 23% of the campaign days, followed by a rapid removal (from 21:00-22:00 LT) at rate of 0.57 hr −1 during the day-tonight transition period. During the end-of-day peak, in-canopy isoprene concentrations increase by 77% (from 3.5 to 6.2 ppbv) on average. Stratification and weak winds (<3.4 m s −1 at 46 m) significantly suppress turbulent exchanges between in-and above-canopy, leading to accumulation of isoprene emitted at dusk. A critical standard deviation of the vertical velocity (σ w) of 0.14, 0.2, and 0.29 m s −1 is identified to detect the end-of-day peak for the height of 13, 21, and 34 m, respectively. In 85% of the end-of-day cases, the wind speed increases above 2.5 m s −1 after the peak along with a shift in wind direction, and turbulence is reestablished. Therefore, the wind speed of 2.5 m s −1 is considered as the threshold point where turbulence switches from being independent of wind speed to dependent on wind speed. The reinstated turbulence accounts for 80% of the subsequent isoprene removal with the remaining 20% explained by chemical reactions with hydroxyl radicals, ozone, and nitrate radicals. Observed isoprene fluxes do not support the argument that the end-of-day peak is reduced by vertical turbulent mixing, and we hypothesize that horizontal advection may play a role.

Abstract. The FORCAsT (FORest Canopy Atmosphere Transfer) model version 1.0 is updated to FORCAsT 2.0 by implementing five major changes, including (1) a change to the operator splitting, separating chemistry from emission and dry deposition, which reduces the run time of the gas-phase chemistry by 70 % and produces a more realistic in-canopy profile for isoprene; (2) a modification of the eddy diffusivity parameterization to produce greater and more realistic vertical mixing in the boundary layer, which ameliorates the unrealistic simulated end-of-day peaks in isoprene under well-mixed conditions and improves daytime air temperature; (3) updates to dry deposition velocities with available measurements; (4) implementation of the Reduced Caltech isoprene mechanism (RCIM) to reflect the current knowledge of isoprene oxidation; and (5) extension of the aerosol module to include isoprene-derived aerosol (iSOA) formation. Along with the operator splitting, modified vertical mixing and dry deposition, RCIM improves the estimation of first generation isoprene oxidation products (methyl vinyl ketone and methacrolein) and some second generation products (such as isoprene epoxydiols). Inclusion of isoprene in the aerosol module in FORCAsT 2.0 leads to a 7 % mass yield of iSOA. The most important iSOA precursors are IEPOX and tetrafunctionals, which together account for > 86 % of total iSOA. The iSOA formed from organic nitrates are more important in the canopy, accounting for 11 % of the total iSOA. The tetrafunctionals compose up to 23 % of the total iSOA formation, highlighting the importance of the fate (i.e. dry deposition and gas-phase chemistry) of later-generation isoprene oxidation products in estimating iSOA formation.

No abstract