-10
0
10
20
30
40
50
CO
2 (pp
m)
midday-10
0
10
20
30
40
50
60
CO
2 (pp
m)
gasoline natural gas biosphere morning
-10
0
10
20
30
40
50
7/1 10/1 1/1 4/1 7/1 10/1 1/1 4/1 7/1
CO
2 (pp
m)
date2013 2014
evening
2015
-0.2
0
0.2
0.4
0.6
0.8
frac
tion
in C
O2x
s midday-0.2
0
0.2
0.4
0.6
0.8
1
frac
tion
in C
O2x
s gasoline natural gas biosphere morning
-0.2
0
0.2
0.4
0.6
0.8
7/1 10/1 1/1 4/1 7/1 10/1 1/1 4/1 7/1
frac
tion
in C
O2x
s
date2013 2014
evening
2015
2015
-80
-60
-40
-20
0
7/1 10/1 1/1 4/1 7/1 10/1 1/1 4/1 7/1
δ13C
(‰)
date2013 2014
morning midday evening
-80
-60
-40
-20
0
7/1 10/1 1/1 4/1 7/1 10/1 1/1 4/1 7/1
δ13C
(‰)
date
morning midday evening
2013 2014 2015
R2 > 0.9
Diurnal and Seasonal Attribution of Anthropogenic CO2 Emissions Over Two Years in the Los Angeles Megacity���Sally Newman1, Xiaomei Xu2, Ying K. Hsu3, Arlyn Andrews4, and Yuk L. Yung1
1California Institute of Technology, Pasadena, CA; 2University of California, Irvine, CA; 3California Air Resources Board, Sacramento, CA; 4NOAA, Boulder, CO
IV. Conclusions • The rela2ve propor2ons of the different emission sources for CO2 may not be the same at different 2mes of day. At least
during the summer, this is not obviously due to diurnal changes in wind direc2on bringing emissions from different regions. • In the megacity of Los Angeles, the rela2ve contribu2on from natural gas combus2on is higher during the summers and lower
during the winter at midday and during the evenings, when orbi2ng satellites take measurements, but is higher in winter-‐spring than in summer during mornings. This is probably because of shiEing wind direc2ons seasonally.
• The biosphere’s contribu2on is higher during the cooler months than during the warmer months for all 2mes of day, as expected for this low mid-‐la2tude semi-‐arid region. However, the early mornings always have the highest biospheric contribu2ons of the day.
• Emissions from gasoline combus2on do not have a clear seasonal cycle for these two years, during midday and evening. • To improve this analysis, we must determine the emission ra2o (RCO/CO2ff) for different 2mes of day. We must also improve
modeling at 2mes other than mid-‐day, in order to quan2ta2vely account for transport of emissions.
References: Draxler, R.R. and Rolph, G.D. HYSPLIT (HYbrid Single-‐Par2cle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website (hYp://www.arl.noaa.gov/HYSPLIT.php). NOAA Air Resources Laboratory, College Park, MD. Eldering, A.; Boland, S. W.; Bowman, K. W.; Crisp, D.; Duren, R. M.; Fisher, J. B.; Frankenberg, C.; Gunson, M. R.; Menemenlis, D.; Miller, C. E.; Kaki, S., Eos AGU Abstract, A51J-‐04, 2012. Eldering, A., Kaki, S., Crisp, D., and Gunson, M.R., Eos AGU Abstract A21G-‐0134, 2013. Keeling, C. D., Geochim Cosmochim Acta, 13(4), 322–334, 1958. Keeling, C., Geochim Cosmochim Acta, 24, 277–298, 1961. Levin, I., Kromer, B., Schmidt, M. and Sartorius, H., Geophys Res LeY, 30(23), 2194, 2003. Miller, J. B., Lehman, S. J., Montzka, S. A., Sweeney, C., Miller, B. R., Karion, A., Wolak, C., Dlugokencky, E. J., Southon, J., Turnbull, J. C. and Tans, P. P., J. Geophys. Res, 117, 2012. Newman, S., Xu, X., Gurney, K. R., Hsu, Y. K., Li, K. F., Jiang, X., Keeling, R., Feng, S., O'Keefe, D., Patarasuk, R., Wong, K. W., Rao, P., Fischer, M. L., and Y. L. Yung, Y.,, Atmos Chem Phys Discuss, 15(20), 29591–29638, doi:10.5194/acpd-‐15-‐29591-‐2015, 2015. Turnbull, J., Miller, J., Lehman, S., Tans, P., Sparks, R. and Southon, J., Geophys Res LeY, 33(1), L01817, 2006. Turnbull, J. C., Karion, A., Fischer, M. L., Faloona, I., Guilderson, T., Lehman, S. J., Miller, B. R., Miller, J. B., Montzka, S., Sherwood, T., Saripalli, S., Sweeney, C. and Tans, P. P., Atmos Chem Phys, 11(2), 705–721, 2011. Vogel, F. R., Hammer, S., Steinhof, A., Kromer, B. and Levin, I., Tellus B, 62(5), 512–520, 2010. Xi, X., Natraj, V., Shia, R.-‐L., Luo, M., Zhang, Q., Newman, S., Sander, S., and Yung, Y., Atmos Meas Tech Discuss, 8(6), 5809–5846, doi:10.5194/amtd-‐8-‐5809-‐2015..
Contact informa2on: S. Newman: [email protected]
Figure 2. Examples of diurnal paYerns (leE column) and Keeling plots for morning, midday, and evening on individual days (right column). Keeling plot intercepts and correla2on coefficients are shown for each correla2on line. Ver2cal lines on the diurnal plots indicate the 2mes chosen between morning and midday and evening.
118°30 118°0 117°30
34°0’N
33°30’N
118°30 118°0 117°0’W117°30
33°30’N
34°0’N
San Gabriel Mountains
Santa Monica
Mountains
San GabrielValley
HillsPuente
Santa Ana
Mountains
Santa Catalina
Island
Palos Verdes
Peninsula
Los Angeles
Basin
Pasadena
117°0’W
I. Introduc2on In order to understand the role of anthropogenic greenhouse gases (GHGs), of which CO2 is the most abundant, in climate change we must understand their sources. Since ci2es produce >70% of the anthropogenic GHGs, large signals there can be used to study the details of emission paYerns. Trea2es are being signed to regulate emissions, and there must be verifica2on of compliance. The easiest way to get a broad picture of the distribu2on of CO2 and its changes through 2me is to collect measurements of total column concentra2ons from space. Satellite-‐borne instruments measure during midday, but is the distribu2on of emissions among the sources the same at all 2mes of day for all 2mes of year? Understanding the diurnal varia2on of the sources is important for comparison with boYom-‐up inventories, modeling, and results from new instruments such as OCO-‐3 (Eldering et al., 2012, 2013) and the Geosta2onary Carbon Process Inves2ga2on (Xi et al., 2014) which will see more of a diurnal cycle. We have shown that CO2, δ13C, and ∆14C can be used to dis2nguish among gasoline combus2on, natural gas combus2on, and biosphere contribu2ons to CO2 in the atmosphere (Newman et al., 2015) in a top-‐down approach. Here we present in situ data for different 2mes of day from the megacity of Los Angeles, CA, specifically from the Caltech campus in Pasadena (Figure 1), in order to understand this diurnal varia2on using this method with CO as a proxy for ∆14C, for morning, midday, and evening for June 2013 through May 2015.
Figure 1. Map showing the loca2on of the Pasadena site in the Los Angeles basin
II. Data and Methods • The data sets involved in this study include con2nuous measurements of CO2 and δ13C (Picarro Isotopic CO2 Analyzer) and CO (LGR N2O/CO EP Monitor) in Pasadena and
background measurements on Mt. Wilson (CO, CO2; flasks). ∆14C composi2ons are determined for flask samples collected on alternate aEernoons at 14:00 PST in Pasadena. • We use mul2ple mass balance calcula2ons on monthly averages to dis2nguish among the sources: fossil fuel combus2on (CO2ff), including gasoline (CO2pet) and natural gas
(CO2ng), and the biosphere (CO2bio). • The ∆14C values of the flask samples give the frac2on of CO2ff vs CO2bio in the total local enhancement over background (CO2xs) (Levin et al., 2003). • We use the CO2ff/CO2xs from the flask ∆14C measurements and COxs/CO2xs from the con2nuous measurements for 13:00-‐15:00 to determine the emission ra2o, RCO/CO2ff,
values for each month (Figure 4; e.g., Turnbull et al., 2006, 2011; Vogel et al., 2010; Miller et al., 2012). This assumes that RCOxs/CO2ff does not vary diurnally, but only seasonally.
• The y-‐intercept of correla2ons between δ13C and 1000/CO2 (Figure 2; Keeling plots; Keeling, 1958, 1961) provide the composi2on of the high-‐CO2, local enhancement of the CO2. The summary of the daily morning, midday, and evening intercepts are shown in Figure 3. Since these can be quite scaYered we filter the data by rejec2ng intercepts from correla2ons with R2 < 0.90. Monthly averages of the retained intercepts are shown in Figure 5.
• The stable isotopic composi2on of the carbon in the CO2 (δ13C, ‰ rela2ve to the standard VPDB) is then used to dis2nguish between gasoline and natural gas combus2on within CO2ff.
Figure 3. Top: all of the Keeling plot intercepts determined for Pasadena; morning values for correla2ons for hourly averages during 0:00-‐8:00, midday for 9:00-‐16:00, and evening for 17:00-‐23:00. BoYom: Keeling plot intercepts from the top panel that have been filtered by removing all results from correla2ons with R2 < 0.9.
Figure 5. Monthly averages of the Keeling plot intercepts for morning, midday, and evening on individual days. The isotopic signatures of natural gas and petroleum/biosphere are labeled to the right of the plot.
Figure 4. Monthly averages of COxs/CO2xs for morning, midday, and evening on individual days (colored dots) and of emission ra2os (RCO/CO2ff; +) from ∆14C data from midday flasks.
Figure 6. Source alloca2on as frac2on of CO2pet, CO2ng, and CO2bio in CO2xs (leE column) and as CO2 (ppm; right column) contributed to the local atmosphere.
III. Results • The monthly averages of the Keeling plot intercepts (Figure 5) show significantly different paYerns at different 2mes of day, especially between early mornings and later.
The lowest values of δ13C are centered on warmer months during midday and evening, whereas they are centered in the fall in the morning. This suggests that gasoline combus2on or the biosphere dominate the signal at different 2mes of day.
• When these δ13C values are combined with values of the frac2on of CO2 from fossil fuels in CO2xs, this dichotomy propagates through to different paYerns for the propor2ons of natural gas and gasoline combus2on of total CO2xs at different 2mes of day (Figure 6 leE). Although natural gas is the dominant source of local CO2 emissions during the summer middays, when the satellite-‐borne instruments observe, gasoline combus2on is the dominant source during summer mornings and evenings. Winter early morning emissions are mostly from the biosphere, very different from midday, when the biosphere contributes at most ~20%. As expected, the biosphere is a sink for CO2 during the spring and summer, although this sink is much less intense than elsewhere.
• The absolute contribu2ons in CO2 ppm (Figure 6 right) are very similar to the paYerns for the frac2ons for the morning period, but are higher for mornings and evenings than for middays, when the atmosphere is most well-‐mixed and therefore mixing ra2os are closest to the background values, resul2ng in lower CO2xs. The seasonal paYern for the evening absolute contribu2ons is similar for all three sources.
• We have shown that the paYern of higher fossil fuel emissions observed in Pasadena during summer is due to the annual wind direc2on paYern (Newman et al., 2015). However, the different paYerns for different 2mes of day during the summer cannot be due to the winds, since the wind direc2ons are the same at all 2me of day.
A21H-‐0248
-15
-14
-13
-12
-11
-10
-9
-8
δ13C
(‰)
6-6-13
morning: -11.9 ± 12.7 ppm; R2 = 0.01
midday: -29.7 ± 5.4 ppm; R2 = 0.67
evening: -15.0 ± 18.4 ppm; R2 = 0.02400
450
500
550
600
-15
-14
-13
-12
-11
-10
-9
-86-6-13
CO
2 (pp
m) δ
13C (‰
)
Diurnal Plots Keeling Plots
400
420
440
460
480
500
520
-15
-14
-13
-12
-11
-10
-9
-8
0 4 8 12 16 20 24
12-1-13
CO
2 (pp
m) δ
13C (‰
)
hour
400
450
500
550
-15
-14
-13
-12
-11
-10
-9
-8
8-15-13
CO
2 (pp
m) δ
13C (‰
)
400
450
500
550
-15
-14
-13
-12
-11
-10
-96-28-13
CO
2 (pp
m) δ
13C (‰
)
-15
-14
-13
-12
-11
-10
-9
-8
1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5
δ13C
(‰)
1000/CO2 (ppm-1)
12-1-13
morning: -28.8 ± 1.0 ppm; R2 = 0.98
midday: -18.3 ± 9.0 ppm; R2 = 0.18
evening: -30.5 ± 1.0 ppm; R2 = 0.99
-15
-14
-13
-12
-11
-10
-9
-8
δ13C
(‰)
morning: -24.5 ± 2.1 ppm; R2 = 0.87
midday: -32.4 ± 1.4 ppm; R2 = 0.94
evening: -39.0 ± 2.3 ppm; R2 = 0.96
8-15-13
-15
-14
-13
-12
-11
-10
-9
δ13C
(‰)
6-28-13
morning: -29.1 ± 0.5 ppm; R2 = 0.994
midday: -31.8 ± 0.5 ppm; R2 = 0.997
evening: -26.9 ± 0.4 ppm; R2 = 0.996
Figure 7. Wind back trajectories calculated using HYSPLIT (Draxler and Rolph, 2015) at different 2mes of day showing seasonal, but not significant diurnal, varia2ons in the source regions. The red dots indicate the loca2on of Pasadena.
Funded in part by California Air Resources Board Contract 13-‐329, and a grant from the W. M. Keck Ins2tute for Space Science.
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N January 2014 10 m ending in Pasadena at 0400 PST 24hr
0400#PST#
January#2014#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N January 2014 10 m ending in Pasadena at 1200 PST 24hr
1200#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N January 2014 10 m ending in Pasadena at 2000 PST 24hr
2000#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N July 2013 10 m ending in Pasadena at 0400 PST 24hr
0400#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N July 2013 10 m ending in Pasadena at 1200 PST 24hr
1200#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N July 2013 10 m ending in Pasadena at 2000 PST 24hr
2000#PST#
July#2013#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N March 2014 10 m ending in Pasadena at 0400 PST 24hr
0400#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N March 2014 10 m ending in Pasadena at 1200 PST 24hr
1200#PST#
121° W 120° W 119° W 118° W 117° W 116° W 115° W
33° N
34° N
35° N
36° N March 2014 10 m ending in Pasadena at 2000 PST 24hr
2000#PST#
March#2014#
0
0.005
0.01
0.015
7/1 10/1 1/1 4/1 7/1 10/1 1/1 4/1 7/1
COxs
/CO
2xs
date
morning midday evening RCOxs/CO2ff
2013 20142013 2015 2014
-40
-35
-30
-25
5/1 11/1 5/1 11/1 5/1
δ13C
(‰)
date
morning midday evening
2013 2014
gasoline combustion/biosphere
natural gascombustion
each data set fit to 2-3° sine curves
2015 2013 2014
