isotopes at continental sites in Germany Long-term observations of atmospheric C0 and carbon

(1)

TELLUS

ISSN 0280-6509

Long-term observations of atmospheric C 0 ² and carbon isotopes at continental sites in Germany

B y I N G E B O R G L E V I N * , R O L F G R A U L1a n d N E I L B. A. T R I V E T T2, Institut fur Umweltphysik, University of Heidelberg, Germany; 1 Umweltbundesamt, Mefistelle Schauinsland, Germany; 2 Atmospheric

Environment Service, Toronto, Canada

(Manuscript received 20 December 1993; in final form 1 August 1994)

A B S T R A C T

A network for regional atmospheric C 02 observations had already been established in Germany by 1972, consisting of 5 stations with basically different characteristics: Westerland, a coastal station at the North Sea, 2 regional stations, Waldhof and Deuselbach, as well as 2 mountain stations, Brotjacklriegel at the eastern boarder of Germany and Schauinsland in the Black Forest. In addition to C Oz concentration observations, from 1977 onwards quasi-continuous

1 3C 02 and l 4C 02 measurements were performed on samples from the Schauinsland site, and for the short period 1985-1988, 1 4C 02 measurements were also made on Westerland samples. C 02

data selection based on wind velocity allows for an estimate of the representative continental COz level over Europe. The peak-to-peak amplitude of the seasonal cycles is between 12.1 ppmv (Schauinsland) and 17.6 ppmv (Waldhof). The phase of the seasonal cycles at the German sites is shifted if compared to maritime background sites with the concentration maxima occuring already between beginning of February and beginning of April, the minima in August. The long- term mean C 02 increase rate in the last 20 years at Westerland and Schauinsland is 1.49 and

1.48 ppmv y r- 1, respectively. The mean dnC of the seasonal source C 02 at Schauinsland is calculated from unselected dl3C and C 02 data to be — 25.1 °/oo- From the 14C observations in unselected C 02, we derive yearly mean fossil fuel contributions at Westerland of 4 ppmv, and at Schauinsland of only 2.5 ppmv. Based on the seasonality of the fossil fuel C 02 component at Schauinsland and on concurrently observed atmospheric 222Radon activities, we derive a seasonal amplitude of the fossil fuel C 02 source which is higher by a factor of 3 compared to emission estimates for Europe.

1. Introduction

Since the pioneering measurements o f a t m o spheric C 0² started b y Keeling in 1957/58 at M a u n a L o a , H a w a i i , and at the South P o l e (Keeling etal., 1989), n u m e r o u s sites for C 02

m o n i t o r i n g have been established all over the w o r l d ( T a n s , 1990). T h e principal a i m of this global network is to d o c u m e n t the a b u n d a n c e o f C O2 in the remote atmosphere, and to gain a better insight into the sources a n d sinks o f this important atmospheric greenhouse gas, b y using

* Corresponding author.

the spatial a n d temporal variations o f C 02 in c o m b i n a t i o n with atmospheric transport models.

O n e shortcoming o f the present global C 02

network is, however, the lack o f representative observations over the continents. T h e C 02

climatology in the remote maritime atmosphere seems to be well d o c u m e n t e d t o d a y , which allows for budgeting the large-scale C 02 fluxes between atmosphere a n d the surface reservoirs (oceans a n d terrestrial biosphere) to within ± 2 0 % . A better quantification of the uptake o f anthropogenic C 02 b y different reservoirs, however, needs an i m p r o v e d knowledge of the sources a n d sinks o n smaller scales, a n d particularly o f the diverse terrestrial biosphere. O n e a p p r o a c h to gain this

(2)

i n f o r m a t i o n uses observations over continents describing the C 02 climatology o n the continental or even the regional scale.

A G e r m a n network o f regional atmospheric C O 2 observations has been established since 1972 (see Fig. 1). T h i s n e t w o r k consists o f five stations with basically different characteristics: T h e n o r t h - ernmost station Westerland ( 5 5 ° N , 8 ° E , 8 m a.s.l.) is situated o n the island Sylt in the N o r t h Sea.

Particularly during northwesterly winds, the air collected here m o s t directly represents maritime c o n d i t i o n s in this latitudinal belt. T h e stations W a l d h o f ( 5 3 ° N , 11°E, 73 m a.s.l.), in a flat agricultural area o f northern G e r m a n y , a n d D e u s e l b a c h ( 5 0 ° N , 7 ° E , 480 m a.s.l.), in a hilly terrain at the western border o f G e r m a n y are strongly influenced b y regional biogenic a n d a n t h r o p o g e n i c sources or sinks. Their diurnal C 02

cycle is also strongly m o d u l a t e d b y the changing stability o f the atmospheric b o u n d a r y layer. T h e m o u n t a i n stations Brotjacklriegel ( 4 9 ° N , 13°E, 1 0 1 6 m a.s.l.) at the eastern border o f G e r m a n y , a n d Schauinsland ( 4 8 ° N , 8 ° E , 1205 m a.s.l.) in the Black Forest in south west G e r m a n y are less influenced b y c o n t a m i n a t i o n t h r o u g h local sources. In particular the Schauinsland station, during moderate or strong winds is s h o w n to be representative for m e a n atmospheric C 02 conditions over Western E u r o p e at this elevation o f a b o u t 1000m a b o v e m e a n sea level.

In addition to c o n t i n u o u s concentration observations, f r o m 1977 o n w a r d s quasi-continuous

Fig. J. Map of Central Europe with C 02 and carbon isotope measurement sites.

1 3C 02 and 1 4C 02 measurements were m a d e at the

Schauinsland site. F o r the period o f a b o u t t w o years 1 4C 02 measurements are also available for the station Westerland. These supplementary isotopic data provide insight into the nature o f the d o m i n a n t sources a n d s i n k s — b i o g e n i c or anthropogenic—influencing the station in question.

T h e c o m p l e x problems and the challenges associated with the interpretation o f highly variable continental C 02 and isotope records are discussed in this paper. T h e a i m is to develop a m e t h o d of interpretation a n d selection of continental C 0² data. Finally, representative continental data records for the Schauinsland a n d the Westerland station will be provided, and m a d e available for validation of regional as well as global scale carbon cycle models.

2. Techniques

2.1. CO2 concentration measurements

C o n t i n u o u s atmospheric C 02 measurements by N D I R were performed with U R A S - 2 ( H a r t m a n n

& B r a u n ) analysers at all five sites till 1982. F r o m there on, U l t r a m a t - 3 (Siemens) analysers were used. C 02- i n - N2 calibration gases provided by the Scipps Institution of O c e a n o g r a p h y ( S I O ) were used (1959 A d j . Index Scale) until 1991.

T h e standards in the concentration range o f 290-355 p p m v were recalibrated with C 02- i n - N2

calibration gases in the concentration range of 290-375 p p m v k i n d l y provided b y N O A A / C M D L . T h i s allowed us to transfer all our concentration values into the W M O 1985 C 02- i n - N2

scale. A linear transformation function was used:

c ( W M O 1985) = c(1959 adj. ind.) x 1.0404 - 1 2 . 8 6 [ p p m v ] . (1)

F r o m 1992 onwards, C 02- i n - s y n t h e t i c - a i r m i x - tures, calibrated b y S I O in 1990 ( W M O 1987 m o l e fraction scale) have been used in the network. All data obtained b y U R A S - 2 a n d U l t r a m a t - 3 instruments with C 02- i n - N2 gases h a d to be corrected for carrier-gas-effect (Griffith et al., 1982). T h e carrier-gas-effect o f one U R A S - 2 instrument which was still available in the network has been deter- m i n e d experimentally using C 02- i n - N2 ( W M O

(3)

ATMOSPHERIC C02 AND CARBON ISOTOPES 25

1985 scale) a n d C 02- i n - a i r ( W M O 1987 m o l e fraction scale) gases:

CURAS(WMO 1987) = c^{U R A S}( W M O 1985) + 4.39 [ p p m v ] . ( 2 ) T h e correction function (2) has then been applied to all records derived with U R A S - 2 instruments (the same correction was used at all 5 stations).

F o r the U l t r a m a t - 3 instruments individual carrier gas corrections have been applied o n the basis o f experimental determinations. T h e five analysers a n d o n e additional instrument o f the Institut fur U m w e l t p h y s i k fell into t w o groups s h o w i n g carrier gas effects differing b y 0.5 p p m v over the concentration range of 3 2 0 - 3 6 0 p p m v . F o r the stations Schauinsland, W e s t e r l a n d , a n d Brotjacklriegel the correction function Cu,tra^mat(WMO 1987) = cU I t r a m a t( W M O 1985)

x 1.0044 + 1.882 [ p p m v ] ( 3 ) was used. F o r D e u s e l b a c h a n d W a l d h o f the correction function

Cuitrama,(WMO 1987) = cu l t r a m a t( W M O 1985) x 0 . 9 9 9 + 4.47 [ p p m v ]

( 4 ) has been applied t o all measurements performed with this instrument f r o m 1982-1991. T h e errors associated with the scale a n d carrier gas corrections are estimated t o be less than + 0 . 5 p p m v . T h e overall precision o f daily m e a n concentration values is estimated to be + 1 p p m v .

2.2. Isotope measurements

C a r b o n isotope measurements have been performed o n large v o l u m e C 02 samples continuously collected by d y n a m i c quantitative a b s o r p t i o n o f atmospheric C 02 in carbonate free s o d i u m h y d r o x i d e solution. 1 3C analyses o f the C 02 are b y mass spectrometry, 1 4C analyses b y high precision p r o p o r t i o n a l c o u n t i n g o f the purified C 02

sample. A l l sampling a n d analysis procedures are described b y L e v i n et al. (1980) a n d Schoch et al.

(1980). <513C values are given relative t o the V - P D B standard ( H u t , 1987), the overall precision o f a single analysis is typically + 0.1 ° /o a. A1 4C data

are given relative to N B S oxalic acid activity corrected for decay (Stuiver & P o l a c h , 1977), the precision o f a single A1 4C measurement is typically

+ 5°/

-!- / OO •

3. Results and discussion

3.1. Diurnal concentration variations

T h e C 02 c l i m a t o l o g y o f the 5 stations can be derived f r o m their short term ( d i u r n a l ) C 0² c o n - centration variability, a n d f r o m their absolute concentration level c o m p a r e d to b a c k g r o u n d c o n - ditions. Fig. 2 shows the typical m e a n diurnal cycles at all 5 stations for J a n u a r y , A p r i l , J u l y , a n d September, calculated f r o m unselected h o u r l y m e a n values during the period o f 1973-1980. T h e straight solid lines represent the m o n t h l y m e a n values calculated for the same period f r o m the selected Schauinsland data, assumed t o represent the respective b a c k g r o u n d concentration level over Western E u r o p e . D u r i n g the winter m o n t h s ( c . f , Fig. 2a), w e observe almost n o diurnal cycle at a n y o f the stations. T h e t w o m o u n t a i n sites s h o w a m e a n concentration level w h i c h is o n l y several p p m v higher than the respective b a c k - g r o u n d level. T h e m e a n concentration increases t o w a r d s l o w elevation sites, a n d is highest at W a l d h o f , p r o b a b l y due to frequent strong g r o u n d level inversions at that site.

D u r i n g spring, a n d m o r e p r o n o u n c e d during the s u m m e r m o n t h s , the stations W a l d h o f a n d D e u s e l b a c h experience large diurnal cycles with m e a n night time concentrations rising u p t o 50 p p m v a b o v e the assumed b a c k g r o u n d level.

T h e d a y t i m e values at these sites are o n l y slightly lower than the b a c k g r o u n d concentration. T h i s b e h a v i o u r is a consequence o f the large diurnal variation o f vertical m i x i n g in the continental b o u n d a r y layer, particularly during s u m m e r , in c o m b i n a t i o n with m a x i m u m C 02 fluxes f r o m soil a n d plant respiration. O b v i o u s l y , the signal of the g r o u n d level plant assimilation sink during d a y t i m e is strongly w e a k e n e d b y enhanced vertical mixing. A t the m o u n t a i n sites Schauinsland a n d Brotjacklriegel the b e h a v i o u r is just opposite:

night time concentrations are close t o the b a c k - g r o u n d level whereas the d a y t i m e concentration values, due to the influence o f local vegetation o n the upslope winds, are depleted in their C 02

concentration. T h i s b e h a v i o u r o f the diurnal

(4)

C O 2 cycle is a c o m m o n p h e n o m e n o n generally observed at m o u n t a i n sites (Schmitt et al., 1988).

T h e diurnal C 02 cycles at the coastal station Westerland s h o w amplitudes which are between those observed at the regional sites W a l d h o f a n d Deuselbach a n d those at the t w o m o u n t a i n sites. Here we have basically t w o meteorological regimes, maritime or continental. T h e maritime air masses s h o w b a c k g r o u n d concentration levels whereas the continental air masses typically show similar diurnal amplitudes as observed at the t w o regional sites.

F r o m this C 0² c l i m a t o l o g y it seems o b v i o u s that only the t w o m o u n t a i n sites Schauinsland and Brotjacklriegel are potential candidates for select- ing C O 2 concentration data representative for the large scale continental C 02 level over Western Europe. Westerland also experiences b a c k g r o u n d concentrations w h e n the w i n d is off the N o r t h Sea, w h i c h are c o m p a r a b l e t o observations over the N o r t h Atlantic.

3.2. Long-term CO2 records: data selection

A s could already be inferred f r o m the m e a n diurnal cycles, due to the influence f r o m continental sources and sinks the unselected m o n t h l y m e a n C 02 concentrations at all 5 stations deviate c o n - siderably f r o m b a c k g r o u n d conditions. D u r i n g high w i n d speeds, a n d thus strong atmospheric mixing conditions, the observations even at regional sites become m o r e representative of a larger area. A s a first step towards m o r e representative data records, the raw hourly values have been selected for high w i n d speeds. T h e typical w i n d velocity distribution differs very m u c h from station to station; individual criteria have thus been used at the 5 sites (see T a b l e 1). H o u r l y C 02

values were assumed as representative for a site when the w i n d velocity was higher than the long term m e a n seasonal value at the station in the summer or in the winter half-year, respectively.

D a i l y m e a n values a n d standard deviations (1<T) have then been calculated f r o m the selected hourly

3 9 0

3 8 0

3 7 0

| 3 6 0 Q_

Q . 3 5 0

O 3401 O

330^-

3 2 0

310,

J a n u a r y Diurnal Cycle 1 9 7 3 - 8 0 ( a )

o o Q Q Q Brotjacklriegel S c h a u i n s l a n d t - n - n Waldhof

Hour 20 2 2 24 20 22 24

Hour

10 12 14 16

Hour 20 2 2 2 4

October Diurnal Cycle 1 9 7 3 - 8 0 (d)

Westerland Brotfacklriegel S c h a u i n s l a n d Deuselbach Waldhof

10 12 14

Hour 20 22 2 4

Fig. 2. Mean diurnal cycle of unselected atmospheric C 02 at the 5 German sites calculated for the period of 1973-1980 for January (a), April (b), July (c), and October (d). The straight solid lines represent the respective back- ground level over Europe determined from the selected Schauinsland data (compare Fig. 4a).

(5)

ATMOSPHERIC C 02 AND CARBON ISOTOPES 2 7

T a b l e 1. Selection criteria for "background data" and parameters of the selected C02 records from the German stations

Minimum velocity Peak to peak Seasonal Cycle

[ m s % of "background" amplitude max. min.

Station summer winter observations [ p p m v ] [ day No. ] [ day No. ]

Waldhof 2.0 2.5 11.1 17.6 31 217

Deuselbach 4.0 5.0 13.1 15.7 46 229

Brotjacklriegel 2.5 3.0 16.2 12.5 71 229

Schauinsland 2.5 3.5 17.1 12.1 90 239

Westerland 7.5 7.5 17.6 16.5 57 239

means if m o r e than six h o u r l y values were selected for a day. T h e calculated daily m e a n value was finally accepted as representative when the standard deviation (\a) o f the selected hourly values of the d a y was less than + 1 . 5 p p m v .

H a r m o n i c curves have then been fitted through the selected daily m e a n values (Trivett, 1989). T h e mean seasonal cycles s h o w significant differences in amplitude and phase at the individual sites (Fig. 3): the seasonal amplitude is largest (15.7 to 17.6 p p m v , T a b l e 1) at the l o w elevation sites W a l d h o f , Westerland, and Deuselbach. Smallest

yearly amplitudes are observed at the m o u n t a i n stations Schauinsland a n d Brotjacklriegel which is in accordance to the generally observed decrease o f the seasonal C 02 amplitude in the troposphere with height in the northern hemisphere ( B o l i n &

Bischof, 1970; N a k a z a w a et al., 1993). T h e c o n - centration decrease at W a l d h o f and Deuselbach occurs already in February, at the northernmost station Westerland only at the beginning o f April.

T h e dates of m a x i m u m a n d m i n i m u m concentrations are listed in T a b l e 1. A l s o the absolute concentration levels at the five stations differ

Mean Seasonal Cycle (Fit)

J a n Mar May J u l S e p Nov J a n Mar May MONTH

Fig. 3. Mean seasonal cycles determined from selected daily values of atmospheric C 02 over the period of 1972-1990 at the 5 German sites. Peak-to-peak amplitudes are largest at the regional stations Waldhof and Deuselbach and decrease towards high elevation sites (e.g., Schauinsland). Whereas the C 02 minimum is generally observed in August, the maximum concentration at the regional sites occurs already in February/March, at the mountain station

Schauinsland only in April (compare Table 1).

(6)

considerably: the yearly m e a n concentrations at Brotjacklriegel, Deuselbach a n d W a l d h o f are higher b y a p p r o x i m a t e l y 3 p p m v , 4 p p m v , a n d 4.5 p p m v respectively w h e n c o m p a r e d t o Schauins- land a n d Westerland. T h e m e a n yearly increase rates are similar at all sites a n d lie between 1.38 p p m v y r- 1 a n d 1.49 p p m v y r_ 1 for the 20 years period o f 1972 to 1992.

N o complete record o f clean air C 02 data in c o m p a r a b l e latitudes ( a b o u t 4 5 ° N to 5 5 ° N ) is available for the measurement period; between a b o u t 50° N a n d 75 ° N , however, the yearly m e a n meridional concentration gradient is usually less than 1 p p m v (e.g., T a n s e t a l . , 1989). F o r the last 10 years, the individual yearly m e a n concentrations observed at the coastal site Westerland a n d at the Schauinsland station (Figs. 4a, b ) have

therefore been c o m p a r e d with data from (clean air) sites in higher latitudes o f the northern hemisphere, namely B a r r o w ( 7 6 ° N , 1 1 9 ° W ) a n d M o u l d B a y ( 7 1 ° N , 1 5 7 ° W ) (see T a b l e 2). T h e agreement o f yearly m e a n concentrations at these four stations is usually better than 1 p p m v . T h i s is partly due to the fact that the data selection for high w i n d speeds results in an implicit selection o f predominately westerly winds (maritime air masses). Nevertheless, at Schauinsland, and even at Westerland there still remains a significant per- centage ( a b o u t 10 a n d 2 0 % , respectively) of

"continental d a t a " in the record.

3.3. 13C02 at the Schauinsland

A s a b y - p r o d u c t f r o m the 1 4C 02 analyses o f large v o l u m e C 0² samples, the stable isotope ratio

3 8 0 f

3 7 0 \

3 6 0 r

>

E

CL 3 5 0 r Q_

3 4 0 \ O O

3 3 0 r

3 2 0 \

3 1 0 ^

Schauinsland 1205 m a.s.l. (a)

'- s e l e c t e d f o r w i n d v e l o c i t y

( s u m m e r > 2 . 5 m / s , w i n t e r > 3 . 5 m / s )

L 8

«

£% j£*L r'% <" %'j \J^c)i'

1*1' Vl \/ •¥ *f »'• * 1 °o o o o o o No n b a s e l i n e d a t a Fitted t r e n d - H a r m o n i c fit

71 7 2 7 3 7 4 7 5 7 6 7 7 7 8 7 9 8 0 81 8 2 8 3 8 4 8 5 8 6 Y E A R

8 7 8 8 8 9 9 0 91 9 2 9 3

3 8 0

3 7 0

3 6 0 r~>"'

E Q. 3 5 0 Q_

3 4 0 O

3 3 0

3 2 0

3 1 0

Westerland 8 m a.s.l. (b)

s e l e c t e d f o r w i n d v e l o c i t y

( s u m m e r > 7 . 5 m / s , w i n t e r > 7 . 5 m / s )

B a s e l i n e d a t a Non b a s e l i n e d a t a Fitted t r e n d H a r m o n i c fit

71 7 2 7 3 7 4 7 5 7 6 7 7 7 8 7 9 8 0 81 8 2 8 3 8 4 8 5 8 6 87

Y E A R

89 9 0 91 9 2 93

Fig. 4. Long-term records of selected daily mean values for Schauinsland (a) and Westerland (b). The solid lines are harmonic fit curves through the selected values (Trivett, 1989). D a t a are considered as baseline data if they fall into the ICT range of all deviations from the curve, non-baseline data fall into the respective 2a range. The mean increase rates (dashed lines) have been calculated to 1.48 and 1.49 ppmv y r- 1 for Schauinsland and Westerland, respectively.

(7)

ATMOSPHERIC C 02 AND CARBON ISOTOPES 2 9

T a b l e 2. Yearly mean CO2 concentrations at the German stations Schauinsland and Westerland {selected) in comparison with sites from the NOAA/CMDL network^ (Conway et ah, 1990; WMO, 1992)

Mould Baya) Barrow a) Schauinsland Westerland

Year (76°N, 119°W) (71°N, 157°W) (48°N, 8 ° E ) (55°N, 8°E)

1982 342.4 342.7 343.4 343.3

1983 343.6 343.7 345.0 342.1

1984 345.6 345.3 345.1 345.8

1985 346.7 346.4 346.9 346.3

1986 348.6 348.6 348.1 346.8

1987 349.8 349.7 350.4 348.7

1988 353.5 353.4 353.2 352.1

1989 355.5 355.0 354.3 356.2

1990 356.0 356.0 354.9 354.6

1991 357.6 357.5 357.7 357.5

? > - 8 . 0 0

2.82 2.86 2.90 2.94 C02"1 [ p p m v " ' * 103]

Fig. 5. Monthly mean values of SnC (a) in atmospheric C 02 (b) at the Schauinsland site (unselected data), the dashed curves represent harmonic fits calculated through the data sets; the correlation of <513C and in vers C 02 concen- tration (c) (calculated from the detrended fit curves) leads to an isotopic signature of the mean source responsable for the observed Sl3C seasonality at Schauinsland of S13C = — 25.1°/00.

(8)

1 3C /1 2C in C 02 has been measured at Schauins- land during the period o f 1977-1992 (Fig. 5a).

A large seasonal cycle with a m e a n peak to peak amplitude o f <513C = 0.8°/Oo is observed which is strongly anti-correlated with the unselected C 02 concentrations (Fig. 5b). T h e m e a n dl3C is decreasing f r o m 1977 to 1992 b y a b o u t 0.02°/^{o o} y^r~\ with a larger decrease in the first part o f the record, a n d nearly n o trend in the second part. H a r m o n i c fits have been obtained for b o t h unselected C 02 a n d <513C m o n t h l y means.

W i t h the assumption that o n l y one source or sink is responsible for the observed C 0² seasonality, the isotopic c o m p o s i t i o n o f C 02 f r o m this apparent source (resp. the fractionation o f the sink) can be calculated f r o m the correlation o f the S^UC values a n d the inverse concentrations ( L e v i n , 1987). F r o m the detrended data a m e a n

<513C = — 25.1°/oo ° f the source is derived (Fig. 5c).

T h i s value represents the m e a n isotopic c o m p o s i - tion o f all sources, biogenic a n d anthropogenic, contributing to the C 02 variations at Schauins- land o n the regional as well as o n the continental a n d hemispheric scale. It is characteristic for exchange with terrestrial biota, a n d somewhat greater than the <513C o f fossil fuels ( - 2 7 ° /0 0 to

— 28°/0 0). It compares very well with values observed at b a c k g r o u n d sites (Keeling etal., 1989), and the source Sl3C derived f r o m recent observations o f the concentration and stable isotopes on flask samples collected at the Schauins- land during supposedly g o o d m i x i n g conditions (Neubert, unpublished data).

3.4. 14CO2 at the Schauinsland

T h e 1 4C /1 2C ratio, expressed as A1 4C , in unselected atmospheric C 02 at the Schauinsland station shows a steady a n d approximately exponential decrease f r o m 1977 until t o d a y with a time constant o f a b o u t T= 16 years (Fig. 6, histogram). O v e r l y i n g this trend is a seasonality with m i n i m u m values occurring during the winter half year. T h e decline is the consequence o f b o m b

1 4C still equilibrating with the w o r l d oceans and the biosphere, as well as an ongoing input of

14C-free fossil fuel C 02 into the atmosphere. T h e seasonality is m a i n l y attributed to a seasonally varying contribution o f fossil fuel C 02 at the measurement site being largest in the winter half year ( L e v i n etal., 1989).

T h e 1 4C b a c k g r o u n d level in m i d latitudes of the northern hemisphere can be derived from observa-

Fig. 6. Monthly mean A1 4C in atmospheric C 02 at Schauinsland (histogram) compared to values in background air over Europe (fitted curve). The seasonal cycle of the background fit has been derived from continuous l 4C 02 observa- tions at Jungfraujoch (1986-1991). The long-term trend of the background fit curve is the same as observed at the Schauinsland, but we added a constant value of A1 4C = +6°/oo (mean AI 4C difference between Schauinsland and Jungfraujoch in 1986-1991).

(9)

ATMOSPHERIC C 02 AND CARBON ISOTOPES 31

tions at the H i g h A l p i n e Research Station J u n g f r a u j o c h in the Swiss A l p s ( 4 7 ° N , 8 ° E , 3450 m a.s.l.; see Fig. 1). A t this site quasi-continuous 1 4C 02 samples have been measured f r o m 1986 o n w a r d s (Fig. 6, s m o o t h curve). A h a r m o n i c fit curve a n d the m e a n trend have been calculated from the J u n g f r a u j o c h data. In the period o f 1986 to 1991 the m e a n Schauinsland 1 4C 02 level was constantly lower b y a b o u t A^{I 4}C = 6 ° /^{0 0} if c o m - pared to Jungfraujoch. T h i s offset is attributed to a general pile u p o f fossil fuel C 02 over central Europe. T o derive the 1 4C 02 b a c k g r o u n d level in m i d latitudes of the northern hemisphere for the time period of 1977 to 1986 we used the trend curve calculated for the Schauinsland data and added the m e a n difference between J u n g f r a u j o c h and Schauinsland to this curve. T h e seasonal variation of the 1 4C 02 b a c k g r o u n d was derived f r o m the mean seasonality at J u n g f r a u j o c h f r o m 1986 to 1991 ( L e v i n et al., 1992).

F r o m the 1 4C depletion at Schauinsland relative to the b a c k g r o u n d curve derived f r o m the J u n g f r a u j o c h data we calculated m o n t h l y m e a n

14C-free fossil fuel contributions at the Schauins- land site according to Levin et al. (1989). A m e a n seasonal cycle for 1977-1992 was derived f r o m these m o n t h l y data (Fig. 7). D u r i n g winter the

Jan Mar May Jul Sep Nov Jan Mar May

MONTH

Fig. 7. Mean fossil fuel C 02 contribution derived from the respective negative AI 4C deviations from the back- ground fit (c.f., Figs. 6 and 8) at Schauinsland for the period of 1977-1992 (smooth curve) and the period of 1985-1988 (cycles) in comparison with Westerland data for the same time span (stars). The fossil fuel C 02

contribution during the summer months is similar at both sites. During the winter months, probably due to suppressed vertical mixing over the continent, the Schauinsland mountain station experiences much lower fossil fuel contributions than the coastal station Westerland.

fossil fuel c o m p o n e n t is a b o u t a factor of t w o higher than during summer. Part of this seasonality at the Schauinsland is due to a seasonality o f the fossil fuel source (see Fig. 9), but also the atmospheric mixing height is considerably larger during the summer than during the winter season.

T h e^{1 4}C d r a w d o w n during winters o f 1987/88, 1988/89 and 1989/90 is m u c h smaller than in the earlier and later years. T h i s can be explained b y the extraordinary w a r m E u r o p e a n winters coupled with a high frequency of maritime ( u n p o l l u t e d ) air masses reaching the Schauinsland station. T h i s special situation in the late 1980s is also manifested in l o w^{2 2 2}R a d o n concentrations observed at this site during these winter half years (Sartorius, I A R - Freiburg; see also Subsection 3.6).

3.5. 14C02 at Westerland

Q u a s i - c o n t i n u o u s observations o f 1 4C 02 at the coastal site Westerland in the period o f 1985-1988 s h o w a m u c h larger seasonality and also bigger deviations f r o m the J u n g f r a u j o c h 1 4C b a c k g r o u n d level than the Schauinsland data (Fig. 8). D u r i n g winter, the fossil fuel c o m p o n e n t at Westerland is, thus, a factor o f t w o higher than at the m o u n t a i n site Schauinsland whereas during the summer half year the fossil fuel contributions at b o t h sites are comparable (Fig. 7). A l t h o u g h the Schauinsland station is situated further inside the continental source area, its elevation a b o v e g r o u n d level leads to generally smaller c o n t a m i n a t i o n b y g r o u n d level sources, particularly during winter. T h i s behaviour is also confirmed b y the significantly smaller seasonal amplitude o f the selected C 02

Westerland

Westerland monthly m e a n s

Background fit ( J u n g f r a u j o c h extrapolated)

YEAR

Fig. 8. Monthly mean values of A1 4C in atmospheric C 02 at Westerland (histogram). The smooth curve represents background air over Europe, as in Fig. 6.

(10)

concentrations at Schauinsland if c o m p a r e d to Westerland ( c o m p a r e Fig. 3).

3.6. Determination of the fossil fuel source strength A tracer t o estimate the residence time of an air mass over the continent as well as the atmospheric m i x i n g conditions is the trace gas 2 2 2R a d o n ( D o r r et al., 1983; L e v i n , 1987). T h e radioactive n o b l e gas 2 2 2R a d o n emanates f r o m all continental surfaces rather h o m o g e n e o u s l y a n d constantly with time. W e used c o n t i n u o u s atmospheric 2 2 2R a d o n observations ( k i n d l y p r o v i d e d b y H . Sartorius, Institut fur A t m o s p h a r i s c h e R a d i o a k t i v i t a t , F r e i b u r g ) a n d the fossil fuel C 02 c o m p o n e n t derived f r o m 1 4C data at the Schauinsland station to estimate the source strength o f fossil fuel C 02 in the catchment area o f the Schauinsland site (e.g., G e r m a n y a n d S o u t h - W e s t E u r o p e ) . W e m a d e the simple a s s u m p t i o n that the 2 2 2R a d o n flux f r o m continental surfaces is h o m o g e n e o u s a n d constant with time, a n d that the flux f r o m ocean surfaces is negligible ( D o r r & M u n n i c h , 1990). T h e observed atmospheric 222 R a d o n activity ( cR n) then m a i n l y reflects the residence time o f the respective air mass over the continent divided b y the vertical m i x i n g height. W i t h the further assumption that the fossil fuel sources are similarly h o m o g e n e o u s l y distributed at g r o u n d level o n the continents, f r o m the k n o w n 2 2 2R a d o n flux density o f jR n = ( 5 3 ± 2 0 ) m B q m -2h -1 ( D o r r & M u n n i c h , 1990), a n d the fossil fuel c o m p o n e n t c[oss, the fossil fuel flux density y'foss is estimated according to:

7foss — 7Rn X cfoss/CRn- (5;

F o r the catchment area o f the Schauinsland site (i.e., G e r m a n y a n d S o u t h Western E u r o p e ) a mean flux density o f 2 m M o l e m ~2h_ 1 is calculated with a sinusoidal seasonal a m p l i t u d e o f + 3 0 % (Fig. 9). T h e m e a n flux density is considerably smaller than the 3 ^ m M o l e C 02 m ~2h_ 1

expected f r o m statistically derived emission data o f the relevant E u r o p e a n countries ( M a r l a n d , 1990).

Possibly, the J u n g f r a u j o c h station is still partly influenced b y the fossil fuel sources o n the E u r o p e a n continent, a n d the 1 4C 02 b a c k g r o u n d level was estimated t o o low. H o w e v e r , the m e a n

1 4C 02 level at the maritime b a c k g r o u n d station I z a n a , Tenerife, is o n l y higher b y a b o u t 2°/⁰⁰, suggesting an underestimation o f the fossil fuel c o n t a m i n a t i o n at Schauinsland b y 3 0 % at most.

^ 5.C

'E

E 2-0

S c h a u i n s l a n d ( 1 9 8 2 - 1 9 9 1 )

Jan Mar May Jul Sep Nov Jan Mar May MONTH

Fig. 9. Monthly mean fossil fuel C 02 flux densities (calculated from parallel 222Radon observations) in the catchment area of the Schauinsland station. The error bars show the variability of flux densities calculated in individual years.

A l s o the m e a n 2 2 2R a d o n flux density has large error margins o f + 3 0 - 5 0 % . W i t h i n these uncer- tainties the estimated m e a n fossil fuel flux density is in reasonable agreement with direct emission statistics.

A n i m p o r t a n t finding is, however, that the seasonality of the fossil fuel flux density from E u r o p e is a b o u t three times higher than the seasonality estimated f r o m emission statistics ( R o t t y , 1987). Rotty's estimate h a d been derived from sales statistics which m a y be considerably different t o the seasonality of the actual fossil fuel emission. A large seasonality o f the fossil fuel emis- sions is also needed to explain the seasonal cycle of

1 4C 02 observed at northern hemispheric b a c k - g r o u n d sites like Izana, Tenerife or Alert, C a n a d a ( L e v i n et al., 1992). If we attribute the 1 4C 02

seasonality at these sites m a i n l y to fossil fuel contributions, a significant part (at least 1 0 % ) o f the seasonal cycle in C 02 concentration in m i d to high latitudes of the northern hemisphere has to be attributed to this source a n d not to atmosphere- biosphere exchange, as has generally been assumed.

4. Summary and conclusions

T h e long-term records of atmospheric C 02 care- fully conducted at continental sites in G e r m a n y provide an i m p o r t a n t contribution to establishing the global atmospheric C 0² budget:

• there is clear observational evidence for differences in seasonal cycles driven by the continental sources a n d sinks;

(11)

ATMOSPHERIC C 02 AND CARBON ISOTOPES 33

• a c c o m p a n y i n g isotopic observations (1 3C a n d

1 4C) at sites subject to largely different pollution levels are shown to provide a quantitative iden- tification o f sources;

• the relative anthropogenic C 02 contribution can be furthermore related to respective fluxes b y the 2 2 2R a d o n approach.

D a t a selection based on meteorological p a r a m - eters have been s h o w n to provide representative continental records which n o w can be used to validate regional, continental, and also global scale C 02 models.

All data presented in this paper will be m a d e available to other investigators through the C a r b o n D i o x i d e I n f o r m a t i o n Analysis Center, P . O . B o x 2008, O a k Ridge, T N 37831-6335, U S A .

5. Acknowledgements

T h i s w o r k has been partly funded by the G e r m a n Umweltbundesamt, Berlin, under contract

N o . F E 104 02 627, a n d has been supported by the Bundesminister fur F o r s c h u n g u n d Technologie under contract N o . 0 7 K F T 1 1 , a n d via G K S S / I B in the framework o f the Science a n d T e c h n o l o g y Agreement between G e r m a n y and C a n a d a , K A N E N V 25. W e wish to thank the personnel at the J u n g f r a u j o c h , Schauinsland a n d Westerland stations for collecting the large v o l u m e C 02

samples for isotope analysis, a n d K . U h s e ( U B A - Pilotstation, O f f e n b a c h ) for re-evaluation of the continuous C 02 records f r o m the G e r m a n C 02

network. H. Sartorius, Bundesamt fur Strahlen- schutz, Institut fur Atmospharische Radioaktivitat, Freiburg, F R G , provided continuous^{2 2 2}R a d o n data from the Schauinsland site. T h e carbon cycle g r o u p at N O A A / C M D L , Boulder, U S A , kindly provided C 02- i n - N2 standards for re-calibration of our station standards. T h e staff o f the

Heidelberg R a d i o c a r b o n a n d M a s s Spectrometer Laboratories are kindly acknowledged for their careful w o r k d o i n g the n u m e r o u s isotopic analyses. W e thank R. Francey for helpful sugges- tions concerning this manuscript.

R E F E R E N C E S

Bolin, B. and Bischof, W. 1970. Variations of the carbon dioxide content of the atmosphere in the northern hemisphere. Tellus 12, 431^*42.

Conway, T. J., Tans, P. and Waterman, L. S. 1990.

Atmospheric C 02 at Barrow and Mould Bay. In:

Trends '90, a Compendium of data on global change (eds. T. A. Boden, P. Kanciruk and M. P. Farrell).

O R N L / C D I A C - 3 6 , Oak Ridge, T E , U S A , 54-55, 62-63.

Dorr, H., Kromer, B., Levin, I., Miinnich, K. O. and Volpp, H. J. 1983. C 02 and Radon-222 as tracers for atmospheric transport. J. Geophys. Res. 88, C2, 1309-1313.

Dorr, H. and Miinnich, K. O. 1990. 222Rn flux and soil air concentration profiles in West Germany. Soil 222Rn as tracer for gas transport in the unsaturated soil zone.

Tellus 42B, 20-28.

Griffith, D. W. T., Keeling, C. D „ Adams, J. A., Guenther, P. R. and Bacastow, R. P. 1982. Calcula- tions of carrier gas effects in non-dispersive infrared analyzers. II. Comparisons with experiment. Tellus 34, 385-397.

Hut, G. 1987. Stable isotope reference samples for geochemical and hydrological investigations. In:

Report to the Director General, I A E A , Vienna.

Keeling, C. D., Bacastow, R. B., Carter, A. F., Piper, S. C , Whorf, T. P., Heimann, M., Mook, W. G. and

Roeloffzen, H. 1989. A three-dimensional model of atmospheric C 02 transport based on observed winds:

1. Analysis of observational data. Geophysical Monograph 55, 165-236.

Levin, I., Miinnich, K. O. and Weiss, W. 1980. The effect of anthropogenic C 02 and l 4C sources on the distribu- tion of 14C in the atmosphere. Radiocarbon 22, 379-391.

Levin, I. 1987. Atmospheric C 02 in continental Europe, an alternative approach to clean air C 02 data. Tellus 39B, 21-28.

Levin, I., Schuchard, J., Kromer, B. and Miinnich, K. O.

1989. The continental European Suess-Effect. Radio- carbon 31, 431^40.

Levin, I., Bosinger, R., Bonani, G., Francey, R., Kromer, B., Miinnich, K. O., Suter, M., Trivett, N. B. A. and Wolfli, W. 1992. Radiocarbon in atmospheric carbon dioxide and methane: global distribution and trends.

In: Radiocarbon after four decades: an interdisciplinary perspective (Eds. R. E. Taylor, A. Long and R. Kra).

Springer-Verlag, New York.

Marland, G. 1990. C 02 emissions. In: Trends '90, a Com- pendium of data on global change (eds. T. A. Boden,

P. Kanciruk and M. P. Farrell). O R N L / C D I A C - 3 6 , Oak Ridge, TE, U S A , 92-133.

Nakazawa, T., Morimoto, S., Aoki, S. and Tanaka, M.

1993. Time and space variations of the carbon isotopic

Tellus 47B (1995), 1/2

(12)

ratio of tropospheric carbon dioxide over Japan. Tellus 45B, 258-274.

Rotty, R. 1987. Estimates of seasonal variation of fossil fuel C O 2 emissions. Tellus 39B, 184-202.

Schmitt, R., Schreiber, B. and Levin, I. 1988. Effects of long-range transport on atmospheric trace con- stituents at the baseline station Tenerife (Canary Islands). Atmos. Chem. 7, 335-351.

Schoch, H., Bruns, M., Miinnich, K. O. and Miinnich, M. 1980. A multi-counter system for high precision carbon-14 measurements. Radiocarbon 22, 442-447.

Stuiver, M. and Polach, H. A. 1977. Discussion: Report- ing of 14C data. Radiocarbon 19, 355-363.

Tans, P. P., Conway, T. J. and Nakazawa, T. 1989.

Latitudinal distribution of the sources and sinks of atmospheric carbon dioxide derived from surface

observations and an atmospheric transport model.

J. Geophys. Res. 94, D4, 5151-5172.

Tans, P. P. 1990 (Ed.). Report of the W M O meeting of experts on carbon dioxide concentration and isotopic measurement techniques. WMO Report No. 77.

Trivett, N. B. A. 1989. A comparison of seasonal cycles and trends in atmospheric C Oz concentration as deter- mined from robust and classical regression techniques.

Proceedings of the Third International Conference on Analysis and evaluation of atmospheric CO 2 data, present and past. W M O Report No. 59.

W M O 1992. Atmospheric C 02 concentrations from Barrow and Mould Bay. WMO WDCGG Data Report Part A (Carbon Dioxide), W D C G G No. 2, Published by Japan. Meteorol. Agency in Cooperation with W M O .