Seasonal evaluation of tropospheric CO 2 over the Asia-Pacific region observed by the CONTRAIL commercial airliner measurements

Measurement of atmospheric carbon dioxide (CO2) is indispensable for top-down estimation of surface CO2 sources/sinks by an atmospheric transport model. Despite the growing importance of Asia in the global carbon budget, the 10 region has been monitored for atmospheric CO2 only sparsely and our understanding of atmospheric CO2 variations in the region (and thereby that of the regional carbon budget) is still limited. In this study, we present climatological CO2 distributions over the Asia-Pacific region obtained from the CONTRAIL (Comprehensive Observation Network for Trace gases by Airliner) measurements. The high-frequency in-flight CO2 measurements over 10 years reveal a clear seasonal variation of CO2 in the upper troposphere (UT), with a maximum occurring in April–May and a minimum in August– 15 September. The CO2 mole fraction in the UT north of 40° N is low and highly variable in June–August due to the arrival of air parcels with seasonally low CO2 caused by the summertime biospheric uptake in boreal Eurasia. For August–September in particular, the UT CO2 is noticeably low within the Asian summer monsoon anticyclone associated with the convective transport of strong biospheric CO2 uptake signal over South Asia. During September as the anticyclone decays, a spreading of this low CO2 area in the UT is observed in the vertical profiles of CO2 over the Pacific Rim of continental East Asia. 20 Simulation results identify the influence of anthropogenic and biospheric CO2 fluxes in the seasonal evolution of the spatial CO2 distribution over the Asia-Pacific region. It is inferred that a substantial contribution to the UT CO2 over the northwestern Pacific comes from continental East Asian emissions in spring, but in the summer monsoon season, the prominent air mass origin switches to South Asia and/or Southeast Asia with a distinct imprint of the biospheric CO2 uptake. The CONTRAIL CO2 data provide useful constraints to model estimates of surface fluxes and to the evaluation of the 25 satellite observations, in particular for the Asia-Pacific region.

Thank you for your thorough corrections.We have made corrections for all the above places according to the suggestions.
P7 line 21 onwards.Reviewer 1 has suggested the location of airports near cities as one explanation for t he difference between the Asian profiles in summer compared to those from North America, and your additional paragraph discussing this is useful.I wonder whether it would also be appropriate to mention here that coastal/inland differences may also be important.I think most of the North American sites are inland, and are consistent with the features you see at DEL which is also inland.The other Asian sites are coastal, and may be influenced by sea-breezes in summer which could also reduce the signal from biospheric uptake.
We thank the editor for the important suggestion about the coastal/inland differences.
All of our measurement sites, except DEL, are indeed in coastal areas and under considerable influence of maritime air with a limited imprint of terrestrial uptake.We have added the following sentence (underlined) in the paragraph to concisely adopt your suggestion.More detailed discussion, including influence of sea breeze, needs further analysis of data for individual sites, given that even the Asian coastal sites presented here have variety of meteorological conditions."The seasonal CO 2 cycles with spring maxima and summer minima, typical for the northern hemispheric troposphere (Stephens et al., 2007;Sweeney et al., 2015), are to some degree obvious across regions over the 12 airports (Fig. 4) in Asia.However, a clear difference from those outside Asia is the general absence of a dramatic decrease of CO 2 near the ground in the summer.In other words, the contoured low ΔCO 2 in the summer is apparently "floating" in the FT and not connected to the ground, implying that the observed vertical profiles in the summer are not dictated by overwhelming uptake underneath.This feature is observed at all airports except DEL.It is noteworthy that our measurement sites are all located in coastal areas except DEL, while previous measurements at inland sites generally showed more pronounced decreases of CO 2 near the ground in the summer (Stephens et al., 2007;Sweeney et al., 2015).In contrast, the springtime maximum ΔCO 2 extends from the ground to the UT, indicating that the surrounding or upwind regions of most airports are strong sources of CO 2 during that season."Sincerely, Taku Umezawa

Introduction
Actions for mitigating climate change require accurate knowledge of global budgets of greenhouse gases.It has been estimated that approximately one-half of CO 2 emissions had remained in the atmosphere during the period 1959-2010, with the rest taken up by land and ocean sinks (Ballantyne et al., 2012).With a rapidly growing economy in recent decades, Asia has become increasingly important in the global carbon budget.China is now the world's largest CO 2 emitter, with and India, Japan, and the Republic of Korea are all in the world's top 10 emitting nations (Boden et al., 2016).At the same time, Asia has gone through significant land use and land cover changes, impacting the magnitude and the spatial distribution of terrestrial carbon fluxes (e.g.Calle et al., 2016;Cervarich et al., 2016).However, there are still large uncertainties in the estimates of every component of the Asian carbon budget.
To estimate surface CO 2 fluxes, atmospheric transport models have been conventionally constrained by various surface measurement networks (e.g.Gurney et al., 2002;Patra et al., 2008).But due to the sparseness of the surface measurement sites in Asia, an increasing number of modeling studies that have focused on the Asian carbon budget (e.g.Patra et al., 2011;Niwa et al., 2012;Zhang et al., 2014;Jiang et al., 2014Jiang et al., , 2016) ) in recent years started to incorporate CO 2 data taken by commercial airliners, such as CARIBIC (Civil Aircraft for the Regular Investigation of the atmosphere Based on an Instrument Container) (Brenninkmeijer et al., 2007) and CONTRAIL (Comprehensive Observation Network for TRace gases by AIrLiner) (Machida et al., 2008).It has been demonstrated that by incorporating the CARIBIC and CONTRAIL data, model estimates of the Asian CO 2 fluxes have been significantly improved (Patra et al., 2011;Niwa et al., 2012;Shirai et al., 2017).
The dominant seasonally-varying atmospheric circulation regime that has an important influence on the variations of atmospheric trace gases throughout the troposphere over Asia is the monsoon circulation (e.g.Lawrence and Lelieveld, 2010).Seasonal variations in trace gases observed at ground stations, as well as in the upper troposphere (UT), have been found to be influenced by the monsoon circulation (Xiong et al., 2009;Park et al., 2009;Randel et al., 2010;Schuck et al., 2010).In this study, we focus on some of the less-well studied features of the CO 2 distributions in that are association associated with the Asian monsoon.In this respect, measurements from commercial airliners that fly in the UT are analyzed to provide invaluable insight into the seasonality of the vertical dynamical connection between atmospheric CO 2 and the surface flux.
The CONTRAIL project has obtained high-frequency CO 2 measurements along flight tracks, as well as vertical profiles during the ascent and descent over airports, providing a more comprehensive time-dependent three-dimensional spatial distribution of atmospheric CO 2 .Analyses of seasonal variations and meridional transport of CO 2 in the free troposphere (FT; including the UT) and in the lowermost stratosphere using data from CONTRAIL have been presented by Sawa et al. (2008Sawa et al. ( , 2012)).Sawa et al. (2012) analyzed the CONTRAIL CO 2 data for the period 2005-2010; the number of flights used in the study exceeded 5000, giving nearly 3 million CO 2 measurement values.By the end of 2015, we have more than doubled the amount of measurement values, allowing us not only to update their results but also to explore additional spatiotemporal CO 2 variations.The present study addresses climatological CO 2 distributions over the Asia-Pacific region and the influence of Asian surface fluxes under varying seasonal atmospheric conditions, as well as to provide a baseline for future optimal use of the CONTRAIL CO 2 data.In Section 2, we describe the CONTRAIL CO 2 measurements, as well as data analysis procedures, and model simulations to aid in the interpretation of the observations.In Section 3, we evaluate seasonal distributions of CO 2 in both observation and model data.In Section 4, we discuss three interesting features found by our measurements: the summertime low CO 2 associated with the Asian summer monsoon, another zone of low CO 2 originating in the boreal summer biospheric uptake, and the springtime high CO 2 observed in East Asia.Concluding remarks are given in Section 5.

Experimental
The CONTRAIL project (http://www.cger.go.jp/contrail/) deploys two types of instruments onboard aircraft: Continuous CO 2 Measuring Equipment (CME) and Automatic air Sampling Equipment (ASE).We refer to Machida et al. (2008) for details, and only a brief description on the CME unit is given here.The CME measures CO 2 mole fractions onboard the aircraft using a non-dispersive infrared gas analyzer (NDIR; LI-840, LI-COR Biogeosciences).As of May 2018, installation of the CME is certified for eight Boeing 777-200ER and two Boeing 777-300ER aircraft of Japan Airlines (JAL).
Once installed, the CME is operated automatically using the aircraft's flight navigation data until it is unloaded from the aircraft two months later.The CME samples air from the air conditioning system on the aircraft.The flow rate and the absolute pressure of the sample air in the NDIR cell are maintained at a constant level to minimize signal drift.The measured sample values are compared with two working standard gases (CO 2 in air) in high-pressure cylinders (2 L) installed inside the CME and the measurements are traceable to the NIES (National Institute for Environmental Studies)-09 CO 2 scale.Mole fraction of CO 2 in dry synthetic air in µmol mol −1 is reported in ppm in this paper.The latest results from the Round Robin intercomparison experiment show that the NIES-09 CO 2 scale differs from the WMO-CO2-X2007 scale by less than 0.1 ppm (http://www.esrl.noaa.gov/gmd/ccgg/wmorr/wmorr_results.php).The standard gases are currently introduced into the NDIR cell every 14 min during the ascent/descent portion of the flight and every 60 62 min during the constant altitude portion of the flight (cruise) typically at 8-12 km.i.e. during the ascent/descent (cruise) measurement cycle, sample air is measured for 12 (60) minutes, then standards 1 and 2 are measured for 1 minute each.These standard gas intervals were initially 10 min during ascent/descent and 20 min during cruise until December 2005; the 20-min interval was then changed to 40 min until October-November 2011.The CME data are recorded as 10-s averaged measurements during ascent/descent (~100-m intervals in altitude) and at 1-min intervals during cruise (~15 km intervals horizontally).The data are rejected for 40 s after switching the gas stream and also when a standard deviation for the average period exceeds 3 ppm and when any failure in pressure/flow control is observed in the CME data record.To avoid heavy pollution around airports, the CME is not operated within 2000 ft (609.6 m) of the ground surface (this altitude was initially set to 1200 ft until March-June 2007).
The overall analytical precision of the CME is estimated to be < 0.2 ppm.
For the 10-year period from 2005 to 2015, we collected > 7 million CO 2 data points from > 12 thousand flights all over the world.The CME measurements over the Asia-Pacific region are shown in Fig. 1.Flights from Japan to Southeast Asia (Bangkok (BKK), Singapore (SIN) and Jakarta (CGK)) provide measurements over the East China Sea, the South China Sea, the Indochina Peninsula and the maritime continent.These measurement areas are substantially overlapped by flights to the continental East Asia (Incheon (ICN), Shanghai (SHA) and Hong Kong (HKG)) and to Taipei (TPE).Flights to Delhi (DEL) provide a unique opportunity for observations over the continental Asia.In addition, extensive measurements from Japan to the north, to the east and to the south are achieved from flights to Europe, to the North America and Hawaii, and to Australia, respectively.The major airports where CONTRAIL CME measurements in Asia are made, along with the number of vertical profile measurements of CO 2 over each airport, are listed in Table 1.Vertical profile data with less than 10 CO 2 data points are not used in this study.As indicated in the table, the largest number of CO 2 data has been obtained over the Tokyo Narita (NRT) airport with over 7000 vertical profiles, followed by Tokyo Haneda (HND) with over 3600 profiles.Figure 1b shows the number of monthly vertical profiles taken over the airports listed in Table 1.As seen in this figure, the CME measurements have acquired over 30 vertical profiles per month (colored red; i.e. at least one or multiple ascent/descent flights every day on average) over NRT and HND.Although measurements over other airports are less regular compared to these airports, data from sites where a substantial number of vertical profiles have also been taken over other airports and those data that cover much of the year are presented included in this study.

Data Analysis
In this study, we focus on CO 2 variations in the troposphere.Observations in the UT are, however, quite often influenced by stratospheric air that has distinct characteristics in the atmospheric composition (e.g.Hoor et al., 2002;Sawa et al., 2004Sawa et al., , 2008Sawa et al., , 2015)).These data are excluded from the dataset based on potential vorticity (PV) values.PV at the location and time of each CO 2 measurement taken by CONTRAIL is calculated from the JCDAS (Onogi et al., 2007) and the JRA-55 (Kobayashi et al., 2015) reanalysis datasets (the latter being used since 2014), and any data accompanied by PV values of > 2 PVU (1 PVU = 10 −6 m 2 s −1 K kg −1 ) are excluded.It has been found that the 2-PVU criteria is relatively robust in separating out the CO 2 measurements in the UT that are stratospherically influenced from those that are not (Sawa et al., 2008(Sawa et al., , 2015)).
In total, 33% of the CONTRAIL CME CO 2 data points collected at altitude > 8 km have been identified as stratospheric, although this fraction varies with altitude, latitude and season (i.e.flight routes).In this study, the UT is defined as the region at altitudes of > 8 km and with PV of < 2 PVU.Note that most commercial airliners cruise at altitudes of 9-12 km, and that this cruising altitude region is deemed in large part stratospheric at higher latitudes (e.g.86% and 64% of the data taken at > 40° N was stratospheric in January and July, respectively), whereas it mostly resides in the UT at lower latitudes (< 10% of the data at < 30° N was stratospheric throughout the year).
To calculate climatological distributions of CO 2 in the troposphere, we apply a method similar to Sweeney et al. (2015).
(1) The long-term trend of the flask-based CO 2 mole fraction data at Mauna Loa (MLO; 19.54°N, 155.58°W, 3397 m.a.s.l.), Hawaii, obtained from NOAA/ESRL/GMD (National Oceanic and Atmospheric Administration/Earth System Research Laboratory/Global Monitoring Division; available at ftp://aftp.cmdl.noaa.gov/data/) is calculated using a digital filtering technique (Nakazawa et al., 1997).The dataset goes to the end of 2015.In general, the long-term CO 2 trend at MLO is representative of the large-scale clean atmosphere and thus has been used as a reference site (Sweeney et al., 2015).( 2) Deviations of individual CO 2 data points from the long-term trend (ΔCO 2 ) are calculated as where lat, lon, alt, t are latitude, longitude, altitude and time of individual CONTRAIL CME data points, respectively, and Trend CO 2 at MLO is the long-term trend curve derived as described above.The CONTRAIL CO 2 data over 12 airports in Asia colorcoded by altitude are presented in Fig. 2, together with the MLO CO 2 data and the calculated long-term trend.In this study, we present results from the statistical analysis of the ΔCO 2 data (i.e.deviations of the individual data points from the black line in each panel of Fig. 2) for the years 2005-2015.

Model simulation
To better understand processes that generate the observed tropospheric distribution of CO 2 over the Asia-Pacific region, we analyze CO 2 simulated by the model NICAM-TM (Nonhydrostatic Icosahedral Atmospheric Model-based Transport Model) (Satoh et al., 2014).Details of the NICAM-TM CO 2 simulation and the evaluation of its performance have been presented by Niwa et al. (2011Niwa et al. ( , 2012)).The atmospheric CO 2 transport is calculated using the 6-hourly meteorological data nudged to the JRA-55 reanalysis.The horizontal model grid interval is about 240 km and the number of vertical model layers is 40.For CO 2 simulation, fossil fuel (FF) emissions are obtained from the CDIAC (Carbon Dioxide Information Analysis Center) database (version 2013) (Andres et al., 2013), while fire emissions are from the GFED (Global Fire Emission Database version 3.1) (van der Werf et al., 2010).A priori terrestrial biospheric (BIO) fluxes are derived from the CASA (Carnegie-Ames-Stanford Approach) model (Randerson et al., 1997).The air-sea exchange is based on the JMA (Japan Meteorological Agency) ocean flux data (Iida et al., 2015).The BIO fluxes are optimized in the NICAM-TM model inversion by using the GLOBALVIEW data (http://www.esrl.noaa.gov/gmd/ccgg/globalview/)and the CONTRAIL data in the FT (Niwa et al., 2012).Thus, the simulated atmospheric CO 2 is obtained from the optimized fluxes.We also examine simulated CO 2 fields driven by two different emission fluxes: one by FF and the other by BIO (hereafter referred as FF CO 2 and BIO CO 2 , respectively).For comparison, the simulated data are sampled at times and locations coincident with the individual CONTRAIL CME data points, and processed in the same manner; stratospheric data are excluded by the model PV values; all CO 2 data points are detrended by the MLO long-term trend in the model.

Seasonal cycle of CO 2 in the UT over the Asia-Pacific region
Figure 3 presents monthly averaged distributions of the UT ΔCO 2 over the Asia-Pacific region (left panels) along with histograms in the respective 5° latitude bands (right panels).In the left panels, the black arrows indicate monthly averaged horizontal wind at 250 hPa pressure surface from the JCDAS reanalysis.We note that monthly CO 2 distributions from the CONTRAIL data previously presented by Sawa et al. (2012) were calculated as averages in 20° (longitude) × 10° (latitude) bins.In this study we were able to increase the spatial resolution to 5° × 5° since we have more data.As seen in Fig. 3, the UT ΔCO 2 undergoes a clear seasonal cycle that varies significantly with latitude and longitude.
In January-February, the UT ΔCO 2 is relatively uniform in space (Figs.3a and 3c), except in regions > 35° N where the histograms show occurrences of higher ΔCO 2 values (Figs.3b and 3d).In March, high ΔCO 2 values are apparent in regions > 30° N over northern Japan and downwind (Fig. 3e) where significantly increased frequency of high ΔCO 2 up to 6 ppm are observed (Fig. 3f).This feature becomes more pronounced in April (Fig. 3h) with expanded areas of high ΔCO 2 around Japan (Fig. 3g).By May, regions with high ΔCO 2 extend to > 20° N (Figs.3i and 3j).
By June, the observed high ΔCO 2 values over Japan and the northwestern Pacific nearly disappear (Fig. 3k).A significant fraction of the low ΔCO 2 values down to −6 ppm and lower is observed at latitudes > 35° N (Fig. 3l).Due to these low CO 2 values appearing at northern latitudes, the latitudinal gradient of the UT ΔCO 2 starts to reverse (i.e.northward positive to negative) after June, aided by a moderately elevated CO 2 observed at 15°-30° N. In July, we begin to see very low ΔCO 2 values below −6 ppm in high latitude regions (> 40° N), particularly over the boreal Eurasia (Figs. 3m and 3n).To the south, only very small spatial gradients are observed.
In August, we see the CO 2 decrease broadly at all latitudes over the Asia-Pacific region, with a distinctly low ΔCO 2 values forming over South Asia to Southeast Asia (Fig. 3o).The UT wind field shows anticyclonic wind circulation pattern over this region.This wind structure is coincident with the distinct low ΔCO 2 observed over the continent, indicating that the low CO 2 air mass is confined within the UT anticyclone.This clear CO 2 spatial structure associated with the anticyclone is for the first time depicted by the improved spatial resolution of the CONTRAIL data since Sawa et al. (2012).It is noted that such distinct low-CO 2 structure does not appear until July, despite the fact that the anticyclonic wind pattern starts in June (Figs.3k and 3m).
Moving into September, we see a further decrease in ΔCO 2 across the wider Asia-Pacific region (Figs.3q and 3r).The persistent UT anticyclonic structure is still observable in both ΔCO 2 and wind fields, but the sharp boundary along the East Asian coast that was seen in August (i.e.longitudinal gradient or contrast between the continent and the ocean) is now to some degree blurred (see also Fig. 7a).In October, the anticyclonic low-ΔCO 2 feature diminishes and ΔCO 2 is now relatively uniform in the observation region.Thereafter ΔCO 2 increases as a whole during the winter until the return of the spring.

Vertical gradients of CO 2 over Asian cities
Figure 4 presents a climatology of seasonal variations and vertical profiles of ΔCO 2 over 12 airports in Asia, as uniquely obtained by the CONTRAIL observation.We consider these figures to represent large-scale (regional) features in the FT as a result of detrending and binning of the data (500-m altitude and 14-day averages from multiple-year data).At lower altitudes, relatively local features can be visible due to boundary layer (BL) processes and flight route biases near the airports, but examining such smaller-scale phenomena in detail is out of beyond the scope in of this study.We also calculate for each airport, altitude variation of the standard deviation (SD) of ΔCO 2 using two-weekly values obtained at each altitude bin (Fig. 5), as an extended update of Shirai et al. (2012) who addressed synoptic-scale CO 2 variability over NRT.This type of analysis is made possible due to CONTRAIL's high-frequency measurements during ascent/descent over the airports.Stephens et al., 2007), the CONTRAIL observations provides a greater spatiotemporal insight into regional carbon cycling processes.One of the remarkable features found in vertical CO 2 profiles from other regions is the dramatic decrease in CO 2 toward the ground in the summer period at mid-continental sites of the northern hemisphere (see Fig. S3 of Stephens et al., 2007 andFig. 5 of Sweeney et al., 2015).Below we show that the vertical CO 2 profiles and their seasonal changes observed by CONTRAIL in Asia are interestingly different from those reported by the previous measurements in other regions.
The seasonal CO 2 cycles with spring maxima and summer minima, typical for the northern hemispheric troposphere (Stephens et al., 2007;Sweeney et al., 2015), are to some degree obvious across regions over the 12 airports (Fig. 4) in Asia.
However, a clear difference from those outside Asia is the general absence of a dramatic decrease of CO 2 near the ground in the summer.In other words, the contoured low ΔCO 2 in the summer is apparently "floating" in the FT and not connected to the ground, implying that the observed vertical profiles in the summer are not strongly influenced dictated by overwhelming uptake underneath.This feature is observed at all airports except DEL.It is noteworthy that our measurement sites are all located in coastal areas except DEL, while previous measurements at inland sites generally showed more pronounced decreases of CO 2 near the ground in the summer (Stephens et al., 2007;Sweeney et al., 2015).In contrast, the springtime maximum ΔCO 2 extends from the ground to the UT, indicating that the surrounding or upwind regions of most airports are strong sources of CO 2 during that season.
It is likely that some features shown in Fig. 4, especially in the BL, are due to the influence of nearby CO 2 emissions.
Indeed, at some airports, large elevations of CO 2 values have been observed frequently in the BL.In order to reduce possible bias due to such pollution events, we did redraw Figure 4 with median ΔCO 2 values, instead of averaged values.We found no clear visual difference in the overall features discussed below.In fact, differences between average and median are mostly < 1 ppm even below 2 km at all airports, except SHA and HKG where the value is ~1.5 ppm on yearly average.Although pollution events are observed frequently over these two airports (as described below), we consider such "airport bias" in the climatological vertical profiles to be small within the scope of this study.Influence of nearby city emissions on the CONTRAIL observations will be addressed in our future publication.
NRT and HND, Japan, are the two airports over which the largest number of CO 2 measurements has been collected by the CONTRAIL CME, giving relatively smooth climatology of ΔCO 2 (Figs.4b and 4c).Seasonal and vertical characteristics of CO 2 over HND and NRT are quite similar to each other.In the FT, ΔCO 2 reaches its seasonal maximum in the spring (April-May) and minimum in the late summer to early autumn (September-October), with the seasonal amplitude in general decreasing with altitude.We also find substantially enhanced SD below ~2 km over HND and NRT in the winter (November-April) and summer (June-August) (Figs.5b and 5c).The high summer variability propagates up to higher altitudes (~6 km), presumably associated with enhanced vertical mixing in the summer.The vertical gradient in CO 2 is small (< 2 ppm) during the summer period (June-September), but a clear gradient is detectable for the rest of the year.These features are commonly observed over the other two Japanese airports Nagoya (NGO) (~260 km west of Tokyo) and Fukuoka (FUK) (~880 km west-southwest of Tokyo and ~850 km east-northeast of Shanghai).Also notable is that, in September, CO 2 decreases with altitude, this feature being observed widely over these four Japanese cities. ΔCO 2 undergoes a seasonal cycle with spring maximum and summer minimum also over ICN (~ 570 km northwest of FUK) (Fig. 4a), but the minimum occurs in late August to early September, about a month earlier than observed over the aforementioned Japanese airports.
The low ΔCO 2 in the BL is a characteristic that is not observed over Japan.
Along the east coast of the continental East Asia, measurements are obtained over three cities: SHA, HKG and TPE (Figs. 4d,4h and 4i,respectively).ΔCO 2 increases from September until May when it reaches a seasonal maximum.The seasonal minimum in the UT appears in September-October, lagging the lower troposphere (LT) minimum by about a month.
We see remarkably high ΔCO 2 values in the BL over SHA and HKG, these phenomena being particularly pronounced over SHA where we frequently observe ΔCO 2 enhancements of > 20 ppm below 1 km.The elevated ΔCO 2 in the winter season (November-April) is also characterized by high variability (Figs.5d and 5h).Although the seasonal and vertical characteristics of CO 2 over TPE appear to be essentially similar to those over SHA and HKG, our measurements are sparse during May-October.
ΔCO 2 over DEL shows a unique seasonal variation.We note that DEL is the only inland site, whereas the all other sites presented in this study are located near the coast.Prominent is the strong CO 2 drawdown throughout the troposphere in August-September, with very little vertical gradient in the FT due to vigorous vertical mixing (Fig. 4g).Another interesting feature is the relatively low ΔCO 2 in the BL (< ~3 km) during January-March.This wintertime CO 2 stagnation over DEL was recently attributed to uptakes by crops (mainly wheat) grown in the winter season in the surrounding region (Umezawa et al., 2016).
Clear seasonal CO 2 variations are also visible over BKK (Fig. 4j).The seasonal maximum happens in March-April in the LT and propagates upward.These 2 months correspond to a period of enhanced ΔCO 2 variability near the ground (Fig. 5j).Over SIN in the Southeast Asia, ΔCO 2 exhibits a measureable seasonal variation (Fig. 4k).The seasonal variation in the FT over SIN is similar in phase with that observed over BKK, but with comparatively reduced magnitude.It should be also noted that, over SIN, the vertical gradient of ΔCO 2 is small throughout the year.A maximum vertical ΔCO 2 difference is only ~2 ppm observed in the boreal spring.Lastly, for CGK in the tropical Asia (Fig. 4l), the observed seasonality in ΔCO 2 in the LT is hard to characterize due to relatively large variability.But interestingly, the seasonal phases are apparently different below and above 2.5 km.Below that height, relatively high ΔCO 2 values appear during August-October, while, over the same period, ΔCO 2 in the FT decreases until the October minimum.

Simulated CO 2 distributions in the UT over the Asia-Pacific region
In Fig. 6, simulated (second column) monthly CO 2 distributions in the UT are compared to the observations (first column).The model outputs are sampled at location and time coincident with the observation and analyzed in the same manner as the measurements.The third and fourth column panels show the simulated FF CO 2 and BIO CO 2 , respectively.
We do not present a contribution from biomass burning, since it is relatively minor (though not negligible) in evaluating the seasonal variation.Also shown are monthly CO 2 distributions at 250 hPa pressure surface for 2011 (last column).We have chosen the model year 2011 as a representative year whose seasonal CO 2 distribution patterns are not exceptional, although the simulated CO 2 exhibits interannual variation due to year-by-year changes in meteorology and CO 2 fluxes.Note that the model data in the last column are simple monthly averages at model resolutions; thus, both the UT and stratospheric model data are included and avoids sampling bias that might result from data availability as in the observations.However, similarity in the features between the model and observed results, together with the model monthly averages, attest to the fact that the CME-based CO 2 distribution is representative of the seasonal CO 2 climatology in the UT.
By comparing with the observations, we see that NICAM-TM (second column) is able to reproduce the overall general seasonal features of the observed CO 2 distribution pattern in the UT over the Asia-Pacific region.The model simulation (second column) shows seasonal CO 2 elevations centered at 20°-40° N in April-June, depletions of CO 2 over the boreal Eurasia starting from June, and a distinct decrease in CO 2 over South Asia to Southeast Asia in August-September, all of which are in agreement with the observation (first column).In Section 4 below, we discuss how these features constitute the large-scale seasonal CO 2 distributions, as depicted in the last column.One notable feature that is not well reproduced by the model is the high ΔCO 2 values observed over northern Japan in April, the cause of which is yet to be determined.

Summertime CO 2 drawdown
In Section 3.1, we presented two major features in the CO 2 distribution in the UT over the Asia-Pacific region in the boreal summer: (1) the distinct low CO 2 values associated with the monsoon anticyclone over South Asia to Southeast Asia during August-September and (2) the highly variable low CO 2 values at northern latitudes (> 40° N) during June-August (Fig. 3).These summertime low-CO 2 phenomena are hereafter referred to as the "monsoon low CO 2 " and "boreal low CO 2 ", respectively.

Monsoon low CO 2
In August, a distinct circular-shaped distribution of low CO 2 over the Asian continent is prominent in both the observed and simulated ΔCO 2 (Figs.6u and 6v).The model reproduces the observation well in terms of the location of the spatially minimum CO 2 (i.e. the low-CO 2 over South Asia and northern Southeast Asia).Although the CONTRAIL data are not available over inland China (in particular over the Tibetan Plateau), the model simulation (Fig. 6y) offers a complete picture of the UT low CO 2 distribution associated with the monsoon anticyclone.Interestingly, the anticyclonic CO 2 pattern is mainly composed of low BIO CO 2 (Fig. 6x).The region of lowered CO 2 in the monsoon anticyclone expands until September, as the confinement of the low CO 2 in the anticyclone becomes less distinct than in August as region-wide decrease of CO 2 occurs (Figs.6z, 6aa and 6ad).The simulation indicates spreading of low BIO CO 2 to the northwestern Pacific from the anticyclone.In October, the UT CO 2 becomes nearly uniform again over the entire Asia-Pacific region (Figs. 6ae, 6af and 6ai).
As described above, the monsoon low CO 2 in the UT anticyclone is seasonally most distinct in August (Figs.3 and 6), which is in fact coincident with the dynamical development of the summer monsoon anticyclone.Previous studies have shown that dynamical strengths of the monsoon anticyclone and convective activity reach their seasonal maxima in July-August (Randel and Park, 2006;Garny and Randel, 2013), and that, consequently, the confinement of the air mass within the UT anticyclone is seasonally strongest in August (Rauthe-Schöch et al., 2016).The CONTRAIL flights have only DEL where vertical profiles inside the monsoon low CO 2 could be collected (see Fig. 3o).The DEL measurements clearly illustrate CO 2 well-mixed in the FT with pronounced decrease in the BL (Fig. 4g).This feature is consistent with the interpretation that the neighbouring region of DEL (i.e.northwestern India) is part of the vertical conduit core that effectively transports surface flux signals upward to the upper tropospheric part of the summer monsoon anticyclone (Bergman et al., 2013).Our simulation shows that the BIO uptake in South Asia plays a dominant role in lowering the UT CO 2 (Fig. 6).In this connection, model studies have demonstrated that aircraft data within the anticyclone have a significant impact in constraining surface CO 2 fluxes in South Asia (Patra et al., 2011;Niwa et al., 2012).In August, over other Asian cities, such as SIN, BKK, HKG and SHA, the summertime ΔCO 2 values are not as low as those in the monsoon anticyclone (Fig. 3o), which means that these cities are outside the monsoon vertical conduit.
In September, the vertical profiles over DEL (i.e. the core of the monsoon low CO 2 ) retain vertically well-mixed low CO 2 as in August (Fig. 5g4g).This is indicative of strong BIO CO 2 uptake, as reflected in the optimized flux (see Fig. 5d of Niwa et al., 2012).At the same time, we see a region-wide CO 2 decrease, as the August sharp CO 2 gradient at the edge of the UT anticyclone becomes blurred (Fig. 3q).This implies a broad propagation of the monsoon low CO 2 in the UT, as the anticyclonic confinement weakens (Garny and Randel, 2013;Rauthe-Schöch et al., 2016).The expansion of the monsoon low CO 2 in the UT (Fig. 3q) is reflected in the vertical CO 2 profiles over HKG and SHA where substantial CO 2 decreases are observed in the UT in September (i.e. the vertical gradients of CO 2 over both cities increase from August to September; see Fig. 4).A similar, but less pronounced, feature is observed further downwind over cities in Japan (FUK, NGO, NRT and HND), as it is advected by strong westerly winds to the western Pacific in October.The decreasing CO 2 with altitude in the late summer is unique over the Asia-Pacific region where outflow from the monsoon low CO 2 in the UT is a significant contributing factor.The same process involving the Asian summer monsoon anticyclone can be invoked to explain the elevated methane (CH 4 ) values of South Asian origin observed in the UT over the western Pacific in the summer (Umezawa et al., 2012).The high CH 4 values in the Asian summer monsoon anticyclone, its formation mechanism, and outflow from the anticyclone were recently discussed by Chandra et al. (2017).
Figure 7a compares seasonal variations of ΔCO 2 in the UT over the South Asian continent (75°-100° E) and the western Pacific Ocean (130°-150° E) at latitudes 20°-30° N.Here we define longitudinal gradient as the difference in ΔCO 2 between these two continental and oceanic areas (black line).The observed longitudinal gradient is nearly zero in July, increasing rapidly to 2.5 ppm in August, decreasing to 1.8 ppm in September, and then disappearing in October.This seasonal change is reproduced well by the NICAM-TM simulation (Fig. 7b).We also show a break down of the longitudinal gradient into BIO and FF CO 2 contributions.Clearly, BIO CO 2 is the predominant driver of the seasonal CO 2 variation in the UT over both areas and contributes to the longitudinal gradient in August-September due to the monsoon anticyclone.Over South Asia, seasonal maximum contribution of BIO CO 2 to the summertime decrease is seen in August.This effect is not observed until September over the western Pacific, a lag on the order of a month.
It is interesting to note the difference in the timing between the CO 2 drawdown and the accumulation of other pollutants inside the UT monsoon anticyclone.As clearly seen in Fig. 3m, no enhancement/depletion in CO 2 is observed in the UT monsoon anticyclone in July.This is in contrast to studies that have indicated elevation of pollutant species within the monsoon anticyclone starting in June to July (e.g. Park et al., 2009;Xiong et al., 2009;Schuck et al., 2010;Randel et al., 2010).The difference between atmospheric CO 2 and other pollutants lies in the fact that these "other pollutants" are mostly of anthropogenic origins that essentially have no seasonal cycle.The observed enhancement of these pollutants are therefore driven mostly by the anticyclone dynamics (Randel and Park, 2006;Park et al., 2009;Bergmann et al., 2013), and not by the seasonal variation in the surface emission, as in CO 2 .Therefore, the absence of the anticyclonic structure in CO 2 in July is attributable to its surface flux characteristics.In July, the region's terrestrial biosphere might be still in transition from overwhelming respiration (net source) to photosynthesis (net sink) (Niwa et al., 2012;Patra et al., 2013), since substantial precipitation arrives 1-2 months after the onset of the monsoon (i.e.prevailing southwest wind) in June (India Meteorological Department at http://imd.gov.in/pages/monsoon_main.php).From August to September, strong biospheric CO 2 uptake in South Asia takes place (Niwa et al., 2012), giving rise to the observed monsoon low CO 2 in the UT (Fig. 6) that is simulated well in our model.

Boreal low CO 2
In July, a prominent feature that is common in the CONTRAIL measurements and the NICAM-TM simulation (Figs.6p   and 6q) is a sharp north-south CO 2 gradient at 40°-50° N, with low values to the north and relatively uniform CO 2 to the south.In the NICAM-TM simulation, much of the BIO CO 2 uptake in the boreal Eurasia propagates to the northern Pacific (Fig. 6s).
The boreal low CO 2 (i.e. the deeper drawdown of CO 2 and its earlier phase at higher latitudes) in the UT has been understood in the context of BIO CO 2 uptake propagating from mid to high latitudes (Tanaka et al., 1988;Nakazawa et al., 1991;Matsueda and Inoue, 1996;Matsueda et al., 2002).It is estimated that a substantial CO 2 uptake by boreal biosphere starts in June and peaks in July to early August (e.g.Randerson et al., 1999;Saeki et al., 2013;Zhang et al., 2014), therefore the occurrence of the boreal low CO 2 in the UT are is consistent in phase with the atmospheric propagation of boreal BIO uptake.Sawa et al. (2012) showed that, in summer, convective uplift of surface low-CO 2 air lowers CO 2 in the FT at the NH mid to high latitudes.As clearly seen in Fig. 3, we observed the large CO 2 variability (the wide spreads of the histograms, see panels l, n and p) in the UT north of 40° N in June-August.This large CO 2 variability can be explained most likely by sporadic occurrences of convection over boreal Eurasia, as well as to a lesser extent by seasonally strongest and heterogeneous BIO CO 2 uptake; such an example from the CONTRAIL measurement flights has been presented in Fig. 5 of Sawa et al. (2012).Miyazaki et al. (2008) pointed out that the boreal low CO 2 in the summer is isolated from the lower latitudes due to slow mean meridional circulation and weak cyclonic activity during the season.This can be seen in the CONTRAIL data (Fig. 6p, see also Fig. 6 of Sawa et al. 2012) and the NICAM-TM simulation (Fig. 6q).It is also noted that the spread of the histogram over the boreal Eurasia decreases in September (Fig. 3r), inferring implying that the convective activity and BIO CO 2 uptake over the continent seasonally weakens and the UT resumes "background" CO 2 after the summer period of large fluctuations.
The extent to which the boreal low CO 2 is advected has significant impact on the observed seasonal cycles of CO 2 in the LT over East Asian cities.As described earlier, the seasonal CO 2 minimum over ICN occurs about a month earlier than over Japan at similar latitudes (Fig. 4).Based on the NICAM-TM model analysis of the ICN measurements (Niwa et al., 2017), it is found that air masses observed in the LT over ICN in the summer are influenced by surface fluxes in the boreal Eurasia.
As mentioned earlier, the boreal BIO CO 2 uptake peaks in July, relatively earlier than at mid latitudes.Accordingly, larger contributions of air masses from the north in the early summer would lower atmospheric CO 2 , shifting earlier the occurrence of the seasonal CO 2 minimum at mid latitudes.

Seasonally elevated and highly variable CO 2 in spring
We have shown in Section 3 that seasonally elevated CO 2 is observed throughout the whole troposphere over the East Asia region in April-May (Figs. 3 and 4).This elevated CO 2 is accompanied by increased spread of ΔCO 2 values in the UT north of 30° N in .As shown by the NICAM-TM simulation, seasonally high CO 2 can be explained mostly by the BIO emission fluxes, but a significant portion of the associated variability is due to enhanced synoptic-scale meteorological variability.
One of the most likely factors in meteorology is the active passage of eastward-tracking synoptic systems.In East Asia, cyclonic activity is most frequent in the spring (Chen et al., 1991;Adachi and Kimura, 2007).In association with the eastward moving springtime cyclonic activity, two major transport pathways have been suggested for pollutant outflow from the continental East Asia to different tropospheric layers over the northwestern Pacific.
(1) The first mechanism involves the advection of polluted BL air behind the cyclonic cold front as it moves eastward over the East Asian continent out to the Pacific (Liu et al., 2003;Sawa et al., 2007).Consequently, periodic passages of cyclones produce episodic variations of anthropogenic trace gases in the BL across the northwestern Pacific (Liu et al., 1997;Liang et al., 2004;Sawa et al., 2007;Tohjima et al., 2010Tohjima et al., , 2014)).( 2) The second mechanism involves frontal uplift of air in front of a moving cold front, in what is called the warm conveyor belt.The uplift frequently takes place over South and Central China and the plume travels northeastward along the warm conveyor belt to the northwestern Pacific (Bey et al., 2001;Liu et al., 2003;Miyazaki et al., 2003;Liang et al., 2004).Convective uplift along a frontal zone over Central China and Southeast Asia also transports BL air to the FT (Miyazaki et al., 2003;Oshima et al., 2004).In addition to the above transport processes associated with cyclones, orographic forcing over South and Central China has also been observed to uplift the BL air to the FT (Liu et al., 2003).Once in the FT, the plume can be easily exported to the Pacific by the midlatitude westerly winds (Bey et al., 2001;Liu et al., 2003).
In summary, periodic and episodic cyclonic uplifting of the BL air over the continental East Asia, with strong surface CO 2 emissions could explain the seasonal maximum level of CO 2 and increased variability in the FT in the spring (Fig 3).
Using the CONTRAIL data, Shirai et al. (2012) also showed that the observed synoptic-scale variability of the FT CO 2 over NRT increases in the spring, as air influenced by the continental East Asian CO 2 emissions is advected towards Japan.

Concluding Remarks
We have presented spatiotemporal variations of tropospheric CO 2 over the Asia-Pacific region observed uniquely by the CONTRAIL commercial airliner measurements.High-frequency in-flight CO 2 measurements by the CONTRAIL CME cover large part of the Asia-Pacific region and contribute to an enhanced characterization and understanding of the climatological distribution of CO 2 over the region.Some of the highlights in this study are summarized as follows.
In summer, the region-wide low CO 2 across the Asia-Pacific region is primarily due to the CO 2 drawdowns in two distinct regions: the monsoon low CO 2 and the boreal low CO 2 .The monsoon low CO 2 reflects South Asian biospheric CO 2 uptake and its propagation in the UT in association with the development and decay of the Asian summer monsoon anticyclone.This process contributes significantly to the observed horizontal and vertical variations in CO 2 over the Asia-Pacific region.The monsoon outflow increases in September as the anticyclone decays, delivering low CO 2 (from South Asian biosphere) to the UT over the northwestern Pacific.In contrast, the boreal low CO 2 is driven by boreal terrestrial biospheric uptake.Heterogeneous spatial distributions of the biospheric flux, combined with the sporadic convective vertical transport over the Eurasian continent, cause seasonally large variability in the UT CO 2 at north of 40° N.
In spring, active passages of the eastward-tracking synoptic system sweep the continental East Asia and transport the region's CO 2 emissions up to the UT, elevating atmospheric CO 2 over the northwestern Pacific.These synoptic systems also increase variability in CO 2 .Given the high-density CONTRAIL measurements over Asia, and in particular around Japan, the CONTRAIL data provide promising opportunity for diagnosing detailed transport processes by midlatitude cyclones of CO 2 emitted from the continental East Asia.
The CONTRAIL commercial airliner measurements over the Asia-Pacific region can be exploited in constraining emissions of various trace gases from East Asia and South Asia, particularly in the context of the role of the Asian summer monsoon.Also, given the unique spatiotemporal measurements along high altitude cruise and vertical flightsprofiles, the CONTRAIL data can be used to evaluate emerging greenhouse gas data obtained by satellites.
Thank you for your suggestions.Our responses are decribed below.Thank-you for the work that you and your co-authors have undertaken to address the reviewers comments on your manuscript.Overall I think you have answered the concerns sufficiently.In reading the manuscript again, I would like to recommend the following technical corrections.The page and line numbers refer to the version of the manuscript which shows the tracked changes.P1, line 20: You appear to have misunderstood the correction requested by reviewer 1 and have deleted 'the' earlier in the sentence than was required.The sentence should read '... observed in the vertical profiles of CO2 over the Pacific Rim of continental East Asia.' P1, line 23: delete 'the' in two places so that the sentence reads '...over the northwestern Pacific comes from continental East Asian emissions in spring, ...' P1, line 25: add 'a' before 'distinct imprint' P4, line 7: delete 'the' before 'North America'

Figure 1 :
Figure 1: (a) A map showing flight tracks of the aircraft carrying the CME during 2005-2015.Airports highlighted in this study are shown by open squares with airport codes (Table1).The coloured bins are climatological annual average ΔCO 2 values in the UT; note that the colour scale is different from that in Fig.3and the annual averages were calculated only for bins where the monthly values are available for the all months.(b) Number of monthly vertical profiles taken over each airport.The airports are

Figure 2 :
Figure 2: Temporal variations of CO 2 over various airports in Asia.See Table 1 and Fig. 1 for the airport codes.Individual CO 2 data points are colored by altitude.The CO 2 data over the two Shanghai airports (SHA and PVG) are merged and designated as SHA, and same for ICN (ICN and GMP) and TPE (TPE and TSA).Also shown in each panel for comparison are the flask-based CO 2 data (black circles) and the long-term trend (black line) at the Mauna Loa Observatory (MLO; 19.54° N, 155.58°W, 3397 m

Figure 3 :
Figure 3: (Left) Monthly climatological CO 2 mole fraction (ΔCO 2 ) in the UT over the Asia-Pacific region.The CO 2 data taken at altitudes > 8 km are averaged in each 5°×5° bin.The CO 2 data influenced by stratospheric air (PV > 2 PVU) were excluded.Also shown are monthly averaged wind vectors at 250 hPa from the JCDAS/JRA-55 reanalysis data (averaged for the observation years).(Right) Histograms of ΔCO 2 in each 5° latitude bands colorcoded in the same manner as in the left panels.Every histogram

Figure 5 :
Figure 5: Same as in Fig. 4 but for standard deviations of ΔCO 2 in each bin.The standard deviation is calculated only when the bin has > 5 data points.

Figure 6 :
Figure 6: Comparison of the observed and simulated distributions of CO 2 in the UT.Column 1 shows ΔCO 2 Climatological distributions of (first column) CO 2 in the UT (> 8 km) observed by CONTRAIL CME.Columns 2-4 show in comparison to (second column) ΔCO 2 , Δ(third column) FF CO 2 , and (fourth column) ΔBIO CO 2 simulated by NICAM-TM.The CONTRAIL data are simply averaged for each grid, and the NICAM-TM data are sampled at locations and times corresponding to the Sweeney et al. (2015))compiled CO 2 data from flask-based aircraft observations at 12 sites around the world for comparison with model simulations.Flask CO 2 data at 16 sites from the NOAA/ESRL aircraft program were reported bySweeney et al. (2015), including some of the data analyzed byStephens et al. (2007).The measurements bySweeney et al. (2015)have revealed climatological CO 2 variations over North America, whereas the present study focuses on Asia with more frequent in-flight observations.Since vertical profile measurements are relatively scarce over Asia (see supporting online material by