An updated tropospheric chemistry reanalysis and emission estimates, TCR-2, for 2005-2018

This study presents the results from the Tropospheric Chemistry Reanalysis version 2 (TCR-2) for the period 20052018 at 1.1◦ horizontal resolution obtained from the assimilation of multiple updated satellite measurements of ozone, CO, NO2, HNO3, and SO2 from the OMI, SCIAMACHY, GOME-2, TES, MLS, and MOPITT satellite instruments. The reanalysis calculation was conducted using a global chemical transport model MIROC-CHASER and an ensemble Kalman filter tech5 nique that optimizes both chemical concentrations of various species and emissions of several precursors, which was efficient for the correction of the entire tropospheric profile of various species and its year-to-year variations. Comparisons against independent aircraft, satellite, and ozonesonde observations demonstrate the quality of the reanalysis fields for numerous key species on regional and global scales, as well as for seasonal, yearly, and decadal scales, from the surface to the lower stratosphere. The multi-constituent data assimilation brought the model vertical profiles and inter-hemispheric gradient of OH closer 10 to observational estimates, which played an important role in improving the description of the oxidation capacity of the atmosphere and thus vertical profiles of various species. The evaluation results demonstrate the capability of the reanalysis products to improve understanding of the processes controlling variations in atmospheric composition, including long-term changes in near-surface air quality and emissions. The estimated emissions can be employed for the elucidation of detailed distributions of the anthropogenic and biomass-burning emissions of co-emitted species (NOx, CO, SO2) in all major regions, as well as their 15 seasonal, and decadal variabilities. The datasets are available at: https://doi.org/10.25966/9qgv-fe81 (Miyazaki et al., 2019a). Copyright statement. Copyright 2020, California Institute of Technology. Government sponsorship acknowledged. 1 https://doi.org/10.5194/essd-2020-30 O pe n A cc es s Earth System Science Data D icu ssio n s Preprint. Discussion started: 3 April 2020 c © Author(s) 2020. CC BY 4.0 License.

ing of emission variability and the processes controlling the atmospheric composition, chemical reanalysis products have been generated by integrating various satellite measurements. Using an ensemble Kalman filter (EnKF) data assimilation technique, Miyazaki et al. (2015) simultaneously estimated concentrations and emissions of various species for an eight-year tropospheric chemistry reanalysis (TCR-1) for the years 2005-2012. The TCR-1 framework based on the AGCM-CHASER (Sudo et al.,55 2002) and MIROC-CHASER (Watanabe et al., 2011) models has been used to provide comprehensive information on atmospheric composition and emission variability (Miyazaki et al., 2012a(Miyazaki et al., , 2014Miyazaki and Eskes, 2013;Ding et al., 2017). Apart from the TCR systems, employing the ECMWF s Integrated Forecasting System (IFS), three recent reanalyses have also been released: the MACC reanalysis for the years 2003-2012 (Inness et al., 2013), the CAMS-Interim reanalysis for the years 2003-2018 (Flemming et al., 2017) and recently the CAMS reanalysis for the years 2003 to present (Inness et al.,60 2019). A decadal reanalysis of CO was conducted at NCAR (Gaubert et al., 2016). Miyazaki et al. (2020) developed a multi-constituent multi-model chemical data assimilation (MOMO-Chem) framework that directly accounts for model error in transport and chemistry by integrating a portfolio of forward chemical transport models into an EnKF system. The MOMO-Chem framework generates an ensemble of data assimilation analyses to provide integrated unique information on the tropospheric chemistry system including precursor emissions and their uncertainty ranges

Forecast model
The forecast model, MIROC-CHASER (Watanabe et al., 2011), contains detailed photochemistry in the troposphere and stratosphere by simulating tracer transport, wet and dry deposition, and emissions. The model calculates the concentrations of 92 85 chemical species and 262 chemical reactions (58 photolytic, 183 kinetic, and 21 heterogeneous reactions). Its tropospheric chemistry considers the fundamental chemical cycle of O x -NO x -HO x -CH 4 -CO along with oxidation of non-methane volatile organic compounds (NMVOCs) to properly represent ozone chemistry in the troposphere. MIROC-CHASER has a T106 horizontal resolution (1.1 • x 1.1 • ) with 32 vertical levels from the surface to 4.4 hPa. This is coupled to the atmospheric general circulation model MIROC-AGCM version 4 (Watanabe et al., 2011). The simulated meteorological fields were nudged toward 90 the six-hourly ERA-Interim (Dee et al., 2011).
The a priori surface emissions of NO x , CO, and SO 2 were obtained from bottom-up emission inventories. Anthropogenic NO x , CO, and SO 2 emissions were obtained from the HTAP version 2 for 2010 (Janssens-Maenhout et al., 2015), which combines regional inventories of the European Monitoring and Evaluation Programme (EMEP), Environmental Protection Agency (EPA), Greenhouse Gas-Air Pollution Interactions and Synergies (GAINS), and Regional Emission Inventory in Asia (REAS).

95
For biomass burning were emissions, we employed the monthly Global Fire Emissions Database (GFED) version 4 (Randerson et al., 2015). Emissions from soils were based on monthly mean Global Emissions Inventory Activity (GEIA) (Graedel et al., 1993). Lightning NO x sources were simulated using the convection scheme of MIROC-AGCM and the relationship between lightning activity and cloud top height (Price and Rind, 1992). Methane concentrations were scaled on the basis of present-day values with reference to the surface concentration.

Data assimilation method
Data assimilation applied here is based upon on an EnKF approach, the Local Ensemble Transform Kalman Filter (LETKF) (Hunt et al., 2007). The EnKF uses an ensemble forecast to estimate the background error covariance matrix and generates an analysis ensemble mean and covariance that satisfy the Kalman filter equations. In the forecast step, a background ensemble, x b i (i = 1, ..., k), is obtained from the evolution of an ensemble model forecast, where x represents the model variable, b is the 150

Assimilated data sets
An observation operator is applied to assimilate individual measurements to map the model fields into the retrieval space.
The operator includes the spatial interpolation operator, a priori profile for the satellite retrievals, and averaging kernel. See Miyazaki et al. (2020) for more details.

155
The tropospheric NO 2 column retrievals used were the QA4ECV version 1.1 level 2 (L2) product for OMI (Boersma et al., 2017a), GOME-2 (Boersma et al., 2017b), and SCIAMACHY (Boersma et al., 2017c). The ground pixel sizes of the OMI, GOME-2, and SCIAMACHY retrievals are 13km×24km, 80km×40km, and 60km×30km, with local equator overpass times of 13:45, 09:30, and 10:00, respectively. Since December 2009, approximately half of the pixels of the OMI measurements have been compromised by the so-called row anomaly, which were excluded before data assimilation. The GOME-2 measurements 160 were assimilated after January 2007, whereas the SCIAMACHY retrievals were assimilated before February 2012. Lowquality data were excluded by applying the provided quality flag. A super-observation approach was employed to generate representative data with a horizontal resolution of the forecast model for OMI, GOME-2, and SCIAMACHY observations, following the approach of Miyazaki et al. (2012a). Super-observations were generated by averaging all data located within a super-observation grid cell. The retrieval uncertainty of individual pixels was calculated based on error propagation in the 165 retrieval. The detailed error characteristics and validation results of the NO 2 products are described by Boersma et al. (2018).

TES ozone
The Tropospheric Emission Spectrometer (TES) ozone retrievals used are the version 6 level 2 nadir data obtained from the global survey mode (Bowman et al., 2006;Herman and Kulawik, 2013) (https://tes.jpl.nasa.gov/data/products/level-2). This data set consists of 16 daily orbits with 5×8 km footprints spaced approximately 200 km apart along the orbit track, with 170 the equator crossing local times of 13:40 and 02:29. Low-quality data were excluded using the quality flag information. The availability of TES measurements is strongly reduced after 2010, which can affect the reanalysis performance (Miyazaki et al., 2015). Super observations were not generated for the TES retrievals, because of relatively large spatial representativeness of the vertically-integrated information primarily within the free troposphere.

175
The MOPITT total column CO data used were the version 7 L2 TIR/NIR product . The TIR/NIR product provides the greatest sensitivity to CO in the lower troposphere and increases sensitivity to near surface CO compared to the TIR-only product. We excluded MOPITT data in polar regions (>65 • latitude), where the quality deteriorates and the information content lowers. because of potential problems related to cloud detection and icy surfaces. We also excluded the night-time data using a filter based on solar zenith angle, because daytime conditions typically provide better thermal contrast 180 conditions for the retrievals. The total column averaging kernel was used in the observation operator. The reported retrieval error was used in the observation error. The super-observation approach was also applied to MOPITT observations.

MLS ozone and HNO 3
The Microwave Limb Sounder (MLS) data used were the version 4.2 ozone and HNO 3 L2 products (Livesey et al., 2011(Livesey et al., , 2018. We used MLS data for pressures of lower than 215 hPa for ozone and 150 hPa for HNO 3 , while excluding tropical-185 cloud-induced outliers. The provided accuracy and precision of the measurement error were used in the observation error.

OMI SO 2
The OMI SO 2 data used were the planetary boundary layer vertical column SO 2 L2 product obtained with the principal component analysis algorithm (PCA) . Only clear-sky OMI SO 2 data (cloud radiance fraction < 20%) with solar zenith angles less than 70 • were used, following the procedure of Fioletov et al. (2016Fioletov et al. ( , 2017. Because of the lack of 190 information regarding the observation error, we assumed the OMI SO 2 error to be a constant value of 0.25 DU, which is about half of the standard deviation of the retrieved columns over remote regions (Li et al., 2013). The super-observation approach was applied to OMI SO 2 observations. 195 We use version 7 TES PAN retrievals (Payne et al., 2014;TES Science Team, 2016;Payne et al., 2017) to evaluate tropospheric profiles of PAN for years [2005][2006][2007][2008][2009]. TES PAN data have provided information on the long-range transport of NO x at low temperatures and ozone production in warmer regions of the remote troposphere (Jiang et al., 2016). Low-quality data were excluded using the provided quality flag and information. Payne et al. (2014) showed that the detection limit for a single TES measurement is dependent on atmospheric and surface conditions as well as on the instrument noise. For observations where 200 the cloud optical depth is less than 0.5, the TES detection limit for PAN is within the region of 200 to 300 pptv.

AIRS/OMI ozone
We used the joint AIRS/OMI version 1 L2 tropospheric ozone profile product (Fu et al., 2018(Fu et al., ) for 2006(Fu et al., -2010(Fu et al., and 2015(Fu et al., -2018 to evaluate decadal changes in tropospheric ozone. The ozone profile retrievals were performed by applying the JPL MUlti-SpEctra, MUlti-SpEcies, Multi-Sensors (MUSES) algorithm to both AIRS and OMI level 1B (L1B) spectral radiances (Fu 205 et al., 2018). The AIRS/OMI ozone profile products have been produced with a spatial sampling and the retrieval characteristics of ozone profiles equivalent to TES L2 standard data product, demonstrating the feasibility of extending the TES L2 data record by a multiple spectral retrieval approach. The retrievals show reasonable agreement with WOUDC global ozonesonde measurements (Fu et al., 2018). The AIRS/OMI data has been successfully assimilated to improve the tropospheric ozone analysis over East Asia during the KORUS-AQ campaign (Miyazaki et al., 2019b) and could be used to improve decadal ozone 210 reanalyses. polluted areas (over Mexico City and Houston) were removed from the comparison, as they could cause significant errors in the representativeness (Hains et al., 2010).

WOUDC Ozonesonde data
The Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) mission (Jacob et al., 2010) was executed during two three-week deployments based in Alaska (April 2008, ARCTAS-A) and western Canada (June- mid latitudes, whereas it is positive in the upper troposphere and lower stratosphere (UTLS) for the globe. The large positive biases in the extratropical UTLS could be associated with errors in the stratosphere-troposphere-exchange (STE) processes and chemical processes such as halogen chemistry, in addition to errors in the prescribed ozone concentrations above 70 hPa in the model.
The reanalysis shows improved agreement with the ozonesonde observations over the globe. The data assimilation generally 280 decreased the ozone concentration in the extratropics UTLS (200-90 hPa) for the globe and in the middle and upper troposphere (500-200 hPa) at high latitudes of both hemispheres throughout the year. In the lower troposphere (850-500 hPa), the data assimilation increased the ozone concentrations and removed most of the model biases for the globe. Consequently, the reanalysis mean bias became nearly zero in the extratropical UTLS regions and less than 15% in the free troposphere for the globe. At high latitudes, the tropospheric ozone is not directly constrained by any measurements. Nevertheless, the reanalysis 285 ozone shows improved agreements with the ozonesonde measurements through atmospheric transport from lower latitudes and from the stratosphere. In the lower troposphere, the annual mean reanalysis ozone bias is less than 1.2 ppbv, except for the tropics (4.2 ppb), which is 70-94% smaller than the bias in the control run. In the middle and upper troposphere, the mean ozone bias is less than 5.7 ppbv for the SH high latitudes and 3.1 ppbv for other regions, which is 74-99% lower than the bias in the control run. The RMSE is also reduced by 6-50% for 850-500 hPa and 500-200 hPa, with large reductions for the SH mid significantly with year, which suggests that the TES measurements provide constraints on making stable long-term analysis of the free-tropospheric ozone. The observed trend is positive at the NH mid-latitudes in the lower troposphere (+0.9 ppb/year), 300 corresponding to increased concentrations after 2012, but the significance of this trend is not very high. The reanalysis (+0.4 ppb/year) shows better agreement with the observed slope than the control run (-1.4 ppb/year). The long-term trends will further be discussed in Section 6.
The mean ozone biases in TCR-2 are reduced from those in TCR-1 for many regions, especially for the NH mid and high latitudes (e.g., from -3.9 to -1.2 ppb and from -8.0 to -0.2 ppb at the NH high-latitudes between 850 and 500 hPa and 305 between 500 and 200 hPa, respectively) and SH mid latitudes (from -1.0 to 0.4 ppb and from -1.9 to -0.2 ppb between 850 and 500 hPa and between 500 and 200 hPa, respectively). An exception is the tropics, where the reduced number of ozonesonde observations for the most recent years used in the TCR-2 validation affected the evaluated performance. Huijnen et al. (2019) and Christophe et al. (2019) compared tropospheric ozone reanalysis products from CAMS, CAMS-Interim, TCR-1 and TCR-2. The updated reanalyses (CAMS-Rean and TCR-2) showed substantially improved agreements with independent ground 310 and ozonesonde observations over their predecessor versions (CAMS-iRean and TCR-1) for the diurnal, synoptical, seasonal, and decadal variability. The improved performance can be attributed to a mixture of various upgrades, such as revisions in the chemical data assimilation, including the assimilated measurements and the forecast model performance. The updated chemical reanalyses agree well with each other in most cases, which highlights the usefulness of the current chemical reanalyses in a variety of studies.  the AIRS/OMI observations are greatly reduced in the reanalysis (e.g., from -9.9 ppb in the control run to 0.6 ppb over central Africa). The estimated reanalysis errors are mostly within the AIRS/OMI retrieval uncertainty. These improved agreements in the reanalysis, along with the good agreements between the reanalysis and ozonesonde observations (c.f., Section 4.1.1), demonstrate the great potential of AIRS/OMI data to further improve decadal ozone reanalysis, as will be discussed in Section 7.3.

Aircraft
The reanalysis captured the observed latitudinal-vertical distributions by the HIPPO aircraft measurements over the Pacific.
( Fig. S1 and Table S1). On average, the control run shows negative biases in the lower troposphere (850-500 hPa) from the SH high latitudes to NH high latitudes (-4.6 to -3.6 ppb), whereas the model bias is positive in the middle and upper troposphere (19.6 to 42.6 ppb between 500-200 hPa) except in the tropics (-2.6 ppb). The negative model biases in the lower lower troposphere, probably associated with corrections made to precursors' emissions over East Asia and the stratospheric concentrations. The positive model biases in the middle and upper troposphere (500-200 hPa) are also reduced by 44-92% except for the tropics and SH low latitudes, as commonly suggested by comparisons to the ozonesonde measurements (c.f., Section 4.1.2). These results demonstrate that the assimilation of multiple-species data sets is a powerful way to globally 340 constrain the entire tropospheric ozone profile, including that over remote oceans.
The comparison with the NASA aircraft data (Fig. 4) shows that the control run generally underestimates ozone in the free troposphere, with largest biases (up to about 15 ppb) for the DC3-GV over the United States and KORUS-AQ profiles over South Korea. In turn, it is overestimated in the lower stratosphere for the ARCTAS-A and -B profiles over the Arctic. Near the surface, the control run overestimates ozone for the DISCOVER-AQ profile by 15 ppb and for the KORUS-AQ profile by    6 shows the time series of regional mean tropospheric NO 2 concentrations. The regional error statistics compared to the 375 OMI retrievals are summarized in Table 4. Over East China, the model negative bias is relatively large in winter, particularly in comparison with SCIAMACHY and OMI during 2010-2014 when the observed NO 2 concentrations are relatively high.
In contrast, the model bias against OMI and GOME-2 is negative during 2015-2018, when the observed concentrations are relatively low. The reanalysis captures the observed decadal changes (r = 0.99 for OMI using montly mean concentrations), through corrections made to NO x emissions. Slight negative biases remain during the 2010-2014 winters compared with OMI 380 and SCIAMACHY.
Over Europe, the negative model bias is persistent against the three retrievals throughout the reanalysis period. The data assimilation reduced about 30-60% of the model negative bias compared with OMI (by 54 % for mean) and most of the biases against SCIAMACHY. In contrast, the reanalysis reveals excessively high NO 2 compared with GOME-2 during summer. The observed negative trend by OMI (-1.2%/year) is efficiently captured by the reanalysis (-1.2%/year, r = 0.95). represented by the reanalysis (-2.1%/year) than by the control run (0.6%/year). The model negative biases compared with the OMI measurements remain partially in late-winter and spring. Data assimilation also increased the temporal correlation with OMI from 0.54 to 0.88.

390
Over India, the model negative bias increased with year because of the lack of the emission increases in the a priori emissions.
The a posteriori NOx emissions in 2018 are up to 90% larger than the a priori emissions over polluted areas at grid scale, whereas the remaining negative NO 2 biases suggest that the NOx emission analysis increments are insufficient. We applied a covariance inflation to the emission factors to prevent covariance underestimation caused by the application of a persistent forecast model, by inflating the spread to a minimum predefined value (i.e., 30% of the initial standard deviation (=40%)) 395 at each analysis step. The inflation was essential to maintain emission variability and continue to increase the emissions. The remaining model biases suggest requirements for a stronger covariance inflation, although too large inflation can cause unstable analysis increments. The reanalysis shows positive trends over the 14 years (+1.3%/year) consistent with the OMI observations (+1.6%/year), with high temporal correlations with respect to all the retrievals (r = 0.96 for OMI). The mean bias was reduced by about 80% compared to OMI. while about 50% of the model negative bias is removed by data assimilation. The remaining model errors can be partially attributed to the limitations in assimilated measurements (e.g., coverage and uncertainty) and persistent model errors, such as too-short lifetime of NO x through processes such as NO 2 + OH reactions and the reactive uptake of HO 2 and N 2 O 5 by aerosols (e.g. Lin et al., 2012;Stavrakou et al., 2013). Further, any errors in the location of individual sources such as power plants in the bottom-up inventories could prevent data assimilation improvements in our approach. in the boundary layer is mostly removed by data assimilation.

430
We used the WDCGC in-situ measurements in 59 stations to evaluate the reanalysis CO concentrations. The comparison results are summarized in Table 5 and shown in Fig. 7 for selected sites. The control run underestimated the mean CO concentrations by 9.4 and 19.8 ppbv at NH mid and high latitudes, with the largest negative biases in winter. The model CO underestimations in the NH are commonly reported in many models (e.g. Stein et al., 2014). The model bias is positive in the tropics and SH by about 13-14 ppbv. After data assimilation, the model biases are greatly reduced in the SH, the tropics, and NH mid-latitudes (by 435 66-88%), while reproducing the observed seasonal and inter-annual variations for many sites. In contrast, at NH high latitudes, the reanalysis CO in TCR-2 reveals small corrections. For instance, over Barrow, Heimaery, and Cold Bay, most of the negative model biases remain. This is different from the substantial improvements found for the entire globe in TCR-1 (Miyazaki et al., 2015). There are several potential reasons for the remaining negative biases, as will be discussed in Section 7.4.

440
Both the control run and reanalysis captured latitudinal variations in CO over the Pacific acquired by HIPPO observations, including maximum gradients around the equator and the subtropical jet (Fig. S2). As summarized in Table S2, the control run underestimates CO concentrations in the NH and overestimates them in the tropics and SH almost for the entire troposphere over the Pacific. The assimilation decreased CO concentrations and removed the model positive bias by about 63-79% in the lower troposphere and by 56-67% in the middle and upper troposphere in the tropics and SH. In the NH, data assimilation 445 improvements are small, which can be attributed to remaining errors in the surface emissions, chemical productions and losses (i.e., OH), long-range transport from the Eurasian continent, and stratosphere-troposphere exchange (STE) (c.f., Section 7.4).
The control run generally captured the observed profiles for most NASA aircraft flights, except for an up to 50-130 ppb underestimation in the lower and middle troposphere for the ARCTAS-A, ARCTAS-B, and KORUS-AQ profiles (Fig. 4).
Substantial reductions in the model negative bias are found for the KORUS-AQ profile because of increased local and remote 450 (mainly China) CO emissions (Miyazaki et al., 2019b). In contrast, the bias reduction is small for the ARCTAS profiles.   China. Thus, adding constraints in the reanalysis framework, especially on VOCs emissions, would benefit improving PAN and chemically-related species including ozone. Further investigation on the detailed PAN distributions using aircraft and satellite measurements would be helpful to comprehend the possible mechanisms and error sources in the reanalysis PAN fields.

OH
OH is directly linked to the concentrations of species determining the primary production (O 3 and H 2 O), removal (CO and 510 methane), and regeneration of OH (NO x ). Because of the multi-constituent constraints for many key species, a positive impact is expected on global OH fields, given that the reactions are reasonably well described by the model. As shown in Fig. 9, the global tropospheric OH distribution is substantially modified in the reanalysis. Data assimilation mostly increased OH, with the largest increases in the SH tropics. The mean OH concentration in the SH tropics is increased over the reanalysis period by 20-25% at 700 hPa and 30-45% at 500 hPa. In the NH extratropics, the OH increases are about 15-20% at 700 hPa and 20-515 30% at 500 hPa. These increases are found throughout the reanalysis period, with the largest increases during spring-summer in both hemispheres. Both the concentration assimilation and the emission optimization were important in introducing these OH changes. The 14-year mean NH/SH OH ratio in the chemical reanalysis is 1.19±0.015 (1σ inter-annual variability), in contrast to 1.30 in the control run, which is closer to the estimates of 0.97±0.12 based on methyl chloroform observations (Patra et al., 2014). The NH/SH ratio is maximum in 2016 (1.23), reflecting relatively high OH concentrations over East Asia The tropospheric mean OH concentrations averaged during the reanalysis periods are estimated at 8.7×10 5 molec cm −3 for the control run and 11.5×10 5 molec cm −3 for the reanalysis. By applying the obtained tropospheric OH burden to the 525 ACCMIP multi-model mean estimates of tropospheric chemical methane lifetime (τ OH (chemical)) from Voulgarakis et al.

530
The model bias against the aircraft profiles varies largely among the campaigns. For the INTEX-B profile, the control run captured the observed profile well, whereas the data assimilation puts too high OH throughout the troposphere, likely corre-sponding to increased ozone. For the ARCTAS-A, ARCTAS-B, and KORUS-AQ profiles, the model negative bias is strongly reduced by data assimilation in the free troposphere, mainly due to the increased NO x emissions and resultant increased ozone.
The large negative bias near the surface remains for the ARCTAS-B profiles. Remaining large errors in HO 2 could influence the 535 performance of the simulation of OH for some profiles, including ARCTAS-B. Observed OH concentrations are also largely uncertain (e.g. Heard and Pilling, 2003;Stone et al., 2012). Brune and Thames (2019) estimated the absolute accuracy for aircraft HOx measurements to be ± 32 % at 2 sigma confidence.
The ATom measurements provide great data to evaluate remote tropospheric OH, for instance, that derived from OMI CH 2 O measurements (Wolfe et al., 2019). Compared with all the profiles during the ATom-1 and ATom-2, the RMSE is reduced by up 540 to 30% above about 600 hPa in the reanalysis than in the control run (Fig. 10a). Improved agreements can be found for many profiles throughout the troposphere (Fig. S5), whereas a few profiles (e.g., on August 6, 2018, February 2, 2017, and February 5, 2017) led to a degradation in the agreements in the lower troposphere. The ATom measurements provide comprehensive pictures of inter-hemispheric ratios of OH and its seasonal changes over remote oceans (Fig. 10b,c). The observed interhemispheric ratio is about two near the surface and exceeds seven in the middle troposphere in boreal summer (ATom-1),

545
whereas it is 0.4-0.8 throughout the troposphere in boreal winter (ATom-2). The control run mostly overestimated the ratios by a factor of up to 2.5 for ATom-1 and by up to 1.6 for ATom-2, with the largest overestimation in the lower troposphere.
Data assimilation decreased the ratio and shows improved agreements from the surface to the upper troposphere. Because the chemical lifetimes of many species are affected by the amount of OH, the improved representation of OH profiles and its global distributions suggests that multi-constituent assimilation improves the simulation of concentrations and emissions of various 550 species. Decadal changes in the tropospheric OH derived from the reanalysis will be discussed in Section 6.

Aerosols
Although no aerosol observations were assimilated, improved representations of aerosol fields can be expected through corrections made to trace gas concentrations, such as NO x and SO 2 , that affect the formation of secondary aerosols. Figure 11 shows the scatter plots of ammonium (NH 4 ), nitrate (NO 3 ), and sulfate (SO 4 ) aerosols from in-situ observations, control run, and 555 reanalysis. The control run overestimates ammonium and sulfate aerosol concentrations and underestimates the nitrate aerosol concentrations for most of the CASTNET (the US), EANET (East Asia), and EMEP (Europe) sites, while the estimated mean biases (Table 6) are dominated by large biases for a few stations. The median biases are lower than the mean biases for many cases. The multi-constituent data assimilation substantially modified the aerosol concentrations. The RMSE is decreased by 7-61% for ammonium aerosols, 2-11% for nitrate aerosols, and by 5-38% for sulfate aerosols, by data assimilation while the 560 correlation improved for many cases, for instance, from 0.27 to 0.42 for ammonium aerosols compared to the EMEP observations. The median bias also became smaller (by up to 75 %) for most cases. For urban stations, the model representativeness errors may prevent data assimilation improvements, which may have caused degradation for some cases. An assessment of global particulate nitrate and ammonium aerosols in the MIROC-CHASER simulation is also given in (Bian et al., 2017).
Substantial changes in the aerosol concentrations suggest considerable potential of trace gas data assimilation for constrain-565 ing secondary aerosol formation processes. Among numerous factors, optimizations of NO x and SO 2 emissions are considered to be essential to improve secondary aerosol formation in our framework. Our comparisons show improved agreements against various aircraft measurements for many key species relevant to aerosol formations, such as NO 2 , HNO 3 and SO 2 (c.f., Section 4.8). Meanwhile, assimilation of AOD and aerosol concentration observations are required to further improve the representation of primary aerosol emissions and concentrations (e.g. Yumimoto et al., 2017). Simultaneous assimilation of trace gas 570 and aerosol observations would be a powerful approach to fully represent aerosol-gas interactions in the data assimilation framework, which would improve both trace gas and aerosol data assimilation analysis.

Other reactive species
As shown in Fig. 4, the observed main structures of HNO 3 are generally captured well by the control run, with increasing errors toward the surface for some profiles. The increase in HNO 3 toward the surface is driven mainly by the oxidation of NO In this section, we briefly describe the estimated emissions from the TCR-2 calculations. Further detailed analyses of the 14- year variations in the estimated emission sources and its influences on global air quality and health impacts will be discussed in a separate study. The global distribution of a priori and a posteriori emissions and its time series are shown in Figs 12 and 13 and summarized in Table 7. The estimated linear trends are shown in Fig. 14. The regional total emission statistics for surface emissions and lightning NO x sources are summarized in Table 8 and 9, respectively.

Surface NO x emissions
The multi-constituent data assimilation framework has been used to improve estimates of global NO x emissions (Miyazaki and Eskes, 2013;Miyazaki et al., 2014Miyazaki et al., , 2015Miyazaki et al., , 2019b. In this framework, the simultaneous optimization of concentrations and emissions of many species reduces the model-observation mismatches that arise from model errors other than those related to emissions. Meanwhile, the simultaneous assimilation of multiple satellite measurements obtained at different overpass times  Table 7. Data assimilation largely increased surface NO x emissions over major polluted areas such as most parts of China, Southeast Asia, and Europe (Fig. 12). The increments vary from year-to-year over these regions. For instance, they decreased in more recent years over China. This is associated with the assumptions applied to the a priori emissions, such as the use of 2010 an-

Surface CO emissions
The 14-year mean of global total emissions of CO is increased by 26% by data assimilation (1104 Tg CO/yr vs. 877 Tg CO/yr),

645
which is attributable to a 35% increase in the NH and 18% increase in the tropics. The large positive increments are found over eastern and southern China, northern parts of Southeast Asia, India, and central Africa (Fig. 12). The emissions increase in the NH is large in the boreal late winter-spring period, especially over polluted areas at NH mid-latitudes, which enhanced the seasonal amplitude for industrialized countries.
In the multi-constituent data assimilation framework, the assimilation of non-CO observations influences various chemical species including OH, which provides additional constraints on the CO emission estimation. As suggested in Section 4.6, possible underestimations in OH in the control run could lead to underestimations in the estimated CO emissions for many 660 regions. Assimilation of ozone and NO 2 measurements exerts a substantial influence on OH and thus on CO emission estimates.
Nevertheless, insufficient corrections for the NH extratropical CO suggest requirements for further improving CO emission estimates, as will be discussed in Section 7.4.

Surface SO 2 emissions
The 14-year mean global total surface SO 2 emissions are decreased by about 30% by data assimilation from 50.9 to 35.1 TgS, 665 with large reductions in the NH (by 37%). The negative data assimilation increments are also large over China (by -50%), India  Table S5), along with changes in NO x emissions, have strong implications into the secondary aerosol formation processes for many polluted regions.
The a posteriori SO 2 emissions seem excessively high in 2011 for many regions (c.f., Fig. S8), which seems unrealistic 680 and could be due to potential problems in data assimilation setting or assimilated retrievals. Volcanic eruptions also affected a temporal increase in the estimated emissions, as shown by Carn et al. (2017) using the OMI SO 2 measurements. This requires additional careful verification before used in detailed trend analysis. The estimated emissions should have a large uncertainty associated with large retrieval uncertainty (e.g., random noise of 0.5 DU for remote areas, as described in Li et al. (2013)) and the assumed constant retrieval errors and air mass factor. Because the optimized emission factors were applied to the a priori 685 emissions, any missing sources in a priori inventories (e.g. Liu et al., 2018) could also lead to systematic biases in the estimated emissions.

Lightning NO x sources
The multi-constituent data assimilation with different vertical sensitivities provides strong constraints to distinguish between surface and lightning NO x sources and to correct the vertical profiles of lightning NO x sources. The a posteriori global total 690 lightning NO x source is 7.5 TgN, which is about 32% higher than in the control run (5.7 TgN). The estimated global total emission is about 17% larger than our previous estimates (6.4 TgN) based on the TCR-1 system (Miyazaki et al., 2014).
The differences between two estimates can primarily be attributed to change in the forecast model and its resolution. The resolution improvement affected the representation of cumulus convection and lightning frequency distributions. Nevertheless, both estimates suggest common problems of the lightning parameterizations such as requirements to modify the C-shape 695 assumption and land-ocean contrasts.
The long-term trends of lightning NO x are mostly insignificant and dominated by multi-year scale variability rather than linear increase or decrease ( Fig. S9 and Table S6). The inter-annual variability of lightning NO x during 2005-2018 is large over Southeast and South Asia, central and southern Africa, Central Africa, and the Amazon (Fig. 14). variability, which would provide important implications into chemistry-climate interaction processes. The regional total values of the estimated lightning NOx sources for major source regions.

Trend diagnostics
The reanalysis reveals substantial changes in concentrations of various species, which provides an important framework to comprehend the roles of natural and human activities on atmospheric composition. We evaluated long-term atmospheric com- According to changes in concentrations of various species including ozone, the reanalysis reveals a general positive trend in OH during the reanalysis period (Figs. 17, S3 and S4). The tropospheric OH from the noTES reanalysis exhibits strong increases over the tropical western and eastern Pacific by up to +1.2%/year, and 0.9-1.4%/yr over southern India, southern 735 Vietnam, west coast of Saudi Arabia, and western Iran. Annual and zonal mean OH concentrations in the noTES reanalysis are increased over 10 • N-20 • N, 700-500 hPa by 0.5-0.6%/yr and at the SH low and mid-latitudes in the lower troposphere by 0.3-0.4%/yr. These trends are commonly found in both data sets, but with weaker trends in the standard reanalysis. At the NH mid latitudes in the free troposphere, only the noTES reanalysis reveals substantial increases in OH by 0.5-0.7%/year. Based on a sensitivity calculation, these significant changes in OH were found to be strongly driven by surface NO x emission variations, 740 with strong increases from 2007 to 2012. These results highlight substantial impact of human activity on the oxidation capacity of the atmosphere and chemical lifetime of many species such as methane (e.g. Rigby et al., 2017), as previously suggested by Wang and Jacob (1998).

Assimilated data biases and availability 745
Significant temporal changes in the reanalysis quality can partly be attributed to discontinuities in the observing systems. As discussed in Section 6, the reduced number of assimilated TES ozone retrievals after 2010 substantially influenced the usability of the reanalysis products for trend analyses. Meanwhile, changes in the NO 2 observing system, including the OMI row anomaly after December 2009 and the limited temporal coverage of SCIAMACHY and GOME-2, are also considered to affect long-term consistency. The reanalysis ozone bias against the ozonesonde measurements was relatively large in the 750 tropical lower and middle troposphere, which could partly be attributed to the positive biases in the assimilated TES measurements. Miyazaki et al. (2015) tested a bias correction scheme for assimilation of TES ozone based on evaluation results using ozonesonde measurements (Boxe et al., 2010;Verstraeten et al., 2013), however the results were not always positive because of the difficulty in estimating the detailed bias structure. The reanalysis ozone bias can also be affected by biases in ozone precursors measurements such as NO 2 measurements. Nevertheless, we did not apply any bias correction to any assimilated 755 measurements in the reanalysis, because of the difficulty in estimating the bias structure including inter-measurement biases.
To improve the temporal consistency, a detailed assessment of biases in individual retrievals (e.g. Compernolle et al., 2020) and between different retrieval products would be helpful, as already tested in the CAMS reanalysis (Inness et al., 2019) The availability of the ozonesonde measurements for the most recent years was also limited at the time of this research, which limits the evaluation of the reanalysis performance. The mean ozonesonde concentrations at SH mid latitudes show 760 rapid changes after 2017, which were associated with the reduced number of available ozonesonde observations at the time of this research and consequent increased representativeness errors of the ozonesonde network for the large domain. The current ozonesonde network is also too sparse to capture the regional and monthly representative ozone fields especially in the tropics, which can lead to substantial sampling biases in the reanalysis performance, as discussed for evaluations of chemistry-climate models ).

Impact of forecast model performance
Even though the assimilation of multi-species data influences the representation of various chemical fields including precursor emissions, the remaining model errors, such as chemical reaction rates, deposition rates, the limited representation of atmospheric chemistry, as well as meteorology, limit the data assimilation improvements. Miyazaki et al. (2020) developed a MOMO-Chem framework using four global CTMs and an EnKF data assimilation that directly accounts for model error in 770 transport and chemistry. They demonstrated that the observational density and accuracy was sufficient for the assimilation to reduce the influence of model errors in data assimilation analysis; i.e., multi-model spread of ozone analysis is reduced by 20-85% in the free troposphere. Model negative biases in tropospheric NO 2 column and surface CO in the NH are also greatly reduced by more than 40% in all models. MOMO-Chem provides integrated unique information on the tropospheric chemistry system and its uncertainty ranges, which would benefit future development of chemical reanalysis.

775
Meanwhile, a strong reanalysis dependence on forecast model performance was found on the near surface concentrations and precursor emissions, associated with insufficient observational constraints Miyazaki et al., 2020).
The ozone response to precursor's emissions was also found to be strongly sensitive to the chemical mechanisms in the model, which varied by a factor of 2 for end-member models, revealing fundamental differences in the representation of fast chemical and dynamical processes (Miyazaki et al., 2020). The emissions of ozone precursors other than NO x and CO, such as VOCs, 780 have a pronounced influence on the tropospheric chemistry. Adjusting additional model parameters such as VOC emissions, deposition, and/or chemical reactions rates could help reduce model errors. Furthermore, a simultaneous assimilation of trace gas and aerosol measurements would also reduce systematic model errors and provide more comprehensive information on various applications. Meanwhile, high-resolution modeling is also essential for accurate modeling of non-linear chemistry and resolving rapid variations in air pollutions and emissions around cities (Valin et al., 2011;Sekiya et al., 2018), which is also 785 needed to improve the reanalysis performance.

Challenges with next generation satellite data
Next generation satellite data products, that have improved vertical sensitivity and accuracy, as well as improved spatial sampling, have great potential to further improve emissions and surface ozone analyses. The exploitation of existing sounders and development of multispectral retrievals is expected to add constraints on the reanalysis and to remove remaining model (c.f., Section 4.1.2). Cross-Track Infrared Sounder (CrIS) on Suomi-NPP also provides tropospheric PAN retrievals with improved coverage and accuracy compared with TES (Payne et al., 2019). Assimilating PAN retrievals from TES and CrIS can be expected to improve the representation of the global nitrogen cycles, which would also benefit surface and lightning NO x emission estimates combining with tropospheric NO 2 column measurements. Meanwhile, TROPOMI provides global maps of the tropospheric NO 2 column on a daily basis with improved accuracy and higher spatial resolution compared with OMI 800 (Griffin et al., 2019). Assimilating TROPOMI NO 2 has potential for improved evaluation of the changing landscape of emissions on urban-to-regional and regional-to-global scales (Lorente et al., 2019). Assimilation of other retrievals such as OMI and TROPOMI CH 2 O, CrIS Isoprene (Fu et al., 2019), and TES, CrIS, and IASI NH 3 (Shephard and Cady-Pereira, 2015) would also help improve the model chemistry and tropospheric ozone reanalysis.

Under-constrained CO emissions 805
The validation results of CO concentrations suggested under-corrected surface emissions of CO, especially in the NH extratropics (c.f., Section 4.3). There are several reasons for this. First, while our previous estimates in TCR-1 used MOPITT TIR-only CO profile data at 700 hPa, TCR-2 used TIR/NIR total column retrievals. The truly optimal settings of data assimilation parameters probably differ between the two setups. The TCR-2 setting might require further optimization. Second, the chi-square and observation-minus-forecast statistics suggested underestimated background errors of CO for many regions. Considering 810 different systematic model errors and the increased model resolution between TCR-1 and TCR-2, background error inflation settings need to be further optimized for TCR-2. Third, the data assimilation window (two-hour) used is clearly too short for CO emission estimates, considering its relatively long chemical lifetime and the coverage and limited near-surface sensitivity of MOPITT measurements. A longer data assimilation window for CO emission estimates, while keeping the short window for short-lived species such as NO x and ozone, would be required. Finally, CO is produced by the oxidation of methane and bio-815 genic NMHCs (Duncan et al., 2007). These components can account for part of the missing CO concentrations. Adding more observational constraints, such as for CH 2 O and methane, would help improve CO emission estimates (e.g. Stein et al., 2014;Zheng et al., 2018). We have already tested some of the developments and obtained improved estimates of CO concentrations and emissions, which will be implemented in the next generation chemical reanalysis.

Uncertainty estimation
820 Important information regarding the reanalysis product is provided by the error covariance. Within the EnKF assimilation framework, the analysis ensemble spread is estimated from the standard deviation across the ensemble and provides a measure of the uncertainty of the reanalysis product. The information on the analysis uncertainty is included in the TCR-2 reanalysis products. For instance, as shown in Fig. S14, the analysis spread for ozone is about 1-3 ppb in the tropics and subtropics  Stevenson et al., 2013;Kuai et al., 2020), attribution of radiative forcing (Bowman and Henze, 2012), and emergent constraints on future projections Bowman et al., 2018). Validation of short-lived species can be used to identify potential sources of error in model fields and is also important for evaluating the radiative forcing because simulated OH fields influence simulated climates through their influences on methane Voulgarakis et al., 2013). The optimized precursor emission fields can be used to validate bottom-up emission inventories and lightning param-845 eterizations. As changes in tropospheric ozone burden and NO x emissions show a close relation in different future scenarios , evaluations using the estimated emissions and evaluated model response to emissions (Miyazaki et al., 2020) have the potential to evaluate preindustrial, present day, and future model simulations. Short-lived climate pollutants (SLCP) are an increasingly important component of greenhouse gas budgets that limit warming to target temperatures, e.g., 1.5C or 2C (Rogelj et al., 2019). Chemical reanalysis can play a crucial role in assessing the changes and efficacy of SLCPs. The evaluation results for various species reveal the benefit of the assimilation of multiple-species data on the analysis of both observed and unobserved species profiles on both regional and global scales, for seasonal and decadal variations, and 860 from the surface to lower stratosphere. The global statistics of the NO 2 , ozone, and CO evaluation results are summarized in Table 10. The reanalysis ozone bias against the ozonesonde measurements was less than 1.2 ppb in the lower troposphere except for the tropics and less than 3.1 ppb in the middle and upper troposphere except for the SH high latitudes, with temporal correlations greater than 0.85 for most regions. The improved agreements in TCR-2 ozone from TCR-1 can be attributed to a mixture of various upgrades, including assimilated measurements and the forecast model performance and resolution. The The combined analysis of concentrations and emissions is considered an important development in the tropospheric chemistry reanalysis. Our comparisons suggest that improving the observational constraints, including the continued development of satellite observing systems, together with the optimization of model parameterizations, such as deposition and chemical reactions, will lead to increasingly consistent long-term reanalyses in the future. An increase in the forecast model resolution 900 and an extension of data assimilation to aerosols are expected to improve the capability of chemical reanalysis for air quality and climate applications. Techniques to reduce the influence of discontinuities in the assimilated measurements and to employ next generation satellite retrievals would also be important developments in future chemical reanalyses. Satellite data sets from a new constellation of LEO sounders and GEO satellites (e.g., GEMS, TEMPO and Sentinel-4) will provide more detailed knowledge of ozone and its precursors for East Asia (Bowman, 2013).
of Technology, under contract with the National Aeronautics and Space Administration (NASA). We would like to thank the editor and the two reviewers for their valuable comments.                 Table 7. The global and regional mean surface NOx (in Tg N yr −1 ), CO (in Tg CO yr −1 ), and SO2 emissions (in Tg S yr −1 ) and lightning   Table 10. Summary of global statistics of the NO2, ozone, and CO evaluation results for the reanalysis and the control run (in brackets) against tropospheric NO2 in 10 15 molec cm −2 from Table 3, ozonesonde measurements in ppb from Table 2, and surface CO measurements in ppb from Table 5.