identifying the european fossil fuel plumes in the ... - Springer Link

Environmental Monitoring and Assessment (2006) 117: 387–409 DOI: 10.1007/s10661-006-0775-4

c Springer 2006

IDENTIFYING THE EUROPEAN FOSSIL FUEL PLUMES IN THE ATMOSPHERE OVER THE NORTHEAST ATLANTIC REGION THROUGH ISOTOPIC OBSERVATIONS AND NUMERICAL MODELLING C. GEELS1,∗ , J. H. CHRISTENSEN1 , A. W. HANSEN2 , J. HEINEMEIER3 , S. KIILSHOLM4 , N. W. LARSEN5 , S. E. LARSEN6 , T. PEDERSEN5 , L. L. SØRENSEN6 , J. BRANDT1 , L. M. FROHN1 and S. DJURHUUS6 1

National Environmental Research Institute, Department of Atmospheric Environment, Denmark; Department of Geophysics, University of Copenhagen, Denmark; 3 AMS 14 C Dating Laboratory, Department of Physics and Astronomy, University of Aarhus, Denmark; 4 Danish Meteorological Institute, Denmark; 5 Institute of Chemistry, University of Copenhagen, Denmark; 6 Risø National Laboratory, Denmark (∗ author for correspondence, e-mail: [email protected])

2

(Received 19 July 2004; accepted 5 July 2005)

Abstract. As part of the Danish NEAREX project the origin and variability of anthropogenic atmospheric CO2 over the Northeast Atlantic Region (NEAR) has been studied. The project consisted of a combination of experimental and modelling activities. Local volunteers operated CO2 sampling stations, built at University of Copenhagen, for 14 C analysis at four locations (East Denmark, Shetland Isles, Faroe Isles and Iceland). The samples were only collected during winter periods of south-easterly winds in an attempt to trace air enriched in fossil-fuel derived CO2 due to combustion of fossil fuels within European countries. In order to study the transport and concentration fields over the region in detail, a three-dimensional Eulerian hemispheric air pollution model has been extended to include the main anthropogenic sources for atmospheric CO2 . During the project period (1998–2001) only a few episodes of transport from Central Europe towards NEAR arose, which makes the data set for the evaluation of the method sparse. The analysed samples indicate that the signal for fossil CO2 , as expected, is largest (up to 3.7 ± 0.4% fossil CO2 ) at the Danish location closest to the European emissions areas and much weaker (up to ∼1.5 ± 0.6% fossil CO2 ) at the most remote location. As the anthropogenic signal is weak in the clean atmosphere over NEAR these numbers will, however, be very sensitive to the assumed background 14 CO2 activity and the precision of the measurements. The model simulations include the interplay between the driving processes from the emission into the boundary layer and the following horizontal/vertical mixing and atmospheric transport and are used to analyse the meteorological conditions leading to the observed events of high fossil CO2 over NEAR. This information about the history of the air masses is essential if an observed signal is to be utilised for identifying and quantifying sources for fossil CO2 . Keywords: atmospheric transport, 14 C, fossil fuel CO2 , numerical modeling, the North East Atlantic Region

1. Introduction The overall picture of the carbon balance of the Earth’s atmosphere is still uncertain due to the complex interplay of processes in and between the oceans, land

388

C. GEELS ET AL.

biosphere and atmosphere. The numerous processes driving the carbon sinks and sources vary on a multitude of temporal and spatial scales depending on both natural and anthropogenic disturbances. The atmospheric CO2 signal bears a signature of the underlying processes, which through observations and modelling can be used to separate the different source types. However, the atmosphere acts as a fast mixer/integrator and the net carbon exchange is small compared to the gross fluxes between the carbon pools, which makes the relevant atmospheric signal small. Additional information can be extracted from the atmospheric CO2 signal if also the isotopic ratios like 13 C/12 C, 14 C/12 C and 18 O/16 O are analysed (e.g. Ciais et al., 1995; Zondervan and Meijer, 1996; Remonet et al., 2002; Pataki et al., 2003). In combination with high-resolution atmospheric transport models this type of data gives us a better understanding of the formation of the three-dimensional time dependent CO2 field and of e.g. the relative importance of the various source types and regions. Such knowledge is essential if national and regional agreements on emission control and carbon sequestration methods are to be verified by independent measurements in the future (Tans and Wallace, 1999). The newly initiated Integrated Project CarboEurope-IP (FP6) will among other things focus on both simulations and measurements of the isotopic components of CO2 in order to assess the present European carbon balance. Field measurements (Takahashi et al., 1997) indicate that the North East Atlantic Region is an important net sink for atmospheric CO2 due to the combination of extensive ocean dynamics (deep water formation), low sea surface temperatures and high biological productivity. Within this region the thermohaline circulation drives a continuous northward and downward transport of surface water, which is cooled at the high northern latitudes. The solubility of CO2 in water is enhanced at lower temperatures and this circulation maintains a gradient in the CO2 concentration at the air-sea interface and hence a net flux from the atmosphere to the ocean. Furthermore the biological uptake of CO2 and the formation of particulate organic carbon and subsequent sinking detritus, brings carbon from the surface waters to larger depths (Sweeney et al., 2000; Antoine and Morel, 1995). The NEAREX project for studying “concentrations and fluxes of atmospheric CO2 and particulate matter in the North East Atlantic Region” was initiated in 1998 as a co-operation of Danish scientists and funded by the Global Change programme of the Danish Natural Science Research Council. The NEAREX project focuses on the part of the North East Atlantic Ocean that covers the area in the Atlantic Ocean north of 60◦ N. The abbreviation used for the region is henceforth NEAR. The combination of the NEAR area as a net ocean sink for CO2 and the atmospheric transport from the surrounding land based sources and sinks for CO2 offers a good opportunity for studying the carbon budget and fluxes on a regional scale. The primary objective of the project has been to study the origin and variability of the atmospheric CO2 and the air–sea flux in this region. The objective has been approached through a combination of experimental and numerical modelling activities. Some of the activities took place during ship cruises in the Greenland Sea. Estimates of the

THE EUROPEAN FOSSIL FUEL PLUMES IN THE ATMOSPHERE

389

Figure 1. pCO2 in seawater measured in the northern North Atlantic during two cruises as part of the NEAREX project. The Dana cruise took place in June 1999 and Lance took place in October 2003. Also the location of the four measuring sites is given: Hoejerup, at Stevns, Denmark; Lerwick, Shetland Isles; Torshavn, Faroe Isles and Reykjavik, Iceland.

local air-sea CO2 fluxes were based on observed differences in the partial pressure of CO2 , between the air and the surface water, as well as on direct micrometeorological flux measurement techniques. Generally large spatiotemporal differences were observed in the oceanic CO2 concentration and hence in the net air-sea flux during these cruises. Only small regions around 64◦ N were in early spring found to be net source areas for atmospheric CO2 , with oceanic CO2 levels higher than the atmospheric levels, see Figure 1 (colour figures are included in the online version of the paper). The observed pattern of the flux correlated well with surface water fluorescence, which is an indicator for biological activity. In remote oceanic regions the atmospheric CO2 content is typically more constant than the oceanic part. However, concentration variability due to synoptic changes in atmospheric circulation can lead to changes in the atmospheric concentration and hence in the driving concentration gradient between the atmosphere and the sea surface. As part of the project this variability has been studied by applying a three-dimensional hemispheric transport model including simple parameterisations of the main sink and source types for CO2 . Simulations for the period October 1990 to December 1997 were compared with observed records from within the NEAR region and the variability on seasonal to synoptic time scales was well captured by the model (Geels et al., 2001). This and other studies have shown that the short-time variability in atmospheric CO2 is highest during wintertime in this region where long-range transport from the continents leads to mainly positive anomalies in the total atmospheric CO2 concentration

390

C. GEELS ET AL.

(e.g. Yuen et al., 1996; Brandefelt and Holmén, 2001). In order to study and possibly quantify the origin and magnitude of these events in more detail, observations of 14 CO2 were conducted as part of the project. The radioactive 14 C isotope is a useful tracer for differentiating between the different sources for atmospheric CO2 , as only the CO2 related to fossil fuel source types is fully free of 14 C (Zondervan and Meijer, 1996; Levin et al., 2003). Air samples taken from four locations on the border or within NEAR could thereby help identifying events of atmospheric transport from primarily fossil fuel heated land regions. Especially it was assumed that large portions of non-background CO2 reaching NEAR from eastern and central parts of Europe would be derived from recent fossil fuel burning and hence would be depleted in 14 C. As a supplement to the 14 C-analysis, a three-dimensional transport model has simulated the fossil fuel component of CO2 . The experimental and modelling activities are described in the following sections, followed by an analysis and discussion of the results focusing on events from April 1998 and January 2001. 2.

14

C-Isotope Analyses at Four Selected Locations

Fossil fuels are devoid of 14 C (T1/2 = 5730 years) i.e. CO2 emitted directly from the smokestacks lacks 14 CO2 completely. However, as the smoke plumes move away from their origins more and more air – and hence modern CO2 – is entrained into the plumes, depending on the meteorological conditions. Consequently, the signature of the fossil fuel is to be found in the deviation of the 14 C-content from that of the well-mixed background atmosphere’s CO2 . The collection of CO2 for 14 C-analysis aimed at following air-parcels passing Central Europe on their way (notably Poland and The Czech Republic, where the main heating source is coal and therefore the winter emissions are large) up to the NEAR area. Sampling stations from Denmark to the NEAR area were established on four locations along the route (see also Figure 1): • • • •

Denmark, at Hoejerup on the east coast of Zealand, 70 km south of Copenhagen. Shetland Isles, coastal site close to Lerwick. Faroe Isles, coastal site close to Torshavn. Iceland, coastal site close to Reykjavik.

“Close to” means that the operators (volunteers) were living at the places mentioned. They assured that the measurements were carried out sufficiently far away from populated areas to ensure that local emissions were not sampled. Sampling was carried out during on-shore flows in order to sample marine air. 2.1. SAMPLING 14

CONDITIONS

C-analyses are expensive and sampling was therefore carried out only during meteorological conditions deemed ideal for a significant atmospheric transport to


391

Figure 2. Back trajectories for air predicted to arrive at the four sampling sites 6 (left figure) and 24 (right figure) hours after midnight on 23 January 2001.

NEAR from Europe. This was accomplished by means of meteorological prognoses provided by the Danish Meteorological Office. When a favourable situation for south-easterly winds was approaching, the HIRLAM prognostic model (Sørensen et al., 1998) was started and air parcel back trajectories for every 6 h, up to 72 h in advance, was estimated. Examples of the predicted trajectories for the four locations are given in Figure 2 for a day in January 2001. Here the predicted trajectories for air parcels arriving 6 h ahead (left figure) indicates a favourable situation with transport across Europe 24 h ahead (right figure) the wind pattern is predicted to have changed to a more unfavourable situation resulting in transport of air from the North Atlantic towards three of the locations. The operators were put on the alert by fax-messages at least 24 h in advance of a prospective event, but asked to sample only if the final prognoses were affirmative. Unfortunately, far fewer favourable events than anticipated occurred during the course of the project. 2.2. S AMPLING

PROCEDURE

Four identical sampling devices were constructed at Chemical Laboratory V at the University of Copenhagen. They were shipped to the operators only after these had gone through a short course in the operation of the devices. Our primary concerns regarding the sampling were twofold: (1) A sufficient amount of BaCO2 (obtained by letting CO2 from the air react with Ba(OH)2 ) was to be obtained over a reasonable sampling time. We ended up sampling 100 mg (6 mg C) over 20 min. (2) The sampling devices were to be handled by volunteers with no relevant background; they therefore had to be easy to handle, also in terms of transportation to and from the sites. In fact all the samplings proceeded without problems or hazards.

392

C. GEELS ET AL.

Solutions of Ba(OH)2 were prepared at the beginning of the campaign by a small medical firm specialised in infusion liquids. Infusion bags of aluminium coated heavy-duty plastic were selected for the purpose. They were supposed to be absolutely airtight, and indeed there was no trace of BaCO3 in the bags after three years. Rinsing water was likewise prepared ahead of time, it was millipore water and it was also stored in heavy-duty plastic infusion bags. The firm undertook the task of testing the water once a year by conductivity measurements, and again no trace of CO2 was found. The central unit in the sampling device is the collection filter. This removable filter is a so-called “glass fritte” which is a plate of porous, sintered glass, 2 cm in diameter. The filter separates two 1 litre glass bulbs from each other. At the beginning of a sampling 0.5 litre 0.1M Ba(OH)2 -solution was poured into the upper bulb. A pump serves the dual purpose of (1) sucking air through the glass fritte into the Ba(OH)2 -solution for 20 min during the sampling phase. The glass fritte causes the air to enter the solution as tiny bubbles ensuring good air/liquid contact. (2) Filtering the BaCO3 formed (see below) during the filtering phase. This is accomplished by first applying suction to the upper bulb, allowing air to enter through the lower bulb and the filter (from below) causing the air to enter into the Ba(OH)2 -solution bubbles. Subsequently suction is applied to the lower bulb until all liquid has come through the filter, while the filter retains the BaCO3 . The combined unit is contained in a metal toolbox 75 cm × 40 cm × 40 cm mounted on a wheelbarrow for convenience. The unit has a rechargeable 12 V battery that runs the vacuum pump, which can accomplish a vacuum of about 1/2 bar. About 100 mg BaCO3 is formed during the 20 min a sampling takes. It is first rinsed twice with CO2 -free water and next packed in India paper. The package is stored in a plastic canister with a tight-fitting lid. The canister contains two silica capsules that will dry out the sample until it is unpacked 24 h later. It is now ready to be scraped off the filter plates and shipped to 14 C analysis in small glass containers. 2.3.

14

C

ANALYSIS

The 14 C analyses were performed at the AMS 14 C Dating Laboratory at the University of Aarhus, Denmark. The BaCO3 samples described above were treated with phosphoric acid to liberate CO2. Most of the CO2 was converted to graphite for 14 C measurement in the tandem accelerator by the AMS method (Accelerator Mass Spectrometry). The rest was used for stable carbon isotope measurement (δ 13 C) ´ E. Sveinbjörnsdóttir at the Science Institute, University of performed by Dr. A. Iceland, in order to correct the 14 C results for isotope fractionation according to international convention (Stuiver and Polach, 1977). For details of the AMS method, see Andersen et al. (1989). The AMS 14 C Dating Laboratory’s measurement results for the TIRI and FIRI intercalibrations (Gulliksen and Scott, 1995; Scott, 2003 (participating laboratory no. 88)) showed that the precision of the 14 C data reflects the


393

actual accuracy. In fact, in the latter intercalibration, the weighted average deviation of the Aarhus laboratory from the consensus value on 18 measurements was only 0.8 ± 12 y.

3. Numerical Modelling of Atmospheric Transport of CO2 The Danish Eulerian Hemispheric Model (DEHM) was applied for the modelling part of the project. DEHM is a three-dimensional Eulerian transport model, which initially was developed at the Danish National Environmental Research Institute with the main purpose of investigating the atmospheric transport of pollutants to the Arctic (Christensen, 1997). The model domain is centred at the North Pole and covers the majority of the Northern Hemisphere and hence the main anthropogenic sources for arctic pollution. Within the last ten years the model has been further developed in order to broaden its usage in relation to various air pollution problems (see Frohn et al., 2003; Hansen et al., 2004; Christensen et al., 2003; Geels, 2003). Many of these air pollutants have relatively long lifetimes in the atmosphere and consequently a large model domain is necessary in order to include the effect of long range transport from distant source areas. To be able to study the spatiotemporal variability of the concentration field in detail, a high model resolution is necessary. The combination of a large domain and a high resolution is, however, very computer demanding. The DEHM model has therefore been developed with two-way nesting capabilities, which allows for a large mother domain as well as a nested domain with a three times higher horizontal resolution. 3.1. M ODEL

SET - UP

In the current model set up the Fifth-Generation NCAR/Penn State Mesoscale Model, MM5 (Grell et al., 1995) is applied as the meteorological driver for the DEHM model. Both models are applied on a polar stereographic projection with a resolution of 150 km in the mother domain and a resolution of 50 km in a nested domain over Europe. Vertically the models are divided into 20 sigma levels. The state-of-the-art meteorological model, MM5, is here run with the ECMWF/TOGA Basic Level III Consolidated data (2.5◦ × 2.5◦ , 12-h resolution) as input. Within the DEHM model the basic continuity equation for CO2 is divided into sub-models, as the physical processes represented by the different terms vary over different time scales. Thereby the most appropriate numerical technique can be applied for each sub-model separately at each time step during the integration. The horizontal and vertical advection is solved by a modified accurate space derivative scheme and a finite element scheme, respectively, while a Taylor series expansion is applied for the time integration of the advection. The three-dimensional diffusion is solved by applying a finite-element scheme for the spatial discretizations and the Crank–Nicolson method for the integration in time. The time step in the integration

394

C. GEELS ET AL.

is controlled by the Courant–Friderich–Levy stability criterion and will for the mother domain typically be between 10 to 20 min. A detailed description of the model, the numerical methods as well as the testing of the methods is given in Christensen (1997), Frohn et al. (2002), and references herein. The emission of CO2 due to combustion of fossil fuels is included as a lower boundary condition for the model. The data applied is based on the 1◦ × 1◦ global map of emissions available through the Emission Database for Global Atmospheric Research, EDGAR (Olivier et al., 1996). This map is constructed from a combination of three types of information: • Inventories of the CO2 emission from the different components (solids, liquids and gases) of fossil fuel combustion and cement manufacturing in the individual countries. • A 1◦ × 1◦ data set of the population density and distribution within each country for 1984. • A 1◦ ×1◦ data set of political units for 1993, defining the borders of 186 countries. Within each country the emission is distributed according to the population density and activity levels, which should secure a precise geographical distribution including large point sources as well as a few ship routes. The resulting data set represents the yearly emission of 1990, see Figure 3. Information on national

Figure 3. The anthropogenic emission of fossil fuel CO2 (Tg C/grid square) within the hemispheric model domain. The emissions are for 1990 and based on the EDGAR database.


395

emission levels from the CDAIC database (Marland et al., 2003) has been introduced in order to update the map to the year of 1998. As the information is only available at the national level, this simple approach will not include variations due to e.g. redistribution of emissions within the individual countries. Since 1998 the total emission of fossil fuel CO2 from the western and eastern European countries has been almost constant. Following Rotty (1987) a weak seasonality of the emissions are assumed, with approximately 17% higher emissions during the winter months than during the summer months north of 30◦ N. This estimate is, however, based on statistics on the sale of fossil fuel in the 1980s and more recent studies of atmospheric 14 CO2 and 222 Rn in Germany indicate that this seasonality could be significantly higher (Levin et al., 1995, 2003). Other studies indicates that the use and hence emission of fossil fuels also varies from day to day and on diurnal times scales (e.g. Takahashi et al., 2002). However, the documentation is sparse and this type of variations has been neglected in the present study. The model has been run for the period 1997 (December) to 2001 (December), including the fossil fuel emissions for 1998 and driven by meteorological data from the MM5 model. An initial and background concentration of 0 ppm CO2 is assumed and the resulting model simulations calculate deviations from this background mean. December 1997 is used as a spin up period and only the model results for 1998–2001 will be discussed in the following. 3.2. MODEL

VALIDATION

The CO2 version of the DEHM model including the main anthropogenic as well as natural (the land biosphere and the oceans) sources and sinks for atmospheric CO2 has previously been validated against observations from more than 40 sites on the Northern Hemisphere (Geels, 2003). Model simulations for the period 1991–1998 have been evaluated and the seasonal to diurnal variability observed at both oceanic and continental sites are well captured by the model with significant correlation coefficients in the range of ∼0.6–0.95.

4. Results and Discussion The experimental results for all samples are given in Table I. The 14 C content is given in the pMC unit, meaning percent of Modern Carbon. The term “modern carbon” denotes the nominal 14 C content in atmospheric CO2 in 1950 i.e. before the onset of the so-called nuclear bomb pulse. The atmospheric nuclear bomb tests during the 1950s and 1960s resulted in approximately a doubling of the atmospheric 14 C content followed by a nearly exponential decrease. Thus, the present atmospheric 14 C content is about 10% higher than the 1950 level (see Table I) and this so-called “normal background” content is still decreasing. The term “normal

Site

1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2

Date time

26-03-1998 14:40 03-04-1998 14:55 04-04-1998 10:10 20-04-1998 12:40 21-04-1998 11:50 24-11-1998 10:20 12-12-1998 12:05 01-01-1999 11:20 04-03-1999 16:00 18-03-1999 15:30 04-04-2000 11:00 20-01-2001 13:40 23-01-2001 13:40 01-03-2001 09:55 03-04-1998 06:20 04-04-1998 10:15 20-04-1998 06:15 21-04-1998 06:15 22-04-1998 06:15

∗

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Sample no. 4298 4300 4301 4302 4303 4738 4739 4740 6254 6253 6252 7046 7047 7048 4304 4305 4306 4307 4308

Analysis no.

C pMC 110.0 109.3 109.1 110.1 109.1 107.1 106.6 107.4 107.3 109.5 107.7 106.3 106.8 107.2 109.8 111.0 109.2 108.8 109.6

−20.6 −20.9 −18.9 −18.2 −18.8 −21.7 −19.6 −19.2 −18.1 −19.5 −18.6 −18.3 −19.4 −18.0 −22.1 −22.2 −21.2 −21.7 −21.3

14

C (%VPDB)

13

0.5 0.6 0.7 0.6 0.6 0.5 0.4 0.5 0.5 0.6 0.6 0.4 0.4 0.5 0.6 0.6 0.6 0.6 0.5

σ 111.0 111.0 111.0 111.0 111.0 110.7 110.7 110.7 110.6 110.6 110.2 109.9 109.9 109.8 111.0 111.0 111.0 111.0 111.0

NNH-bg pMC 110.3 109.9 109.9 109.9 109.9 110.7 110.4 110.2 110.1 110.1 109.2 109.2 109.2 108.7 109.9 109.9 109.9 109.9 109.9

JFJ bg pMC 0.9 1.5 1.7 0.8 1.7 3.3 3.7 3.0 3.0 1.0 2.3 3.2 2.8 2.4 1.1 0.0 1.6 2.0 1.2

0.2 0.6 0.8 −0.1 0.8 3.2 3.5 2.6 2.5 0.5 1.3 2.6 2.2 1.4 0.1 −1.0 0.7 1.0 0.3

Fossil CO2 with JFJ-bg (%)

(Continued on next page)

Fossil CO2 with NNH-bg (%)

TABLE I C, 14 C and % fossil CO2 content for all BaCO3 samples. The % fossil CO2 content in the samples is estimated by applying the Northern Hemisphere background (NNH-bg) and the continental background from Jungfraujoch (JFJ-bg). The uncertainty related to the background values is assumed equal to 0.2 pMC. The sites are 1. Hoejerup; 2. Lerwich; 3. Torshavn; 4. Reykjavik

13

396 C. GEELS ET AL.

Site∗

2 2 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4

Date time

19-01-2001 20:05 20-01-2001 08:45 03-04-1998 09:45 04-04-1998 12:00 04-04-1998 20:00 19-04-1998 14:00 20-04-1998 14:00 21-04-1998 14:00 30-12-1998 15:00 01-01-1999 14:05 22-01-2001 17:15 23-01-2001 15:10 24-01-2001 13:30 02-04-1998 17:50 19-04-1998 14:20 21-04-1998 17:25 30-12-1998 15:25 01-01-1999 13:15

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Sample no. 7049 7050 4309 4310 4311 4312 4313 4314 6251 6250 7052 7053 7051 4315 4316 4317 6249 6248

Analysis no.

14 C pMC

107.7 108.2 110.3 110.7 110.0 110.4 110.6 109.6 109.6 111.3 107.2 108.6 107.8 110.3 110.5 111.0 109.1 109.6

13 C (%VPDB)

−21.5 −21.8 −21.0 −20.6 −20.8 −21.1 −20.9 −21.0 −21.2 −21.4 −19.2 −19.5 −20.2 −21.2 −21.8 −21.3 −21.5 −21.4 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.5 0.6 0.5 0.4 0.5 0.5 0.5 0.6 0.6 0.7

σ

TABLE I (Continued)

109.9 109.9 111.0 111.0 111.0 111.0 111.0 111.0 110.7 110.7 109.9 109.9 109.9 111.0 111.0 111.0 110.7 110.7

NNH-bg pMC 109.2 109.2 109.9 109.9 109.9 109.9 109.9 109.9 110.4 110.2 109.2 109.2 109.2 109.9 109.9 109.9 110.4 109.2

JFJ bg pMC 2.0 1.5 0.6 0.3 0.9 0.5 0.3 1.2 1.0 −0.5 2.4 1.2 1.9 0.6 0.4 0.0 1.4 1.0

Fossil CO2 with NNH-bg (%) 1.3 0.9 −0.3 −0.7 −0.1 −0.4 −0.6 0.3 0.8 −1.0 1.8 0.5 1.3 −0.3 −0.5 −1.0 1.2 −0.4

Fossil CO2 with JFJ-bg (%)


397

398

C. GEELS ET AL.

background” applies to the 14 C level in the well-mixed atmosphere in the absence of distinct fossil fuel plumes. As we are studying the weak fossil signal in the clean air over the North Atlantic the results will be very sensitive to the applied background level. Temporal and spatial variations in the overall carbon cycle including the use of fossil fuels and variations in the stratosphere/troposphere exchange of 14 C as well as in the atmospheric transport will influence the 14 C level at a given location and are known to lead to seasonal or inter-annual variations in measured 14 C curves (see e.g Levin et al., 2003; Goodsite et al., 2001 and references herein). We have included two different background levels representing (1) the expected background level in the northern part of the Northern Hemisphere (NNH) and (2) the free troposphere continental background level at Northern Hemisphere mid latitudes, see Table I. The NNH background is determined from an assumed value of 111.1 pMC on 1 January 1998, and an annual decrease of 0.4 pMC estimated from the Northern Hemisphere bomb-pulse curve displayed in Goodsite et al., 2001. This curve is based on a combination of several 14 C records obtained from either annually averaged atmospheric 14 CO2 measurements or from tree-rings and seeds in the region north of 30◦ N. The applied NNH background will therefore not include e.g. seasonal variations, but the overall level will in our opinion be representative of the clean background air within NEAR. The continental background is based on a more detailed 14 CO2 data set of quasicontinuous observations from the high altitude site Jungfraujoch (46◦ 33 N, 7◦ 42 E, 3450 m a.s.l.) in the Swiss Alps (Levin and Kromer, 2005). The 14 C data at this location show a seasonal cycle (peak to peak amplitudes of 5 to 8‰ with a minimum in March and maximum in August. In Table I this background is given as monthly means based on bi-weekly integrated 14 CO2 samples from Jungfraujoch (JFJ). In Figure 4 the measured 14 C values at the four sites are compared with the two different background levels. It is seen that almost all measured values are depleted in 14 C relative to the NNH background. This was expected since all samples were collected when the trajectory calculations predicted that the air had passed over central Europe and therefore should contain a fossil fuel contribution to the atmospheric CO2 concentration. Also the continental background throughout the periods in focus here is seen to be lower (by between 0.1 to 1.0 pMC) than the NNH background and the samples from the three locations in NEAR during a few episodes are seen to be slightly higher than the JFJ background. Levin and Kromer (2005) compared the continental JFJ background with the marine background levels at Izaña and Mace Head and found the JFJ annual means to be slightly lower (order of 0.2–0.3 pMC) due to occasional influence by European emissions. All in all this indicates that the data from JFJ might not be representative of the marine atmosphere in the NEAR area and hence not an ideal background value for this study. The fossil fuel contribution is seen as the difference between the brown (the background) and the coloured (measured) columns in Figure 4. The relative fossil


399

Figure 4. Observed 14 C content in the BaCO3 samples collected at: Zealand (red), the Shetland Isles (blue), the Faroe Isles (yellow) and in Iceland (green). The brown columns indicate the two applied background levels representing well-mixed atmospheric air in the northern part of the Northern Hemisphere (NNH) and at the continental station Jungfraujoch (JFJ), respectively.

fuel contribution to the observed CO2 signal given in Table I is defined as: %fossil =

pMC(background) − pMC(obs) 100% pMC(background)

High/low values of this number are hence related to air with relatively more/less CO2 from fossil fuel related sources. Two of the sampling periods include measurements from all four locations (April 1998) or from three locations (January 2001). In order to relate the observed data to the model results, time series and a scatter plot of the simulated fossil fuel CO2 component (in percentage of the global CO2 level) are displayed in Figures 5 and 6 together with the measured values. The uncertainty ranges given in the plots include

400

C. GEELS ET AL.

Figure 5. Hourly time series of the simulated fossil fuel CO2 component during April 1998 and January 2001, as well as the observed values. Both observations and model results are given as a percentage of the background level. The model result for the Icelandic site is from the hemispheric model domain, while the results for the other sites are from the nested model domain over Europe.

the combined uncertainties of our measurements (see Table I) and of the applied background levels. From both the measured values and the simulated time series, a negative gradient is seen in the fossil signal from the Danish location towards the Atlantic stations. This is likely reflecting the gradual dilution of the continental plume during the transport over the ocean. From the model simulation it is clear that the temporal variability in fossil fuel CO2 is largest at Hoejerup, due to the proximity to the European fossil fuel emissions. At the more remote locations the atmospheric transport processes smooth out the signal and only a few anomalies are seen during the two


401

Figure 6. Observations and model results of % fossil fuel CO2 during the months April 1998 and January 2001. The overall correlation coefficient is 0.64 and 0.61 using NNH (left) and JFJ (right) background levels, respectively.

months displayed in Figure 5. In the following the agreement between model and observations during these events and the associated meteorological conditions will be discussed. Due mainly to the proximity to local emissions and variations in the planetary boundary layer (PBL) depth, instantaneous and discrete samples at a given location can be more or less spatially representative. This is important to keep in mind in the following where the measured fossil fuel percentage is compared to the results (hourly means) of the model simulations with a horizontal resolution of 50 km × 50 km. As the events are the result of long-range transport even small errors in the simulated transport (the wind field) and mixing during the several thousand kilometres of transport can lead to small leads and lags of the simulated events compared to the observed values. The scatter plot in Figure 6 displays the overall agreement between the observations and the model results for April 1998 and January 2001. When applying the NNH (JFJ) background the correlation coefficient is 0.64 (0.61) and the observed and simulated means are 1.26% (0.4% )and 1.67% fossil fuel CO2 , respectively. Both in case of the NNH and the JFJ background the agreement between observations and model simulations are highest during January 2001. In April 1998 the model overestimates the CO2 signal over the NEAR area, indicating that the applied emission field is too strong in this month. The model especially overestimates the % fossil CO2 values based on the JFJ background. This continental background in April is close to the annual minimum in the 14 CO2 level and is, as mentioned above, lower than some of the 14 CO2 samples from NEAR in April. Based on measurements from only two months it is, however, not reasonable to discuss if this is due to a difference in the seasonal cycle of 14 CO2 in the NEAR area and over the continent. 4.1. APRIL 1998 In the beginning of April 1998 a low-pressure system developed just west of Ireland, which led to advection of air from western and central part of Europe towards the

402

C. GEELS ET AL.

British Islands and the North Atlantic. Around 4 April enhanced fossil CO2 concentrations were seen in the simulations for Hoejerup with an increase from ca. 1.3 to 2.6%, which is in reasonable accordance with the slightly lower observed1 values of 1.5 ± 0.6%–1.7 ± 0.7%. The conditions led; however, not to transport towards the more northern locations where both observed and calculated concentrations are in the range 0.0 to 1.2% fossil CO2 . A new potentially favourable situation for south-easterly winds arose towards 19–20 April. The combination of a large high-pressure system over northern Russia, a low-pressure system over the Black Sea region and a front over the British Islands resulted in atmospheric transport from Siberia and central Europe towards NEAR. This can be seen in the simulated fossil fuel CO2 maps for Europe, which are given as six hour means at 18 GMT for the period 19–22 April in Figure 7. In the corresponding time series (Figure 5) the gradual increase/decrease in the fossil CO2 component towards/after the 21st is

Figure 7. The simulated surface concentration (ppm) of fossil fuel CO2 displayed as six hour means at 18:00 in the period 19 to 22 April 1998. The concentration is given as the deviation from a background mean. Mean sea level isobars are also shown.


403

most clearly seen at Lerwich and Torshavn. At Lerwich the measurements and the model results agree on the level on the 20th and 22nd, while the model estimates a maximum fossil fuel CO2 of 3.2% compared to the measured number of 2 ± 0.6%. At Torshavn the highest value (1.2 ± 0.4%) of the six samples in this month is from the afternoon on the 21st April, which also corresponds to a maximum in the simulated time series of 2.3%. From both the time series and the simulated maps in Figure 7 it is seen that the fossil fuel plume also reaches Iceland and results in slightly elevated levels. This is, however, not in agreement with the analysed samples from Reykjavik.

4.2. J ANUARY 2001 During January 2001 the meteorological forecast indicated a period with southeasterly winds and several samples were collected at Hoejerup, Lerwich and Torshavn 19 to 24 January (Table I and Figure 5). The meteorological conditions from midJanuary were characterised by a stable period with high-pressure systems moving from the North Sea towards central Europe with weak winds over the continent. The simulated fossil fuel concentrations and the mean sea level pressure conditions in the period 19 to 23 January are displayed in Figure 8. On the 19th high CO2 levels are seen over Western Europe where stable conditions led to accumulation of pollution in the shallow wintertime PBL in this high emission region (also seen as a gradual increase in fossil CO2 at Hoejerup in Figure 5). On 20–21 January the combination of a low-pressure system moving in from the North Atlantic and a high over Scandinavia disturbs this stable period and forms a trough over the North Sea. The resulting stronger south-easterly winds lead to the advection of air from Europe towards NEAR, where especially the simulated concentrations over the Shetlands are high. Unfortunately the samples from this location were taken in the evening and morning on 19 and 20 January, respectively and thereby missed the maximum of the event that occurred about 24 h later according to the simulations. At the Thorshavn site the sampling was initiated later and the model results are in agreement with the observed levels (2.4 ± 0.5 and 1.2 ± 0.4%) on 22 and 23 January, but lower than the value observed (1.9 ± 0.5%) on the 24th. Again only small variations are seen in the model simulations at Iceland. At the Danish location the two samples from this period correspond to a fossil fuel contribution of 3.2 ± 0.4% and 2.8 ± 0.4%, which should be compared to the modelled levels of 4.3 and 2.9% at the same date and hour. The model can also be used to get a view of the vertical distribution of CO2 and examples of vertical profiles at the Torshavn and Hoejerup locations are given in Figure 9 for January. At the Danish location during the days around the 20th CO2 is seen to gradually build up in shallow boundary layer and is therefore confined to the lowest 800–900 m of the atmosphere. At Torshavn the long-rang atmospheric transport leads on the other hand to an effective vertical mixing of CO2 and the

404

C. GEELS ET AL.

Figure 8. The simulated surface concentration (ppm) of fossil fuel CO2 displayed as six-hour means at 18:00 in the period 19 to 23 January 2001. The concentration is given as the deviation from a background mean. Mean sea level isobars are also shown, with L and H specifying low- and highpressure centres, respectively.

episode of high fossil CO2 is in the simulated profile seen to reach up to about 1800 m over the ground with only a weak vertical gradient. 4.3. THE

ORIGIN OF FOSSIL FUEL

CO2

OVER

NEAR

By dividing the fossil fuel map into geographical regions and running the model with the emission activated in these regions separately, the dominating source areas for the NEAR region can be studied. Here the EDGAR emissions map (Figure 3) has been subdivided into four main areas corresponding to Europe, Russia, Asia and North America. In Figure 10 the contribution from each of the four regions to the total fossil fuel level is presented at Iceland and Shetland. The results are based on model runs for 1998 and of the total emission in the hemispheric domain approximately 11, 30, 27 and 32% are allocated in the Russian, Asian, European and North American region, respectively.


405

Figure 9. Simulated vertical profiles of fossil CO2 at Hoejerup and Torshavn for January 2001. Note the different colour scales in the two profiles.

Within the NEAR region the anthropogenic emissions in the Russian and Asian sector both contribute with about 15% to the well-mixed background concentration of fossil fuel CO2 . As expected the influence of the European sources is high close to Europe (∼70% at the Danish location), while the influence of the North American emissions increases in the western part of the region. According to the model estimates, the air above Iceland thereby contains about 32 and 36% anthropogenic CO2 from North America and Europe, respectively. Further to the north the Russian emissions play a larger role. Mainly during winter the general westerly flow can be blocked by persistent high-pressure systems over Siberia leading to periods with a northern flow (Christensen, 1997). Likewise, episodes where polluted air masses

406

C. GEELS ET AL.

Figure 10. A pie chart for Iceland (left) and Shetland (right) indicating the percentage that each of the four regions (Russia, North America, Asia and Europe) contributes to the total fossil fuel CO2 field simulated by DEHM. The results are obtained as an average of a model simulation including fossil fuel emissions and meteorological data for 1998.

originating from the East Coast of the USA are transported towards the NEAR are regularly seen in the current model simulations as well as in other studies.

5. Conclusions One of the objectives of the Danish project NEAREX was to test the feasibility of identifying the European fossil fuel plumes over the carbon sink region in the Northeast Atlantic Region (NEAR). The hypothesis was that isotopic (14 C) observations sampled during winter periods with southeasterly winds combined with mesoscale transport modelling over NEAR could be used to identify and possibly quantify the fossil fuel emissions in Central Europe. However, during the project period 1998–2001 only a few ideal episodes with transport from Central Europe arose during the winter seasons and in total less than 40 samples were collected at four locations within or at the boarder to NEAR. The 14 C analyses were performed at the AMS 14 C Dating Laboratory at the University of Aarhus, Denmark, with a precision of about ±0.4–0.7 pMC. The 14 C observations have in the present paper been compared to simulations with the DEHM model including fossil fuel emissions. Due to the sparseness of the data set it has not been possible to make detailed statistical analyses of the comparison and it is accordingly not possible to make any firm conclusions. The 14 C signatures of fossil fuel burning are nevertheless evident in the analyzed samples with the expected gradient of strongest signatures at the Danish location closest to the emission areas and weakest signatures at the sampling site on Iceland. The simulations of the fossil fuel component of atmospheric CO2 confirm this pattern and capture most of the observed events of enhanced fossil fuel CO2 . The model results give insight into the interplay be-


407

tween the emissions of fossil CO2 , the subsequent mixing processes in the boundary layer and the long-range atmospheric transport leading to the observed events. As an example the most evident episode in the period 1998–2001 arose in the days 20–24 January 2001 due to stable atmospheric conditions leading to accumulation of CO2 over the continent followed by a period of southeasterly winds and advection towards the North Atlantic. A given signal will always be a result of the interplay of such processes, while their relative importance will vary depending on time and location. Information about origin and transport of a given air mass is therefore crucial if the observed 14 C signal is to be used for e.g. evaluation of international emission reduction agreements. The model simulations for 1998 showed that the main part of the fossil CO2 over the NEAR region originated in North America and Europe, but e.g. interannual variations in the main atmospheric transport routes could potentially lead to changes in the relative importance of the main emission areas. This emphasizes the importance of knowledge of the origin of the air masses analyzed, as changes in transport otherwise might be interpreted as changes in emissions. In this study locals operated a small monitoring network including four sites and sampling was only initiated when predictions of air parcel trajectories indicated transport from Central Europe. In this way the total expenses for the costly 14 C analysis could be minimized. Unfortunately, some of the events of enhanced fossil CO2 were missed and the usefulness of this sampling strategy needs to be improved by increasing the number of samples during each event, and/or by a more detailed analysis of the trajectories and hence a better prediction of the arrival time for the polluted air masses at each location. Periods with sampling during clean-air conditions are also recommended, as this would indicate the background level of 14 CO2 in the NEAR area and thereby support the choice of background curve used in the analysis. Acknowledgments The Danish Research Agency funded the NEAREX project, Grant number 201201-0002. We want to express our gratitude to our volunteers, operating the CO2 sampling stations: F. Birc (Denmark), B. E. Jamieson (Shetland Isles), K. Meitil (Faroe Isles), and A. Olafsson (Iceland). Part of the development of the DEHM-CO2 model was done as part of the EU cluster project CARBOEUROPE, sub-project AEROCARB. We also acknowledge the CARBOEUROPE-IP project for financial support.

Note 1. In this comparison the measured %fossil CO2 based on the NNH background is used.

408

C. GEELS ET AL.

References Andersen, G. J., Heinemeier, J., Nielsen, H. L., Rud, N., Thomsen, M. S., Johnsen, S., ´ and Hjartarson, A.: ´ 1989, ‘AMS 14 C dating on the Fossvogur sediments, Sveinbjörnsdóttir, A. Iceland’, Radiocarbon 31(3), 592–600. Antoine, D. and Morel, A.: 1995, ‘Modelling the seasonal course of the upper ocean pCO2 (I). Development of a one-dimensional model’, Tellus 47B, 103–121. Brandefelt, J. and Holmén, K.: 2001, ‘Anthropogenic and biogenic winter sources of Arctic CO2 – A model study’, Tellus 59B, 10–21. Christensen, J. H.: 1997, ‘The Danish Eulerian Hemispheric Model – A three-dimensional air pollution model used for the Arctic’, Atmos. Environ. 31, 4169–4191. Christensen, J., Brandt, J., Frohn, L. and Skov, H.: 2003, ‘Modelling of mercury with the Danish Eulerian Hemispheric Model’, Atmos. Chem. Phys. Discuss. 3, 3525–3541. Ciais, P., Tans, P. P., White, J. W., Trolier, M., Francey, R. J., Berry, J. A., Randall, D. R., Sellers, P. J., Collatz J. G. and Schimel, D. S.: 1995, ‘Partitioning of ocean and land uptake of CO2 as inferred by δ 13 C measurements from the NOAA Climate Monitoring and Diagnostics Laboratory Global Air Sampling Network’, J. Geophys. Res. 100(D3), 5051–5070. Frohn, L. M., Christensen, J. H. and Brandt, J.: 2002, ‘Development of a regional high resolution air pollution model – The numerical approach’, J. Comp. Phys. 179(1), 68–94. Frohn, L. M., Christensen, J. H., Brandt, J., Geels, C. and Hansen, K. M.: 2003, ‘Validation of a 3-D hemispheric nested air pollution model’, Atmos. Chem. Phys. Discuss. 3, 3543– 3588. Geels, C.: 2003, ‘Simulating the current CO2 content of the atmosphere: Including surface fluxes and transport across the Northern Hemisphere’, Ph.D. Thesis, National Environmental Research Institute, Roskilde, Denmark. Geels, C., Christensen, J. H., Hansen, A. W., Kiilsholm, S., Larsen, N. W., Larsen, S. E., Pedersen, T. and Sørensen, L. L.: 2001, ‘Modelling concentrations and fluxes of atmospheric CO2 in the North Atlantic Region’, Phys. Chem. Earth 26(10), 763–768. Goodsite, M. E., Rom, M., Heinemeier, J., Lange, T., Ooi, S., Appleby, P. G., Shotyk, W., van der Knaap, W. O., Lohse, C. and Hansen, T. S.: 2001, ‘High-resolution AMS 14 C dating of post-bomb peat archives of atmospheric pollutants’, in: I. Carni and E. Boaretto (eds), Proceedings of the 17th International 14 C Conference, Radiocarbon 43(2B), 495–515. Grell, G. A., Dudhia, J. and Stauffer, D. R.: 1995, ‘A description of the fifth-generation PennState/NCAR Mesoscale model (MM5)’, NCAR/TN-398+STR. NCAR Technical Note, Mesoscale and Microscale Meteorology Division. National Center for Atmospheric Research. Boulder, Colorado, USA. Gulliksen, S. and Scott, M.: 1995, ‘Report on the TIRI workshop’, Radiocarbon 37, 820–821. Hansen, K. M., Christensen, J. H., Brandt, J., Frohn, L. M. and Geels, C.: 2004, ‘Modelling atmospheric transport of a-hexachlorocyclohexane in the Northern Hemisphere with a 3-D dynamical transport model: DEHM-POP’, Atmos. Chem. Phys. 4, 1125–1137. Levin, I. and Kromer, B.: 2005, ‘The tropospheric 14 CO2 level in mid-latitudes of the Northern Hemisphere (1959–2003)’, Radiocarbon 46(3), 1261–1272. Levin, I., Graul, R. and Trivett, N. B. A.: 1995, ‘Long-term observations of atmospheric CO2 and carbon isotopes at continental sites in Germany’, Tellus 47B, 23–34. Levin, I., Kromer, B., Schmidt, M. and Sartorius, H.: 2003, ‘A novel approach for independent budgeting of fossil fuel CO2 over Europe by 14 CO2 observations’, Geophys. Res. Lett. 30(23), 2194, doi:10.1029/2003GLO18477. Marland, G., Boden, T. A. and Andre, R. J.: 2003, ‘Global, Regional, and National CO2 Emissions’, in: Trends: A Compendium of Data on Global Change. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, TN, USA.


409

Olivier, J. G. J., Bouwman, A. F., van der Maas, C. W. M., Berdowski, J. M., Veldt, C., Bloos, J. P. J., Visschedijk, A. J. H., Zandveld and Haverlag, J. L.: 1996, ‘Description of EDGAR Version 2.0: A set of global emission inventories of greenhouse gases and ozone-depleting substances for all anthropogenic and most natural sources on a per country basis and on 1◦ × 1◦ grid’, National Institute of Public Health and the Environment (RVIM) Report Nr. 771060 002/TNO-MEP Report Nr. R96/119. Pataki, D. E., Bowling, D. R. and Ehleringer, J. R.: 2003, ‘Seasonal cycle of carbon dioxide and its isotopic composition in an urban atmosphere: Anthropogenic and biogenic effects’, J. Geophys. Res. 108(D23), 4735, doi:10.1029/2003JD003865. Ramonet, M., Ciais, P., Nepomniachii, I., Sidorov, K., Neubert, R. E. M., Langendörfer, U., Picard, D., Kazan, V., Biraud, S., Gusti, M., Kolle, O., Schulze, E.-D. and Lloyd, J.: 2002, ‘Three years of aircraft-based trace gas measurements over the Fyodorovskoye southern taiga forest. 300 km north-west of Moscow’, Tellus 54B, 713–734. Rotty, R. M.: 1987, ‘Estimates of seasonal variation in fossil CO2 emissions’, Tellus 39B, 184–202. Scott, E. M.: 2003, ‘The fourth international radiocarbon intercomparison (FIRI)’, Radiocarbon 45(2), 249–275. Sørensen, J. H., Rasmussen, A., Ellermann, T. and Lyck, E.: 1998, ‘Mesoscale influence on long-range transport – evidence from ETEX modelling and observations’, Atmos. Environ. 32, 4207–4217. Stuiver, M. and Polach, H. A.: 1977, ‘Reporting of 14 C data’, Radiocarbon 19(3), 355–363. Sweeney, C., Smith, W. O., Hales, B., Bidigare, R. R., Carlson, C. A., Codispoti, L. A., Gordon, L. I., Hansell, D. A., Millero, F. J., Park, M.-O. and Takahashi, T.: 2000, ‘Nutrient and carbon removal ratios and fluxes in the Ross Sea, Antarctica’, Deep-Sea Res. II 47, 3395–3421. Takahashi, T., Feely, R. A., Weiss, R. F., Wanninkhof, R. H., Chipman, D. W., Sutherland S. C. and Takahashi, T. T.: 1997, ‘Global air-sea flux of CO2 : An estimate based on measurements of sea-air pCO2 difference’, Proc. Natl. Acad. Sci. 94, 8292–8299. Takahashi, H. A., Konohira, E., Hiyama, T., Minami, M., Nakamura, T. and Yoshida, N.: 2002, ‘Diurnal variation of CO2 concentrations and 14 C and δ13C in an urban forest: Estimate of the anthropogenic and biogenic CO2 contributions’, Tellus 54B, 97–109. Tans, P. P. and Wallace, D. W. R.: 1999, ‘Carbon cycle research after Kyoto’, Tellus 51B, 562–571. Yuen, S. A., Higuchi, K., Trivett, N. B. and Cho, H.-R.: 1996, ‘A simulation of a large positive CO2 anomaly over the Canadian Arctic Archipelago’, J. Met. Soc. Japan 74(6), 781–795. Zondervan, A. and Meijer, H. A. J.: 1996, ‘Isotopic characterisation of CO2 sources during regional pollution events using isotopic and radiocarbon analysis’, Tellus 48B, 601–612.