SEARCH

SEARCH BY CITATION

Keywords:

  • FLEXPART model;
  • Shangdianzi regional atmospheric background monitoring station;
  • GFS and ECWMF data;
  • characteristics of source distribution

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

The FLEXPART model is a Lagrangian particle dispersion model developed by the Norwegian Institute for Air Research. In this paper, first the model simulation driven by GFS and ECWMF meteorological data is compared and evaluated with the observations of Shangdianzi station, then, taking CO as the tracer and adopting the method of screening background and non-background concentration, the FLEXPART model is used in tracing clean and polluted source areas of observation station. The comparisons of the simulations using GFS and ECWMF with the observations show that the two results using GFS and ECWMF are alike, and both can demonstrate the trend of CO concentration and better reflect the daily and monthly variation trend of CO. The correlation co-efficients of simulation value and actual measurement value of CO concentration are 0.67 and 0.71 for GFS and ECWMF respectively. Both simulations are slightly lower than the observations. Also the simulation discovers that the north-westnorthwest-northwest sector is the most important background sector of CO while the westsouthwest-south sector is the most important non-background sector for Shangdianzi station. The model can capture the distribution characteristics of background and pollution source areas of Shangdianzi station and better reflect the distribution of CO source areas in Beijing and the surrounding area. Copyright © 2013 Royal Meteorological Society


1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

The FLEXPART model is a Lagrangian particle dispersion model developed by the Norwegian Institute for Air Research (Stohl et al., 2005). It can describe the processes of long-range mesoscale transport and dispersion, dry and wet deposition and radiation attenuation of the tracer in the atmosphere by calculating the trajectory of large amounts of particles emitted by sources such as point, line, area or bulk. Though models can either simulate the dispersion of tracer from source area to the surrounding area by forward simulation, or determine the distribution of potential source area which impact the fixed stations by backward calculation, particularly when the quantity of observation stations in the research area is less than the quantity of emission sources, the backward simulation is more advantageous (Seibert and Frank, 2004).

The FLEXPART model has been widely used in recent years around the world, and it was originally developed for studying the problem of long-range mesoscale dispersion of air pollutants from point sources (Stohl et al., 1996) and evaluation of the simulation effect of the early version of the FLEXPART model (version 2.0) has been made according to the data of large tracing experiments (Stohl et al., 1998). Later, this model was developed into a comprehensive tool for simulating and studying atmospheric transport in addition to air pollution, and was used to prove the trans-state transport process of pollution air mass (Stohl et al., 1999, 2003). Besides, the model was used to analyse the atmospheric exchange process of stratosphere and troposphere (Cristofanelli et al., 2003). The model was also used to evaluate the direct effect of aerosols in the atmospheric transport and exchange process in combination with satellite imagery and aircraft survey data (Avey et al., 2007). Hirdman et al. (2010) used this model and statistical analysis data to analyse the sources of air pollutants with short life cycle in the Arctic region. In addition, the FLEXPART model can be used jointly with other models. For example, Foy et al. (2006) took the output of the MM5 model as the initial atmospheric field to study the flows of air pollution in the basin of Mexico City, and Jerome et al. (2006) made a slight modification on FLEXPART model and used it in the application of the WRF model.

Current studies show that the use of the FLEXPART model has been extended from air pollutant dispersion and transport to the combination with other models (such as MM5 and WRF). At present, most work with FLEXPART is conducted on the basis of ECWMF data (Stohl et al., 2009, 2010; Vollmer et al., 2009). From this, what is the FLEXPART result driven by the meteorological data of GFS and can it be used in the FLEXPART model? The aim of this research is to use carbon monoxide (CO) observations at Shangdianzi observation station to compare and evaluate the simulation results of FLEXPART driven by the GFS and ECMWF data, then apply the FLEXPART model in tracking the distribution of background and non-background source areas of the observation station. In this paper, Section '2. Observation station and model descriptions' introduces the station and the model method. Section '3. Model evaluation and comparison' includes the model evaluation and comparisons with GFS and ECWMF. Section '4. Model application in tracing distribution of CO sources' is the result of the source distribution characteristics in different seasons and the final section provides a summary and discussion of the conclusions.

2. Observation station and model descriptions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

2.1. Observation site description

The observation data used in this paper are from the Shangdianzi site, the regional atmospheric background station, which is located in the northeast of Beijing (40.65°N, 117.12°E, 293.91 m a.s.l.), and about 120 km away from the urban area of the city (Figure 1). Because Shangdianzi station is far from the densely-populated urban area, its air pollution level can represent the regional atmospheric background concentrations of the North China region, and the longer time-series observations can reflect the effect of human activities on regional atmospheric background concentrations (Lin et al., 2008; Zhang et al., 2010). The observation data of CO in this paper come from the GC-ECD in situ system at Shangdianzi with relative measurement precisions typically < 5%. The system was installed in a laboratory in Shangdianzi.

image

Figure 1. The location of Shangdianzi station and the surrounding region

Download figure to PowerPoint

2.2. FLEXPART model introduction and model establishment

The core of the FLEXPART model is to study the source-receptor relationship (SRR) of air pollutants. The pollution emission is the ‘source’, and the observation station is the ‘receptor’. Through study of the processes of transportation, dispersion, convection, dry and wet deposition and radiation attenuation, the pollution concentration of grid varying with time series (forward simulation) or the residence times at grid (also called the sensitivity co-efficient or the footprint, backward simulation) can be obtained.

The kernel of the FLEXPART model adopts the zero acceleration approach to calculate the trajectory of particles, and its expression formula is:

  • equation image(1)

in which Δt is the time increment; X is the position vector; equation image expresses the wind vector that is composed of grid scale component (equation image, turbulent wind fluctuations (νt) and mesoscale wind fluctuations (νm).

The calculation of turbulent wind fluctuations is based on the Langevin equation (Thomson, 1987):

  • equation image(2)

The drift term a and the diffusion term b in the above equation are the functions of space position, turbulent velocity and time. dWj are incremental components of a Wiener process with mean zero and variance, dt, which are uncorrelated in time (Legg and Raupach, 1982).

Through transition of the source-receptor relationship, the formula for calculating the residence times at grid point can be expressed as:

  • equation image(3)

in which ΔT is the time resolution; N is the quantity of sampling in the scope of ΔT;J is the total particles emitted; fijn is a function that decides the quantity of particles with ‘contribution’ at designated grid point.

According to the research characteristics, the parameters of the FLEXPART model in this paper are set up as follows: the simulation direction of the model is backward, and relative to the forward simulation, the backward simulation can more effectively reflect the distribution of potential source areas that have impact on the designated stations (Stohl et al., 2005). The source of emission in the model is set as a point source, i.e., the Shangdianzi background station (40.65°N, 117.12°E). For each 3 h, 50 000 particles were released in a layer reaching from 0 to 100 m above the model ground at the measurement location and tracked backwards in time for 7 and 20 days. The model output results are the residence times, and ps kg−1 is the unit used to express the residence times of pollution gases of unit mass at horizontal grid. The horizontal grid resolution of the model is 1°× 1°, and the time resolution is 3 h.

CO is selected as the tracer because CO is one of the atmospheric pollutants with the largest emission, and is an important tracer for studying the transfer, transport and redistribution of pollutants in atmosphere, so it has the better representation. The 1°× 1° grid GFS data provided by NCEP are used as the initial field for FLEXPART model. The meteorological data used to drive FLEXPART was 1°× 1° resolution data of the GFS (Global Forecast System model) from NCAR/NCEP (The National Center for Atmospheric Research/National Centers for Environmental Prediction). GFS data were available with 3 h resolution (analysis at 0000, 0600, 1200, 1800 UTC and T = + 3 h forecast at 0300, 0900, 1500 and 2100 UTC). Also, the ECWMF data are used to compare with the GFS data and evaluate the model performance. ECWMF is another set of meteorological data with 3 h and 1°× 1°. The simulation time is from 10 February 2009 to 31 December 2009, in accordance with the observation data.

2.3. Simulation concentration method

The FLEXPART backward simulation is an emission sensitivity, which refers to the residence times at grid point. The simulated mixing ratio at the receptor can be obtained by multiplying the footprint emission sensitivity with the emission inventory. The CO grid emission is from the INTEX-B2006 inventory of the East Asia region (Zhang et al., 2009). The space resolution of the inventory is 0.5°× 0.5°. Therefore, the model output value of each grid multiplied with the CO emission of such grid amounts to the contribution of such grid to CO concentration of Shangdianzi station and by summating the contributions of all grids on the same time level, the simulation value of CO concentration of Shangdianzi observation station is obtained.

Before comparison with the simulation, the observed CO concentration is screened into background and non-background. The background concentration represents the characteristics of homogenized atmospheric composition after straining out the direct influence of local conditions and human activities. Therefore, when discussing what influence human sources from the surrounding region have on CO concentration of the observation station, the result of comparison between non-background concentration and simulated value of FLEXPART ought to be more logical and closer to the actual status. The time range is from 10 February 2009 to 31 December 2009. In reference to the scheme of Stohl et al. (2009), the algorithm of robust extraction of background signal (REBS) (Ruckstuhl et al., 2010) is used to separate the CO time series concentrations of Shangdianzi station into background (also called baseline concentration) and non-background concentration. The REBS is a statistical method based on robust local regression that is well suited for the selection of background measurements and the estimation of associated baseline curves. The REBS method does not simply take the lowest values of the time series, but identifies pollution events in an iterated series of local regression fits using robust weights for values above the baseline. The partial data of abnormally high CO concentration caused by local sources is eliminated according to the observation records and the surface meteorological data. CO data eliminated take up 1.76% of the total quantity of data. The comparison of non-background time series concentration with model simulation value is also a base for carrying out the study on applicability of the FLEXPART model.

3. Model evaluation and comparison

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

3.1. Comparison between 7 and 20 day backward simulation

Usually, the 20 day length of the backward simulations is suggested (Stohl et al., 2009). In this paper, first the CO simulations based on different backward simulation lengths are compared. Figure 2 shows the comparison of the CO model results based on 7 and 20 day backward calculations with the FLEXPART model driven by GFS data. It can be seen that the difference between the two simulations is very slight and the correlation co-efficient is nearly 1. So in order to save calculation time, 7 day backward simulation is enough for CO. The following model results are all based on a 7 day backward simulation.

image

Figure 2. Comparison of 7- and 20-day FLEAPRT backward simulations of CO concentrations driven by GFS data

Download figure to PowerPoint

3.2. Comparisons between simulations driven by ECMWF and GFS data and observation

Comparisons of the two simulations with non-background observations of the Shangdianzi site are shown in Figure 3. The missing CO observation data in the figure are mainly caused by the relocation of the Shangdianzi station laboratory in March 2009, damage to the power board and installation of a replacement in May, and UPS failure in July. From the figure, it can be seen that the trends of the two simulation variations and observations are basically identical, and both the simulations with ECMWF and GFS can reflect the daily variation rule of CO observations. The correlation co-efficient statistics between the observations and the simulations are both higher, up to 0.71 and 0.69 for ECWMF and GFS, respectively (Figure 4). The shortage is that the simulation capacity of the model for peak value is insufficient, and the simulation value of CO concentration at time level of high pollution is comparatively low. Moreover, the simulation with GFS is slightly less than that with ECWMF.

image

Figure 3. Comparisons of observed (line) and simulated (sqaures: ECMWF, triangles: GFS) CO time series at Shangdianzi site in 2009

Download figure to PowerPoint

image

Figure 4. Correlation coefficient of the observed and simulated CO concentration driven by ECMWF and GFS data at Shangdianzi site in 2009

Download figure to PowerPoint

Figure 5 shows the monthly variations of the observed and simulated CO concentrations, and the correlation co-efficients in each month. It can be seen that both of the simulations can capture monthly variations of the observed CO concentration at the Shangdianzi site. Both simulations catch the characteristics of the relatively low values of CO monthly mean concentrations, and the observed and simulated CO values are accordantly relatively low from March to June. The model results with ECWMF and GFS capture the relatively higher CO monthly mean values in February and September. The correlation co-efficients in each month for ECWMF and GFS are all more than 0.5, except in July: both exceed 0.7 in June, September, November and December.

image

Figure 5. Comparisons of observed (circles) with simulated (squares: ECMWF, triangles: GFS) CO monthly mean concentrations at Shangdianzi site in 2009, and correlation coefficient of the observed and simulated driven by ECMWF (light grey bar) and GFS (dark grey bar)

Download figure to PowerPoint

4. Model application in tracing distribution of CO sources

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

After comparing the FLEXPART model results of GFS with the ECWMF simulation and observations, it is found that both simulations are accordant with the observations, and there is no obvious difference between the two simulations. So backward simulation of FLEXPART with GFS is used to trace the potential source distribution influencing on background and non-background concentration of the Shangdianzi site.

The sensitivity co-efficient distribution produced by the model simulation is the potential source or footprint, which reflects the degree of relative contribution of different grids to fixed stations, and it has no relation to whether it has true CO emission source at that grid or not.

In this research, the REBS method is used to extract the observed CO concentration time series of background and non-background at the Shangdianzi site in spring (March to May), summer (June to August), autumn (September to November) and winter (December to February) respectively. According to the time series of background and non-background, the footprint distributions of potential sources in the four seasons are then calculated by the model are shown in Figures 6 and 7. The actual statistical emission sources of CO from INTEX-B2006 inventory (Zhang et al., 2010) are used to compare with the simulated results (Figure 8).

image

Figure 6. Footprint distributions of CO background concentration in four seasons of 2009 at Shangdianzi site, and (a) spring, (b) summer, (c) sutumn, and (d) winter. (Black dot in the figure represents the Shangdianzi station)

Download figure to PowerPoint

image

Figure 7. Footprint distribution of CO non-background concentration in four seasons of 2009 at Shangdianzi station, and (a) spring, (b) summer, (c) autumn, and (d) winter. (Black dot in the figure represents the Shangdianzi station)

Download figure to PowerPoint

image

Figure 8. Statistical CO emission distribution at Shangdianzi station and the surrounding areas from the INTEX-B2006 emission inventory

Download figure to PowerPoint

It can be clearly seen from Figure 6 that the potential source areas for background concentration with comparatively high sensitivity co-efficient in four seasons are concentrated in the North-westnorthwest-northwest sector of the Shangdianzi station. As the atmospheric circulation in four seasons in the Beijing region is different, the footprint distribution is slightly different. In winter, the Beijing region is mostly subject to the control of the Mongolian high pressure and air masses from the northwest prevail; in summer, the Pacific subtropical high pressure moves northwards, the Mongolian high retreats and the easterly air flow is intensified in the Beijing region. It can also be seen from the figure that the footprint distribution in summer has the trend of an easterly direction, but in the other seasons it is mostly concentrated in the north-westnorthwest-northwest sector.

Figure 8 shows the statistical CO emission distribution at Shangdianzi station and the surrounding areas from the INTEX-B2006 emission inventory. It can be seen from the figure that CO emission source in the north-westnorthwest-northwest sector of Shangidanzi is much less than the other regions. In fact, Beijing is bordered by hills to the west, north and east, the underlying surface of the region of Hebei and Inner Mongolia covered by the north-westnorthwest-northwest sector is comparatively high, and the most part is hilly, belonging to the non-urban industrial region with relatively less population. Therefore, the north-westnorthwest-northwest sector is the most important background sector of CO for Shangdianzi station.

With respect to the background sector, people may be concerned more about the influence of a pollution air mass from the non-background sector on the CO concentration at Shangdianzi station. Figure 7 reflects the distribution of air mass footprints inverted by the model according to the non-background time series concentration. The sources of air masses for the non-background in spring are mainly distributed in the westsouthwest sector of Shangdianzi station. The southwest side of Shangdianzi station is the Beijing urban area with dense population and high urbanization, so the air mass from the westsouthwest sector is not clean. The sources of air masses for the non-background in summer are mainly distributed in the westsouthwest-to-southsoutheast sector of Shangdianzi station, but in autumn and winter they are in the west-south sector, covering the industrial cities of Beijing, Tianjin and Baoding of Hebei province. When combined with Figure 8, the regional CO emission in the west-southwest-south sector of Shangdianzi station is comparatively high, thus CO concentration in the air mass transported from the direction of this sector to Shangdianzi station is comparatively high. Therefore the west-southwest-south sector is the most important non-background sector of CO at Shangdianzi station.

The above analysis suggests that taking CO as the tracer and adopting the REBS method to screen the observation into background and non-background concentration, the FLEXPART model driven by GFS can capture the distribution characteristics of background and non-background source areas of the observation station.

5. Conclusions and discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

By adopting CO as the tracer, the FLEXPART model results driven by ECWMF and GFS are compared and evaluated by the on-line observations at Shangdianzi station. The comparisons and statistics show that both model results with ECWMF and GFS can depict daily and monthly trends of observed CO concentration and the correlation co-efficients between the daily observations and simulations for GFS and ECWMF are higher, up to 0.67 and 0.71 respectively. Moreover, the simulation with GFS data has the same good performance as that of the usually-used ECWMF data, and GFS simulation can be used to catch the characteristics of CO source distribution.

Using the REBS method to screen the observation into background and non-background concentration, the FLEXPART model with GFS is used to backward simulate the source area distribution of background and non-background time series concentration, respectively. The results indicate that the north-westnorthwest-northwest sector is the most important background sector for CO, while the west-southwest-south sector is the most important non-background sector. The comparison of the simulation results with the emission inventory shows that the model has ability in tracing the source distribution of CO background and non-background concentration. In addition, the model results can reflect the source distribution around Shangdianzi station and the surrounding areas.

In general, the model results with two input meteorological data of ECWMF and GFS can better capture the trend of daily and monthly variations of CO concentration at Shangdianzi station, but the model cannot capture the abnormally high values caused by the passage of polluted air mass of high concentration and the local occasional factor. On one hand, it may be for the reason that the adopted INTEX-B2006 emission inventory has not been updated in time. CO emission is mainly influenced by human activities. After experiencing the fast development of economy and urbanization as well as the drastic increase of motor vehicles in 2007 and 2008, the CO emission in Beijing and the surrounding areas has undergone some changes, particularly an increase in the proportion of traffic sources. Thus, when calculating CO concentration in Shangdianzi in 2009 by using the inventory of emission sources from the emission inventory of 2006, it may cause the situation that the simulated peak concentration is comparatively low. On the other hand, the reason may be that the ground observation station is liable to be influenced by the local environment. The observations are often influenced greatly by the local pollution events, and when there is an occasional emission process in the region around the observation station, such as leakage of plume or combustion of garbage, its impact on CO observations will be high, so creating the abnormally high observation values. This part of the emissions has some randomness both in time and position, the emission amount cannot be taken into the conventional emission inventory, so the model simulation result cannot reflect the pollution events, causing the simulation value to be lower than the observed value. Although there is no ideal improved measure for this problem at present, along with the diversification of monitoring measures and the real-time application of satellite information, which will help trace and analyse the causes of high pollution events and update the inventory on time. Along with the further improvement of physical process and model resolution, this will help improve the simulation capacity of the model for the peak value.

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES

This study was supported by the Natural Science Foundation of China (41030107) and Chinese Ministry of Science and Technology (2010CB950601). EU S & T Cooperative Project 2SMONG and Sino-Swiss Science and Technology Cooperation project 2SMONG. Thanks Andreas Stohl through SOGG-EA of the Research Council of Norway. The authors gratefully acknowledge S. Henne for offering ECWMF results and A. Stohl for offering FLEXPART model.

REFERENCES

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observation station and model descriptions
  5. 3. Model evaluation and comparison
  6. 4. Model application in tracing distribution of CO sources
  7. 5. Conclusions and discussion
  8. Acknowledgements
  9. REFERENCES
  • Avey L, Garrett TJ, Stohl A. 2007. Evaluation of the aerosol indirect effect using satellite, tracer transport model, and aircraft data from the International Consortium for Atmospheric Research on Transport and Transformation [J]. J. Geophys. Res. 12: D10S33, DOI: 10.1029/2006JD007581.
  • Cristofanelli P, Bonasoni P, Collins W, Feichter J, Forster C, James P, Kentarchos A, Kubik PW, Land C, Meloen J, Roelofs GJ, Siegmund P, Sprenger M, Schnabel C, Stohl A, Tobler L, Tositti L, Trickl T, Zanis P. 2003. Stratosphere-to-troposphere transport∼:A model and method evaluation[J]. J. Geophys. Res. 108(D12): 85258547.
  • Foy B, Varela JR, Molina1 LT, Molina MJ. 2006. Rapid ventilation of the Mexico City basin and regional fate of the urban plume[J]. Atmos. Chem. Phys. Discuss. 6: 839877.
  • Hirdman D, Sodemann1 H, Eckhardt S, Burkhart JF, Jefferson A, Mefford T, Quinn PK, Sharma S, Ström J, Stohl A. 2010. Source identification of short-lived air pollutants in the Arctic using statistical analysis of measurement data and particle dispersion model output[J]. Atmos. Chem. Phys. 10: 669693.
  • Jerome D, Richard Fast, Easter C. 2006. A lagrangian particle dispersion model compatible with WRF [Z]. 7th Annual WRF User's Workshop, 19-22 June 2006, Pacific Northwest National Laboratory, Richland, WA.
  • Legg BJ, Raupach MR. 1982. Markov-chain simulation of particle dispersion in inhomogeneous flows: the mean drift velocity induced by a gradient in Eulerian velocity variance. Boundary Layer Meteorol. 24: 313.
  • Lin W, Xu X, Zhang X, Tang J. 2008. Contributions of pollutants from North China Plain to surface ozone at the Shangdianzi GAW station. Atmos. Chem. Phys. 8: 58895898.
  • Ruckstuhl AF, Henne S, Reimann S, Steinbacher M, Buchmann B, Hueglin C. 2010. Robust extraction of baseline signal of atmospheric trace species using local regression. Atmos. Meas. Tech. Discuss. 3: 55895612, DOI: 10.5194/amtd-3-5589-2010.
  • Seibert P, Frank A. 2004. Source-receptor matrix calculation with a Lagrangian particle dispersion model in backward mode[J]. Atmos. Chem. Phys. 4: 5163.
  • Stohl A. 1996. Trajectory statistics- a new method to establish source-receptor relationships of air pollutants and its application to the transport of particulate sulfate in Europe [J]. Atmos. Environ. 30: 579587.
  • Stohl A, Forster C, Eckhardt S, Spichtinger N, Huntrieser H, Heland J, Schlager H, Wilhelm S, Arnold F, Cooper O. 2003. A backward modeling study of intercontinental pollution transport using aircraft measurements [J]. J. Geophys. Res. 108: 4370, DOI: 10.1029/2002JD002862.
  • Stohl A, Forster C, Frank A, Seibert P, Wotawa G. 2005. Technical note: the Lagrangian particle dispersion model FLEXPART version 6.2 [J]. Atmos. Chem. Phys. 5: 24612474.
  • Stohl A, Hittenberger M, Wotawa G. 1998. Validation of the lagrangian particle dispersion model FLEXPART against large-scale trace experiment data [J]. Atmos. Environ. 32(24): 42454264.
  • Stohl A, Kim J, Li S, O'Doherty S, Mühle J, Salameh PK, Saito T, Vollmer MK, Wan D, Weiss RF, Yao B, Yokouchi Y, Zhou LX. 2010. Hydrochlorofluorocarbon and, hydrofluorocarbon emissions in East Asia determined by inverse modeling. Atmos. Chem. Phys. 10: 35453560.
  • Stohl A, Seibert P, Arduini J, Eckhardt S, Fraser P, Greally BR, Lunder C, Maione M, Mühle J, O'Doherty S, Prinn RG, Reimann S, Saito T, Schmidbauer N, Simmonds PG, Vollmer MK, Weiss RF, Yokouchi Y. 2009. An analytical inversion method for determining regional and global emissions of greenhouse gases: sensitivity studies and application to halocarbons. Atmos. Chem. Phys. 9: 15971620.
  • Stohl A, Trickl T. 1999. A textbook example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and North American origin in the free troposphere over Europe [J]. J. Geophys. Res. 104(D23): 30,44530,462.
  • Thomson DJ. 1987. Criteria for the selection of stochastic models of particle trajectories in turbulent flows [J]. J. Fluid Mech. 180: 529556 (ISSN 0022-1120).
  • Vollmer MK, Zhou LX, R Greally B, Henne S, Yao B, Reimann S, Stordal F, Cunnold DM, Zhang XC, Maione M, Zhang F, Huang J, Simmonds PG. 2009. Emissions of ozone-depleting halocarbons from China. Geophys. Res. Lett. 36: L15823, DOI: 10.1029/2009GL038659.
  • Zhang Q, Streets DG, Carmichael GR, He K, Huo H, Kannari A, Klimont Z, Park I, Reddy S, Chen D, Duan L, Lei Y, Wang L, Yao Z. 2009. Asian emissions in 2006 for the NASA INTEX-B mission. Atmos. Chem. Phys. 9: 51315153.
  • Zhang F, Zhou LX, Yao B, Vollmer MK. 2010. Analysis of 3-year observations of CFC-11, CFC-12 and CFC-113 from a semi-rural site in China. Atmos. Environ. 44: 44544462.