Auto MAX-DOAS measurements around entire cities : quantification of NO x emissions from the cities of Mannheim and Ludwigshafen ( Germany )

Auto MAX-DOAS measurements around entire cities: quantification of NOx emissions from the cities of Mannheim and Ludwigshafen (Germany) O. Ibrahim, R. Shaiganfar, R. Sinreich, T. Stein, U. Platt, and T. Wagner Max-Planck-Institute for Chemistry, Mainz, Germany Institute for Environmental Physics, University of Heidelberg, Germany now at: University of Colorado at Boulder, USA now at: Ernst & Young, Luxembourg, Luxembourg Received: 2 February 2010 – Accepted: 2 February 2010 – Published: 11 February 2010 Correspondence to: T. Wagner (thomas.wagner@mpch-mainz.mpg.de) Published by Copernicus Publications on behalf of the European Geosciences Union.


Abstract
We present Auto Multi-Axis (MAX-) DOAS observations of tropospheric NO 2 carried out on circles around the cities of Mannheim and Ludwigshafen (Germany) on 24 August 2006.Together with information on wind speed and direction, the total emissions of the encircled source(s) can be quantified from these measurements.In contrast to recent similar studies based on of zenith scattered sun light (elevation angle of 90 • ), we use a MAX-DOAS instrument mounted on a car, which observes scattered sun light under different elevation angles (here 45 • , and 90 • ).Compared to simple zenith sky observations, MAX-DOAS observations have higher sensitivity and avoid systematic offsets in the determination of the vertically integrated trace gas concentration.
Auto MAX-DOAS observations are especially well suited for the determination of the total emission of extended emission sources (e.g.whole cities), for which typically no sharply defined plumes are formed.In such cases, the trace gas concentrations can be rather small and thus even small systematic offsets in the observed integrated tropospheric trace gas concentration can have a large effect on the determined total emissions.However, such measurements are still affected by several uncertainties which need to be further investigated and minimised.The largest error source is probably the variability and imperfect knowledge of the wind field.In addition -depending on the trace species observed -also chemical transformations between the emission sources and the measurement location have to be considered.In this study we use local observations within the encircled area to quantify and correct these errors.From our observations we derive a total NO x emission from the Mannheim/Ludwigshafen area of (7.2±1.7)×10 24molecules/s (or 17350±4100 t, calculated with the mass of NO 2 ), which is in surprisingly good agreement with existing emission estimates.

Introduction
The precise knowledge of the emissions of natural and anthropogenic trace gases (e.g.pollutants or greenhouse gases) is important for many applications.Emission estimates are used as input to atmospheric chemistry and transport models.By varying the strengths of the emission sources in the model, it is e.g.possible to predict which emission reductions would have the strongest impact on air quality in a given area.Similar arguments hold for the emission inventories of greenhouse gases, which are important input in models predicting future climate.Accurate knowledge of the emission strengths from different sources is especially important to quantify the influence of anthropogenic emissions in comparison to natural emissions.
Usually, emission inventories are built using bottom-up strategies.Emission strengths of individual sources are quantified and summed up according to the frequency of the respective source type.Uncertainties of the calculated emission inventories result from both errors in the estimate of the number of individual sources and errors in the emission strength of individual sources.
An alternative to bottom-up emission inventories are top-down emission inventories.They are based on measurements of the atmospheric concentration of a pollutant, which is then related to the emission strength of an individual source or an integrated source strength within a specified area.For the determination of the emission strength of an observed trace gas additional knowledge on the atmospheric transformation processes (transport and chemistry) is required.Depending on the complexity of the emission source (e.g. a point source or a mixture of different sources with different spatio-temporal emission patterns) and the information content of the measurement (e.g. point measurements at fixed locations, or satellite observations), either simple assumptions (e.g. on the atmospheric lifetime and wind fields), or complex inverse models are required.
In this study we estimate the total urban NO x emissions from Auto MAX-DOAS observations performed on circular driving routes around complete cities.This method Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version

Interactive Discussion
was recently introduced and applied to determine emissions of SO 2 , NO 2 and HCHO (Johansson et al., 2008(Johansson et al., , 2009;;Rivera et al., 2009).In contrast to these zenith sky observations we use MAX-DOAS observations, which have several advantages: first because of the slant path through the troposphere they have a higher sensitivity for tropospheric trace gases.Second, the results of MAX-DOAS observations are less affected by light path modifications caused by multiple scattering in thick clouds (see e.g., Johansson et al., 2008), because the measurements at different elevation angles are affected by such clouds in a similar way.Third, MAX-DOAS observations allow to determine the vertically integrated tropospheric trace gas concentration (vertical column density, VCD) above the instrument location without systematic biases (Wagner et al., 2010).In contrast, from zenith looking instruments only the difference of the tropospheric VCD compared to a reference measurement can be determined.This complicates the interpretation of the measurements especially in cases when emission sources are not completely encircled (Johansson et al., 2009).MAX-DOAS observations allow the determination of absolute tropospheric VCDs, which is also important for the quantitative comparison with model results and for the validation of satellite observations.In addition to the inherent vertical integration of the tropospheric trace gas concentration by MAX-DOAS observations, a horizontal integration can be carried out if MAX-DOAS observations are performed from aircrafts or cars.Such observations eventually allow to determine the complete flux F of trace gas molecules across the area span given by the driving route s and the vertical (see Fig. 1).For that purpose the knowledge of the wind speed and direction is required: Here W s is the average wind vector within the trace gas layer.The integral is evaluated along the driving route s.We carry out Auto MAX-DOAS measurements along closed driving routes around large emissions sources, e.g.whole cities (Fig. 2).If the emission sources are completely encircled, the total flux entering the encircled area (on the side from which the wind blows) as well as the total flux leaving the encircled area (at the opposite side, see Fig. 2) can be determined (see also Johansson et al., 2008).In simple cases, the difference between both fluxes yields the integrated total emission of the encircled area.To determine the total flux F total of the encircled area the integral in Eq. (1) has to be evaluated along the complete circle around the area of interest: It should be noted that for the complete surrounding of extended areas the driving period is typically between a few ten minutes and more than one hour.Thus temporal variations on time scales below that period can not be resolved and the resulting emission estimates therefore are averages over the period of the observations.Two main effects complicate the determination of the total emission and have to be taken into consideration: A) Transport by wind In the simplest case, a wind field with constant speed and direction is present over the complete area of interest, and the wind speed is also high enough that the transport across the encircled area is fast compared to the atmospheric lifetime of the trace gas.In more complicated cases, the wind speed is low and varies with time and location within the encircle area.In such cases, emission estimates are difficult and might only be possible with the additional use of atmospheric model simulations.A further general problem is caused by the fact that the wind speed usually increases with altitude (also the direction changes systematically), but the vertical distribution of the observed trace gas is usually not well known.Thus assumptions on the vertical distribution and appropriate wind speed have to be made.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion B) Chemical transformations Chemical transformations and deposition processes change the trace gas abundance after the emission process.Depending on the speed, two cases can be distinguished.First, rapid chemical reactions can change the partitioning of the emitted species.In this study, MAX-DOAS observations of NO 2 are investigated, while most of the NO 2 was primarily emitted as NO.The partitioning of both species depends on the ozone concentration and the NO 2 photolysis rate.In order to quantify the total NO x (NO 2 +NO) emissions, the so called Leighton ratio (L=[NO]/[NO 2 ]) has to be known (or assumed).Second, the atmospheric concentration of the emitted species can be changed by processes, which are slow compared to the transport time between the emission source and the location of the measurement.Thus, the measured trace gas concentration represents only a fraction of the emitted trace gas abundance.These destruction processes are usually described by an exponential decrease with an efolding lifetime τ, after which the trace gas concentration is reduced to a fraction 1/e of the initial value.It should be noted that in cases of substantial trace gas destruction between the emission source and the location of the measurement (short atmospheric lifetime compared to the transport time), also the spatial distribution of the emission sources within the encircled area becomes important.
These potential error sources are discussed in detail and quantified for the Auto MAX-DOAS measurements presented in this study (see Sect. 4).
The paper is structured as follows: in Sect. 2 the MAX-DOAS method is briefly introduced and the performed Auto MAX-DOAS measurements are described.Section 3 provides details on the different steps of the data analysis.In Sect. 4 the results of our measurements are presented and the uncertainties caused by the different error sources are discussed and quantified.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc

Printer-friendly Version
Interactive Discussion

MAX-DOAS observations
In recent years Multi-Axis-Differential Optical Absorption Spectroscopy (MAX-DOAS) observations have become a widely and successfully used technique for the remote sensing of tropospheric trace gases and aerosols.MAX-DOAS instruments observe scattered sun light under different slant viewing angles, which make them especially sensitive to tropospheric trace gases and aerosols (e.g., H önninger et al., 2002, 2004a,b;Leser et al., 2003;Van Roozendael et al., 2003;Wittrock et al., 2004;Wagner et al., 2004;Sinreich et al., 2005Sinreich et al., , 2007;;Heckel et al., 2005;Frieß et al., 2006;Fietkau et al., 2007;Theys et al., 2007;Brinksma et al., 2008;Wagner et al., 2007a,b;Irie et al., 2009).From MAX-DOAS observations, vertical profiles of tropospheric trace gases and aerosol extinction can be retrieved yielding up to a few pieces of information, with the highest vertical resolution close to the surface (see e.g., Theys et al., 2007).However, profile retrievals are typically restricted to MAX-DOAS measurements made at fixed locations and under cloud free conditions.For MAX-DOAS observations made from moving platforms, horizontal gradients of the trace gas concentration often prevent a meaningful profile retrieval.In addition, unlike ship measurements, for car measurements the observations at low elevation angles are usually affected by obstacles (e.g.trees or houses) in the field of view.Thus from mobile MAX-DOAS observations usually only the (absolute) tropospheric VCD is determined.

Determination of the tropospheric VCD
For the retrieval of the tropospheric VCD, typically observations at rather high elevation angles (>about 10 • ) are used.For such elevation angles the effects of atmospheric aerosols are rather small and the atmospheric light paths can be geometrically well approximated (Brinksma et al., 2008;A. Richter, personal communication, 2005).In addition, the retrieval of tropospheric VCDs is usually even possible in the presence of clouds (at least for trace gases located below the cloud base).
For the analysis of MAX-DOAS observations it is usually assumed that the concen-Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc

Printer-friendly Version
Interactive Discussion tration field does not change during the time period needed for the observations at the different elevation angles.This assumption is roughly valid for MAX-DOAS observations made at fixed locations.In contrast, mobile MAX-DOAS observations are typically strongly affected by horizontal concentration gradients of air masses probed along the driving route.For these platforms, the usual way of the MAX-DOAS data analysis can thus lead to large errors, in extreme cases even negative concentrations might be obtained.Recently a technique was developed which overcomes this problem and allows to determine absolute tropospheric VCDs along the driving route (Wagner et al., 2010).

Auto MAX-DOAS instrumentation for the Mannheim/Ludwigshafen measurements
The Mini-MAX-DOAS instrument is a fully automated, light weighted spectrometer (13 cm×19 cm×14 cm) designed for the spectral analysis of scattered sunlight (e.g., Bobrowski et al., 2003;H önninger et al., 2004).It consists of a sealed aluminium box containing the entrance optics, a fibre coupled spectrograph and the controlling electronics.A stepper motor mounted outside the box rotates the whole instrument to control the elevation of the viewing angle (angle between the horizontal and the viewing direction).The entrance optics consists of a quartz lens of focal length f =40 mm coupled to a quartz fibre bundle which leads the collected light into the spectrograph (field of view is ∼1.2 • ).The light is dispersed by a crossed Czerny-Turner spectrometer (USB2000, Ocean Optics Inc.) with a spectral resolution of 0.7 nm over a spectral range from 320-460 nm.A one-dimensional CCD (Sony ILX511, 2048 individual pixels) is used as detector.Before the signal is transferred to the 12 bit analog-to-digital converter, an electronic offset is added.After conversion, the signal is digitally transmitted to a laptop computer via one USB cable and stored for subsequent analysis.
For the mobile measurements the Mini-MAX-DOAS instrument was mounted on the top of a car (Auto MAX-DOAS) with the telescope pointing in driving direction and was powered by the 12 V car battery.The rest of the set-up was inside the car and both parts were connected via two electric cables.The measurements are controlled from Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion a laptop using the DOASIS software (Kraus, 2004).
On 24 August 2006, measurements were carried out around the Mannheim-Ludwigshafen industrial/urban area.The sequence of elevation angles was chosen to: 45 • , 45 • , 45 • , 45 • , 90 • , and the duration of an individual measurement was about 20-25 s.

Data analysis
The measured spectra are analysed using the DOAS method (Platt and Stutz, 2008).
A wavelength range of 415-435 nm was selected for the analysis.Several trace gas absorption cross sections (NO 2 at 297 K (Vandaele et al., 1998) et al., 2003) as well as a Fraunhofer reference spectrum, a Ring spectrum (calculated from the Fraunhofer spectrum) and a polynomial of second order were included in the spectral fitting process using the WinDOAS software (Fayt and van Roozendael, 2001).The wavelength calibration was performed based on a high resolution solar spectrum (Kurucz et al., 1984).The output of the spectral analysis is the slant column density (SCD), the integrated trace gas concentration along the light path through the atmosphere.Following to the retrieval technique of Wagner et al. (2010) the tropospheric NO 2 VCD is determined from each observation made at 45 • elevation angle using the zenith measurement of the respective sequence.In the retrieval process, tropospheric air mass factors (AMF) are needed, for which the geometric approximation is used: Depending on the aerosol load and on the vertical distribution of the trace gas, the geometrical approximation for the tropospheric AMF can deviate to some degree from the true value.We investigated these deviations using the Monte Carlo radiative transfer model TRACY-2 (Wagner et al., 2007;Deutschmann and Wagner, 2008) and found that for aerosol loads with optical depth <0.5 for surface-near trace gases (<200 m) Introduction

Conclusions References
Tables Figures

Back Close
Full the errors of the geometrical approximation are typically below 15%.In the presence of clouds the uncertainties can in principle become larger.However, especially for the part of the trace gas profile below the cloud (e.g.freshly emitted NO x ) the geometrical approximation is in general a very good choice, because the lower boundary of the cloud acts as well defined illumination source and scattering events below the cloud (within the trace gas layer) become less important.The total error of the tropospheric NO 2 VCD is estimated from the typical fit residual and the uncertainty of the geometric approximation to >3×10 15 molecules/cm 2 or >15%.

Results and discussion
Results of the Auto MAX-DOAS observations made on trips around the cities of Ludwigshafen and Mannheim (Southern Germany, see Fig. 3) on 24 August 2006 are shown in Figs. 4 and 5.The sky was mainly overcast and temperatures were around 20 • C. Both cities were surrounded in four successive "circles" with approximate extensions of 26 km in north-south direction and 18 km in east-west direction.During these circles, a repeating pattern with high and low tropospheric NO 2 VCDs was found.In general the highest NO 2 VCDs were measured in the north-east corner, which can be directly related to the prevailing wind direction on that day (mainly from south-west).However, not exactly the same patterns are obtained for the different circles reflecting the variation of emissions as well as wind speed and direction.
From the MAX-DOAS observations the integrated NO x emission within the encircled area can in principle be calculated according to Eq. (2).However, to account for the finite atmospheric lifetime and for the partitioning between NO and NO 2 , two corrections have to be applied to Eq. (2):

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version

Interactive Discussion
Here c L is a correction factor which accounts for the partitioning of NO x into NO and NO 2 , which is typically expressed as Leighton ratio (L=[NO]/[NO 2 ]).c τ is a correction factor which accounts for the destruction of NO x while it is transported from the emission sources to the locations of the measurements.
In the following sub-sections the three terms of Eq. 4 are determined separately and applied to the measurements on 24 August 2006.

Integration of the tropospheric NO 2 flux along the driving route
In this section the integral of Eq. ( 4) is evaluated for the MAX-DOAS measurements around the cities of Mannheim and Ludwigshafen.Because of the finite integration time of the individual spectra, the integral is substituted by a sum which is evaluated for all observations made at 45 • elevation during the individual circles: The distance between two measurements ∆s i is the geometric difference between the locations at the beginning of two successive measurements.For the wind direction and wind speed W constant values for all measurements during the respective circle were assumed.They were calculated from half hour averages at three meteorological stations within the encircled area: Mannheim south, Mannheim center, Mannheim north (http://mnz.lubw.baden-wuerttemberg.de/messwerte/aktuell/, the locations are indicated in Fig. 5).Also the angles β( s i ) between the driving route and the wind direction were calculated in discrete steps (the angle of the driving route was determined according to the starting points of two successive spectra).The average wind speeds and directions are listed in Table 1.
As an example, Fig. 6 shows the individual terms of the sum in Eq. ( 5) for the fourth circle on 24 August 2006.In the top panel the tropospheric NO 2 VCD is shown.The second panel shows the length of the driving route for the individual measurements (the difference of the instrument position between the start of two consecutive spectra).The Introduction

Conclusions References
Tables Figures

Back Close
Full path length varies with the driving speed.In addition, according to the sequence of elevation angles systematically different path lengths between successive measurements are also found.Larger distances occur for the end of an elevation sequence because of the gap during the record of the zenith spectrum.In addition, also some time is needed to change the elevation angle.The third panel shows the sine of the angle between the driving route and the wind direction, respectively.The bottom panel shows the resulting tropospheric NO 2 flux above the driving route for the segment between two measurements taking into account also the wind speed (assumed to be constant, see Table 1).According to the relative orientation of the driving direction and wind direction, the NO 2 flux is either positive or negative.The high positive values at the beginning of the circle indicate the main emissions from the Mannheim/Ludwigshafen area.At the end of the circle both substantial negative and positive fluxes occur.They are probably caused by emissions from the city of Speyer south of the encircled area.They first enter and then leave the encircled area at the south-west edge (see Fig. 5).Besides this example, the influx (negative flux) of NO 2 into the encircled area is typically rather small.Summing-up the NO 2 fluxes along the driving route, yields a total emission of 5.8×10 24 molecules/s NO 2 .In Table 2 also the results for the other cycles are shown.
The uncertainty of the derived total emission is dominated by the variation and the imperfect knowledge of the wind field within the encircled area (see also discussion in Johansson et al., 2008).Three main uncertainties can be distinguished:

Errors caused by variations of the wind speed
We estimate errors caused by variations of the wind speed from the standard deviation of the wind speed at the different stations within the time of the measurements (see Table 1).In principle, fluctuations of the wind speed should cancel each other, and the knowledge of the average wind speed should be sufficient for the determination of the total emissions.However, this would only be true for short fluctuations and long measurement duration.If the wind speed changes systematically during the measurement period, the associated inhomogeneities of the wind field can lead to an over-or Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc

Printer-friendly Version
Interactive Discussion underestimation of the total emissions, depending on the details of the distribution of individual emission sources within the encircled area.The resulting error is difficult to quantify without complete knowledge of the wind field and the distribution of emission sources.Here we estimate the relative error of the determined NO 2 emission associated with inhomogeneities of the wind field using half of the standard deviation (see Table 2).

Errors caused by uncertainties of the wind direction
Especially for measurements with the average wind direction almost parallel to parts of the driving route (e.g. as in circle 2) the precise knowledge of the wind direction becomes especially important.We estimated the associated errors by varying the wind direction by ±20 • (see Table 2).For most circles, the errors are rather small, typically <15%.However, for the second circle (with the average wind direction almost parallel to larger parts of the driving route) the error reaches up to 60%.

Errors caused by the vertical variation of the wind
The wind speed and direction usually depend systematically on altitude with higher wind speeds at higher altitudes.Thus surface wind data typically underestimate the effective wind speed relevant for the NO 2 layer.According to Eq. ( 2) this directly leads to an underestimation of the determined NO 2 emissions.Here we apply no correction for this effect, because no wind profile data are available and also the vertical distribution of the NO 2 concentration is unknown.However, since most emission sources are located at the surface, it can be assumed that most of the freshly emitted NO x should be present close to the surface (exceptions are e.g.power plants).Thus the resulting underestimation should be small (at least compared to other uncertainties).
We want to point out that our current way of estimating the errors associated to variations and missing knowledge of the wind field is rather crude and should be improved in future investigations (e.g. by using higher resolved wind data from measurements Introduction

Conclusions References
Tables Figures

Back Close
Full and/or model simulations).Nevertheless, our current error estimate takes into account the most important errors caused by inhomogeneities of the wind field in a semi quantitative way.

Estimation of the NO x lifetime
In this section the effect of the finite atmospheric lifetime of NO x (factor c τ in Eq. 4) is estimated.The NO x lifetime depends on meterological parameters (e.g.temperature) and on the photochemical state of the atmosphere, and is difficult to determine for the specific situation of our measurements.Thus we use an average value of 6 h for the NO x lifetime which is derived from several experimental studies (Beirle et al., 2003(Beirle et al., , 2004a,b),b).According to the extension of the circles and the average wind speeds (Table 1) the time the air masses to cross the encircled area is about 2 h.Because most NO x is emitted close to the center of the encircled area, we assume an effective transport time of 1 h.Within that duration the NO x concentration decreases by about 15% for an assumed lifetime of 6 h.This results in a correction factor c τ of 1.18.For assumed lifetimes of 4 and 8 h the correction factors would be 1.28 and 1.13, respectively.From this variation we estimate the uncertainty caused by the finite atmospheric liftetime of NO x to about 10%.

Leighton ratio during the measurements
For the determination of the second correction factor c L we use measurements of NO and NO 2 made at two meteorological stations (Mannheim North and Mannheim Center, see Fig. 7).For the period of the MAX-DOAS measurements Leighton ratios of about 0.35 were found.Thus we use a correction factor c L of 1.35 to obtain to determine the total NO x emissions.From the variation of the Leighton ratio during the time of our measurements we estimate the related uncertainties of the total emission to be about

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion 10%.

Final results of NO x emissions for the different circles
Table 3 summarises the total emissions and respective errors for the four circles.The calculated errors include the contributions associated with variations and imperfect knowledge of the wind field, NO x lifetime, NO x partitioning and retrieval of the tropospheric VCD.
The average of total NO x emission calculated from the results of the four circles is (7.2±1.7)×10 24molecules/s.This equals a NO x emission of 0.55 kg/s (using the mass of NO 2 ).

Additional error sources
In some cases, additional complications can occur, e.g. if the NO concentrations are higher than the background ozone concentrations.Such situations might occur close to the stacks of power plants.Then the reaction of NO with ozone will eventually consume all available ozone, which prevents the further conversion of NO into NO 2 .Only after additional ozone-rich air is mixed with the ozone-depleted air masses, the Leighton ratio expected for ozone-rich conditions can be established.For our measurements the ozone mixing ratio measured at the different stations (see Fig. 7) is found to be higher that the NO x concentration.However, especially for emissions from the large power plant in the southern part of Mannheim these conditions might not be fulfilled and the measured NO 2 might underestimate the total NO x emissions.Unfortunately, with our current knowledge this effect can not be quantified.Since the ozone concentrations increase during the day (Fig. 7), the possible underestimation should be smaller for the later measurements.The effect of suppressed NO to NO 2 conversion should be investigated in more detail in future studies.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion

Conclusions
We applied Auto MAX-DOAS observations for the determination of the total NO x emissions from the Mannheim/Ludwigshafen industrial and urban area.This method is similar to that introduced by Johansson et al. (2008), but instead measuring scattered sun light under zenith sky observations, we use a MAX-DOAS instrument mounted on a car.MAX-DOAS observations measure scattered sunlight under various slant elevation angles and thus provide increased sensitivity for tropospheric species.Moreover, MAX-DOAS observations allow to determine the absolute value of the vertically integrated tropospheric trace gas concentration (tropospheric VCD).This makes the method especially well suited for observations of emissions from extended areas, for which the emission plumes might not be characterised by well defined and sharp gradients.Observations of the tropospheric VCD without systematic offsets are also important if the measurements are not performed on closed loops, but only along selected transects through the emission plume.In such cases Auto MAX-DOAS observations allow to quantify the absolute value of the total trace gas flux, which can be directly compared to model simulations or used for satellite validation.Auto MAX-DOAS observations provide a rather simple and cheap method for the determination of total emissions of extended areas like complete cities.Thus they allow to check existing emission estimates with a completely independent method.As shown by Johansson et al. (2008) besides NO 2 also the emissions of other species like e.g.SO 2 and HCHO can be quantified.However, the accuracy of Auto MAX-DOAS should be further improved.Major uncertainties are caused by temporal and spatial variations of the wind field within the encircled area.Additional errors are related to chemical transformations between the emission source and the location of the measurement.In both cases more detailed information on the meteorological conditions and chemical composition would reduce the error sources.In this study we quantified the respective errors from in-situ observations within the encircled area. On

Vertical integration for individual measurement
Horizontal integration along the driving route s 2-dimensional integration of the tropospheric trace gas concentration

Vertical integration for individual measurement
Horizontal integration along the driving route s 2-dimensional integration of the tropospheric trace gas concentration

Fig. 1 Fig. 1 .Fig. 3
Fig. 1 Form MAX-DOAS observations the vertically integrated tropospheric trace gas concentrations can be determined.If MAX-DOAS observations are performed on mobile platforms, a second integration along the driving route becomes possible (indicated by the blue arrows).The black arrows indicate the viewing directions of the MAX-DOAS instrument.

Fig. 3 .Figure 4
Fig. 3.The red rectangle indicates the location of the encircled area around the cities of Mannheim and Ludwigshafen in Southern Germany (see also Fig. 5).(The map was taken from http://www.freeworldmaps.net/europe/germany/political.html.)

Fig. 5 .
Fig. 5. NO 2 VCDs for the four individual cycles around the Mannheim-Ludwigshafen area on 24 August 2006.The letters "N", "M", and "S" indicate the locations of the in-situ measurement stations Mannheim North, Mannheim Center, and Mannheim South.Arrows indicate the average wind direction.

Fig. 6
Fig. 6 Different components used for the net flux estimation shown for loop #4 on 24.08.2006.In a) the observed NO2 VCD are presented.Fig. 6b shows the driving distance during one observation and c) the sine of the angle between the driving route and the wind direction.In d) the resulting NO2 flux is shown: negative fluxes indicate import and positive fluxes export from the encircled area.

Fig. 6 .
Fig. 6.Different components used for the net flux estimation shown for loop #4 on 24 August 2006.In (a) the observed NO 2 VCD are presented.Figure (b) shows the driving distance during one observation and (c) the sine of the angle between the driving route and the wind direction.In (d) the resulting NO 2 flux is shown: negative fluxes indicate import and positive fluxes export from the encircled area.

Table 1 .
24 August 2006 we performed Auto MAX.DOAS observations on 4 consecutive Introduction Wind data calculated from three stations (Mannheim center, south, north, see Fig.5).For each cycle, the average and standard deviation of wind speed and direction are calculated.

Table 2 .
Total NO 2 emissions for the different circles.Also shown are the relative errors related to variations and imperfect knowledge of the wind field.

Table 3 .
NO x emissions from the encircled area after corrections for the lifetime and NO x partitioning.The total error includes the sun of all error contributions (retrieval of the tropospheric VCD, wind field, NO x lifetime, NO x partitioning).