Figure 1. Overlay of smoothed CO and OH in the Northern Hemisphere. Biweekly CO mixing ratios were obtained by averaging CMDL network data. Biweekly OH concentrations were obtained from monthly concentrations provided by Spivakovsky et al. [1990]. Adapted from Novelli et al. [1998].

Introduction

Carbon monoxide (CO) is produced as a by-product of incomplete combustion of carbon-containing materials as well as due to photochemical conversion of atmospheric methane and other hydrocarbons. The main sources of CO are located at continental surfaces; the most important of them are human-related emissions (year-round) and biomass burning (dry seasons in tropics and warm part of the year in boreal areas). Photochemical conversion of hydrocarbons, e.g., methane, is an important source of CO, especially in summertime and the tropics. Reaction with hydroxyl radical is widely considered as the most significant sink for carbon monoxide (contributions of continental soil and ocean are very small for the CO global budget). At the same time, CO is usually the dominant reactant for OH, although other species in polluted or forested areas can be important [Brune, this issue; Wang, this issue]. As a result, OH distributions and trends should influence CO concentration dramatically. Dependence of OH concentration on the CO field is considerable as well. However, a detailed understanding of these relationships is not clear yet. The aim of this article is to consider available data on long-term variability of CO in relation to OH trends.

CO was discovered in the atmosphere through its absorption at the fundamental band near 4.6 mm [Migeotte, 1949] measured in the solar spectrum in late 1940s in the US and again at the International Scientific Station, Jungfraujoch, Switzerland. Systematic recording of CO bands in solar spectra started in Russia in 1970, at Kitt Peak Observatory in the US in 1976, and at the Jungfraujoch station in 1984. In situ measurements of this trace gas were made possible with the advent of the hot mercuric oxide technique [Robbins et al., 1968; Seiler, 1974] and gas chromatography (GC) [Khalil and Rasmussen, 1994; Novelli et al., 1998].

	Figure 2. Tropospheric mean mixing ratio over Zvenigorod, Russia, compared to free tropospheric mixing ratio, sampled by CMDL/NOAA at the altitude of 3475 m above sea level at Niwot Ridge, Colorado, US.

It should be noted that solar spectroscopy and gas sampling techniques monitor CO in different atmospheric domains, namely, in the entire tropospheric layer above the site and in the surface layer, respectively. Moreover, most GC monitoring is conducted at remote islands or at coastal stations so as to minimize continental influence (e.g., NOAA/CMDL network); therefore the data mainly characterize the boundary layer over oceans. Only a few mountain observatories sample free tropospheric air over continents (e.g., Niwot Ridge, Colorado). An advantage of the solar spectroscopic method is an opportunity to apply new retrieval algorithms and updated line parameters to spectra measured in the past (e.g., re-analysis of Migeotte's spectra of early 1950s by Zander et al., [1989]). In contrast, various in situ techniques require reference calibration mixtures and accuracy strongly depends on the consistency and validity of calibration procedures. A re-analysis of former results is usually questionable.

An important feature of atmospheric CO is its seasonal cycle (Figure 1, from Novelli et al. [1998]). CO gradually accumulates in the atmosphere during the dark period of the year with low OH concentrations between early fall and late spring. In the Northern Hemisphere it is rapidly consumed from April–June due to intensive OH removal. Figure 1 illustrates a strong seasonal dependence of CO concentration on [OH]; interannual [OH] variations also should influence CO. A decline of CO after the Pinatubo eruption is also visible in Figure 2, which shows mean monthly values for the layer 0-10 km in Zvenigorod plotted together with concurrent monthly mean local mixing ratios at 3.7 km altitude asl at the Niwot Ridge station in Colorado, US.

	Table 1. A review of available estimates for CO trend in % per year. Arial is for spectroscopic measurements in the total column above the site (H is the height above sea level) or columns in some layers in the atmosphere (see the Layer column). Italic is for sampling in the surface layer (which is assumed to represent the marine boundary layer). Trend for Jungfraujoch, 1984–1997 according to WMO [1999].

CO trend before and after the mid-1980s.

Both total column and in situ methods provided evidence in support of a positive CO trend between the 1950s and mid-1980s (Table 1). First, measurements of CO total column in Russia revealed a 1.3 % per year increase in January–October means between 1970 and 1982 [Dvoryashina et al., 1984; Dianov-Klokov and Yurganov, 1989] (see Figure 3). Spectra measured sporadically by M. Migeotte at the Jungfraujoch station in 1950-1951 were re-analyzed by Zander et al. [1989] and compared to similar measurements of 1985-1987. A positive trend of 0.85% per year was found. Khalil and Rasmussen [1994] derived a 0.8 % per year trend between 1981 and 1986. Analysis of air bubbles incorporated in Greenland glaciers revealed a 0.35 ppb per year (0.3 % per year) increase between 1850 and 1950 [Haan et al., 1996].

	Figure 3. Total column CO abundance over Zvenigorod (right scale) and corresponding mean tropospheric mixing ratio (left scale). Regression lines for 1970-1984 and for 1985-1997 are shown.

However, estimates for CO trends after 1985 changed dramatically. A global CO decline of –2.6% per year in the boundary layer was observed by Khalil and Rasmussen [1994] between 1987 and 1992. Even higher rates of global decline (–5.6%/yr) were reported by Novelli et al. [1994] for the boundary layer during a relatively short period of 1991–1993 (explained mostly by a perturbation from the Pinatubo eruption, see below).

Total column CO also ceased to grow. Practically stable CO (-0.08 ± 0.5 %/yr) was found over Russia between 1983 and 1993 [Yurganov et al., 1995]. Generally stable CO with irregular fluctuations was observed there after 1993 as well (see Figure 2). Total column CO in New Zealand also had no significant trend [0.37 ± 0.57%/yr, 1993-1997, Rinsland et al., 1998]. Small negative trends were observed at Switzerland [–0.18 ± 0.16%/yr, 1984 - 1995, Mahieu et al., 1997] and the US (–0.27 ± 0.17%/yr, 1978 -1997, Rinsland et al., 1998].

One can conclude that stabilization or decrease of global CO concentration after mid-1980s after a long period of growth seems to be a real phenomenon. The question is how large is this deceleration: 2-3%/yr (as found by GC in the boundary layer) or 1-1.5 %/yr (for the total column)?

Possible explanations for changing trend

Several attempts have been made to explain the observed CO deceleration. Yung et al. [1999] concluded that a significant part of CO (and CH₄) change in trend can be explained by a change in biomass burning. Changes in anthropogenic CO production and UV increase due to the thinning of the ozone layer were considered by Novelli et al. [1994] and Mahieu et al. [1997] as possible explanations of the phenomenon.

and to total ozone after major eruptions. There were two volcanic eruptions since the early 1980s that significantly perturbed stratospheric aerosol on a global scale: El Chichon in April 1982 and Mt. Pinatubo in June 1991. The problem is that both aerosol and ozone are changed after major eruptions; aerosol is increased and ozone is decreased. Therefore we must take both effects into account. Fortunately, ozone changes after the El Chichon eruption were small compared to AOD changes. So the slope of the CO-AOD regression line after the El Chichon event (Figure 4a) can be used for correction of CO values after the Pinatubo eruption. The slope of the CO (corrected)–O₃ regression line, 0.55 ppb CO/DU (Figure 4b) can be considered as an "empirical partial sensitivity" of CO to changes in total ozone. Further, this sensitivity can be applied to the long-term ozone trend after 1980: –1.3 DU/yr for the mid-latitude belt; a deceleration of CO of 0.7 ppb/yr (-0.6%/yr) can be expected due to ozone depletion. For total column measurements (Table 1), this change in the CO trend explains about half of the observed value (1.0 –1.5%/yr). The rest of the change may be attributed to lowering of man-made emissions in the NH.

	Figure 4. a) Correlation between CO and aerosol optical depth (AOD) at 550 nm for 17 months after the El Chichon eruption.	b) Correlation between monthly mean CO, reduced to AOD=0, using the slope from Figure 4a, and total ozone for 38 months just after the Pinatubo eruption.

These conclusions agree with the most recent CO data. The CO minimum for the last 20 years in 1997 (Figure 1) corresponds to very low ozone in the boreal spring of that year, while the record high CO in 1998 seems to coincide with normal ozone and abnormally intense boreal forest fires [Wotawa et al., 2000].

There may be several explanations for the discrepancy in decelerations observed by GC and spectroscopy. First, it may be a real phenomenon, caused by different chemistry of the boundary layer and total troposphere. Second, it may be an artifact due to incomplete coverage of the hemisphere by observing sites. Total column measurements are much less dependent on the site location and local contamination.

Conclusions

1. The CO trend changed in mid-1980s from ~1% increase per year increase to stability or decline. Boundary layer concentration was observed to diminish faster (up to a few % per year) than CO total column. The latter may be an artifact associated with a sparse network or limited periods of GC measurements.
2. About half of the change in CO long-term trend (~0.6% per year) can be explained by a growth of CO sink (OH concentration) caused by ozone decline and increase of UV.
3. The rest of the change in CO trend may be attributed to a change in CO sources (human-induced emissions, biomass burning or photochemical conversion of hydrocarbons).

References

1. Dianov-Klokov, V.I., and L.N. Yurganov, Spectroscopic measurements of atmospheric carbon monoxide and methane. 2: Seasonal variations and long-term trends, J. Atmos. Chem. 8, 153-164, 1989.
   
2. Dlugokencky, E. J., E.G. Dutton, P.C. Novelli, P.P. Tans, K.A. Masarie, K.O. Lantz, and S. Madronich, Changes in CH4 and CO growth rates after the eruption of Mt.Pinatubo and their link with changes in tropical tropospheric UV flux, Geoph. Res. Lett. 23, 2761-2764, 1996.
   
3. Dvoryashina E.V., V.I. Dianov-Klokov and L.N. Yurganov, On the variations of total column carbon monoxide during 1970-1982, Izv. Acad. Sci. USSR, Atmos. Oceanic Phys., Engl. Transl. 20, 27-33, 1984.
   
4. Haan, D., P. Martinerie, and D. Reynaud, Ice core data of atmospheric carbon monoxide over Antarctica and Greenland during the last 200 years, Geophys. Res. Lett., 23, 2235-2238, 1996.
   
5. Khalil, M.A.K., and R.A.Rasmussen, Global decrease in atmospheric carbon monoxide concentration. Nature, 370, 639-641, 1994.
   
6. Mahieu, E., Zander, R., Delbouille, L., Demoulin, P., Roland, G., and Servais, C. Observed Trends in Total Vertical Column Abundances of Atmospheric Gases from IR Solar Spectra Recorded at the Jungfraujoch. Journal of Atmosph. Chem.. 28, 227-243, 1997.
   
7. Migeotte, M., The fundamental band of carbon monoxide at 4.7 m in the solar spectrum Phys. Rev. 75, 1108-1109, 1949.
   
8. Novelli, P.C., K.A. Masarie, P.P.Tans, and P.M. Lang, Recent changes in atmospheric carbon monoxide, Science., 263,1587-1589, 1994.
   
9. Novelli, P.C., K.A. Masarie, and P.M. Lang, Distributions and recent changes in atmospheric carbon monoxide, J. Geophys. Res., 103, 19015-19033, 1998.
   
10. Robbins, R.C., Borg, K.M., and Robinson, E., Carbon monoxide in the atmosphere, J. Air Pollut. Contr. Assoc., 18, 106-109, 1968
    
11. Seiler W., The cycle of atmospheric CO, Tellus, 26, 116-135, 1974.
    
12. Spivakovsky C.M., R. Yevich, J.A.Logan, S.C.Wofsy, M.B.McElroy, and M.J.Prather, Tropospheric OH in a three-dimensional chemical tracer model: An assessment based on observations of CH3CCl3, J. Geophys. Res.95, 18441-18471, 1990
    
13. WMO, World Meteorological Organization, Scientific Assessment of Ozone Depletion: 1998, Global Ozone Research and Monitoring Project - Report No. 44, Geneva, P. 2.43, 1999.
    
14. Wotawa, G., P. Novelli, M. Trainer, and C. Granier, Boreal forest fires and their contribution to northern hemispheric summertime CO background concentrations during the 1990s, Nature, 2000 (submitted).
    
15. Yurganov L.N., E.I. Grechko, and A.V. Dzhola, Carbon monoxide and total ozone in Arctic and Antarctic regions: seasonal variations, long-term trends and relationships, The Sci. of the Total Environ. 160/161, 831-840, 1995.
    
16. Yurganov, L.N., Grechko, E.I., and Dzhola, A.V., Zvenigorod Carbon Monoxide Total Column Time Series: 27 years of Measurements. Chemosphere: Global Change Science, 1, 127-136, 1999
    
17. Zander, R., P. Demoulin, D.H. Ehhalt, U. Schmidt, and C.P. Rinsland, Secular increase of total vertical column abundance of carbon monoxide above central Europe since 1950, J. Geophys. Res. 94 , 11021-11028, 1989.