Abstract

The authors investigate the influence of El Niño on midtropospheric CO2 from the Atmospheric Infrared Sounder (AIRS) and the Model for Ozone and Related Chemical Tracers, version 2 (MOZART-2). AIRS midtropospheric CO2 data are used to study the temporal and spatial variability of CO2 in response to El Niño. CO2 differences between the central and western Pacific Ocean correlate well with the Southern Oscillation index. To reveal the temporal and spatial variability of the El Niño signal in the AIRS midtropospheric CO2, a multiple regression method is applied to the CO2 data from September 2002 to February 2011. There is more (less) midtropospheric CO2 in the central Pacific and less (more) midtropospheric CO2 in the western Pacific during El Niño (La Niña) events. Similar results are seen in the MOZART-2 convolved midtropospheric CO2, although the El Niño signal in the MOZART-2 is weaker than that in the AIRS data.

1. Introduction

Atmospheric CO2, an important greenhouse gas in the atmosphere, is increasing globally at a rate of approximately 2 ppm yr−1 mainly as a consequence of fossil fuel combustion (Keeling et al. 1995). In addition to the CO2 trend, atmospheric CO2 also exhibits strong seasonal cycles. Variations of CO2 seasonal cycle amplitudes are closely related to carbon exchange with the biosphere (Pearman and Hyson 1980, 1981; Cleveland et al. 1983; Bacastow et al. 1985; Keeling et al. 1996; Buermann et al. 2007). Atmospheric CO2 also demonstrates intraseasonal and interannual variabilities (Bacastow 1976; Enting 1987; Feely et al. 1987; Keeling and Revelle 1985; Keeling et al. 1995; Dargaville et al. 2000; Dettinger and Ghil 1998; Jiang et al. 2010). Combining satellite/in situ observations and model simulations, we found that there are Madden–Julian oscillation (MJO), semiannual oscillation (SAO), and tropospheric biennial oscillation (TBO) signals in midtropospheric CO2 (Li et al. 2010; Jiang et al. 2012; Wang et al. 2011). Using midtropospheric CO2 data from the Atmospheric Infrared Sounder (AIRS), Jiang et al. (2010) found that El Niño–Southern Oscillation (ENSO) can influence midtropospheric CO2 concentration as the result of a change in the Walker circulation (Julian and Chervin 1978). Midtropospheric CO2 is enhanced in the central Pacific Ocean and diminished in the western Pacific Ocean during El Niño (Jiang et al. 2010). In the high latitudes, midtropospheric CO2 concentration can be influenced by the strength of the polar vortex. In this paper, we investigate the temporal variability of midtropospheric CO2 from AIRS using a multiple regression method, and we compare the results with those from a chemistry-transport model.

2. Data and model

Mixing ratios of midtropospheric AIRS CO2 are retrieved using the vanishing partial derivative (VPD) method (Chahine et al. 2005, 2008). The maximum sensitivity of AIRS midtropospheric CO2 is between 500 and 300 hPa. AIRS midtropospheric CO2 is retrieved globally in the midtroposphere during day and night. AIRS, version 5, CO2 retrieval products are available at 2° × 2.5° (latitude by longitude) resolution from September 2002 to February 2011. Validation, by comparison to in situ aircraft measurements and retrievals from land-based upward-looking Fourier transform interferometers, demonstrated that AIRS CO2 is accurate to 1–2 ppm between latitudes 30°S and 80°N (Chahine et al. 2005, 2008). Midtropospheric CO2 retrieved via the VPD method captures the correct seasonal cycle and trend compared with those from the Comprehensive Observation Network for Trace Gases by Airliner (CONTRAIL) (Chahine et al. 2005).

A three-dimensional (3D) chemistry and transport model, the Model for Ozone and Related Chemical Tracers, version 2 (MOZART-2), was used in this paper to simulate the El Niño signal in midtropospheric CO2. MOZART-2 is driven by the European Centre for Medium-Range Weather Forecasts Interim (ECMWF-Interim) meteorological data. The horizontal resolution of MOZART-2 is 2.8° × 2.8° (latitude by longitude). There are 45 vertical levels extending up to approximately 50-km altitude (Horowitz et al. 2003). MOZART-2 is built on the framework of the Model of Atmospheric Transport and Chemistry (MATCH). MATCH includes representations of advection, convective transport, boundary layer mixing, and wet and dry deposition. The surface boundary condition for MOZART-2 is the climatological CO2 surface fluxes from biomass burning, fossil fuel emission, ocean, and biosphere used in Jiang et al. (2008). MOZART-2, driven by the ECMWF-Interim meteorological data and climatological CO2 surface fluxes, is used to investigate the influence of El Niño on midtropospheric CO2. To reveal if ECMWF-Interim simulates the ENSO signal well, we have calculated the Southern Oscillation index (SOI) from ECMWF-Interim Re-Analysis datasets by analyzing the standardized mean sea surface pressure differences between Darwin and Tahiti. The SOI derived from ECMWF-Interim correlates well with the standard SOI, which is defined by the sea surface pressure difference between Tahiti and Darwin. The correlation coefficient between two time series is 0.81. The corresponding significance level is 1%. The significance statistics for the correlation are generated by a Monte Carlo method (Press et al. 1992; Jiang et al. 2004). The ECMWF-Interim Re-Analysis captures the ENSO signal well at the surface.

3. Results and discussion

Before exploring the influence of El Niño on midtropospheric CO2, we first calculated the mean AIRS midtropospheric CO2 abundance from September 2002 to February 2011. Results for the mean AIRS midtropospheric CO2 are shown in Fig. 1. There is more midtropospheric CO2 over the western Pacific and less over the eastern Pacific. This is related to the redistribution of CO2 as a result of the Walker circulation. There is upwelling air over the western Pacific Ocean, which can bring high values of CO2 from the surface to the midtroposphere. Air is sinking over the eastern Pacific Ocean, which can bring low concentrations of CO2 from the high altitude to the midtroposphere.

Fig. 1.

Mean AIRS midtropospheric CO2 averaged from September 2002 to February 2011.

Fig. 1.

Mean AIRS midtropospheric CO2 averaged from September 2002 to February 2011.

Next, we investigated the temporal variations of the ENSO signal in midtropospheric CO2. We calculated the difference of CO2 between the central Pacific (18°S–18°N, 190°–240°E) and the western Pacific (18°S–18°N, 110°–160°E) areas. A linear trend was removed from the AIRS midtropospheric CO2 difference. The detrended AIRS midtropospheric CO2 difference between the central Pacific and the western Pacific is shown by the solid line in Fig. 2a. A linear trend was removed from the standard SOI. The detrended and inverted SOI is shown in Fig. 2a by the dashed line. When there is an El Niño (La Niña) event, the SOI is negative (positive). The CO2 difference (central Pacific − western Pacific) is positive (negative) for El Niño (La Niña) episodes. It suggests that there is more (less) midtropospheric CO2 over the central Pacific than over the western Pacific during El Niño (La Niña) episodes. As shown in Fig. 2a, the detrended AIRS CO2 difference correlates well with the inverted and detrended SOI. The correlation coefficient between the detrended AIRS CO2 difference and the inverted and detrended SOI is 0.61. The corresponding significance level is 1%. To investigate the interannual variability between the two time series, we applied a low-pass filter to the two time series. The low-pass filter is constructed to keep only signals with periods longer than 15 months. The low-pass filtered CO2 difference and low-pass filtered and inverted SOI are shown in Fig. 2b. The correlation coefficient between the two low-pass-filtered time series is 0.94 (1%).

Fig. 2.

(a) Differences in the detrended AIRS midtropospheric CO2 between the central Pacific (18°S–18°N, 190°–240°E) and the western Pacific (18°S–18°N, 110°–160°E) (solid line), and the inverted and detrended SOI (dashed line). Correlation coefficient between the two time series is 0.62 (1% significance level). (b) As in (a), but for low-pass-filtered data. Correlation coefficient between two low-pass-filtered time series is 0.94 (1% significance level).

Fig. 2.

(a) Differences in the detrended AIRS midtropospheric CO2 between the central Pacific (18°S–18°N, 190°–240°E) and the western Pacific (18°S–18°N, 110°–160°E) (solid line), and the inverted and detrended SOI (dashed line). Correlation coefficient between the two time series is 0.62 (1% significance level). (b) As in (a), but for low-pass-filtered data. Correlation coefficient between two low-pass-filtered time series is 0.94 (1% significance level).

To better investigate the temporal and spatial variability of the AIRS midtropospheric CO2 in the tropical region, we applied the multiple regression method to the AIRS midtropospheric CO2 data. We decomposed AIRS midtropospheric CO2 concentrations X at each location using the following empirical model:

 
formula

where is time; is the half length of the time period; and , , and are the first, second, and third Legendre polynomials. The coefficients , , , and are the mean value, the trend, the acceleration in the trend, and the coefficient for , respectively. We added the third Legendre function to better fit the datasets. Seasonal and semiannual cycles are represented by the harmonic functions ( and are the amplitudes of the annual cycle, and and are the amplitudes of the semiannual cycle); is the regression coefficient for the ENSO signal in the midtropospheric CO2; and is the inversed, detrended, and normalized SOI index, which is shown in Fig. 3a and was used for regressing the coefficients of ENSO signal in the AIRS midtropospheric CO2. The standard deviation for the inversed and normalized SOI is 1. Pagano et al. (2011) discussed the midtropospheric CO2 seasonal cycle from AIRS. We have studied the amplitude and mechanism for the semiannual cycle of the midtropospheric CO2 in another paper (Jiang et al. 2012). In this paper, we will mainly focus on investigating the influence of ENSO on midtropospheric CO2.

Fig. 3.

(a) Inversed, detrended, and normalized SOI. (b) Regression map [coefficient B in Eq. (1)] of the ENSO signal in the AIRS midtropospheric CO2 in the tropics. Central Pacific (18°S–18°N, 190°–240°E) and western Pacific (18°S–18°N, 110°–160°E) areas used in Fig. 2 are highlighted by dashed boxes in Fig. 3b.

Fig. 3.

(a) Inversed, detrended, and normalized SOI. (b) Regression map [coefficient B in Eq. (1)] of the ENSO signal in the AIRS midtropospheric CO2 in the tropics. Central Pacific (18°S–18°N, 190°–240°E) and western Pacific (18°S–18°N, 110°–160°E) areas used in Fig. 2 are highlighted by dashed boxes in Fig. 3b.

Regression coefficients for the ENSO signal in the AIRS midtropospheric CO2 are shown in Fig. 3b. The multiplication of positive (negative) values in Fig. 3a and regression coefficients in Fig. 3b represents the El Niño (La Niño) signal in the AIRS midtropospheric CO2. There are positive (negative) CO2 anomalies in the central Pacific and negative (positive) CO2 anomalies in the equatorial western Pacific during El Niño (La Niña) events. The CO2 anomaly is about 0.5 ppm in the central Pacific and −0.5 ppm in the western Pacific during El Niño episodes. During strong El Niño cases (e.g., February 2005 and February 2010), the CO2 amplitude is about 1 to 2 ppm in the central Pacific and −2 to −1 ppm in the western Pacific. During a strong La Niña case (e.g., February 2008), the CO2 amplitude is about −1 ppm in the central Pacific and 1 ppm in the western Pacific, which is consistent with the results obtained in Jiang et al. (2010).

To investigate how well the model could simulate the ENSO signal in the midtropospheric CO2, we convolved MOZART-2 CO2 vertical profiles with the AIRS midtropospheric CO2 weighting function. MOZART-2 convolved midtropospheric CO2 differences between the central Pacific and the western Pacific regions were calculated. Detrended MOZART-2 CO2 differences are shown in Fig. 4a. Detrended model CO2 differences correlate well with the inverted and detrended SOI index. The correlation coefficient between the two time series is 0.48 (1%). A low-pass filter was applied to the detrended MOZART-2 midtropospheric CO2 difference and the inverted and detrended SOI (Fig. 4b). The correlation coefficient between two low-pass-filtered time series is 0.65 (1%).

Fig. 4.

(a) Differences in the detrended MOZART-2 midtropospheric CO2 between the central Pacific (18°S–18°N, 190°–240°E) and the western Pacific (18°S–18°N, 110°–160°E) (solid line), and the inverted and detrended SOI (dashed line). Correlation coefficient between the two time series is 0.48 (1% significance level). (b) As in (a), but for low-pass-filtered data. Correlation coefficient between the two low-pass-filtered time series is 0.65 (1% significance level).

Fig. 4.

(a) Differences in the detrended MOZART-2 midtropospheric CO2 between the central Pacific (18°S–18°N, 190°–240°E) and the western Pacific (18°S–18°N, 110°–160°E) (solid line), and the inverted and detrended SOI (dashed line). Correlation coefficient between the two time series is 0.48 (1% significance level). (b) As in (a), but for low-pass-filtered data. Correlation coefficient between the two low-pass-filtered time series is 0.65 (1% significance level).

We used the SOI to separate MOZART-2 detrended and deseasonalized CO2 into two groups. When the SOI was 1.5 standard deviations below (above) the mean value, we considered it to be an El Niño (La Niña) month. MOZART-2 detrended and deseasonalized CO2 data averaged for 13 El Niño months are shown in Fig. 5a. We also have overlain the vertical velocity in Fig. 5a. During El Niño months, there is rising air over the central Pacific Ocean as shown by the dotted white contours in Fig. 5a. As a result, the surface high CO2 can be lifted into the midtroposphere over the central Pacific region during El Niño months. A low concentration of midtropospheric CO2 is seen in the western Pacific Ocean; however, the low CO2 appears in the subtropical area instead of the tropical area as seen in the AIRS midtropospheric CO2 (Jiang et al. 2010), which might be related to the relatively weak vertical velocity and relatively strong northward winds over the western Pacific Ocean in the ECMWF-Interim Re-Analysis data. In Fig. 5b, MOZART-2 detrended and deseasonalized CO2 data for 14 La Niña months suggest that lower CO2 has been transported from high altitude to the midtroposphere over the central Pacific Ocean. High CO2 is seen over the western Pacific Ocean; however, the position of the high CO2 shifts a little bit northward, which might be related to the relatively strong northward winds in the model. Figure 5c presents the MOZART-2 midtropospheric CO2 differences between the El Niño and La Niña months. The MOZART-2 midtropospheric CO2 differences (El Niño − La Niña) are about 1 ppm over the central Pacific and −0.7 ppm over the western Pacific. These are consistent with changes in the Walker circulation during El Niño and La Niña months. A Student’s t test was used to calculate the statistical significance of the MOZART-2 CO2 concentration differences during El Niño and La Niña months. The CO2 differences between El Niño and La Niña months were statistically significant when t was larger than a certain value t0. CO2 differences with significance levels less than 5% are highlighted by blue areas in Fig. 5d.

Fig. 5.

(a) MOZART-2 detrended and deseasonalized CO2 (color) and vertical velocities (white contours) averaged for 13 El Niño months. (b) As in (a), but for 14 La Niña months. (c) MOZART-2 CO2 differences and vertical velocity differences (white contours) between El Niño and La Niña months. (d) MOZART-2 CO2 differences within a 5% significance level are highlighted in blue. Solid (dotted) white contours refer to sinking (rising) air.

Fig. 5.

(a) MOZART-2 detrended and deseasonalized CO2 (color) and vertical velocities (white contours) averaged for 13 El Niño months. (b) As in (a), but for 14 La Niña months. (c) MOZART-2 CO2 differences and vertical velocity differences (white contours) between El Niño and La Niña months. (d) MOZART-2 CO2 differences within a 5% significance level are highlighted in blue. Solid (dotted) white contours refer to sinking (rising) air.

We applied the multiple regression method to the MOZART-2 midtropospheric CO2. The ENSO component in the MOZART-2 midtropospheric CO2 is shown in Fig. 6b. There is more (less) midtropospheric CO2 over the central (western) Pacific Ocean during El Niño months. The amplitude of the ENSO signal in the MOZART-2 midtropospheric CO2 is about half of the ENSO amplitude in the AIRS midtropospheric CO2. In addition, the spatial pattern of the ENSO signal in MOZART-2 is different than that from AIRS midtropospheric CO2 over the western Pacific Ocean. Differences between MOZART-2 and AIRS midtropospheric CO2 might be related to the climatological surface CO2 emission and transport used in MOZART-2. In one of our previous studies (Jiang et al. 2008), we found that convection in the 3D models is likely too weak in the boreal winter and spring. This will lead to underestimation of midtropospheric CO2 in the models. Similar results were also reported by Yang et al. (2007) when they compared column-averaged dry molar mixing ratios of CO2 at Park Falls with TransCom simulations. In addition, the strong northward winds in ECMWF-Interim Re-Analysis data might lead to the poor simulation of CO2 in the western Pacific Ocean. Previous studies suggested that the ocean flux decreases during El Niño events (Feely et al. 1987, 1999). In a sensitivity study, we reduced the CO2 ocean flux by 20%, and found that the midtropospheric CO2 decreases by 0.02 ppm. The MOZART-2 CO2 results might be improved in the future with better surface emission inventories and transport fields.

Fig. 6.

(a) Inversed, detrended, and normalized SOI. (b) Regression map of the ENSO signal in the MOZART-2 midtropospheric CO2 in the tropics.

Fig. 6.

(a) Inversed, detrended, and normalized SOI. (b) Regression map of the ENSO signal in the MOZART-2 midtropospheric CO2 in the tropics.

4. Conclusions

AIRS midtropospheric CO2 retrievals have been used to investigate the interannual variability of CO2 in the tropics. Detrended AIRS midtropospheric CO2 differences between the central and western Pacific correlate well with the inverted and detrended SOI. There is more (less) CO2 in the central Pacific and less (more) CO2 in the western Pacific for El Niño (La Niña) events. The multiple regression method was also applied to the AIRS midtropospheric CO2 in the tropical region. During El Niño episodes, there is more (less) CO2 in the central (western) Pacific as a result of changes in the Walker circulation. A similar signal was also seen in the MOZART-2 midtropospheric CO2, although the amplitude and spatial pattern in the model was a little different compared with that in the AIRS midtropospheric CO2. These results reveal temporal and spatial variability of midtropospheric CO2 as a response to ENSO. The results may be helpful to modelers who wish to better simulate ENSO signals in the middle troposphere and better constrain vertical transport in the chemistry-transport models. An improved model can be used to better simulate the influence of ENSO on tracers, such as CO2, CO, O3, and H2O.

Acknowledgments

We especially acknowledge Moustafa Chahine, Alexander Ruzmaikin, and Mimi Gerstell, who gave helpful suggestions on this research. XJ was supported by JPL Grant G99694. YLY was supported by the JPL OCO-2 project. Part of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration.

REFERENCES

REFERENCES
Bacastow
,
R. B.
,
1976
:
Modulation of atmospheric carbon dioxide by the Southern Oscillation
.
Nature
,
261
,
116
118
.
Bacastow
,
R. B.
,
C. D.
Keeling
, and
T. P.
Whorf
,
1985
:
Seasonal amplitude increase in atmospheric CO2 concentration at Mauna Loa, Hawaii, 1959-1982
.
J. Geophys. Res.
,
90
,
10 529
10 540
.
Buermann
,
W.
,
B.
Lintner
,
C.
Koven
,
A.
Angert
,
J. E.
Pinzon
,
C. J.
Tucker
, and
I.
Fung
,
2007
:
The changing carbon cycle at the Mauna Loa Observatory
.
Proc. Natl. Acad. Sci. USA
,
104
,
4249
4254
.
Chahine
,
M.
,
C.
Barnet
,
E. T.
Olsen
,
L.
Chen
, and
E.
Maddy
,
2005
:
On the determination of atmospheric minor gases by the method of vanishing partial derivatives with application to CO2
.
Geophys. Res. Lett.
,
32
,
L22803
,
doi:10.1029/2005GL024165
.
Chahine
,
M.
, and
Coauthors
,
2008
:
Satellite remote sounding of mid-tropospheric CO2
.
Geophys. Res. Lett.
,
35
,
L17807
,
doi:10.1029/2008GL035022
.
Cleveland
,
M. S.
,
A. E.
Freeny
, and
T. E.
Graedel
,
1983
:
The seasonal component of atmospheric CO2: Information from new approaches to the decomposition of seasonal time series
.
J. Geophys. Res.
,
88
(
C15
),
10 934
10 946
.
Dargaville
,
R. J.
,
R. M.
Law
, and
F.
Pribac
,
2000
:
Implications of interannual variability in atmospheric circulation on modeled CO2 concentrations and source estimates
.
Global Biogeochem. Cycles
,
14
,
931
943
.
Dettinger
,
M. D.
, and
M.
Ghil
,
1998
:
Seasonal and interannual variations of atmospheric CO2 and climate
.
Tellus
,
50B
,
1
24
.
Enting
,
I. G.
,
1987
:
The interannual variation in the seasonal cycle of carbon dioxide concentration at Mauna Loa
.
J. Geophys. Res.
,
92
(
D5
),
5497
5504
.
Feely
,
R. A.
,
R. H.
Gammon
,
B. A.
Taft
,
P. E.
Pullen
,
L. S.
Waterman
,
T. J.
Conway
,
J. F.
Gendron
, and
D. P.
Wisegarver
,
1987
:
Distribution of chemical tracers in the eastern equatorial Pacific during and after the 1982/1983 ENSO event
.
J. Geophys. Res.
,
92
(
C6
),
6545
6558
.
Feely
,
R. A.
,
R.
Wanninkhof
,
T.
Takahashi
, and
P.
Tans
,
1999
:
Influence of El Niño on the equatorial Pacific contribution to atmospheric CO2 accumulation
.
Nature
,
398
,
597
601
.
Horowitz
,
L. W.
, and
Coauthors
,
2003
:
A global simulation of tropospheric ozone and related tracers: Description and evaluation of MOZART, version 2
.
J. Geophys. Res.
,
108
,
4784
,
doi:10.1029/2002JD002853
.
Jiang
,
X.
,
C. D.
Camp
,
R.
Shia
,
D.
Noone
,
C.
Walker
, and
Y.
Yung
,
2004
:
Quasi-biennial oscillation and quasi-biennial oscillation-annual beat in the tropical total column ozone: A two-dimensional model simulation
.
J. Geophys. Res.
,
109
,
D16305
,
doi:10.1029/2003JD004377
.
Jiang
,
X.
,
Q.
Li
,
M.
Liang
,
R. L.
Shia
,
M. T.
Chahine
,
E. T.
Olsen
,
L. L.
Chen
, and
Y. L.
Yung
,
2008
:
Simulation of upper troposphere CO2 from chemistry and transport models
.
Global Biogeochem. Cycles
,
22
,
GB4025
,
doi:10.1029/2007GB003049
.
Jiang
,
X.
,
M. T.
Chahine
,
E. T.
Olsen
,
L.
Chen
, and
Y. L.
Yung
,
2010
:
Interannual variability of mid-tropospheric CO2 from Atmospheric Infrared Sounder
.
Geophys. Res. Lett.
,
37
,
L13801
,
doi:10.1029/2010GL042823
.
Jiang
,
X.
,
M. T.
Chahine
,
Q.
Li
,
M.
Liang
,
E. T.
Olsen
,
L. L.
Chen
,
J.
Wang
, and
Y. L.
Yung
,
2012
: CO2 semiannual oscillation in the middle troposphere and at the surface. Global Biogeochem. Cycles, 26, GB3006, doi:10.1029/2011GB004118.
Julian
,
P. R.
, and
R. M.
Chervin
,
1978
:
A study of the Southern Oscillation and Walker circulation phenomenon
.
Mon. Wea. Rev.
,
106
,
1433
1451
.
Keeling
,
C. D.
, and
R.
Revelle
,
1985
:
Effects of ENSO on the atmospheric content of CO2
.
Meteoritics
,
20
,
437
450
.
Keeling
,
C. D.
,
T. P.
Whorf
,
M.
Wahlen
, and
J.
Vanderplicht
,
1995
:
Interannual extremes in the rate of rise of atmospheric carbon dioxide since 1980
.
Nature
,
375
,
666
670
.
Keeling
,
C. D.
,
J. F. S.
Chin
, and
T. P.
Whorf
,
1996
:
Increased activity of northern hemispheric vegetation inferred from atmospheric CO2 measurements
.
Nature
,
382
,
146
149
.
Li
,
K.
,
B.
Tian
,
D. E.
Waliser
, and
Y. L.
Yung
,
2010
:
Tropical mid-tropospheric CO2 variability driven by the Madden–Julian oscillation
.
Proc. Natl. Acad. Sci. USA
,
107
,
19 171
19 175
,
doi:10.1073/pnas.1008222107
.
Pagano
,
T. S.
,
E. T.
Olsen
,
M. T.
Chahine
,
A.
Ruzmaikin
,
H.
Nguyen
, and
X.
Jiang
,
2011
: Monthly representations of mid-tropospheric carbon dioxide from the Atmospheric Infrared Sounder. Imaging Spectrometry XVI, S. S. Shen and P. E. Lewis, Eds., International Society for Optical Engineering (SPIE Proceedings, Vol. 8158), doi:10.1117/12.894960.
Pearman
,
G. I.
, and
P.
Hyson
,
1980
:
Activities of the global biosphere as reflected in atmospheric CO2 records
.
J. Geophys. Res.
,
85
(
C8
),
4468
4474
.
Pearman
,
G. I.
, and
P.
Hyson
,
1981
:
The annual variation of atmospheric CO2 concentration observed in the northern hemisphere
.
J. Geophys. Res.
,
86
(
C10
),
9839
9843
.
Press
,
W.
,
S.
Teukolsky
,
W.
Vetterling
, and
B.
Flannery
,
1992
: Numerical Recipes in Fortran 77: The Art of Scientific Computing. 2nd ed. Cambridge University Press, 933 pp.
Wang
,
J.
, and
Coauthors
,
2011
:
The influence of tropospheric biennial oscillation on mid-tropospheric CO2
.
Geophys. Res. Lett.
,
38
,
L20805
,
doi:10.1029/2011GL049288
.
Yang
,
Z.
,
R. A.
Washenfelder
,
G.
Keppel-Aleks
,
N. Y.
Krakauer
,
J. T.
Randerson
,
P. P.
Tans
,
C.
Sweeney
, and
P. O.
Wennberg
,
2007
:
New constraints on Northern Hemisphere growing season net flux
.
Geophys. Res. Lett.
,
34
,
L12807
,
doi:10.1029/2007GL029742
.