Full Terms bState Key Laboratory of NumericalModeling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing,China;cDepartment of Atmospheric and Oceanic Science, and Earth System Science Interdisciplinary Center, University of Maryland, CollegePark, Maryland, USA;dState Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry, Institute of AtmosphericPhysics, Chinese Academy of Sciences, Beijing, China;eSchool of Atmospheric Physics, Nanjing University of Ination Science andTechnology, Nanjing, China;fMeteorological Observation Centre of China Meteorological Administration, Beijing, ChinaABSTRACTThe vertical distribution of carbon dioxide CO2 is important for the calibration and validation oftransport models and remote sensing measurements. Due to the large mass and volume oftraditional instruments as well as supporting systems, in-situ measurements of the CO2verticalprofile within the boundary layer are rare. This study used a miniaturized CO2monitoring instru-ment based on a low-cost non-dispersive infrared NDIR sensor to measure the CO2vertical profileand meteorological parameters of the lower troposphere 0–1000 m in southwesternShijiazhuang, Hebei Province, China. The sensors were onboard a tethered balloon with twoprocesses the ascending process and the descending process. The results showed that the overalltrend of CO2concentration decreased with height. Weather conditions and CO2emission sourcescaused fluctuations in CO2concentrations. The CO2concentration varied from morning to after-noon due mainly to the faster spread of air mass during daytime, with strong convections and theaccumulation of emissions at night. The low-cost sensor produced results consistent with thetraditional gas chromatography . The Weather Research and Forecasting model could notcapture the CO2profiles well due mainly to the bad perances in boundary layer height and thepotential outdated fossil fuel emissions around the experimental site. This experiment is the firstsuccessful attempt to observe the CO2vertical distribution in the lower troposphere by using low-cost NDIR sensors. The results help us to understand the vertical structure of CO2in the boundarylayer, and provide data for calibrating and validating transport models.重污染郊区环境中二氧化碳垂直分布的观测和模拟摘要CO2垂直分布对于大气传输模式和遥感测量的校准和验证非常重要。由于传统仪器质量大, 配套系统体积大, 边界层内CO2的原位测量比较少见。本研究利用基于低成本非分散红外 NDIR传感器的微型化CO2监测仪器, 搭载在系留气艇上, 测量河北省石家庄西南部低对流层 0–1000米 的CO2垂直分布和气象参数。低成本传感器与传统气相色谱分析仪测得的结果一致,CO2浓度的总体趋势随高度增加而下降。研究结果有助于直接分析大气边界层中CO2浓度的垂直结构,为WRF等大气传输模式的校准和验证提供基础数据。ARTICLE HISTORYReceived 7 December 2019Revised 20 February 2020Accepted 24 February 2020KEYWORDSLow cost sensor; co2verticalprofile; tethered balloon;meteorological conditions;non-dispersive infraredNDIR关键词低成本传感器；CO2垂直廓线；系留气艇；气象条件；非分散红外1. IntroductionWith the increased concentrations of greenhouse gases inthe atmosphere, global warming has become a majorscientific and political issue Le Qur et al. 2018;Schneider1989.Amongallgreenhousegases,carbondiox-ide CO2 accounts for the largest share Solomon et al.2007. The concentration of greenhouse gases is risingdue largely to anthropogenic activities. Therefore, accurateassessment of the distribution and changes of CO2concentration in the atmosphere is crucial for the ula-tion of climate policy and prediction of future climatechange.CO2at the ground surface has been observed for a longtime. Since the 1970s, the World MeteorologicalOrganizationWMOhasconductedlong-termglobalmon-itoring of greenhouse gases and reactive gases, accumulat-ingdecadesofobserveddataWMO/GAW2016.TheChinaMeteorological Administration has started CO2in-situ mea-surements at Mt. Waliguan in Qinghai Province as one ofCONTACT Pengfei HAN ATMOSPHERIC AND OCEANIC SCIENCE LETTERShttps//doi.org/10.1080/16742834.2020.1746627 2020 The Authors. Published by Ina UK Limited, trading as Taylor Brenninkmeijeret al. 2007; Chan and Kwok 2000; Tolton and Plouffe 2001.VerticalCO2measurementscanbemadeusinghightowers,tethered balloons, lidar, aircraft, and stratospheric balloonsLi et al. 2014. High tower measurements can provide datafrom the lowest 100–500 m of the atmosphere of theplanetary boundary layer, but the range of measuredheights is often restricted by the tower height generallylower than 500 m Davis et al. 2010; Inoue and Matsueda2001. A sampling plat mounted on aircraft and strato-spheric balloons can measure CO2at higher altitudesDeutscher et al. 2010;Maysetal.2009; Nakazawa,Hashida, and Sugawara 2013. However, although thesetwo types of measurements can obtain CO2concentrationsinthetroposphereandstratosphereupto13kmonaircraftand 35 km on stratospheric balloons Li et al. 2014, theyhavea higher cost Machidaet al. 2008andalowerverticalresolution 100–200 m than a tethered balloon 40 mKarion et al. 2010. Compared with other vertical CO2observation s, tethered balloons are not only low-cost and easy to operate Esteki et al. 2017, but can alsocontinuously observe the vertical distribution of atmo-spheric CO2in the boundary layer.In recent years, some low-cost sensors have beenproven to be practical, feasible, and suitable for environ-mental monitoring and have been put into use, such asfor pollutant gas e.g. sulfur dioxide, nitrogen dioxide,ozone and greenhouse gas e.g. CO2, methane mea-surements Holstius et al. 2014; Piedrahita et al. 2014;Wang et al. 2015. Among these sensors, the SenseAirK30 sensor, which is based on non-dispersive infraredNDIR technology produced by a Swedish manufacturerSenseAir, has been verified to be useful for high spatialdensity CO2monitoring in urban areas Martin et al.2017; Yasuda, Yonemura, and Tani 2012.Inthis study, by using the low-cost K30 sensorcarried bya tethered balloon to monitor the vertical CO2distribution,experiments were conducted to understand the verticaldistribution of CO2in the boundary layer and providecalibration and validation data for transport models e.g.the Weather Research and Forecasting WRF model andremote sensing.2. Materials and s2.1. Sampling siteTheexperimentwasconductedatYuanshiNationalMeteorological Observing Station, Shijiazhuang 11430ʹE,3748ʹN. Shijiazhuang is the capital of Hebei Province,located in the southern part of the North China Plain andat the eastern foothills of the Taihang Mountains, which arelow in the southeast and highinthe northwest, and it istheeconomic, cultural, and transportation center of HebeiProvince. The experimental site was located 28 km southof the city center with an altitude of 68.4 m, as shown inFigure 1. The mean annual temperature and mean annualprecipitation are 13.5C and 576 mm, respectively.2.2. Balloon-based CO2soundingsThe CO2vertical profile and meteorological parameterspressure, temperature, relative humidity of the lowertroposphere 0–1000 m were measured in the winterof 2019 8–16 January by a miniaturized CO2monitor-ing instrument based on NDIR technology. The teth-ered balloon experiment consisted of two processesascending and descending, and the duration of eachflight lasted approximately 1–2 h. The average ascend-ing or descending speed was 0.6 m s−1. The maximumheight of the flights was 1000 m Zhao et al. 2019. Theexperiment procedure is shown in Figure 2. There weretwo types of tethered balloon experiments Type 1involved the balloon rising to a height of 500 m type1.1, remaining for about 1.5 h, and then being pulledback to the ground type 1.2; Type 2 involved theballoon rising to 1000 m type 2.1 and then beingdirectly pulled back to the ground type 2.2. Duringthe study period, we carried out a total of 11 experi-ments, and as the experiments were divided into twoprocesses, i.e. ascending and descending, a total of 22experiment profiles were obtained. For this paper, weselected 10 typical experiment profilestoexplainthevertical distributions and factors controlling such dis-tributions. Table 1 shows the time record of typicalexperiments.2.3. Data processingThe CO2concentration was measured by a miniaturizedCO2monitoringinstrumentequippedwitha SenseAir K30sensor with an initial accuracy of 30 ppm of readinghttps// accessed October 2019. After calibration withstandard gas and environmental correction, the accuracyof K30 sensor was improved to within 5 ppm, as com-pared with the simultaneous and precise greenhouse gasanalyzer Picarro G2401 in the laboratory. The potentialtemperature θ was calculated by Bolton 1980θ TkP0Pk12 Z. BAO ET AL.where P0equals 1000 hPa, P is ambient atmospherepressure, and Tkis the Kelvin temperature, and k 0.286.In experiment 1, the data quality of the miniaturizedinstrument was verified by comparing with the gas chro-matography technology GC-FID, Agilent 7890A, SantaClara, California, USA. The GC-measured CO2data wereobtained through lab analysis after taking back air sam-ples from air bags onboard the tethered balloon. Theplanetary boundary layer height PBLH was measuredby lidar during the experiment period, which was usedto verify the potential temperature gradient .2.4. Spatial and temporal resolution simulation ofCO2concentrationHigh spatial and temporal resolution simulations of CO2concentration were conducted by using WRF-CO2for theperiod of 0600 UTC 7 January to 0600 UTCFigure 1. Location of the experiment site.ATMOSPHERIC AND OCEANIC SCIENCE LETTERS 39 January 2019 and from 0700 UTC 13 January to 0700UTC 15 January 2019. Based on the WRF model withchemistry WRF-Chem, WRF-CO2is a mesoscale, com-pressible model that provides passive tracer transportnetworks and mesoscale weather prediction capabilitiesMartin et al. 2017. It uses the VEGAS Vegetation-Global-Atmosphere-Soil model to simulate urban-scale eco-system carbon emissions/absorption as biosphere car-bon flux and GFS Global Forecast System grid data asthe meteorological boundary field of the CO2nestedsimulation. It also uses high-resolution 0.1 0.1 fossilfuel emissions hourly grid data Fossil Fuel DataAssimilation System FFDAS data http//ffdas.rc.nau.edu/Data.html in 2015 as the prior, which utilizes theKaya Identity, a to estimate emissions based oneconomic factors, and ination on national fossil fuelemissions, satellite-derived nightlights, population den-sity, and power plant ination to extract each gridpoint, to simulate the temporal and spatial distributionof CO2at the observation point Martin et al. 2019. Wealso used GEOS-Chem output as the initial field andboundary field. GEOS-Chem is a global 3D model ofatmospheric chemistry driven by meteorological from the NASA Goddard Earth Observing System http//acmg.seas.harvard.edu/geos/. The physical processparameterization schemes of WRF were configured asfollows The forecast model was configured with fullphysics options. The simulation was conducted usingthe WSM5 class for microphysics, the RRTM scheme forlongwave radiation, the Dudhia scheme for shortwaveradiation, the MM5 scheme for the surface layer, theunified Noah land-surface model for the land surface,and the YSU scheme for PBL parameterizations.3. Results and discussion3.1. Vertical distribution of CO2and comparisonwith traditional Figure 3 shows typical vertical CO2profiles obtainedduring the experiment period. The range of CO2concen-tration from 0–1000 m is 400–600 ppm, and the overallCO2concentration displays a decreasing trend with theincrease in height, which is consistent with the study ofEsteki et al. 2017. We used the near-surface CO2andPM2.5data of these 10 profiles for linear regression ana-lysis, and found a positive correlation between CO2andPM2.5, so as to infer that source emissions would affectthe concentration of CO2. Due to the combined influ-ences of emission sources and meteorological condi-tions, the level of CO2concentration changed acrossdifferent days.This is the first time that we have used miniaturizedinstruments for vertical monitoring in the boundarylayer, and so we compared the results with traditionalinstruments. Specifically, the results of experiments werecompared with the CO2concentration obtained throughgas chromatography GC through air bag sampling, asshown in Figure 4, and it was found that the two meth-ods produced good consistency. The height of the NDIRsensor and air bag taking samples was basically theabout 500mabout 1000m1.5hType 1 Type 21.1 up1.2 down2.1 up 2.2 upFigure 2. Diagram of the experiment. For type 1, the balloon stayed at the height of 500 m for 1.5 h, and for type 2 the balloon wasdirectly pulled back to the ground.Table 1. Typical tethered balloon experiments.Date Time LST Duration min Type1 8 January 2019 1404–1431 27 2.12 8 January 2019 1432–1456 24 2.23 9 January 2019 0655–0707 12 1.14 9 January 2019 0833–0852 19 1.25 9 January 2019 1332–1344 12 1.16 9 January 2019 1425–1530 65 1.27 13 January 2019 1301–1310 9 1.18 13 January 2019 1434–1455 21 1.29 14 January 2019 1317–1327 10 1.110 14 January 2019 1457–1512 15 1.14 Z. BAO ET AL.same. Within the height of 0–800 m, the CO2concentra-tion measured by the two s was within 425–500ppm. The sampling of ground air samples for GCmeasurements might be influenced by human respira-tion due to directly pumped-up ambient air near thesurface and the changes of wind direction. However,Figure 3. Typical vertical profiles of CO2. The profiles marked with triangles are type 1.1 and type 2.1, and the profiles marked withsquare signs are type 1.2 and type 2.2. Each point represents the mean CO2concentration per minute and error bars indicate the rangeof three K30 s values.Figure 4. Comparison of CO2vertical profiles between the low-cost sensor and gas chromatography s on the afternoon of8 January 2019.ATMOSPHERIC AND OCEANIC SCIENCE LETTERS 5the miniaturized instrument was placed in a box, andthus we speculate that the red dots near the surfacehigher than the K30 measurement were due to humanrespiration.3.2. Effects of meteorological conditions on thevertical distribution of CO2The two experimental tethered balloons rose to 706 mon 8 January Figure 5a and 1019 m on 13 JanuaryFigure 5d, and the results of these two experimentswere analyzed with particular emphasis because theexperimental time was relatively long and the heightwas relatively high. The potential temperature gradientdata on 8 January Figure 5c and 13 January Figure 5f were processed into minute data in order to betterdistinguish the position of the boundary layer. The pro-files both have an interval where the potential tempera-ture gradient is almost zero. Therefore, it can be judgedthat the top height of the mixing layer at both ends wasabout 600 m and 200 m, respectively, which has goodconsistency with the observed PBLH measured by lidarFigure 5c,f, dark blue line. Above the mixing layer, theCO2concentration decr