Comparisons of box model calculations and measurements of formaldehyde from the 1997 North Atlantic Regional Experiment
1,2
Frost,
3
Fried,
4
Lee,
3,5
Wert,
3
Henry,
6
Drummond,
7,8
Evans,
1
Fehsenfeld,
1
Goldan,
1,2
Holloway,
1,2
Hübler,
1
Jakoubek,
G. J.
A.
Y.-N.
B.
B.
J. R.
M. J.
F. C.
P. D.
J. S.
G.
R.
1,2
1,2,9
1
1
1
10
1
11
1,2,5
1,2
1
1,2,12
B. T. Jobson, K. Knapp,
W. C. Kuster, D. D. Parrish, J. Roberts, J. Rudolph, T. B. Ryerson, A. Stohl, C. Stroud,
D. T. Sueper, M. Trainer, and J. Williams
1Aeronomy Laboratory, National Oceanic and Atmospheric Administration, Boulder, Colorado.
2Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder.
3Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, Colorado.
4Environmental Chemistry Division, Brookhaven National Laboratory, Upton, New York.
5Dept. of Chemistry, University of Colorado, Boulder.
6Dept.of Physics, University of Toronto, Toronto, Canada.
7Centre for Atmospheric Science, Dept. of Chemistry, Cambridge University, United Kingdom.
8Now at Dept. of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts.
9Now at Dept.of Chemistry, Western State College, Gunnison, Colorado.
10Centre for Atmospheric Chemistry , Dept. of Chemistry, York Univ., York, Ontario, Canada.
11Lehrstuhl für Bioklimatologie und Immissionsforschung, Technische Universität München, Freising-Weihenstephan, Germany.
12Now at Abteilung Chemie der Atmosphare, Max Planck Institut für Chemie, Mainz, Germany.
Abstract
Possible Reasons for Measurement-Model Discrepancies
Formaldehyde (CH2O) measurements from two independent
instruments are compared with photochemical box model
calculations. The measurements were made on the NOAA P-3
aircraft as part of the 1997 North Atlantic Regional Experiment
(NARE 97). The data set considered here consists of air
masses sampled between 0 and 8 km over the North Atlantic
Ocean which do not show recent influence from emissions or
transport.
These air masses therefore should be in
photochemical steady state with respect to CH2O when
constrained by the other P-3 measurements, with methane
oxidation being the predominant source of CH2O. The two
instruments, which agree with each other on average though
with relatively low correlation, both show systematically higher
CH2O levels than the model. The median measured - modeled
CH2O difference is 0.13 or 0.18 ppbv (depending on the
instrument), or about a factor of two. Such large differences
cannot be accounted for by varying model input parameters
within their respective uncertainty ranges. After examining the
possible reasons for the model - measurement discrepancy, we
conclude that there are probably one or more additional
unknown sources of CH2O in the North Atlantic troposphere.
0.6
CH2O
CH2O+OH
Total
-5
0
5
4
-3 -1
Rate, 10 cm s
10
CH3O2+PA
O2
hn
OH,H2O
OH
OH
NO
CH3OH
OH
O2
CH3O
O2
BNL CH2O, ppbv
4
5
-3
10 15
-1
H2O
Diurnally averaged loss (left) and production (right) rates for
CH2O (top) and CH3O2 (bottom) calculated by the box
model. Median values are shown for the full data set of 86
points discussed in this work. PA = peroxy acetyl radical.
0.6
• HO2 + CH3O2 Æ CH2O + H2O + O2?
- up to 40% yield at low T and P
- unlikely for T and P of this data set
1
j(O3ÆO( D)+O2)
j(CH2OÆCHO+H)
j(CH2OÆH2+CO)
-0.4 0.0 0.4 0
s, %/%
BNL vs NCAR
50 100 100 101 102 103
2
2
u, %
(s*u) , %
Sensitivity in [CH2O] to (left), 1s uncertainty in (middle),
and contribution to the square of the [CH2O]
uncertainty from (right) the model input parameters,
where “s” = sensitivity and “u” = uncertainty. All data
are medians for the full data set of 86 points and
represent the input parameters to which the modeled
[CH2O] is most sensitive. The contribution of each
parameter to the square of the total [CH2O] uncertainty
is the square of the product of the parameter’s
sensitivity and uncertainty. “j scaling factor” is a
constant factor applied to all j values to simulate the
effects of clouds.
0.4
0.2
Measured and modeled [CH2O]
comparison scatter plots, showing
the 86 data points (circles), the
weighted least squares bivariate
fit (solid line), and the 1:1 line
(dotted line).
HO2
4) Model sources:
• Need 0.4 ppbv/day extra CH2O source
• Fairly constant from 0 - 8 km
k(OH+CH3OOH)
k(CH3O2+NO)
k(CH2O+OH)
0.0
0.0 0.2 0.4 0.6 0.8
NCAR CH2O, ppbv
CH2OH
O2
HO2
0
k(OH+CO)
k(HO2+CH3O2)
BNL vs Model
35
OH
North Atlantic Ocean
30
-80
-70
-60
-50
-40
Longitude, degrees
-30
Late summer / early fall frontal passages
• Transport of boundary layer air to upper troposphere
• Subsidence of stratospheric air
• Transport of polluted continental air to free troposphere
O2
O2,M
H
HO2
M
Brookhaven National Lab (BNL):
DNPH solution collection
Post-flight separation by HPLC
Quantitation by UV-VIS absorption
• 5 min integration
• 1s precision for 5 min sample: ±40-80 pptv
• 1s accuracy: 9-18%
• Completely independent CH2O determinations
• First extensive CH2O aircraft intercomparison
8
CH3O2
HO2
7
6
2
hn
1
H2
0
HO2
CO
OH
NCAR
20
15
Schematic representation of the methane oxidation cycle
with carbon-containing species shown in boxes. The
shaded boxes indicate those species with tropospheric
lifetimes longer than 1 second when [NO] < 1 ppbv.
Only photochemical reactions are shown; surface
emission and physical removal processes are not
included.
10
1.25*CH3OOH
5
0.6*k(OH+CO)
8
7
4
3
Assumption
• CH2O equilibrated with longer lived species
(lifetime = 6 - 9 hr under NARE 97 conditions)
2
Air mass selection
• Average data over 5 - 50 min legs with constant
conditions
• Eliminate air with recent emissions
• Total of 86 legs with coincident measurements
0
Input
• O3, NO, NOy, CO, NMHCs, organic nitrates, H2O, T, P
from NOAA WP-3
• CH4, H2, methanol, acetone from other observations
• j values from radiative transfer model (clear sky over
full diurnal cycles)
Output
• OH, HO2, RO2 , NO2, and CH2O (diurnal steady state)
Estimated accuracy
• 50 - 60% (1s) from rates and concentrations
BNL-Model
5
0.2
0.3 0.4 0.5
CH2O, ppbv
0.6
0.7
0.8
CH2O mixing ratios as a function of altitude. Gray open
symbols are the individual data points. Black solid
symbols joined by lines are the medians for the 0 - 2, 2 4, and 4 - 8 km altitude bins.
4
6 8
Conclusions
0.6*j(CH2OÆH2+CO)
BNL-NCAR
5
0.1
2
1
10
H2O Mixing Ratio, g/kg
1.5*k(OH+CH3OOH)
0.5*all j values
0.0
0.0
6 8
Effect of adding a 0.4 ppbv day-1 source of CH2O
on the levels of OH, HO2, and CH3O2 as a function
of altitude and H2O mixing ratio, compared with the
base model run.
0.6*j(CH2OÆCHO+H)
0
0
4
1.35*j(O3ÆO(1D)+O2)
10
1
10
0 10 20 30 -10
0
Percent Change with Added CH2O Source
0.7*k(OH+CH2O)
10
5
Base Model
0.5*k(HO2+CH3O2)
0
15
6
Altitude, km
Steady State Box Model
NCAR
BNL
Model
NCAR-Model
0
2
BNL
CO2
National Center for Atmospheric Research (NCAR):
100 m path length cell
20 s integration, 1 sample/min
1s precision for 5 min average: ±50-80 pptv
1s accuracy: 7%
OH
3
CH2O Measurements
Tunable IR diode laser absorption
• Unmeasured long lived NMHCs or oxygenates?
- need additional source equivalent to 0.7 - 2 ppmv CH4
or 5 - 10 ppbv CH3OH
4
H2O
CHO
hn
• CH3OH reaction on aerosols?
- no aerosol measurements on WP-3
- constant with altitude?
5
CH2O
40
•
•
•
•
Total
Rate, 10 cm s
CH3O2
NO2
0.8
NMHC sources
Total
CH3O2
H 2O
CH3OOH
45
M
O( D)+CH4
Number of points
Latitude, degrees
O2,M
1
3) Model sinks:
• Model and measured OH agree in other remote areas
• Model CH2O not very sensitive to j values
1
k(O( D)+M)
1
k(O( D)+H2O)
k(OH+CH4)
0.0
0.0 0.2 0.4 0.6 0.8
Model CH2O, ppbv
OH+CH3OOH
-15 -10 -5
[NO]
j scaling factor
0.4
CH4+OH
CH3O2+CH3O2
H2O
[CH3OOH]
0.2
CH3O2+HO2
NOAA WP-3 Aircraft Flight Tracks
50
0.6
CH3O2
CH3
HO2
0.8
2) Steady state assumption invalid:
• CH2O < 1 day
• Clean air only, no recent emissions
• No recent vertical transport in most air masses
Altitude, km
-10
OH
St. John's,
Newfoundland
[CO]
[CH4]
0.0
0.0 0.2 0.4 0.6 0.8
Model CH2O, ppbv
BNL CH2O, ppbv
CH2O+hnÆH+CHO
1997 North Atlantic Regional Experiment
(NARE 97)
55
[H2O]
NCAR vs Model
0.2
CH3O2+NO
CH3OOH+hn
OH+CH3OOH
CH3O2+CH3O2
OH+CH3OH
NMHC sources
Total
CH2O+hnÆH2+CO
CH4
OH
[O3]
0.4
CH3O2+NO
60
NCAR CH2O, ppbv
0.8
1) Instrument problems:
• Independent instruments, same trend
• Similar discrepancies seen at Cape Grim and in SONEX
0.1
0.2
0.3
median CH2O, ppbv
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
CH2O Difference, ppbv
Distribution of absolute differences between measured
and modeled [CH2O] for the full data set of 86 points.
The solid black vertical lines are the median differences
and the horizontal bars indicate 2 standard deviations.
Median CH2O mixing ratios for the full data set of 86
points are shown for the two measurements, the base
model, and a number of model sensitivity tests. In each
test, the indicated parameter was adjusted by the given
amount and the model was rerun for all points. These
adjustments represent the estimated 1s uncertainty
range for the parameter and were made in the direction
which would increase the modeled [CH2O].
• Steady state point model calculations of CH2O compared with
2 independent aircraft measurements in 86 air masses from
NARE 97.
• Model-measurement discrepancies within uncertainties.
• Measured CH2O systematically larger than model; median
measured/modeled ratio = 2.
• Model uncertainties could be reduced by
1) more laboratory studies:
HO2 + CH3O2 rate and products
CH2O photolysis quantum yields
k(OH + CO)
k(OH + CH3OOH)
2) additional aircraft measurements:
OH
HO2
other VOCs besides C2-C6 hydrocarbons (oxygenates)
• More CH2O measurements with better precision needed in
non-polluted environments.
• Additional sources of CH2O besides CH4 oxidation appear to
be present in non-polluted troposphere.