The seasonal variation of column abundance of atmospheric CH4 and precipitable water derived from ground-based IR solar spectra

Infrared Physics & Technology 41 (2000) 313±319

www.elsevier.com/locate/infrared

The seasonal variation of column abundance of atmospheric CH4 and precipitable water derived from ground-based IR solar spectra Heli Wei * Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, P.O. Box 1125, Hefei, Anhui 230031, People's Republic of China Received 7 June 2000

Abstract Infrared (IR) solar spectra on clear days were measured automatically by an IR solar spectrometer (ISS), capable of 0.4 cmÿ1 resolution, developed by us. A line-by-line computation method was used to calculate theoretical atmospheric absorption. In the wavelength range of 3.410±3.438 lm, the absorption is mainly due to atmospheric methane and water vapor. Column atmospheric methane was retrieved from the recorded IR solar spectra. The seasonal variation of column atmospheric methane in Hefei was obtained from the measuremental data of nearly 18 months since April 1997, and the seasonal variation was found to be similar to that of background data. The cycles of precipitable water were also obtained simultaneously, and the observed precipitable water coincided with that of a radiosonde. The instruments, principles of measurement and some of the results are introduced, and the results are also discussed brie¯y in this paper. Ó 2000 Elsevier Science B.V. All rights reserved. Keywords: Column abundance of methane; Seasonal variation, precipitable water; IR solar spectra

1. Introduction Methane is an important atmospheric trace gas. Its lifetime is about ten years in the atmosphere. The variation of its density plays an important role in atmospheric chemistry and climate. Methane in the troposphere is also an important greenhouse gas. Its contribution to radiative forcing of the atmosphere is estimated to be about 20% of all the greenhouse e€ects [1], next only to CO2 . Methane

*

Tel.: +86-551-5591675; fax: +86-551-5591572. E-mail address: [email protected] (H. Wei).

also takes part in the photochemical reaction with the production of atmospheric ozone. The concentration of atmospheric methane has risen continuously since the industrial revolution, and it has more than doubled since 1800 AD. The growth rate of atmospheric methane was as high as more than 1% yearÿ1 in 1980s [2,3]. However, it was found that the growth rate of methane slowed at recent times, and the abnormal phenomenon of dramatic decrease of the growth rate of methane in 1992 was also found by some researchers [4]. Up to now, it is not clear why and how the sources and sinks of atmospheric methane vary, and so, we cannot predict the variation of atmospheric

1350-4495/00/$ - see front matter Ó 2000 Elsevier Science B.V. All rights reserved. PII: S 1 3 5 0 - 4 4 9 5 ( 0 0 ) 0 0 0 4 9 - 9

314

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

methane accurately. Atmospheric methane has been of intense international interest, and several methods have been used to measure the density of atmospheric methane [2±5,7]. Conventionally, atmospheric methane is measured by air sampling. Clear air at a remote site that is considered as far from the source of methane is sampled weekly with a ¯ask and carried back to the laboratory. Methane measurement is made by a gas chromatograph (GC) with ¯ame ionization detection. The accurate methane mixing ratio is obtained by comparing with the standard calibration gas. By analyzing the air sample, the secular variation of background atmospheric methane is got. CH4 mixing ratio in background troposphere was well documented, but there is no systematic data about the column amounts of CH4 even on one site, and so we developed an e€ective method to remote sense the total vertical column atmospheric CH4 to observe the seasonal variation and its trend. The vertical column abundance of atmospheric methane in Hefei district was observed with an IR solar spectrometer system (ISS) developed by us. The measurements were carried out in Hefei (117.1614°E, 31.904°N, 0.046 km above sea level), which is located in the southeast region of China, at the plain towards the end of Yangzi river; the

main agricultural plants of the surrounding place are rice. Rice ®elds are an important source of atmospheric methane. The Huainan and Huaibei coal mines are located to the northwest of Hefei, and some of the methane is leaked into the atmosphere during the coal exploitation. So, the observation site represents a region that is a€ected by sources of methane. Since April 1997, we recorded the IR solar spectra continually and automatically on every clear day; the column abundances of atmospheric methane and precipitable water were deduced from the solar spectra in the band of 3.425±3.436 lm (2910±2920 cmÿ1 ). The seasonal variation of column atmospheric methane over Hefei district was obtained and it was similar to that of the background. 2. ISS and the experiment details The schematic diagram of ISS is shown in Fig. 1. ISS basically consists of three components: a heliostat to track the sun, a grating spectrometer, and a detecting system including an InSb IR detector cooled by liquid nitrogen, a lock-in ampli®er and a microcomputer used to acquire and process

Fig. 1. The schematic diagram of IR solar spectrometer system.

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

data and control the stepper-motor. A computercontrolled stepper-motor was used to drive the grating scanning in wavelength step by step. The  The focal length of scanning step is about 0.81 A. the grating spectrometer is f ˆ 1000 mm, the blaze wavelength of the grating is 3 lm, the grating constant is 300 lines mmÿ1 . By considering the slit width and scanning speed, the resolution is about 0.4 cmÿ1 , and the detectivity of InSb detector is D ˆ 1:11011 cm Hz1=2 Wÿ1 . A beam of solar radiation is brought into a laboratory by a solar tracking heliostat in conjunction with two 200 mm aluminum front-coated ¯at mirrors, focused by an o€-axis collimator on the slit of monochromator, then dispersed by a grating and imaged on the exit slit, received by the IR detector, and ampli®ed by a preampli®er and a lock-in ampli®er. The data is acquired by a PC through an A/D board. The computer gives the pulse signal to control stepper-motor in wavelength scanning through an I/O port. A 3±5 lm band ®lter was used in the exit slit to cut the short wavelength below 3 lm. The measurements of solar spectra were all performed automatically.

315

the gas to be measured, q0 …z† is the atmospheric density pro®le, R…z† is the initial vertical mixing ratio (VMR) pro®le, and q is the scale factor which we want to get; M…h† is the air mass and, h the solar zenith. I0 …m†, KI and sa are all smooth variables in the wavelength region of IR, sms can be omitted because it is too small compared to absorption. By ignoring the other gasesÕ absorption if they are too small within a narrow band, these variables could all be considered as constants and taken out from the convolution symbol, and Eq. (1) then becomes V …m† ˆ C/…m†   Z  exp ÿ

1 0

 r…m; z†qq0 …z†R…z†M…h† dz ; …2†

3. The principles of measurement

where C is a constant factor related to outer space solar spectra, aerosol extinction and the instruments response function. In order to decrease the e€ect of the factor C on the measurement result, a di€erential spectrum of relative ratio method was applied. The spectra are normalized with respect to the values at a certain reference wavelength m0 at which the absorption is minimal within the selected wavelength range, that is

3.1. The basic principles

R…m† ˆ

According to the equations of radiation propagation, considering the instruments slit function, the solar spectra signal received by ground-based instrument is given by  V …m† ˆ /…m†  I0 …m†KI sa …m†so …m†sms …m†  Z 1   exp ÿ r…m; z†qq0 …z†R…z†M…h† dz ;

where

0

…1†

where m is the wave number and /…m† the instruments slit function. The  symbol represents convolution, I0 …m†, the solar spectra out of the atmosphere, KI , the instrument's constant; sa , so and sms are aerosol extinction, absorption by other interfering gases, and molecular scattering, respectively. r…m; z† is the absorption coecient of

V …m† /…m†  s…m† ˆ ; V …m0 † /…m0 †  s…m0 †

 Z s…m† ˆ exp ÿ

1 0

…3†

 k…m; z†qq0 …z†R…z†M…h† dz :

A triangle function with half-width D ˆ 0:4 cmÿ1 was used as instruments slit function /…m†; M…h† could be calculated from local time, solar declination, local longitude and latitude; q0 …z† was taken from model atmosphere; r…m; z† was calculated with a line-by-line (LBL) algorithm. The Lorentz, Voigt and Doppler line shapes were applied for the height range of 0±25, 25±50, 50±100 km, respectively, and the spectroscopic parameters used here were taken from the 1992 HITRAN compilation [6]. A mean local atmospheric model of every month (averaged with radiosonde data in local meteorological station) was applied in the

316

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

calculated and observed ratio in solar spectra in the whole wavelength band of 3.427±3.436 lm, and then the total column densities of atmospheric methane and precipitable water were obtained: Z 1 WCH4 ˆ qCH4 RCH4 …z†q0 …z† dz 0 Z 1 RCH4 …z†q0 …z† dz; …5† ˆ qCH4 0

Z WH2 O ˆ Fig. 2. The calculated absorption of atmospheric methane and water vapor.

calculation. The atmospheric model included the pro®les of temperature, pressure, water vapor and atmospheric density, in addition to the real time surface meteorological parameters (such as temperature, pressure and relative humidity). Fig. 2 shows the calculated absorption by vertical column atmospheric methane and water vapor. The absorption by other gases such as O3 , NO2 and HCl is too small to be considered within the wavelength range. It can be seen that methane dominates in 3.417 and 3.428 lm bands, and water vapor in 3.413 and 3.434 lm bands. We select the interval of 3.425±3.436 lm (2910±2920 cmÿ1 ) to remote sense the column densities of atmospheric methane and precipitable water simultaneously. In fact, in the wavelength range of 3.425±3.436 lm, Eq. (3) should be written as R…m† ˆ

V …m† /…m†  ‰sH2 O …m†sCH4 …m†Š ˆ : V …m0 † /…m0 †  ‰sH2 O …m0 †sCH4 …m0 †Š

1

qH2 O RH2 O …z†q0 …z† dz Z 1 RH2 O …z†q0 …z† dz: ˆ qH2 O 0

0

…6†

4. Results 4.1. Comparison between the observed and calculated solar spectra Fig. 3 presents the observed solar spectra (solid line) at 14.01 of 16 May, 1998. For comparison, the calculated results (dashed line) using the retrieved column methane and precipitable water in the wavelength region are also presented. One can see that the agreement between the observed and calculated line shape attained is excellent. 4.2. The seasonal cycle of column methane Fig. 4 shows the daily mean vertical column density of CH4 observed from April 1997 to Sep-

…4†

In Eq. (4), the normalized VMR pro®le of CH4 RCH4 …z† was taken from US standard 1976. The simulated result showed that di€erent VMR pro®les did not alter the ®nal column abundance by more than 1%. As for the initial VMR of H2 O, we took the local monthly averaged humidity pro®les, added by the real time surface data. The scale factors of qCH4 and qH2 O in Eq. (4) could be derived simultaneously from the observed ratio of solar spectra R…k†, with the most favored method of adjusting qCH4 and qH2 O in the calculation to obtain an exact agreement between the

Fig. 3. Comparison of the solar spectra between the observed (±±) and the calculated (


).

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

317

Fig. 5. The monthly average column methane. Fig. 4. Daily mean column abundance of CH4 (the solid line denotes the running averaging over 30 days).

tember 1998. The solid line is the running average over 30 days. One can see that there is considerable ¯uctuation of column methane over time. The ratio of minimum to maximum was found to be 82%, the mean square deviation was 3.8%. The variation of atmospheric methane over time might be due to the emission of a local methane source or the transportation of methane from another place or atmospheric convection. The higher amount of column methane might be due to the strong emission of local methane source, and the smaller amount might come from the clear air transported from another place or sunk from the stratosphere. The characteristics of seasonal variation of column methane over Hefei region could be obtained from the running average data obtained for 30 days. The average column methane over Hefei was 0.64 mol mÿ2 , the maximum value (0.66 mol mÿ2 ) reached was at the end of December, which was cold winter, and the minimum data (0.62 mol mÿ2 ) of 1997 appeared at the end of July, and the value of 1998 (0.608 mol mÿ2 ) appeared in mid June, both were in the hot summer period. The mean variation amplitude (maximum±minimum) was 6.1% in 1997 and 8.4% in 1998. The monthly average column CH4 over the whole observed time (18 months) is shown in Fig. 5. One can clearly draw the conclusion from the ®gure that the seasonal variation of column CH4 over Hefei was the maximum column CH4 reached in winter, then it decreased gradually from winter

to summer until the minimum value appeared in summer. Many observations [3,4] with the method of air sampling in background atmosphere on the surface have shown that the mixing ratio of methane had a minimum value in summer and a maximum value in winter in the northern hemisphere. The methane seasonal variation amplitude at 32N°, which is nearly same latitude as our observation site, was about 5%. Compared with the result, the seasonal variation of column methane over Hefei was similar, yet the amplitude was a little larger. The seasonal variations of column methane are results of sources, sinks, atmospheric di€usion and transportation. The principal sources of atmospheric methane are believed to be related with the activities of biomass (such as the release from organic-rich sediments below shallow water and rice paddies, etc.), which become more active in the summer season primarily owing to the high air temperature, which enhances the CH4 emission. It is common knowledge that methane emission sources are stronger in summer. Removal of atmospheric methane is dominated by the reaction with the hydroxyl radical (OH). Destruction of CH4 by reaction with OH in the atmosphere is also enhanced in summer, since OH increases during this season. OH removes more methane in summer than in winter, which results in the higher methane level in winter for background atmosphere where no methane is emitted from its sources. But our observation site is located in a region with a strong methane source, and the seasonal variation trends

318

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

were similar as that of the background; this suggests that there may be some other source of CH4 in winter in the inland region, or the removal of methane by OH dominates the column methane abundance in the atmosphere. Hefei is located in the southeast region of China. During summer, the southeast current from the Paci®c Ocean carrying clear air with low density of atmospheric methane may contribute to the minimum of the curve in the summer season. A north or northwest current system from the mainland possibly bringing large amounts of methane, and the possibility of destruction of methane owing to the reaction with OH becoming slow, resulted in the biggest column methane appearance in winter. The results of other researchers in the region of strong emission of methane sources also showed that atmospheric methane was higher in winter. Shipham et al. [7] observed the atmospheric methane in Massachusetts, USA, where the observation site was subjected to the strong in¯uence of anthropogenic and natural CH4 sources. Although the observed mixing ratio of methane was in¯uenced by the local emission of CH4 sources, the diurnal variation could be as large as 10%, and the average seasonal variation of methane they observed was fairly similar to ours: minimum in summer, maximum in winter. The author also showed that the seasonal and secular trends were completely similar to that of the background. 4.3. Precipitable water Fig. 6 shows the comparison between the retrieved precipitable water from the solar spectra and the results of a radiosonde at the same site and nearly the same time. The di€erence between them is less than 10%. Fig. 7 shows the daily mean precipitable water deduced from ISS at the same time as methane measurements in the last section. One can see that water vapor increased from spring to summer, reached the maximum value at the end of July, and then decreased to a minimum in January, then the next cycles started to appear again. The precipitable water had a large seasonal variation. The maximum was 7.4 cm presented on 13 July 1998,

Fig. 6. Comparison of the observed precipitable water (WR ) to radiosonde (WS ).

Fig. 7. Daily mean precipitable water.

and the minimum of 0.29 cm occurred on 24 January of the same year, a variation of more than one order.

5. Conclusion The column abundances of atmospheric methane and precipitable water over Hefei were observed from the solar spectra in the wavelength region of 3.428 m (2916.7 cmÿ1 ). The seasonal variation of column atmospheric methane over Hefei district was obtained; it varied from 0.62 to 0.66 mol mÿ2 , the average value was 0.64 mol mÿ2 , it was lower in summer and higher in winter. The

H. Wei / Infrared Physics & Technology 41 (2000) 313±319

observed seasonal variation of column methane over Hefei was similar to that of data measured with air sampling at the same latitude in the background troposphere, but the variation amplitude was a little larger. The cycles of precipitable water was also obtained simultaneously, and the observed precipitable water was coincided with that of a radiosonde.

Acknowledgements We thank Ms. Lin Huiqin and Mr. Ma Chengsheng for providing the real time surface meterological data. We render our acknowledgment to Prof. Zhou Jun for reading the manuscript carefully.

319

References [1] E.D. Robert, J.C. Ralph, Future global warming from atmospheric trace gases, Nature 319 (9) (1986) 109±112. [2] M. Wang, et al., The observation of secular variation trends of atmospheric CH4 , Sci. Bull. 34 (9) (1989) 684±687 (in Chinese). [3] L.P. Steele, et al., The global distribution of methane in the troposphere, J. Atmos. Chem. 5 (1987) 125±171. [4] E.J. Dlugokencky, et al., The growth rate and distribution of atmospheric methane, J. Geophys. Res. 99 (D8) (1994) 17021. [5] R. Zander, P.H. Demoulin, Secular increase of vertical column abundance of methane derived from IR solar spectra recorded at the Jungfraujoch Station, J. Geophys. Res. 94 (D8) (1989) 11029±11039. [6] L.S. Rothman, et al., The HITRAN Molecular Database, 1991±1992 ed., JQSRT, vol. 48, 1992, pp. 469±507. [7] C.S. Mark, B.B. Karen, M.C. Patrick, et al., Atmospheric methane measurements in central New England: an analysis of the long-term trend and the seasonal and diurnal cycles, J. Geophys. Res. 103 (D9) (1998) 10621±10630.