Isotopic composition of carbon in atmospheric air; use of a diffusion model at the water/atmosphere interface in Velenje Basin

CO 2 concentrations (partial pressure of CO 2 , pCO 2 ), and isotope compositions of carbon dioxide in air ( δ 13 C CO2 ), temperature (T) and relative humidity (H) have been measured in the atmosphere in the Velenje Basin. Samples were collected monthly in the calendar year 2011 from 9 locations in the area where the largest thermal power plant in Slovenia with the greatest emission of CO 2 to the atmosphere (around 4M t/year) is located. Values of pCO 2 ranged from 239 to 460 ppm with an average value of 294 ppm, which is below the average atmospheric CO 2 pressure (360 ppm). δ 13 C CO2 ranged from -18.0 to -6.4 ‰, with an average value of -11.7 ‰. These values are similar to those measured in Wroclaw, Poland. We performed the comparison of δ 13 C CO2 values in atmospheric air with Wroclaw since researchers used similar approach to trace δ 13 C CO2 around anthropogenic sources. The isotopic composition of dissolved inorganic carbon


Introduction
Investigation of the fate of atmospheric CO 2 is central to efforts to measure and predict global anthropogenic changes and to assess the impact of fossil fuel usage on environmental quality (EEA, 1998(EEA, , 2003. Analyses of the concentration and anisotropic composition of atmospheric CO 2 have been carried out to assess their anthropogenic impact (Kuc et al., 2003;LonginELLi & SELmo, 2005;PAtAKi et al., 2005;Zimnoch et al., 2004). In the atmospheric boundary layer, the concentration and carbon isotope composition of atmospheric CO 2 (δ 13 C CO2 ) is determined by the mixing of tropospheric air with locally derived air that is affected by 36 anthropogenic and/or biogenic CO 2 sources and sinks (Zimnoch et al., 2004). Biogenic CO 2 originates from plant respiration and from heterogenic soil microbes which convert soil organic matter to CO 2 . Because 12 C is taken up preferentially by plants during photosynthesis, soils are lower in 13 C than the atmosphere (BowLing et al., 2008). Where C3 vegetation (e.g. Filipendulion (with dominant and characteristic species Filipendula ulmaria (L.) Maxim.) and Bidention (species from genera Bidens L., Rorippa Scop., Chenopodium L., Polygonum L.,…), Fagus sylvatica L., Picea abies (L.) Karst., Abies alba P. Mill.) dominates, as is the case for the studied area, soil organic matter and CO 2 respired by vegetation exhibit δ 13 C values between -28 and -20 ‰ (SZArAn, 2002). Values of δ 13 C CO2 derived from burning fossil fuels (anthropogenic sources) range from -40.5 (natural gas burning fumes) to -24.6 ‰ (coal burning fumes) (widory & JAvoy, 2003). Combustion of coal produces almost twice as much carbon dioxide per unit of energy as does the combustion of natural gas, while the amount from the combustion of crude oil falls in between (Energy Information administration, Emissions of Greenhouse Gases in the United States 1985-1990). In the vegetative season the anthropogenic input is minimized and the biological input is dominant (LonginELLi & SEmo, 2005). Values of δ 13 C CO2 and pCO 2 in the atmosphere have also been used to determine pollution levels in the atmosphere (ZwoZ ªdZiAK et al., 2010).
Concentrations of dissolved inorganic carbon, DIC, and its isotopic composition (δ 13 C DIC ) in freshwater environments have been widely investigated (AmiottE-SuchEt et al., 1999;AtEKwAnA & KriShnAmurthy, 1998;mArfiA et al., 2004;KAndu^ et al., 2007) and groundwater/ surface water interactions, with evaluation of biogeochemical processes, have been reported for Velenje Basin (KAndu^ et al., 2010. Here we report measurements of pCO 2 (partial pressure) and δ 13 C CO2 in the vicinity of the Šo{tanj thermal plant which is the biggest emitter of CO 2 to the atmosphere in Slovenia. Thus, around 4 Mt of CO 2 are emitted (EMEP/EEA, 2013) into the atmosphere per year. The aim of this study was 1) to measure monthly air concentrations of pCO 2 and to measure δ 13 C CO2 in air to determine the influence of the combustion of lignite on pCO 2 concentrations and to define the origin of the CO 2 in the air masses in Velenje Basin, 2) to compare pCO 2 concentrations and δ 13 C in air with published data (Wroclaw between 1 st January and 31 st December 2008) and 3) using the concentration and isotope diffusion model to calculate the time of equilibration of CO 2 needed to equilibrate concentrations of pCO 2 and δ 13 C DIC values between air/water interface.

Materials and methods
Partial pressure of CO 2 (pCO 2 ) in the atmosphere was measured above surface water at 9 locations ( Figure 1) in Velenje Basin, using an IAQ-CALC Indoor Air Quality Meter, Model 7545, Thrust Science Innovation (TSI) with an accuracy of ±3 % of reading or ±50 ppm. Air samples for measurement of the carbon isotope composition in carbon dioxide in air (δ 13 C CO2 ) were sampled as follows: a Labco ampoule (4 ampoules per location) was opened in the windward direction to let it fill with air. After filling (about 2 minutes), the ampoule was immediately closed and transported to the laboratory for prompt analysis of carbon isotope composition (δ 13 C CO2 ). Air for δ 13 C CO2 analysis was sampled 2 m above surface water. At the same locations, relative humidity (H), and Figure 1. Sampling locations (10 locations) from Velenje Basin area (river locations: 1, 2, 3, 4, 6 and 8, lake locations: 5, 7, 9). Tja�a KANDU^ 37 temperature (T), in the air were measured monthly during the year 2011. δ 13 C CO2 in air was measured with a Europa Scientific 20-20 continuous flow IRMS ANCA-TG preparation module with an estimated precision of ±0.3 ‰. Working standards calibrated to VPDB (Vienna Pee Dee Belemnite) were used during measurements with a defined value of -3.2 ‰ for CO 2 . Since CO 2 concentrations in air are very low, working standards were diluted to air CO 2 concentrations to optimize peak area. At the same locations surface water samples (additionally at location 3, which was not sampled for δ 13 C CO2 air measurements) were collected seasonally for δ 13 C DIC measurements (Table 1, Figure 1).
Surface waters (lakes and rivers) were measured at 10 locations for alkalinity and δ 13 C DIC (Figure 1). Discharge data were obtained from the Slovenian Environment Agency for the gauging stations: Paka at Šo{tanj, Gaberke at Velunja and Lepena at Škale (intErnEt). Total alkalinity of surface waters was measured according to Gran (giESKES, 1974). The stable isotope content of dissolved inorganic carbon (δ 13 C DIC ) in surface waters (lakes and rivers) was determined on an IsoPrime GV isotope ratio mass spectrometer coupled with a MultiflowBio preparation module.
Phosphoric acid (100%) was added (100-200 µl) to a septum tube and then purged with pure He. A water sample (1 ml) was then injected into the tube and CO 2 measured directly from the headspace. Two standard solutions of Na 2 CO 3 (Carlo Erba and Scientific Fisher), with known δ 13 C DIC values of -10.8 ± 0.2 ‰ and -4.8 ± 0.2 ‰, were used to calibrate δ 13 C DIC measurements (SPötL 2005;KAndu^ et al., 2007). When sampling surface waters, pCO 2 immediately above the surface water was measured in an open system and in a closed system. pCO 2 was measured in a closed system above water as follows. A cardboard box with a surface area of 36 cm 2 and a probe for pCO 2 measurements (IAQ-CALC Indoor Air Quality Meter, Model 7545, Thrust Science Innovation (TSI)) was placed through a hole in a cardboard box and, after 10 minutes (górKA et al., 2011) of equilibration between water and air phase, pCO 2 (partial pressure of CO 2 ) was read.

Results and discussion
Atmospheric data: relative humidity (H), temperature (T), δ 13 C CO2 and pCO 2 in calendar year 2011 with notes on weather conditions are presented in Table 1. Locations from 1-10 are labeled in Figure 1. Air temperature ranged from 5.6 to 35.0 °C during 2011 (Figure 2A). Relative humidity ranged from 18.0 to 78.0 % with an average value of 43.6 % ( Figure 2B). CO 2 concentration in the atmosphere, expressed in �ppm� as pCO 2 and carbon isotope signatures of carbon dioxide in the atmosphere (δ 13 C CO2 ) from the Velenje Basin indicate seasonal variation ( Figures 3A and B). Partial pressures (pCO 2 ) in the atmosphere from 9 different locations range from 239 to 460 ppm -average 294 ppm. The lowest pCO 2 value was recorded at Velunja location and the maximum value at Paka River ( Figure 3A). The values of δ 13 C CO2 range from -18 to -6.4 ‰, depending on the source (Figure 3 B). The δ 13 C CO2 values that approach -6.4 ‰ (location Paka, South Preloge mine) could reflect bacterial CO 2 and/or endogenic CO 2 from underground coalmine activity (LAZAr et al., 2014), while values approaching -18 ‰ (Škalsko and Velenjsko jezero in November 2011) could be attributed to anthropogenic  In a coal burning chimney, δ 13 C CO2 values are -24.1 ‰, exhaust from a gasoline propelled car has values of δ 13 C CO2 of -31.7 ‰, from a diesel car -31.9 ‰ and from a liquid petroleum gas car -33.5 ‰ (górKA et al., 2011). The characteristic value of δ 13 C CO2 for a coal-burning chimney is -24.1 ‰ and is much lower in comparison to δ 13 C CO2 values in our study, where δ 13 C CO2 ranges from -18.0 to -6.4 ‰ ( Table 1).
Seasonal variations of total alkalinity, δ 13 C DIC and pCO 2 (ppm) in surface waters, with pCO 2 (closed system, measurements with cardboard box) measured and pCO 2 measured just above surface water during year 2011 are presented in Table 2. Discharge data (Q) were obtained from the Slovenian Environmental Agency gauging stations for the year 2011 for locations Velunja, Lepena and Paka.
Alkalinity in surface waters changes seasonally from 2.2 to 5.7 mM in January 2011, from 2.6 to 5.5 mM in May 2011, from 2.5 to 6.1 mM in August 2011 and from 2.5 to 5.7 mM in October 2011. δ 13 C DIC changes seasonally from -11.0 to -8.8 ‰ in January 2011, from -11.8 to -7.7 ‰ in May 2011, from -13.5 to -7.1 ‰ in August 2011 and from -12.8 to -9.1 ‰ in October 2011 (Table 2). Higher δ 13 C DIC values would be expected in lake water (standing water) since it equilibrates more quickly than surface water (running water), but it is only the case in lake Velenje (δ 13 C DIC = -7.7 ‰ in spring season). The opposite trend is observed between δ 13 C DIC and alkalinities ( Figure 5A), with the lowest δ 13 C DIC value and the highest alkalinity being observed at location Pe~ovnica (location 2) in January 2011.
Since surface water is an open system, its equilibration with the atmosphere is important. Equilibration lines ( Figure 5A) were calculated according to possible biogeochemical processes influencing δ 13 C DIC value as follows: Line 1. Given the isotopic composition of atmospheric CO 2 of -7.8 ‰ (LEvin et al., 1987) and the equilibration fractionation with DIC of +9 ‰, DIC in equilibrium with the atmosphere should have a δ 13 C DIC of about +1 ‰.
Tja�a KANDU^ Line 3. An average δ 13 C value of -26.6 ‰ for particulate organic carbon (POC) was assumed to represent the isotopic composition of POC that was transferred to DIC by in-stream respiration. Open system equilibration of DIC with CO 2 enriches DIC in 13 C by about 9 ‰ (mooK et al., 1974), which corresponds to a value of -17.6 ‰.
Line 4 represents open system equilibration of DIC, with soil CO 2 originating from degradation of organic matter with δ 13 C CO2 of -26.6 ‰.
From Figure 5A it is observed that most of the samples fall between lines 2 and 3: dissolution of carbonates with an average δ 13 C CaCO3 = -2 ‰ and non-equilibrium carbonate dissolution with carbonic acid produced from soil zone with δ 13 C CO2 of -26.6 ‰. The highest pCO 2 is observed at location Paka (location 10) with a value of 460 ppm (open system), pCO 2 measured value is 480 ppm (measured as a closed system) in October 2011 probably due to higher degradation of organic matter at the end of the summer season. Elevated pCO 2 concentrations are also recorded at Table 2. Carbon species in surface waters (alkalinity, δ 13 C DIC , pCO 2 air-opened system, pCO 2 water/air closed system), discharge data (m 3 /s) and surface water temperature (°C) in the year 2011.
Isotopic composition of carbon in atmospheric air; use of a diffusion model at the water/atmosphere interface in Velenje Basin Pe~ovnica (location 2) with value of 404 ppm in surface water measured in opened system above water and 425 ppm as measured in closed system (in cardboard box) May 2011 ( Figure 5B).

Calculation of fluxes
The CO 2 flux between surface water and the atmosphere �DIC� ex based on a diffusion model (two layer model in which the molecules are transported through a gas film and a liquid layer adjacent to the surface) can be calculated according to the following equation (BroEcKEr, 1974): where D is the CO 2 diffusion coefficient in water with value of 1.26 x 10 -5 cm 2 /s at a temperature of 10 °C and 1.67 x 10 -5 cm 2 /s at a temperature of 20 °C (JähnE et al., 1987), z is the empirical thickness of the liquid layer �cm�, �CO 2 � eq and �CO 2 � are the concentrations of dissolved CO 2 at equilibrium with the atmosphere and with the studied water �mol · cm -3 �, respectively. The thickness of the boundary layer z, a thin film existing at the air-water interface, depends largely on wind velocity (BroEcKEr et al., 1978) and water turbulence (hoLLEy, 1977). D/z, therefore, is the gas exchange rate, which gives the height of the water column that will equilibrate with the atmosphere per unit time. Using a mean wind speed of 4 m/s in all sampling seasons (JähnE et al., 1987), D/z was estimated to be 8 cm/h under low turbulence conditions, 28 cm/h under moderate turbulence conditions and 115 cm/h under high turbulence conditions.   Figure 3A. pCO 2 (partial pressure in air) in the calendar year 2011. Figure 3B. δ 13 C CO2 in the calendar year 2011. Numbers from 1-10 refer to sampling locations. At location 3 only surface water was sampled.

43
Calculation of the CO 2 flux between the river water surface and the atmosphere at the Paka River gauging station, according to equation (1), gives values ranging from 2.6 x 10 -8 to 9.0 x 10 -8 mol/cm 2 h in spring 2011, from 6.0 x 10 -8 to 20 x 10 -8 mol/cm 2 h in late summer 2011 and from 2.7 x 10 -8 to 9.4 x 10 -8 mol/cm 2 h in winter 2011. Taking into consideration the river surface area of 0.40 km 2 (mean width of 10 m and length of 40 km), the total loss of inorganic carbon through its surface in the spring ranges from 6.0 x 10 4 mol/day during periods of low wind speeds to 2.0 x 10 5 mol/day during high turbulence storm events. The predicted total loss of inorganic carbon to the atmosphere in the late summer ranges from 1.0 x 10 5 to 5.0 x 10 5 mol/day and from 6.0 x 10 4 to 2.1 x 10 5 mol/day in winter.

Concentration diffusion model
In addition, values of the time evolution of stream pCO 2 and δ 13 C DIC were calculated using available diffusion models (e.g. BroEcKEr 1974;richEy et al. 1990;Aucour et al., 1999). These calculations yield the amount of time needed for CO 2 evasion and for stream -atmosphere isotopic exchange relative to the transit time of stream waters. Such calculations were performed only for two main tributaries: Velunja River (location 4) and Paka River (location 10) for all sampling seasons ( Figure 1, Table 2). The estimated rate of change of DIC concentration due to CO 2 evasion are calculated by: (2) and the DIC concentration in water is expressed as a function of time by: ( 3) where h is the mean depth of the river �cm� and t is the time needed for equilibration �min�, all other parameters having been determined by equation (1). The calculations assume a value of 8 cm/h for D/z (low turbulent conditions due to low discharge) for both locations (4 and 10) (mooK, 1970) and h values of 10 cm. The computed results, according to equation (3), show that between 0.6 and 2.6 hours (January, 2011), 8.8 and 9.2 hours (May, 2011), 5.7 and 6.4 hours (August, 2011), and from 5.7 to 6.4 hours (October, 2011) would be required for equilibrium between atmospheric CO 2 and dissolved riverine CO 2 to be approached.

Isotopic diffusion model
Additionally, the rate of change of δ 13 C DIC resulting from CO 2 exchange between the river and the atmosphere was also estimated by the equation (Aucour et al., 1999): (4) Again, the DIC concentration (�DIC�) is expressed as a function of time (t) by: In equations (4) and (5), δ 13 C a and δ 13 C DIC are the δ 13 C values of atmospheric CO 2 (-7.8 ‰;LEvin et al., 1987) and DIC, δ 13 C 0 is the initial value of DIC and ε is the equilibrium fractionation factor between CO 2 and HCO 3 - (ZhAng et al., 1995).
Starting with the δ 13 C DIC value of -12.5 ‰ (Aucour et al., 1999) and h value of of 10 cm, calculated time of equilibration ranged from 26.2 to 132.6 hours, which would be needed to equilibrate δ 13 C DIC and δ 13 C CO2 values. This time interval was calculated for Velunja River Figure 4. δ 13 C CO2 levels in the calendar year 2011 compared with those at Wroclaw (górKA et al., 2011). Bold lines indicate the potential anthropogenic sources analyzed in Wroclaw (górKA et al., 2011). The δ 13 C CO2 value characteristic for the absence of pollution is taken from Baltic Sea values (whitE & vAughn, 2009)  Isotopic composition of carbon in atmospheric air; use of a diffusion model at the water/atmosphere interface in Velenje Basin (location 4) and Paka River (location 10) and suggests that stream -atmosphere isotopic exchange alone cannot explain the 13 C enrichment of DIC in this carbonate/clastics catchment. Stream -atmosphere isotopic exchange alone cannot explain the 13 C enrichment of DIC since longer time is needed for equilibration than expected. Both models (concentration and isotopic) should provide same values of time of equilibration, but in our case they do not. However, it has been shown that equilibration of CO 2 between water/air boundaries is more significant in impermeable silicate drainages (KAndu^ et al., 2007). Therefore equilibration of atmospheric CO 2 does not influence the value of δ 13 C DIC in surface waters significantly, which is a consequence of low discharge conditions in the catchment area.

Conclusions
Values of the carbon isotope composition of atmospheric CO 2 (δ 13 C CO2 ), at locations in the vicinity of the thermal power plant in Velenje Basin, have been measured in the calendar year 2011. Based on measurements of alkalinity and δ 13 C DIC for surface water, values of δ 13 C CO2 of air samples taken just above water (opened system) and from a closed cardboard box (closed system) it is concluded that combustion of lignite in thermal power plant has little influence on the δ 13 C CO2 value in the atmosphere. Measured CO 2 concentrations (average pCO 2 value of 294 ppm) and δ 13 C CO2 in the atmosphere in the vicinity (few kilometers) of the thermal power plant are in the normal range in the atmosphere (360 ppm) and the influence of lignite combustion is negligible Figure 5B. Seasonal variation of pCO 2; comparison between pCO 2 air (open system) and pCO 2 water/air (closed system) at 9 locations from Velenje Basin. Normal pCO 2 in air is considered to be 360 ppm. Tja�a KANDU^ 45 at the locations investigated in this study. The values of δ 13 C CO2 in air range from -18 to -6.4 ‰, with an average value of -11.7 ‰, indicating the absence of influence of coal combustion, since the characteristic value of coal combustion is -24.1 ‰. δ 13 C CO2 values in our study (observations during year 2011) are similar as obtained for Wroclaw, Poland (observation during year 2008).
The total alkalinity in surface waters ranged from 2.2 to 6.1 mM. Dissolution of carbonates and degradation of organic matter are the most important biogeochemical processes affecting δ 13 C DIC . They range seasonally from -13.5 to -7.1 ‰ in the surface waters (lakes, rivers) investigated in this study. pCO 2 in the air immediately above water (open system) and in the air above the water, measured in the cardboard box (closed system), is similar at all measured locations. The highest pCO 2 in an open system -immediately above water-and in a closed system (measured in a box) were measured at Paka (location 10) and Pe~ovnica (location 2) in May 2011 and in October 2011, respectively. Both locations are located in the vicinity of the thermal power plant. Based on thermodynamic modelling and on previous studies reported for Slovenian watersheds (rivers and lakes), surface waters acted like sources of CO 2 (oversaturated more than 10 times) released to the atmosphere. However, the measurements of pCO 2 reported here were made just above the surface water, where normal values of pCO 2 (around 360 ppm) are present.
Two diffusion models (concentration and isotopic) were applied to obtain the time of equilibration at two locations. Between 0.6 and 6.4 hours were required to equilibrate atmospheric CO 2 and dissolved riverine DIC (concentration diffusion model), and 26.2 to 132.6 hours to equilibrate δ 13 C DIC and δ 13 C CO2 values (isotopic diffusion model) if equilibration with atmospheric CO 2 was the only factor influencing DIC values of surface waters.
Even though Velenje Basin is a natural analogue with very large amounts of endogenic and bacterial CO 2 (with the characteristic value of δ 13 C CO2 � 2 ‰) and with large amounts of CO 2 emitted (around 4 Mt/year) from lignite combustion from the thermal power plant, we conclude from this study that pCO 2 concentrations in air around the thermal power plant are not elevated.