Validation of acetonitrile (CH3CN) measurements in the stratosphere and lower mesosphere from the SMILES instrument on the International Space Station

Acetonitrile (CH3CN) is a volatile organic compound (VOC) and a potential tracer of biomass burning. We evaluated the capability of using observations derived from the Superconducting Submillimeter-Wave Limb-Emission Sounder (SMILES) on the International Space Station (ISS) to measure CH3CN profiles. The error in a CH3CN vertical profile from the Level-2 research (L2r) product version 3.0.0 was estimated by theoretical error analysis and also compared with other instrumental measurements. We estimated the systematic and random errors to be ∼ 5.8 ppt (7.8 %) and 25 ppt (60 %), respectively, for a single observation at 15.7 hPa in the tropics, where the CH3CN measurements are enhanced. The major source of systematic error was the pressure-broadening coefficient, and its contribution to the total systematic error was approximately 60 % in the middle stratosphere (15.7–4.8 hPa). The random error decreased to less than 40 % after averaging 10 profiles in the pressure range of 28.8–1.6 hPa. The total error due to uncertainties in other molecular spectroscopic parameters (2.8 ppt) was comparable to that of CH3CN spectroscopic parameters. We compared the SMILES CH3CN profiles with those of the microwave limb sounder (MLS) on the Aura satellite (version 4.2). The SMILES CH3CN values were consistent with those from MLS within the standard deviation (1σ ) of the MLS observations. The difference between the SMILES and MLS CH3CN profiles increased with altitude and was within 20–35 ppt (20 %–260 %) at 15.7–1.6 hPa. We observed discrepancies of 5–10 ppt (10 %–30 %) between the SMILES CH3CN profiles observed by different spectrometers, and hence, we do not recommend merging SMILES CH3CN profiles derived from different spectrometers. We found that the SMILES CH3CN volume mixing ratio (VMR) in the upper stratosphere has a seasonal maximum in February.


Introduction
Air pollution caused by biomass burning (BB) has become a serious problem with population growth (Marlon et al., 2008). BB events are important sources of various trace gases and particles in the atmosphere (Eagan et al., 1974;Crutzen et al., 1979). The study of atmospheric gas species associated with BB is significant because early estimates of pyrogenic emissions suggested that some atmospheric pollutants from BB could be comparable to those from fossil fuel burning (Crutzen and Andreae, 1990;Seiler and Crutzen, 1980). These emissions could, therefore, significantly affect the global atmosphere and its temperatures (Andreae, 1983). Acetonitrile (CH 3 CN) is a good tracer for BB as 90 %-95 % of CH 3 CN comes from wildfires . The mean lifetime of CH 3 CN in the atmosphere is about 6.6 months, with ocean uptake and reactions with hydroxyl (OH) radicals de Gouw, 2003). Chemical loss of CH 3 CN due to OH radicals occurs primarily in the stratosphere, whereas oceanic loss is dominant in the troposphere. Carbon monoxide (CO) is also a wellknown BB tracer, but it has an atmospheric lifetime of only about 2 months in the free troposphere. CO is also emitted from some anthropogenic sources. Thus, CH 3 CN is not only existent for a longer period of time but also more specific to BB and is, therefore, a better tracer. Arnold et al. (1978) first measured the presence of stratospheric CH 3 CN from the composition of positive ions using active chemical ionization mass spectrometry. CH 3 CN has also been detected using balloon-borne and airborne measurements in the lower stratosphere (Knop and Arnold, 1987;Schneider et al., 1997). More recently, satellite observations of CH 3 CN in the lower stratosphere have been measured using several satellite instruments, such as the microwave limb sounder (MLS) onboard the UARS (Upper Atmosphere Research Satellite) (Barath et al., 1993), the Atmospheric Chemistry Experiment Fourier-transform spectrometer (ACE-FTS) onboard SCISAT-1 (Bernath, 2001), the MLS onboard Aura (Waters et al., 2006), and the Superconducting Submillimeter-Wave Limb-Emission Sounder (SMILES) onboard the JEM (Japanese Experiment Module) of the International Space Station (ISS) (Kikuchi et al., 2010). Previous research reported the volume mixing ratio (VMR) of CH 3 CN mainly in the upper troposphere and lower stratosphere (UTLS; Livesey et al., 2001Livesey et al., , 2004Harrison and Bernath, 2013). However, there are only a few reports of the CH 3 CN VMR in the lower stratosphere to mesosphere.
Here, we derived vertical distribution profiles of CH 3 CN between the lower stratosphere and mesosphere from SMILES observations. We also performed a validation analysis comparing these results with the Aura MLS observation data.

SMILES CH 3 CN observations
The JEM SMILES was operated from 12 October 2009 until 21 April 2010 on the ISS (Kikuchi et al., 2010). The ISS has a non-Sun-synchronous orbit and an inclination angle of 51.6 • to the Equator, which enables it to observe the atmosphere at various local solar times. The antenna field of view of the SMILES instrument was set to point in a 45 • direction leftward from the ISS orbital motion. Lowtemperature system noise (T sys ∼ 350 K) was achieved using 4 K cooled submillimeter-wave superconductive heterodyne receivers . This noise level is 10 times lower than that achieved in previous observations (Kikuchi et al., 2010). A summary of characteristics for SMILES observations is shown in Table 1.
The targeted CH 3 CN transition at 624.82 GHz for (J , K) = (33, 3)-(33, 4) is allocated with a frequency region of Band A (624.32-625.52 GHz) as shown in Fig. 1. The SMILES employed two acousto-optical spectrometers (AOSs) with a bandwidth of 1.2 GHz, denoted as AOS1 and AOS2. The band configuration for AOS1 and AOS2 is summarized in Table 2. The dates of observations made by AOS1 and AOS2 are shown in Fig. 2. The two AOSs detect Band A, B, or C separately, enabling the SMILES to observe two of the three bands simultaneously.
The SMILES Level-2 research (L2r) version 3.0.0 (v3.0.0) product was used in this study. The CH 3 CN VMR profile was retrieved from the measurement spectra data of the Level-1b (L1b) version 008 (v008) product. Major improvements in v3.0.0 over the previous version 2.1.5 included the AOS response function and the a priori temperature pro-  file. This version of the L2r product was derived from the L1b-v008-calibrated spectra, which used version 5.2 of the Goddard Earth Observing System (GEOS-5.2) model as a priori information (e.g., O 3 VMR profile, temperature, and pressure profiles; Rienecker et al., 2008). The details can be found in the JEM SMILES L2r data product guidelines (see http://smiles.nict.go.jp/pub/data/index.html, last access: 20 April 2020). The optimal estimation method (OEM) was used for the retrieval processing. The OEM leads to the maximum a posteriori probability solution (Rodgers, 2000), which minimizes the value of χ 2 described below.
where F(x, b) is the forward model depending on the x state vector and on the known model parameters b, S −1 y the measurement covariance matrix, x a the a priori state of x, and S a the a priori covariance matrix. The detailed retrieval algorithm of the L2r product can be found in Baron et al. (2011) and Sato et al. (2012).
The quality of the retrieval processing was quantified by the chi-squared statistics, or goodness of fit (Eq. 1), and the measurement response (m) was defined as wherex is the solution of the retrieval, A the averaging kernel, D the contribution function, and K the weighting function. K was used to derive m, A, and D (Urban et al., 2004). Details on m are explained by Sato et al. (2014). The χ 2 of CH 3 CN for v3.0.0 had a range of 0.4-0.6. In cases where the measurement response was low, the information was retrieved from the a priori state. Here, the data selection thresholds of χ 2 and measurement response were set as χ 2 < 0.6 and m > 0.80, respectively. Figure 3 shows an example of the retrieval results from a single spectral scan on 4 November 2009 in the tropics at a latitude of less than 20 • , including the retrieved CH 3 CN vertical profile, averaging kernel, and vertical resolution. The vertical resolution was defined as the full width at half maximum (FWHM) for each row of the averaging-kernel matrix. The measurement response of retrieved CH 3 CN, shown as a black solid line in Fig. 3b, is the sum of elements from the averaging kernel on each altitude grid. The measurement response was almost 1 from 30 to 55 km, with a vertical resolution of 7-15 km, decreasing with altitude.

Theoretical error analysis
We theoretically estimated the error in deriving CH 3 CN profiles from SMILES observations by perturbing the model parameters in a forward model (Sato et al., 2012;Kasai et al., 2013;. We used a typical CH 3 CN profile derived using observations from the tropics, where BB (a major source of CH 3 CN) frequently occurs. The total error (E total ) is given by where E n is the error due to spectral noise, E s the smoothing error, and E p the model parameter error. The error due to the spectral calibration was ignored in this study, because the L1b data were updated in this version, and the error due to the spectral calibration was not significant according to previous SMILES error analyses (e.g., Sato et al., 2012). Errors E n and E s were calculated by the following equation: where and where 10 % Here, S n and S s are the error covariance matrices for measurement noise and the errors from S a , respectively. U is the unit matrix. The model parameter error E p includes errors caused by uncertainties in the parameters used in both the forward and inversion calculations. Error sources of the model parameters are summarized in Table 3. The error related to each of the individual model parameters was calculated using the perturbation method following Sato et al. (2012). The total error E p for all of the parameters was calculated using the root sum square of the individual errors. Figure 4 shows the estimated systematic errors. Figure 4a shows the uncertainties in the AOS response function (AOS), the antenna beam pattern (Antenna), the spectral line strength (Strength), the air-pressure-broadening coefficient (γ ), its temperature dependence (n), and their root sum square (Total). The largest error source, ∼ 2 ppt (5 %), was from the airpressure-broadening coefficient (γ ) across the entire pressure range, followed by those from the line intensity (Strength) and temperature dependence of air-pressure-broadening coefficient (n; 1.5 ppt). The error from spectroscopic parameters was more significant than that from the instrumental functions.
In Band A, O 3 and H 37 Cl are observed near the CH 3 CN transition (see Fig. 1). The spectral shapes of O 3 and H 37 Cl should, therefore, influence the retrieval of the CH 3 CN VMR profiles. To estimate the influence from the other spectral lines, the errors due to the spectroscopic parameters γ and its temperature dependence n of the O 3 and H 37 Cl lines were also calculated. γ and temperature dependence of γ were perturbed for each species and are expressed as O 3 γ , O 3 n, H 37 Clγ , and H 37 Cln. As shown in Fig. 4b, H 37 Clγ is the largest error source, in which the maximum absolute difference was 1.1 ppt. Error analyses performed for O 3 and ClO demonstrated that the error caused by other molecular spectral lines was negligible as they have high, isolated line strengths (Sato et al., 2012;Kasai et al., 2013). In the case of CH 3 CN retrieval, however, the total error caused by uncertainties in other molecular spectroscopic parameters was comparable to the error caused by CH 3 CN spectroscopic parameters. The error due to H 37 Cl was larger than that from O 3 at each pressure level.
The measurement noise and smoothing error from a single scan are shown in Fig. 5a. These errors were considered as a random error for a CH 3 CN profile. The SMILES CH 3 CN total error consists of both the systematic and random errors. Figure 5b shows the total systematic error, the random error, and the total error averaged by the number of profiles (N = 1, 10, and 100). For a single scan, the random error was larger than the systematic error. However, the random error averaged by 100 profiles was comparable to the systematic error, except for the highest systematic error found at a pressure level of about 1 hPa.

Comparison with Aura MLS
In this section, we compare SMILES CH 3 CN observations with Aura MLS observations and discuss the validity of SMILES observations. Figure 6 shows CH 3 CN vertical profiles observed by AOS1 and AOS2 (Fig.6a) and the absolute (Fig.6b) and relative (Fig.6c) differences between AOS1 and AOS2 observed in equatorial regions (20 • S-20 • N) from March until April 2010, when AOS1 and AOS2 were alternating at a ratio of 1 : 3. The error bars shown in Fig. 6a are standard deviations (1σ ) of the CH 3 CN VMR observations retrieved from SMILES pressure grids for AOS1 (red) and AOS2 (blue). The relative difference between AOS1 and AOS2 is approximately 12 ppt (30 %) with the maximum at 15.7 hPa. It is likely that these discrepancies between the two AOSs result from sensitivity differences. We believe that the sensitivity differences indicate inherent sensitivity differences between the two AOSs derived from instrumental characterization determined during manufacturing. Kasai et al. (2013) also reported the discrepancies between the two different AOSs, albeit for the analysis of ozone profiles using the SMILES L2 version 2.1.5 product. As mentioned above, in this analysis, we used the SMILES L2r version 3.0.0 product that improves the AOS response function. However, there may still be disagreement between the two AOSs. The relative difference between the two AOSs decreases to less than 10 % at an upper pressure altitude of 4.8 hPa, except at 0.3 hPa, showing a good agreement of the two AOS observations from the middle stratosphere.
We also investigated seasonal variation in SMILES CH 3 CN observations for each altitude grid as shown in Fig. 7. This figure shows daily scatterplots and daily averages for AOS1 (shaded red area) and AOS2 (shaded blue area) observations. The red circles and bars represent the daily mean values and 1σ standard deviations, when more than 100 ob-servation points were obtained in 1 d. As shown in Fig. 2, at lower altitudes (28 to 36 km), the difference between the two AOS observations was significantly larger, especially from December until the beginning of January. However, in the upper stratosphere, there was no difference between the two AOS observations, and the standard deviations decreased with altitude. In terms of seasonality, CH 3 CN levels peaked in February, seen from approximately 40 to 52 km, where the difference between the two AOSs is negligible.

Comparison with Aura MLS v4.2 data
We investigated the difference between CH 3 CN VMRs obtained from SMILES and MLS observations. We set the data quality thresholds and the coincidence selection criteria for the SMILES and MLS observations as summarized in Table 4. The MLS data quality criteria were based on the MLS v4.2 Level-2 data quality and description document.
The geolocation and measurement time criteria were determined as follows: the distance of measurement location within 300 km the difference in the measurement time within 6 h.
We investigated the diurnal variation in SMILES CH 3 CN observations at several altitudes (32, 40, and 48 km) for AOS1 and AOS2 individual observational periods and confirmed that there is no diurnal variation in stratospheric CH 3 CN observations. Figure 8 shows the distribution of coincident points satisfying these criteria at 8.6 hPa. The interpolation of VMRs was carried out using a linear interpolation with respect to the logarithmic pressure levels. There are, on average, 10 coincident points in each bin at this pressure level, and the total coincident data number was 17 910.
For comparing the SMILES and the MLS observations, the mean absolute difference, abs , and relative difference, rel , at the pressure levels, p, between coincident CH 3 CN profiles of the two observations were calculated as follows: where N (p) is the number of coincidences at p, x s (p) and x m (p) are the VMRs at p for SMILES and MLS observations, and the reference (x p ) is x p = 1 2 (x s (p) + x m (p)).

Aura MLS v4.2
The MLS has been onboard the Aura satellite since 2004 and has observed CH 3 CN levels from the lower to upper stratosphere. This satellite was launched in the Sun-synchronous orbit with an Equator-crossing time of 13:45 LT (ascending)     etc.) using thermal emission data from the atmosphere. The CH 3 CN VMR values were retrieved from the MLS observation data using the optimal estimation method. Details on the retrieval algorithm are described in . The MLS uses spectral bands of 118, 190, 240, and 640 GHz and 2.5 THz, observing CH 3 CN from 640 GHz spectral regions (Waters et al., 2006). MLS Level-2 CH 3 CN profiles were observed in 640 GHz spectral regions. Although the pressure range of a retrieved MLS CH 3 CN is 147 to 0.001 hPa, the pressure range of CH 3 CN version 4.2.0 is 46-1.0 hPa . Figure 9 shows the vertical profiles, the absolute differences, and the relative differences between SMILES AOS1 and AOS2 and MLS CH 3 CN observations. Figure 9a indicates good agreement among the three observations in the range from 15.7 to 4.8 hPa. Across the range of the pressure levels, the absolute difference and the relative difference among the three observations were −15 to 25 ppt and 20 % to 80 %, respectively. The difference between the SMILES and the MLS observations increases with altitude, from a pressure level of 8.6 hPa. It should be noted that the discrepancies between the two instruments were exaggerated at upper pressure levels from 2.8 hPa, although negative values of CH 3 CN VMR derived from MLS have no physical meaning. Overall, the variance in SMILES observations is less than that in MLS observations, as SMILES T sys was more than 10 times less than that of MLS, indicating that the SMILES has an advantage in the upper stratosphere. The SMILES was also able to observe the CH 3 CN VMR in the upper stratosphere with a much lower uncertainty of ∼ 20 ppt; the uncertainty in the MLS CH 3 CN VMR was approximately 100 ppt at this altitude. The differences in the CH 3 CN VMRs observed by the two AOSs were sufficiently small in comparison with the difference between SMILES and MLS observations. The theoretical systematic error (dashed blue lines in Fig. 9b) derived in Sect. 3 was less than the differences between SMILES and MLS observations, except at 8.9 hPa. We also investigated latitudinal and seasonal variations between the two observations. Figure 10 shows the seasonal variation in SMILES and MLS CH 3 CN observations and the absolute differences for each pressure level at coincident points, as a function of latitude. The left column represents the SMILES CH 3 CN VMR in units of parts per trillion, which was separated into two AOS observations. The middle column represents the MLS CH 3 CN VMR, and the right column represents the absolute differences between SMILES and MLS observations. At lower altitudes of 15.7 and 8.6 hPa, SMILES observations were overestimated when compared with the MLS observations, while at upper lev-els (< 4.8 hPa) SMILES observations were underestimated by up to 40 ppt. At every pressure level, SMILES CH 3 CN VMRs were higher in the tropics (20 • S ∼ 20 • N). However, in the case of MLS observations at higher pressure levels, MLS observations were negatively biased and the trend was not conspicuous. The MLS CH 3 CN levels in the tropics were ambiguous at pressure levels above 4.8 hPa, indicating that, in the upper stratosphere, it is hard to observe latitudinal and seasonal trends of CH 3 CN, due to the large uncertainty in MLS observations. At pressure levels above 4.8 hPa, SMILES CH 3 CN observations in February were greater than those in the other periods, which can also be seen in the MLS results. This result indicates that CH 3 CN in the upper stratosphere reaches its seasonal maximum in February.

Conclusions
Our analysis demonstrates the validity of using SMILES observations to measure CH 3 CN profiles. We were able to successfully derive vertical profiles and observe seasonal variation in CH 3 CN in the stratosphere, using SMILES observations. This study is the first of its kind to describe the data observed by the satellite instruments for the CH 3 CN VMR from the upper stratosphere to lower mesosphere with a much lower uncertainty of 20 ppt. Error analysis showed that random error was the dominant source of uncertainty (around 25 ppt at 15.7 hPa) in the measurement altitude range, resulting in a better precision by a factor of 2 or more than that of Aura MLS. The uncertainty in air pressure broadening was the dominant systematic error source, with a maximum difference of 2.0 ppt (5 %). The random error from a single-scan spectrum was more than 2 times larger than systematic error at 15.7 hPa, while the random error averaged with 100 spectra was comparable to systematic error. SMILES and Aura MLS observations were in agreement in the stratosphere from 15.7 to 4.8 hPa. At upper pressure levels, the difference between the two observations increased up to 35 ppt (260 %) because of greater uncertainty in Aura MLS observations and because CH 3 CN levels were at their seasonal maximum. The theoretical systematic error and the relative difference in the SMILES measurements compared to MLS measurements were, respectively, 10 and 35 ppt at altitudes between 15.7 and 1.6 hPa (28-44 km). Furthermore, the two AOSs show comparable errors (∼ 10 ppt) at 0.93 hPa to 0.29 hPa (approximately 48-56 km) and at lower pressure levels, implying the reliability of SMILES CH 3 CN observations.
Author contributions. TF designed the study and performed the analysis. YK designed the study and provided the SMILES data. TOS provided the retrieval code and contributed to data analysis and interpretation. TY, SN, YU, and KT contributed to data analysis and reviewed the manuscript. NY supervised and reviewed the manuscript. TF wrote the manuscript with contributions from all coauthors.
Competing interests. The authors declare that they have no conflict of interest.
Acknowledgements. SMILES is a collaborative project of the National Institute of Information and Communications Technology and the Japan Aerospace Exploration Agency (JAXA). We deeply appreciate Hideo Sagawa (Kyoto Sangyo University) for developing the SMILES L2r product v3.0.0. We are grateful to Atsushi Hirakawa and Kota Kuribayashi for supporting this study. We are grateful to Nathaniel Livesey for providing the MLS version 4.2.0 product.
Financial support. This research has been supported by the Ministry of Internal Affairs and Communications (grant nos. 0155-0285 and 0155-0093).
Review statement. This paper was edited by Alyn Lambert and reviewed by Hugh C. Pumphrey and one anonymous referee.