Using global models and chemical observations to diagnose eddy diffusion
Feb 05, 2016
Using global models and chemical observations to diagnose eddy diffusion
•Goal: determine the eddy diffusion rate in the upper mesosphere, including latitudinal and seasonal changes
•Proposal: use a 3-D chemical model to determine which global measurements can best constrain the mean global diffusivity coefficient; use the measurements and model to narrow the range of diffusion rates
•Why do we want/need to know diffusion?•theoretical (how much turbulence & diffusion is generated by gravity wave breaking?)•without knowing diffusive transport, we don’t know if/when our chemical simulations are correct
•What do we know now?•current estimates from observations and numerical models differ widely
•Why use chemicals?•different constituents are sensitive to diffusion over different altitude ranges•matching profiles for multiple constituents provides a stringent test of our estimates
WHAT MAKES A CHEMICAL USEFUL?
• concentration large enough to be measurable• sensitivity to transport because of either
• long lifetime & strong vertical gradient• short lifetime but equilibrium concentration
depends on transported species• reactions and rate coefficients reasonably well known• consistent response (for example, increases
monotonically with increasing diffusivity)
ROSE model
•vertical range: tropopause to thermosphere•driven by meteorological observations at lower boundary•radiation and dynamics can be decoupled from chemistry (as in the present study)•time-dependent chemistry•easily changed for mechanistic studies•simulates
•oxygen: O(1D), O, O2, O3
•hydrogen: H, OH, HO2, H2O, H2O2, H2, CH4
•nitrogen*: N, NO, NO2, NO3, HNO3, N2O5, N2O, HO2NO2
•chlorine: Cl, ClO, HCl, HOCl, ClONO2, CFCl3, CF2Cl2•carbon: CO, CH2O, CO2
*Note: thermospheric NO is specified based on SNOE empirical model (Marsh et al. 2004)
eddy diffusion coefficient Kzz in ROSE model
€
DχDt
=P−Lχ+ 1ρ∂∂z
ρKzz∂χ∂z
⎛
⎝ ⎜
⎞
⎠ ⎟+FM
€
P−Lχ =souρχe&sink
€
FM =moleχulaρdiffusion
Kzz is the eddy diffusion coefficient:calculated by the Hines gravity wave drag parameterization and includes effective Prandtl number
m2/s
chemical continuity eqn for mixingratio χ:
How the model is used
•several model integrations with different levels of eddy diffusion in the chemical continuity eqn; otherwise identical•NOTE: large-scale dynamics is identical in all runs because •the dynamical Kzz does not change•these runs are uncoupled (climatological radiative gases)
•comparison of averaged vertical profiles: global at all local times or day-only and night-only•subjective assessment of which provide the best constraints on diffusion assuming the availability of global measurements over all local times •actual application will depend on extent and accuracy of measurements
basic results• source gases and stratospheric species that are not useful because
concentrations are too small in mesosphere– hydrogen family: H2O2
– nitrogen family: NO3, HNO3, N2O5, N2O, HO2NO2
– chlorine family: ClO, HOCl, ClONO2, CFCl3, CF2Cl2– carbon family: CH2O
• other species with low concentrations– oxygen family: O(1D)
• species that are not useful because of weak vertical gradient– O2 and H2
• species that cannot be tested due to specified thermospheric NO in ROSE model
– nitrogen: N, NO, NO2,
• species considered below– oxygen family: O, O3
– hydrogen family: H, OH, HO2, H2O, CH4
– chlorine family: Cl, HCl– carbon family: CO, CO2
sample for interpretation of the model results
• curves show global mean CO profiles from 4 model runs
• CO increases with altitude due to a source in the thermosphere
• higher diffusion leads to lower mesospheric CO due to upward transport of low-CO air
• the differences among the 4 cases increase with altitude in the mesosphere
• interpretation: CO could provide a good diagnostic of diffusion near the mesopause
profile values range from zero to 2 x 10-4 vmr
CO and CO2 (transport)
• With increasing diffusion, CO2 increases near the mesopause while CO decreases
Kzz=0
Obs (SABER) & model (ROSE) of CO2
SABER v 1.06
ROSE model
CO2: 3 model cases for January-February
Cl and HCl (transport)
• With increasing diffusion, HCl increases near the mesopause while Cl decreasesKzz=0
long-lived hydrogen species
• With increasing diffusion, CH4 and H2O increase in the lower and middle mesosphere and H increases in the upper mesosphere
Kzz=0
atomic oxygen
percentage change with diffusion is small
the altitude of the rapid increase of O in the middle mesosphere moves downslightly with increased diffusion
Kzz=0
ozone (photochemistry)
• both day & night ozone change with diffusion, but the signs are opposite
• at night, lower O3 with higher Kzz is dominated by the impact of eddy diffusion on H
• during day, diffusion increases O3 in the middle mesosphere through the increase in O
• a valuable diagnostic since the response differs in day & night (easier to distinguish from other perturbations)
€
O3[ ] =k1 O[ ] O2[ ]M
J
€
O3[ ] =k1 O[ ] O2[ ]M
k2 H[ ]
Kzz=0
Kzz=0
night ozone: 3 model cases for January-February
high diffusion bringsup water, which leadsto ozone destruction
Obs & model of night ozone
SABER
ROSE model
OH and HO2
• daytime increase with increasing diffusion in the vicinity of the vmr maximum
• nighttime differences not monotonic with changes in Kzz
Kzz=0
Summary of useful chemicals:• Useful in middle mesosphere
– CH4
– H2O– O3, day & night
• Useful in upper mesosphere/mesopause– O3 night– CO– CO2
information about vertical structure of diffusion rate?
high diffusion better above the mesopause?
high diffusion worse at and below the mesopause?
Problems with this approach
• molecular diffusion– these tracers are also sensitive to molecular diffusion– how best to treat the two together?
• numerical formulation of diffusion• at present, models are being used to “validate” the upper
mesosphere chemical observations from SABER