Regional Representativeness of CH 4 and N 2 O Mixing Ratio Measurements at High-Altitude Mountain Station Kasprowy Wierch , Southern Poland

Records of trace gas composition of the atmosphere originating from mountain stations are difficult to interpret in terms of continental gas balances due to complex topography and local wind patterns. Kasprowy Wierch station, located in the High Tatra Mountains, Poland, provides datasets to European GHG databases and can be used for regional and continental balances of trace gases in Earth’s atmosphere. Although it suffers from strong influence of local valley winds, the concentration record obtained at Kasprowy Wierch contains also a discernible fraction of methane (up to 100 ppb) emitted mainly by large regional source of this gas (Silesian Coal Basin) located ca. 150 km north-west of Kasprowy Wierch. Comparison of GOSAT satellite CH4 record in the pixel including High Tatras at the level of 850 hPa with the data obtained at Kasprowy Wierch mountain station revealed surprisingly good correlation, even for relatively short time intervals.


INTRODUCTION
Methane (CH 4 ) is one of the major greenhouse gases actively changing radiative forcing of the Earth's atmosphere (Hofmann et al., 2006;Myhre et al., 2013).During the last three decades, globally averaged trends of methane concentrations revealed three distinctive phases of change: (i) strong increase from 1984 to 1992, (ii) a period of relative stability during 1992-2008, and (iii) renewed significant increase since 2009 (Dlugokencki et al., 1994;Rigby et al., 2008;Dlugokencki et al, 2009;Kai et al., 2011).While the main sources and sinks of CH 4 are relatively well constrained on the global scale, the reasons behind the stabilization of atmospheric CH 4 levels during the first several years of the 21st century are still under debate, underlining the need for better understanding of its sources and sinks on local and regional scales.Therefore, there is a strong need of for dense and reliable CH 4 monitoring networks over the areas with large natural and anthropogenic sources of CH 4 (Zhu et al., 2015).
Nitrous oxide (N 2 O) is another critical component of the Earth's atmosphere as it can affect both the radiation budget (WMO, 2013) and the stratospheric ozone layer.Since the implementation of the Montreal Protocol in 1989, successful reduction of CFCs and HCFCs made N 2 O the most important anthropogenic ozone depleting substance (ODS) (Ravishankara et al., 2009).Global and regional budgets of N 2 O are still uncertain, pointing to the need of further studies of its emission and transport patterns on various temporal and spatial scales.
The United Nations Framework Convention on Climate Change (UNFCCC) reported a high contribution (15.4%) of both gases to the CO 2 -equivalent of total greenhouse gas release in Europe (EEA, 2014).Moreover, the report indicates large reduction of continental emissions of those gases over the European continent since the implementation of the Kyoto Protocol -by 33% in case of methane and 34% (CO 2 -eq.) for nitrous oxide (2012 compared to 1990).
It has been shown that monitoring networks can be a sensitive tool for verifications of the greenhouse gas balance over the European continent (Bergamaschi et al., 2015).There are numerous monitoring stations measuring atmospheric concentrations of non-CO 2 greenhouse gases in Europe.Most of the monitoring sites are ground-level stations or tall towers that are located in the lower troposphere, which makes them sensitive to local surface emissions of pollutants.However, it also makes it hard to discern weak signals from distant sources, and makes it necessary to take into account the diurnal cycle of the planetary boundary layer when analysing the results.On the other hand, there are some examples of using mountain stations for greenhouse gas studies, where high elevation facilitates measurements of the free-tropospheric air composition and allows for significantly larger footprints and sensitivity to changes in continental balances of trace gases (Berchet et al., 2013, Sturm et al., 2014, Graziosi et al., 2015).However, the remote location of such stations usually demands high infrastructure requirements, which in turn decreases the number of potential sites.Currently, European mountain stations with at least a decadal record of methane observations are located in Pyrenees (Pic du Midi), Apennines (Monte Cimone), Alps (Jungfraujoch, Plateau Rosa, Zugspitze and High Tatras (Carpathians), with the (Kasprowy Wierch -KAS) being the easternmost European high-altitude site (Angelbratt et al., 2011;Weaver et al., 2014;Bergamaschi et al., 2015).
The main difficulty in analysing atmospheric composition at high-mountain stations arises from the complexity of mountainous terrain, which is very challenging for numerical modelling of transport processes.High-altitude sites can be affected by valley winds blowing along the terrain slope and advecting the air masses enriched (or depleted) in pollutants emitted at lower altitudes.This effect was shown to be important for diurnal variation of greenhouse gases concentrations measured at KAS (Nęcki et al., 2003;Nęcki et al., 2013), and is not fully reproduced in transport models even if high spatial resolution (2 km × 2 km) is used in the simulations (Pillai et al., 2011;Sheridan and Vosper, 2012).
The central aim of this study was to quantify the influence of local and regional sources of two major non-CO 2 greenhouse gases (methane and nitrous oxide) on the atmospheric concentrations of these gases recorded at the KAS station with the use of several measurement methods and a supporting numerical modelling framework utilizing the STILT backward trajectory model that allowed to assess the regional footprint of the station, and to estimate the impact of individual sources on the measured mixing ratios of those trace gases for the first time.The presented results provide initial information necessary in characterizing the regional representativeness of the results obtained at this station.

METHODS
Kasprowy Wierch is a high-altitude meteorological station established in 1936 in northwestern High Tatra mountain ridge, Western Carpathians, Poland (49°14'N, 19°59'E, 1989 m a.s.l.-Fig.1), named after the peak on which the laboratory building is located.As the station is situated in the convergence area of three large valleys, approximately 300 m above the tree line, anabatic and katabatic wind events are frequently observed.Wind statistics available for the station indicate prevalence of southerly and northeasterly wind directions, caused by two of the valleys originating from Kasprowy Wierch summit (Gąsienicowa valley and Cichej valley) channelling the air masses in these directions.
Regular trace gas measurements at Kasprowy Wierch meteorological station begun in 1994.In 1996, an automated gas chromatograph (Agilent HP5890) equipped with FID (Flame Ionization Detector) and ECD (Electron Capture Detector) detectors was installed and quasi-continuous analyses have been carried out ever since (Nęcki et al., 2003;Chmura et al., 2008;Nęcki et al., 2013).Results obtained until year 2003, hasn't been linked properly to the international scales via set of reference materials used at the station.Currently, in situ analyses of ambient concentrations of CH 4 , N 2 O, are performed approximately every 20 minutes.Since 2003 all results are accurately calibrated and are traceable to the international primary scales (WMO CH 4 X2004 for methane and WMO N 2 O X2006A for nitrous oxide), with inter-comparison programs conducted in the framework of several EU Projects as a part of quality control procedures at the station.KAS station is currently delivering near real time data (NRT) in the framework of the InGOS project (http://ingos-atm.lsce.ipsl.fr/NRT).
Configuration of the measurement system, described in detail in (Trivett et al., 1989, Pepin et al., 2001), has undergone several modifications since 2009, in particular the exchange of the measurement column from Porapak Q (Agilent) to Carboxen1000 (Supelco) on the FID (CH4) line and Porapak N to Hayesep Q on the ECD (N 2 O) line.These modifications allowed for enhanced separation stability and increased resistance to impurities in the carrier gas.
In addition to in situ measurements, flask samples are also collected in biweekly time intervals, with occasional event triggered 24 hours collection campaigns.In those samples, isotopic composition of CH 4 is analysed (Różański et al., 2014).In January 2015, a cavity ring-down spectrometer (CRDS, Picarro, 2201-i) was installed at the station.Methane concentration and isotopic composition (δ 13 C-CH 4 ) is recorded with a frequency of approximately 0.5Hz.
In order to support the interpretation of the measurement results obtained at KAS site, the Stochastic Time-Inverted Lagrangian Transport (STILT) model (Lin et al., 2003) was used, which not only allows to describe the extent of the region that influences the short-term variability of N 2 O and CH 4 concentration record at KAS, but also to estimate the magnitude of that influence.The model calculates backward time particle trajectories to quantitatively describe the sourcereceptor relationship via calculation of a footprint (f) function, which for discretized calculations can be expressed as: where: x r , t r -space and time coordinates of the receptor point x i , y i , t m -space and time coordinates of the emission point, N tot -number of particles released from the receptor, m air -molar mass of air h -turbulent mixing height (below which we assume that air is perfectly mixed); only particles below that height are contribute to the footprint function, p,i,j,m -indexes of particles (p), space coordinates (i,j) and time step (m) in numerical calculations.The total station footprint was calculated as a sum of footprint elements, each derived from 10 days long backward simulations with 100 particles released.The releases were programmed to start every 3 hours, since 01.01.2011 until 01.10.2014(excluding the data from 11th to 26th of June 2013, for which the meteorological driving fields were not available).An aggregation method described in Gerbig et al. (2003) was used to minimize the impact of the low density of particles in the far field.The parameter h was chosen as half of PBL height in Eq. ( 1), allowing to fully resolve diurnal variations in footprint strength due to changes in vertical mixing depth (Gerbig et al., 2003).Meteorological fields from the ECMWF Integrated Forecasting System (IFS, 0.125 deg × 0.125 deg horizontal resolution, 3 hourly temporal resolution, and 137 vertical levels) were used to drive the trajectory computation.The KAS receptor point was located at 700 m above ground level in the model domain to reflect the station's elevation above mean sea level.
Given a grid of flux (F) values in units of micro-mol m -2 s -1 , a single-release footprint connects the concentrations measured at the receptor point with these fluxes by the relationship: | , ,  ( , , ) where:   , ,j , concentration that is added to the lateral boundary conditions, i.e., the initial concentrations of the pollutant in the simulated air parcel.F(x i , y j , t m ) is the spatial and temporal distribution of pollutant emissions.
In the framework of this study, an attempt was made to estimate relative influences of chosen emission sources on respective ambient concentrations of those gases measured at the KAS station.To do that, IER (Institute for Energy Economics and the Rational Use of Energy, Univeristy of Stuttgart) inventory was used (Pregger et al., 2007).This inventory quantifies anthropogenic CH 4 and N 2 O emissions over the European continent with high spatial (10 km × 10 km for European Union area, 1° × 1° other countries) and temporal (hourly) resolution.
Simulated backward simulations calculated for 2011-2014 were coupled to three versions of the initial IER CH 4 and N 2 O datasets, providing three distinct emission scenarios: for "BASE" scenario, no modifications to the IER emission databases were made.For "SCB" scenario, the CH 4 database was modified to include only the sources from the Silesian voivodoship, with the rest of the emission grid values set to zero.For simulation marked as "ZAT", N 2 O database was changed so that only emissions from chemical industrial facility in Tarnów (Zaklady Azotowe -ZAT, 20.9°E, 50.0°N), the largest single industrial source of N 2 O in the region, were considered.Additionally, a more detailed set of numerical experiments was performed for selected events, with hourly releases of 1000 particles simulated for 10-day backwards.In all the simulations performed in the presented study, lateral boundary conditions were set to zero, i.e., it was assumed that the initial concentrations of the pollutant inside each of the simulated air parcels was equal to zero.Thus, the described modelling framework only predicted the concentration increase due to emissions from the surface sources located inside the modelling domain (i.e., Europe).
To test the spatial representativeness of KAS CH 4 data record, a comparison with the Greenhouse Gases Observing Satellite (GOSAT) product L4B has been made.GOSAT was launched in 2009 and since that time it delivers CO 2 and CH 4 data representing spatial and temporal variability of the total column mixing ratios of these gases (level 2 data) in the Earth atmosphere derived from radiance spectra obtained by spacecraft instruments (level 1 data).In the subsequent processing steps, the column averaged mixing ratios data are interpolated, extrapolated and smoothed forming level 3 data product.The most sophisticated GOSAT data product (level 4B) presents global CO 2 and CH 4 concentrations in three dimensions with horizontal resolution of 2.5 × 2.5 degree, several pressure levels and a time step of six hours.This data are the result of inverse estimation from the FTS SWIR Level 2 column averaged mixing ratios and ground based observational data using a global atmospheric transport model (Provided by JAXA/NIES/MOE, http://www.gosat.nies.go.jp).

RESULTS AND DISCUSSION
The 10-year record of atmospheric methane concentration available for KAS mountain station represents both shortand long-term variability of the CH 4 mixing ratios in the continental air masses, superimposed on its global growth rate trend (Fig. 2).Comparison of the KAS long-term trend with analogous trend curve available for the marine background station representing north hemispheric midlatitudes (Mace Head -MHD, west coast of Ireland, 53°20'N, 9°54'W) reveal similar growth rates of CH 4 , though with elevated values at the continental site (KAS).Differences between both stations, possibly explained by enhanced releases of methane from continental sources over European continent indicate how important is a catchment area of the particular station.KAS, situated over the mountain ridge in Central Europe will be influenced by the sources distributed over the continent and regional signal can't be filtered from the record.For the purpose of comparison with Mace Head data, extraction of nocturnal values where applied.Only this part of data set was applied to calculate the trend Fig. 2. Long-term record of atmospheric methane concentration at Kasprowy Wierch.Dots represent daily means of CH 4 concentration filtered with the procedure described in Necki et al. (2003).Black line shows the calculated trend line for Kasprowy Wierch while light grey line represents the trend line for the data record available for Mace Head station.Both trends are calculated using CCGvu 4.40 procedure recommended by NOAA (Thoning et al., 1989).Mace Head data originate from AGAGE database (http://agage.mit.edu;Prinn et al., 2000).line of KAS while MaceHead data set includes only marine fraction.During the day the air over the mountainous area is frequently affected by elevated values of species emitted both, locally and regionally (Forrer et al., 2000;Henne et al., 2008), transported to the mountain top by the meteorological circulation (Forrer et al, 2000;Henne et al., 2004b) affecting representativeness of the station.CH 4 record available for KAS station was also used to test the CHIMERE model sensitivity to methane release over the European continent with satisfying results, indicating high sensitivity of the mountain stations (KAS, Jungfraujoch) to the lateral boundary conditions (Berchet et al., 2013).
Fig. 3 presents a seasonal cycle detectable in the atmospheric CH 4 concentrations at KAS station together with validated GOSAT level 4B dataset retrieved for the 850 hPa level, representing approximate altitude of the KAS site.Comparison of in situ and satellite data reveal the same patterns in the annual cycle and no significant bias, thus providing an independent verification for satellite observations.Short-term variability of methane concentration is also well represented by the GOSAT data.Fig. 3 contains also an example of synoptic variability of CH 4 concentration at KAS station recorded during March 2011.Even small fluctuation are occurring parallel in both records and large part of datasets reveal remarkably high similarity.Importance of lower tropospheric sources is clearly visible in the GOSAT data, even though at single altitude level the most important regional sources of CH 4 are not spatially distinguishable due to size of the grid cell (2.5° × 2.5°).If the retrieval from the surface level (975 hPa level in Fig. 3) is considered, the amplitude of CH 4 variability observed in the record increases approximately by a factor of 2 when compared to KAS data.This is caused by presence of large CH 4 emission sources in far field, such as Silesian Coal Basin (SCB) and Kraków agglomeration in the same spatial grid cell.It is difficult to use the satellite data to interpret the short-term variability of CH 4 at ground level because surface observations performed in Kraków and inside SCB indicate local mixing ratio values reaching even 5 ppm (Zimnoch et al., 2010).
The effect of sudden ups and downs observed in the short-term variability of the CH 4 record occurring on the time scale of several hours to few days can be attributed to two distinctly different transport pathways.The first pathway is linked with local emission sources (near-field).Methane released by those sources is subsequently transported to the measurement site via advection controlled by orographic features, observed at most of mountain stations (Seibert et al., 1998;Forrer et al., 2000;Henne et al., 2004b).It has been shown previously that this phenomenon can cause diurnal amplitude of methane concentration reaching 30 ppb during winter months at KAS (Necki et al., 2003).The second pathway is associated with occurrence of specific synoptic situation that causes the air masses to be transported over the strong regional sources of CH 4 and N 2 O, such as SCB for CH 4 and ZAT for N 2 O.These air masses are subsequently advected toward Carpathians, and while they pass over Tatra Mountains, elevated concentrations of those gases are recorded at KAS.Such situations occurred frequently during the years 2011-2014.Identification of these two classes of events is possible by retrospective analyses of regional atmospheric circulation with the aid of back trajectory modelling or fully-fledged, 3D transport models used to calculate the footprint of the station (Seibert et al., 1998, Kaiser, 2009).
The footprint of KAS station calculated using STILT simulations (Fig. 1) shows that this site can be used for the estimation of the source strength distributed over the region of central Europe, covering Poland, Slovakia, large part of Czech Republic, Hungary, and, to lesser degree -Austria, eastern part of Germany, Romania, Balkan countries and Ukraine.The surface area covered by KAS footprint is comparable with those, calculated for Alpine stations (Kaiser, 2009).
From the perspective of this study the coverage of Silesian Coal Basin (SCB, Fig. 4) is particularly interesting as it is a source of very large methane releases, mostly associated with coal mining activities.The whole region of SCB, located in the Silesian District, has approximately 1.5 × 1.5 degrees of latitudinal and longitudinal dimension and is responsible for between 450-1350 Gg CH 4 of atmospheric methane releases annually (EDGARv4.2.FT 2010-Olivier and G.Janssens-Maenhout, 2012; Patynska, 2013), making it one of the most significant sources of this gas in Europe.Therefore, demonstrating the influence of SCB region on the methane level observed at KAS would confirm the calculated footprint extent of this mountain station.
One of the largest regional sources of N 2 O is large chemical plant ("Zaklady Azotowe Tarnów" (ZAT), part of Grupa Azoty S.A).The plant, located approximately 150 km NE and 1600 m below the KAS station (Fig. 5), produces fertilizers, nitric acid as well as ammonium nitrate, responsible for large amounts fugitive N 2 O releases.Annual The above-outlined sources of CH 4 and N 2 O were considered in the STILT simulations.Fig. 6 shows the SCB CH 4 fraction arriving at KAS station, as simulated by STILT coupled with IER emission database, for the whole surveyed period (2011)(2012)(2013)(2014).The model simulation show that the SCB overlaps with the calculated station footprint for 57.8% of the time.When the enrichment associated with SCB is present, it usually comprises a small part (median: 1.13%) of the total European CH 4 enrichment (base scenario).There are cases, however, when the SCB fraction is larger than 50%, with predicted CH 4 enhancements higher than 20 ppb -well in the detectability range of available GC instrumentation.It is important to note that the obtained simulation results cannot be directly compared with the ambient observations, as a) only anthropogenic sources were considered in the numerical calculations, b) spatial resolution of the driving meteorological fields is insufficient to fully reproduce the local effects (e.g., valley winds), c) lateral boundary conditions were not included, therefore only enhancements arriving from the modelled domain could be reproduced (while changes in the global and background values were not considered).The result of CH 4 simulation prove, however, that STILT-IER can successfully be used to indicate the events for which the peaks in the observed CH 4 concentrations can be correlated with emissions from a particular source, even despite relatively coarse resolution.
An attempt was also made to estimate the influence of emissions from a strong, single point-source on the N 2 O concentrations measured at KAS (Fig. 7).This has proven to be difficult, however, with predicted values of N 2 O mixing ratios, usually remaining below 1 ppb, even for base simulation scenarios, where all available European sources   were included.This is in line with identified difficulties in application of gas chromatography to analyse the European N 2 O balance (Corazza et al., 2011), and underlines need for analysers with enhanced performance, characterised by reproducibility level an order of magnitude lower than that offered by present GC methodology, i.e., in the order of 0.02 ppb.Nevertheless, N 2 O concentrations measured at KAS station indicate presence of large surface sources inside the footprint of the station, which is confirmed in existing bottom-up inventories, e.g.IER and EDGAR (Fig. 5).
The STILT simulations indicate that the events when the air masses arriving at KAS have been influenced by ZAT emissions are not frequent (37.1%,Fig. 1).This is not surprising, as (i) the spatial extent of this particular source is comprised of a single grid cell and (ii) ZAT is located NE of KAS site, not on a preferential, westerly direction.Only several cases when the fraction of ZAT-emitted N 2 O comprised more than 5% to the total enhancement predicted by STILT were recorded during the 2013-2014 period.Median value of that fraction for the presented period was very low -only 0.04%.Even when all European sources of N 2 O present in the IER database were considered, absolute value of the predicted enhancement was below 0.1 ppb, the level which remains imperceptible for the GC set-up operated at KAS site (detection limit of approximately 0.15 ppb increase), for in all but 3 simulations.
The predicted N 2 O ZAT fraction may be underestimated due to the low values of N 2 O releases values attributed to this source by the applied emission database (IER).If the emissions from EDGAR v. 4.2 were used, then the median fraction of ZAT N 2 O would be 0.44%, with the maximum of almost 88%.Regardless of that enhancement, predicted N 2 O enhancement attributed to ZAT would still remain below 0.1 ppb and ZAT emissions would need to be at least an order of magnitude larger to allow its direct detection at KAS.However, STILT in combination with IER can still be used as a qualitative indicator that can identify synoptic conditions which potentially may allow for the occurrence of ZAT-caused events at KAS site.The real emission rates of ZAT might be underestimated by the IER database, but were nevertheless used in the STILT model to calculate the fraction of N 2 O observed at KAS that originated from this plant.Several such cases were identified in the presented period, however no clear N 2 O signal could be detected at KAS during these events.In the subsequent discussion we focus on several events during which air masses were travelling over SCB (CH 4 ) and subsequently towards the KAS site.Fig. 8 shows an example of such event, observed at KAS in June 2014.The upper panel shows methane concentration recorded at KAS between June 16 th and 20 th , together with partitioned fraction originating from SCB as predicted by STILT simulation.For three simulations starting at 12:00 on June 16 th , 01:00 on June 17 th and 09:00 on June 18 th that are marked by arrows, the calculated footprint is displayed in the lower panel.The chosen points in time represent synoptic situations that occurred before, during and immediately after the elevated CH 4 concentrations were recorded at KAS station.Throughout the event, Kasprowy Wierch was under the influence of air masses of Scandinavian origin, crossing European continent in the southward direction.Methane concentrations rose from the initial 1900 ppb level (16 th of June, 14:00) to ca. 1970 ppb (12 hours later) and dropped back to 1900 ppb on June 18 th .Second, more pronounced peak of CH 4 was observed late during the day on June 18 th .
The analysis of synoptic situation at the site alone cannot distinguish if the recorded CH 4 concentration increases occurred due to local or regional influences.Application of STILT, however, allows for a more detailed analysis of the factors controlling the observed changes in CH 4 concentration.As can be seen from calculated footprints, the advection paths for the particles travelling in the lower troposphere, and thus directly influenced by surface emissions, is different for each of the three marked particle releases.Bottom left and bottom right panels in Fig. 9 clearly indicate that the model predicts the surface influence areas mostly to the east and north of SCB, while the middle panel, correlated with high CH 4 observations at KAS, shows high relative influence of the SCB source.Further insight can be gained when spatial and temporal distribution of particles from a single release are analysed (not shown).The second example of an event, observed in July 2013, with a clear identification of SCB contribution to CH 4 levels recorded at KAS is shown in Fig. 9. Again, the modelled fraction of methane originating from SCB shows a good similarity with the CH 4 enhancements observed at the station.The synoptic situation was similar to that in the previous example, with prevailing northerly winds maintained for several days.One notable difference, however, is the observed night time decrease of CH 4 concentration due to an inversion occurring over the SCB region, causing reduced mixing between the observed air mass and the air enriched in CH 4 locked below.This can be detected in a form of an edge in the footprint functions displayed in the lower panel of Fig. 10 (for 06:00, July 23 rd and 04:00 on July 24 th ), and is also confirmed by more a detailed analysis of particle transport (not shown).
Both these events prove that the location of the mountain station in the vicinity of strong surface sources of methane reveals discernible influences of those sources on the concentrations measured at the station.In both events results of the simulations revealed that SCB emission is responsible for approximately 50% of methane enrichment signal originating from all sources included in IER database.This however, doesn't reflect whole variability of methane concentration observed at KAS site.Possibly IER emission database underestimate the efficiency of the sources, Large areas of biogenic emission (e.g., peatlands of Nowy Targ) were not included in simulations.Furthermore, our aim was not to prove that model is coherent with recorded data but to demonstrate that part of the data can be attributed to single, strong surface source of methane.
To complement the information on the variability of CH 4 observed at KAS, we have used a Picarro analyser to monitor short-term fluctuations of its ambient concentrations.It has been previously demonstrated that anthropogenic CH 4 emissions from local sources Podhale region (e.g., peatlands, waste disposal sites) can be traced in the record of atmospheric mixing ratios available for the site (Chmura et al., 2009).
Typically, the influence of these local emissions is pronounced by a sudden change of amplitude of small CH 4 concentrations oscillations superimposed on a rapid increase  in the baseline level.These events are usually observed at breakdown of night time thermal inversion, but can also reflect a change in the synoptic situation, e.g. during low pressure advection or weather front passage.In both cases, the mountainous area experiences an enhanced ventilation through valley winds (Henne et al., 2004), which causes not only an increase in the level of methane concentration observed at KAS but also the amplitude of short time fluctuations.Without a high-frequency signal, the analysis of CH 4 GC record cannot give a full explanation behind the driving phenomena.
An example of application of CRDS analyser, when an event with a strict pattern of methane fluctuations was revealed can be found on January 23 rd (Fig. 10).During the first 9 hours of the day, southerly wind experienced at KAS has stabilized the CH 4 concentration approximately at 1900 ppb level, with a very low noise (smaller than 5 ppb).Around 10:00, the wind pattern has shifted to north-easterly, forcing the lower tropospheric air, enriched in CH 4 , to be advected along the mountain slopes, causing the observed concentrations to increase by almost 100 ppb.Apart from the overall increase in the background level, a significantly higher noise (up to 60 ppb), caused by the turbulent passage of large eddies, is also observed.
Without the detailed, high-frequency information delivered by the state-of-the-art instruments, interpretation of the described event would have been very difficult, and the possibility of attributing the observed variability to the instrumental malfunction cannot be excluded.This example underlines the complexity of the transport phenomena observed at the mountain stations that still cannot be fully represented by neither satellite nor numerical modelling.

CONCLUSIONS
KAS station located in the High Tatra Mountains, southern Poland, is the easternmost high altitude station in Europe where regular measurements of atmospheric mixing ratios of CH 4 and N 2 O are being performed.The footprint of this station calculated using STILT model shows that the site can be used for estimation of the source strength of those gases distributed over the region of central Europe, covering Poland, Slovakia, large part of Czech Republic, Hungary, eastern part of Germany, Romania, Balkan countries and Ukraine.Station is capable to have large influence on determination of European carbon budget.
Simulation of the impact of large regional sources of CH 4 on atmospheric mixing ratios of this gas measured at KAS, reveals that the record may contain discernible fraction of methane (up to 100 ppb) emitted by the Silesian Coal Basin (SCB) located ca. 150 km north-west of the station.At current state of analysis only anthropogenic sources of methane were included in simulations, thus methane mixing ratio reproduced by model remains underestimated.Large area of peatlands situated between SCB and KAS may act as substantial source of this gas.Current investigations are aiming at quantification of methane flux over this area will hopefully result with a better coherence of simulated and measured signals at KAS. Partition of methane signal at KAS can be also strengthened adding isotopic composition of CH 4 to the balance.This should be possible if Picarro analyser will work at the station during next year.
Although nitrous oxide emitted by the large chemical plant located approximately 150 km north-east of Kasprowy Wierch is generally not traceable with the analytical technique in use at this station (gas chromatography), the model simulations suggest that single, short-term episodes of elevated N 2 O levels associated with particularly strong pulses of N 2 O released by the plant can still be detected.
Comparison of GOSAT satellite CH 4 retrievals with the methane record available for KAS station revealed remarkably high correlation, both on seasonal and synoptic scale.The annual CH 4 cycle was well reproduced, thus providing an independent verification for satellite observations.Even short-term variability of methane concentration recorded at KAS appeared to be well-represented in the GOSAT data.Further computational and statistical analysis is required to resolve and understand parts of the records when we observed discrepancies between GOSAT and KAS.This especially might be helpful for determination of local sources influence on KAS record.

Fig. 1 .
Fig. 1.Geographical location of Kasprowy Wierch -KAS mountain station in Western Carpathians at the border of Poland and Slovakia (black triangle).Background of the map illustrates footprint of KAS calculated with STILT model for the years 2009-2014.Total footprint is derived from 10-day long backward simulations with 100 particles released every 3 hours.Relative scale (in ppb × μmole -1 × m 2 × s units) represents the probability of the concentration measured at KAS being influenced by sources located in a given grid cell.Inserted panel shows the wind rose for KAS station, averaged for the years 1945-2014.

Fig. 3 .
Fig. 3. Comparison of three years records of diurnal averages of methane concentration obtained at KAS station (black dots) with the dataset available from GOSAT level 4B (thick grey line for the level 850 hPa and light grey line for the 975 hPa level).GOSAT data provided by JAXA/NIES/MOE, http://www.gosat.nies.go.jp.).The inserted panel shows an example of short-term variability of all three records during March 2010.In this case single dot represent single CH 4 measurement.
N 2 O releases by ZAT reported in various databases for the past 15 years differ significantly: 187 Mg N 2 O for 2000 (IER), 2010 Mg N 2 O for 2010 (EDGARv4.2.-Source: EC-JRC/PBL.EDGAR version 4.2.http://edgar.jrc.ec.europa.eu/,2011., used in Fig. 5) or 1530 Mg N 2 O for 2012 (E-PRTR).It is unclear at this point if this variability comes from physical changes in production rates or different methodologies of bottom-up estimations used in these databases.

Fig. 4 .
Fig. 4. Methane sources according to the EDGAR 4.2 inventory for central Europe (Source: European Commission, Joint Research Centre (JRC)/Netherlands Environmental Assessment Agency (PBL).Emission Database for Global Atmospheric Research (EDGAR), release EDGARv4.2FT2010, http://edgar.jrc.ec.europa.eu,2013.).Black thick borders mark Silesian district, where majority of the Silesian Coal Basin is located.The KAS station is marked with a triangle.

Fig. 5 .
Fig. 5. Emissions of nitrous oxide in the central Europe (main panel) and Malopolska district (upper left small panel) according to EDGAR 4.2 (Source: EC-JRC/PBL.EDGAR version 4.2.http://edgar.jrc.ec.europa.eu/,2011.).Black borders mark Silesian district.KAS station is marked with a triangle.ZAT -industrial facility marked with black arrow.Shading of the squares reflects N 2 O emission rates (scale on the right-hand side).

Fig. 6 .
Fig. 6.Methane concentration measured at KAS station between 2011 and 2014 (black dots -left axis) and modelled fractions of methane originating from all continental sources of this gas located within the footprint of this station (light grey bars -right-hand scale).Dark grey bars represent fraction of CH 4 attributed by the model to Silesia Coal Basin emissions.

Fig. 7 .
Fig. 7. Nitrous oxide concentration record from Kasprowy Wierch station covering the period April 2013-July 2014 (black dots -left-hand scale).Modelled fraction of N 2 O originating from all European sources of this gas (light grey barsright-hand scale) and the fraction attributed to ZAT source (dark grey bars -right-hand scale).

Fig. 8 .
Fig. 8.An example of high CH 4 concentration events recorded at the KAS station in June 2014.Upper panel shows changes in methane concentration during 4 consecutive days, measured at KAS station (left-hand scale) and fractions originating from SCB alone and from and all European sources assigned by the STILT model (right-hand scale).Lower panel shows footprint changes as the event progresses in time (footprint calculated for the point in time marked by arrows).

Fig. 9 .
Fig. 9. Another event recorded at the KAS station in June 2013.Upper panel illustrates methane concentration variability during 3 consecutive days measured at KAS (left scale) and fractions originating from SCB and all sources assigned by the STILT model (right scale).Lower panel illustrates footprint changes for moments marked by the indication arrows.

Fig. 10 .
Fig. 10.Short-term variation of methane concentration recorded at KAS station in 23 January 2015.Black circles represent gas chromatography analyses results.Grey line reflects high frequency analyses generated by CRDS instrument.