Simulation of Non-Homogeneous CO2 and Its Impact on Regional Temperature in East Asia

+ Author Affiliations + Find other works by these authors
  • Corresponding author: Tijian WANG, tjwang@nju.edu.cn
  • Funds:

    Supported by the National Natural Science Foundation of China (41575145, 41621005, and 91544230), National Key Basic Research and Development Program of China (2016YFC0203303 and 2014CB441203), and EU 7th Framework Marie Curie Actions IRSES Project: REQUA (PIRSES-GA-2013-612671)

  • doi: 10.1007/s13351-018-7159-x

PDF

  • Carbon dioxide (CO2) is an important greenhouse gas that influences regional climate through disturbing the earth’s energy balance. The CO2 concentrations are usually prescribed homogenously in most climate models and the spatiotemporal variations of CO2 are neglected. To address this issue, a regional climate model (RegCM4) is modified to investigate the non-homogeneous distribution of CO2 and its effects on regional longwave radiation flux and temperature in East Asia. One-year simulation is performed with prescribed surface CO2 fluxes that include fossil fuel emission, biomass burning, air–sea exchange, and terrestrial biosphere flux. Two numerical experiments (one using constant prescribed CO2 concentrations in the radiation scheme and the other using the simulated CO2 concentrations that are spatially non-homogeneous) are conducted to assess the impact of non-homogeneous CO2 on the re-gional longwave radiation flux and temperature. Comparison of CO2 concentrations from the model with the observations from the GLOBALVIEW-CO2 network suggests that the model can well capture the spatiotemporal patterns of CO2 concentrations. Generally, high CO2 mixing ratios appear in the heavily industrialized eastern China in cold seasons, which probably relates to intensive human activities. The accommodation of non-homogeneous CO2 concentrations in the radiative transfer scheme leads to an annual mean change of –0.12 W m–2 in total sky surface upward longwave flux in East Asia. The experiment with non-homogeneous CO2 tends to yield a warmer lower troposphere. Surface temperature exhibits a maximum difference in summertime, ranging from –4.18 K to 3.88 K, when compared to its homogeneous counterpart. Our results indicate that the spatial and temporal distributions of CO2 have a considerable impact on regional longwave radiation flux and temperature, and should be taken into account in future climate modeling.
  • 加载中
  • Fig. 1.  Model domain with topography (shaded, m) and locations of the observation stations (red dots). See Table 1 for station full names.

    Fig. 2.  Comparison of simulated (blue) and observed (red) monthly mean CO2 mixing ratios (in ppmv) at the seven stations.

    Fig. 3.  Spatial distributions of atmospheric CO2 mixing ratio (ppmv) at surface, 800 hPa, and 600 hPa in four seasons: (a) winter, (b) spring, (c) summer, and (d) autumn.

    Fig. 4.  Differences of the annual mean upward longwave flux (W m–2) under (a, c) clear sky and (b, d) total sky conditions between EXP_CTRL and EXP_REL at (a, b) TOA and (c, d) the surface.

    Fig. 5.  The zonal averaged annual mean net upward longwave radiation under the total sky condition at (a) TOA and (d) the surface, net upward longwave radiation under the clear sky condition at (b) TOA and (e) the surface, and longwave cloud forcing at (c) TOA and (f) the surface for EXP_REL (blue, scales on the left) and for the differences between EXP_CTRL and EXP_REL (green, scales on the right). Units are W m–2.

    Fig. 6.  (a) Difference (K day–1) and (b) relative difference (%) in the annual mean zonal heating rate between EXP_CTRL and EXP_REL.

    Fig. 7.  Difference of annual mean zonal average temperature (K) between EXP_CTRL and EXP_REL.

    Fig. 8.  Surface temperature difference (K) between EXP_CTRL and EXP_REL for (a) the annual mean, (b) the July mean, and (c) the January mean.

    Table 1.  Geographic information of the observation stations

    StationAbbreviationLongitude (°E)Latitude (°N)Altitude (m)
    LulinLLN120.8623.462867
    Mariana Islands, GuamGMI144.6613.39 5
    Mt. WaliguanWLG100.9036.293815
    Sary Taukum, KazakhstanKZD 75.5744.45 412
    Plateau Assy, KazakhstanKZM 77.8843.252524
    Tae-ahn PeninsulaTAP126.1236.74 21
    Ulaan Uul, MongoliaUUM111.1044.451012
    Download: Download as CSV

    Table 2.  Statistical characteristics of simulated (Sim) and observed (Obs) monthly CO2 mixing ratios at the seven stations

    StationMeanAnnual amplitudeStandard deviation
    Sim (ppmv)Obs (ppmv)Bias (%)Sim (ppmv)Obs (ppmv)Sim (ppmv)Obs (ppmv)
    LLN387.12382.51 1.20 6.4611.332.123.84
    GMI385.75385.31 0.11 4.35 6.781.732.46
    WLG386.82386.12–0.18 8.41 9.692.863.11
    KZD387.57389.63–0.5315.3015.795.275.59
    KZM386.55385.67 0.2312.1512.674.204.44
    TAP390.90392.05–0.2911.9212.883.913.62
    UUM386.67386.90–0.0613.6415.254.905.00
    Note: Bias = (simulation – observation)/observation × 100%.
    Download: Download as CSV
  • [1]

    Ahmadov, R., C. Gerbig, R. Kretschmer, et al., 2007: Mesoscale covariance of transport and CO2 fluxes: Evidence from observations and simulations using the WRF-VPRM coupled atmosphere–biosphere model. J. Geophys. Res, 112, D22107. doi: 10.1029/2007JD008552.
    [2]

    Ahmadov, R., C. Gerbig, R. Kretschmer, et al., 2009: Comparing high resolution WRF-VPRM simulations and two global CO2 transport models with coastal tower measurements of CO2. Biogeosciences, 6, 807–817. doi: 10.5194/bg-6-807-2009.
    [3]

    Ballav, S., P. K. Patra, M. Takigawa, et al., 2012: Simulation of CO2 concentration over East Asia using the regional transport model WRF-CO2. J. Meteor. Soc. Japan, 90, 959–976. doi: 10.2151/jmsj.2012-607.
    [4]

    Ballav, S., P. K. Patra, Y. Sawa, et al., 2016: Simulation of CO2 concentrations at Tsukuba tall tower using WRF-CO2 tracer transport model. J. Earth Syst. Sci., 125, 47–64. doi: 10.1007/s12040-015-0653-y.
    [5]

    Chaudhari, P. R., D. G. Gajghate, S. Dhadse, et al., 2007: Monitoring of environmental parameters for CO2 sequestration: A case study of Nagpur City, India. Environ. Monit. Assess., 135, 281–290. doi: 10.1007/s10661-007-9649-7.
    [6]

    Cheng, Y. L., X. Q. An, F. H. Yun, et al., 2013: Simulation of CO2 variations at Chinese background atmospheric monitoring stations between 2000 and 2009: Applying a CarbonTracker model. Chinese. Sci. Bull., 58, 3986–3993. doi: 10.1007/s11434-013-5895-y.
    [7]

    Chevillard, A., U. Karstens, P. Ciais, et al., 2002: Simulation of atmospheric CO2 over Europe and western Siberia using the regional scale model REMO. Tellus B, 54, 872–894. doi: 10.3402/tellusb.v54i5.16737.
    [8]

    Collins, W. D., P. J. Rasch, B. A. Boville, et al., 2004: Description of the NCAR Community Atmosphere Model (CAM 3.0). National Center for Atmospheric Research, Boulder, Colo-rado, USA, 210 pp.
    [9]

    Dickinson, R. E., A. Henderson-Sellers, and P. J. Kennedy, 1993: Biosphere–Atmosphere Transfer Scheme (BATS) Version 1e as Coupled to the NCAR Community Climate Model. Colo-rado, National Center for Atmospheric Research, Climate and Global Dynamics Division, 69 pp.
    [10]

    Dlugokencky, E., and P. Tans, 2017: Trends in Atmospheric Carbon Dioxide. Earth System Research Laboratory, National Oceanic and Atmospheric Administration (NOAA/ESRL). Available at www.esrl.noaa.gov/gmd/ccgg/trends/global.html. Accessed on 8 May 2018.
    [11]

    Dudhia, J., D. Gill, K. Manning, et al., 2004: PSU/NCAR Mesoscale Modeling System Tutorial Class Notes and User’s Guide: MM5 Modeling System Version 3. Boulder, National Center for Atmospheric Research, 390 pp.
    [12]

    Elguindi, N., X. Q. Bi, F. Giorgi, et al., 2010: RegCM Version 4.0 User’s Guide. ICTP, Trieste, Italy, 24 pp.
    [13]

    Feng, L., P. I. Palmer, Y. Yang, et al., 2011: Evaluating a 3-D transport model of atmospheric CO2 using ground-based, aircraft, and space-borne data. Atmos. Chem. Phys., 11, 2789–2803. doi: 10.5194/acp-11-2789-2011.
    [14]

    Ge, C., M. G. Zhang, L. Y. Zhu, et al., 2011: Simulated seasonal variations in wet acid depositions over East Asia. J. Air Waste Manag. Assoc., 61, 1246–1261. doi: 10.1080/10473289.2011.596741.
    [15]

    Giorgi, F., J. S. Pal, X. Bi, et al., 2006: Introduction to the TAC special issue: The RegCNET network. Theor. Appl. Climatol., 86, 1–4. doi: 10.1007/s00704-005-0199-z.
    [16]

    Giorgi, F., E. Coppola, F. Solmon, et al., 2012: RegCM4: Model description and preliminary tests over multiple CORDEX domains. Climate Res., 52, 7–29. doi: 10.3354/cr01018.
    [17]

    Grell, G. A., 1993: Prognostic evaluation of assumptions used by cumulus parameterizations. Mon. Wea. Rev., 121, 764–787. doi: 10.1175/1520-0493(1993)121<0764:PEOAUB>2.0.CO;2.
    [18]

    Grell, G. A., J. Dudhia, and D. Stauffer, 1994: A Description of the Fifth-Generation Penn State/NCAR Mesoscale Model (MM5).NCAR Technical Note NCAR/TN-398+STR, Boulder, Colo-rado, NCAR, 128 pp, doi: 10.5065/D60Z 716B.
    [19]

    Guan, D. B., G. P. Peters, C. L. Weber, et al., 2009: Journey to world top emitter: An analysis of the driving forces of China’s recent CO2 emissions surge. Geophys. Res. Lett., 36, L04709. doi: 10.1029/2008gl036540.
    [20]

    Holtslag, A. A. M., E. I. F. De Bruijn, and H. L. Pan, 1990: A high resolution air mass transformation model for short-range weather forecasting. Mon. Wea. Rev., 118, 1561–1575. doi: 10.1175/1520-0493(1990)118<1561:AHRAMT>2.0.CO;2.
    [21]

    Huang, X. X., T. J. Wang, R. Talbot, et al., 2015: Temporal characteristics of atmospheric CO2 in urban Nanjing, China. Atmos. Res., 153, 437–450. doi: 10.1016/j.atmosres.2014.09.007.
    [22]

    Iacono, M. J., E. J. Mlawer, S. A. Clough, et al., 2000: Impact of an improved longwave radiation model, RRTM, on the energy budget and thermodynamic properties of the NCAR community climate model, CCM3. J. Geophys. Res., 105, 14873–14890. doi: 10.1029/2000JD900091.
    [23]

    Kiehl, J. T., and V. Ramanathan, 1983: CO2 radiative parameterization used in climate models: Comparison with narrow band models and with laboratory data. J. Geophys. Res., 88, 5191–5202. doi: 10.1029/JC088iC09p05191.
    [24]

    Kiehl, J. T., J. J. Hack, G. B. Bonan, et al., 1996: Description of the NCAR Community Climate Model (CCM3). NCAR/TN-420+STR, Boulder, Colorado, National Center for Atmospheric Research, 55–60, doi: 10.5065/D6FF3Q99.
    [25]

    Kou, X. X., M. G. Zhang, and Z. Peng, 2013: Numerical simulation of CO2 concentrations in East Asia with RAMS-CMAQ. Atmos. Ocean. Sci. Lett., 6, 179–184. doi: 10.3878/j.issn.1674-2834.13.0022.
    [26]

    Kou, X. X., M. G. Zhang, Z. Peng, et al., 2015: Assessment of the biospheric contribution to surface atmospheric CO2 concentrations over East Asia with a regional chemical transport model. Adv. Atmos. Sci., 32, 287–300. doi: 10.1007/s00376-014-4059-6.
    [27]

    Kou, X. X., X. J. Tian, M. G. Zhang, et al., 2017: Accounting for CO2 variability over East Asia with a regional joint inversion system and its preliminary evaluation. J. Meteor. Res., 31, 834–851. doi: 10.1007/s13351-017-6149-8.
    [28]

    Krol, M., S. Houweling, B. Bregman, et al., 2005: The two-way nested global chemistry-transport zoom model TM5: Algorithm and applications. Atmos. Chem. Phys., 5, 417–432. doi: 10.5194/acp-5-417-2005.
    [29]

    Lau, C. C. S., J. C. Lam, and L. Yang, 2007: Climate classification and passive solar design implications in China. Energy Convers. Manage., 48, 2006–2015. doi: 10.1016/j.enconman.2007.01.004.
    [30]

    Liu, L. X., L. X. Zhou, X. C. Zhang, et al., 2009: The characteristics of atmospheric CO2 concentration variation of four national background stations in China. Sci. China Ser. D Earth Sci., 52, 1857–1863. doi: 10.1007/s11430-009-0143-7.
    [31]

    Masarie, K. A., and P. P. Tans, 1995: Extension and integration of atmospheric carbon dioxide data into a globally consistent measurement record. J. Geophys. Res., 100, 11593–11610. doi: 10.1029/95JD00859.
    [32]

    Nassar, R., D. B. A. Jones, P. Suntharalingam, et al., 2010: Modeling global atmospheric CO2 with improved emission inventories and CO2 production from the oxidation of other carbon species. Geosci. Model Dev., 3, 689–716. doi: 10.5194/gmd-3-689-2010.
    [33]

    Oleson, K. W., G. Y. Niu, Z. L. Yang, et al., 2008: Improvements to the Community Land Model and their impact on the hydrological cycle. J. Geophys. Res., 113, G01021. doi: 10.1029/2007JG000563.
    [34]

    Olivier, J. G. J., G. Janssens-Maenhout, M. Muntean, et al., 2015: Trends in Global CO2 Emissions: 2015 Report. The Hague, PBL Netherlands Environmental Assessment Agency, 80 pp.
    [35]

    Peters, W., A. R. Jacobson, C. Sweeney, et al., 2007: An atmospheric perspective on North American carbon dioxide exchange: CarbonTracker. Proc. Natl. Acad. Sci. USA, 104, 18925–18930. doi: 10.1073/pnas.0708986104.
    [36]

    Pillai, D., C. Gerbig, R. Ahmadov, et al., 2011: High-resolution simulations of atmospheric CO2 over complex terrain-representing the Ochsenkopf mountain tall tower. Atmos. Chem. Phys., 11, 7445–7464. doi: 10.5194/acp-11-7445-2011.
    [37]

    Sahay, S., and C. Ghosh, 2013: Monitoring variation in greenhouse gases concentration in urban environment of Delhi. Environ. Monit. Assess., 185, 123–142. doi: 10.1007/s10661-012-2538-8.
    [38]

    Stocker, T., D. Qin, G. Plattner, et al., 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom, New York, NY, USA, 1535 pp, doi: 10.1017/CBO9781107415324.
    [39]

    Uppala, S., D. Dee, S. Kobayashi, et al., 2008: Towards a climate data assimilation system: Status update of ERA-Interim. ECMWF Newsletter, 115, 12–18. doi: 10.21957/byinox4wot.
    [40]

    Wang, H. J., and J. W. Liu, 2008: Global Change and Human Adaptation. China Forestry Press, Beijing, 347 pp. (in Chinese)
    [41]

    World Meteorological Organization, 2017: WMO Greenhouse Gas Bulletin: The State of Greenhouse Gases in the Atmosphere Based on Global Observations Through 2015. Switzerland, WMO, 8 pp.
    [42]

    Yang, C. Y., H. J. Wang, S. J. Han, et al., 2012: Climate simulation for dynamic heterogeneous distribution of atmospheric CO2 concentration. Chinese J. Geophys., 55, 2809–2825. doi: 10.6038/j.issn.0001-5733.2012.09.001. (in Chinese)
  • 20180702142540.pdf

  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Simulation of Non-Homogeneous CO2 and Its Impact on Regional Temperature in East Asia

    Corresponding author: Tijian WANG, tjwang@nju.edu.cn
  • 1. School of Atmospheric Sciences, Nanjing University, Nanjing 210023
  • 2. College of Plant Science & Technology, Huazhong Agricultural University, Wuhan 430070
  • 3. Joint Laboratory for Climate Prediction Studies of China Meteorological Administration and Nanjing University, Nanjing 210023
  • 4. Collaborative Innovation Center for Climate Change of Jiangsu Province, Nanjing 210023
Funds: Supported by the National Natural Science Foundation of China (41575145, 41621005, and 91544230), National Key Basic Research and Development Program of China (2016YFC0203303 and 2014CB441203), and EU 7th Framework Marie Curie Actions IRSES Project: REQUA (PIRSES-GA-2013-612671)

Abstract: Carbon dioxide (CO2) is an important greenhouse gas that influences regional climate through disturbing the earth’s energy balance. The CO2 concentrations are usually prescribed homogenously in most climate models and the spatiotemporal variations of CO2 are neglected. To address this issue, a regional climate model (RegCM4) is modified to investigate the non-homogeneous distribution of CO2 and its effects on regional longwave radiation flux and temperature in East Asia. One-year simulation is performed with prescribed surface CO2 fluxes that include fossil fuel emission, biomass burning, air–sea exchange, and terrestrial biosphere flux. Two numerical experiments (one using constant prescribed CO2 concentrations in the radiation scheme and the other using the simulated CO2 concentrations that are spatially non-homogeneous) are conducted to assess the impact of non-homogeneous CO2 on the re-gional longwave radiation flux and temperature. Comparison of CO2 concentrations from the model with the observations from the GLOBALVIEW-CO2 network suggests that the model can well capture the spatiotemporal patterns of CO2 concentrations. Generally, high CO2 mixing ratios appear in the heavily industrialized eastern China in cold seasons, which probably relates to intensive human activities. The accommodation of non-homogeneous CO2 concentrations in the radiative transfer scheme leads to an annual mean change of –0.12 W m–2 in total sky surface upward longwave flux in East Asia. The experiment with non-homogeneous CO2 tends to yield a warmer lower troposphere. Surface temperature exhibits a maximum difference in summertime, ranging from –4.18 K to 3.88 K, when compared to its homogeneous counterpart. Our results indicate that the spatial and temporal distributions of CO2 have a considerable impact on regional longwave radiation flux and temperature, and should be taken into account in future climate modeling.

    • Carbon dioxide (CO2) is a strong greenhouse gas, and its radiative forcing (RF) since industrial era is reported to be 1.82 W m–2 with an average growth rate slightly less than 0.3 W m–2 per decade (Stocker et al., 2013). The World Meteorological Organization (WMO) Greenhouse Gas Bulletin No. 13 states that CO2 has been the dominant component of greenhouse gases (GHGs) and contributes approximately 65% to the total global RF by long-lived greenhouse gases (World Meteorological Organization, 2017). Climate modeling is an effective approach to quantifying the radiative impact of GHGs on climate system. Currently, global or regional climate models have been developed with detailed parameterization schemes to consider the RF of GHGs. However, the spatiotemporal variations of atmospheric CO2 are usually ignored in these models (Kiehl and Ramanathan, 1983). Atmospheric CO2 is often prescribed to be well mixed in these models. For instance, CO2 mixing ratio is assumed as 355 ppmv in a global climate model CAM3.0 (Collins et al., 2004), while it is set as 330 ppmv in a mesoscale model MM5 (Dudhia et al., 2004). In fact, real-time atmospheric CO2 concentrations are constantly modulated by the magnitudes of sources and sinks of atmospheric CO2 fluxes. For example, the averaged daytime ambient concentrations of CO2 in residential areas vary from 495 to 621 ppmv in Delhi (Sahay and Ghosh, 2013), whereas the maximum daytime ambient concentrations of CO2 in Nagpur are only 391 ppmv (Chaudhari et al., 2007). Although annual increments of GHGs during the period 1750–2100 have been considered in radiation schemes of some climate models such as RegCM4, seasonal and spatial variations of CO2 concentrations have been neglected (Elguindi et al., 2010). The constant CO2 concentrations prescribed in climate models undoubtedly influence model performance (Yang et al., 2012). Therefore, it is essential to investigate the effects of non-homogeneous CO2 concentrations on the regional RF and temperature.

      Three dimensional numerical models can capture the impacts of human activities (Cheng et al., 2013), ecosystem (Kou et al., 2015, 2017), mesoscale transport (Ahmadov et al., 2009), and chemical conversion (Nassar et al., 2010) on atmospheric CO2. Therefore, it is possible for these models to capture the spatial and temporal variations of atmospheric CO2. Models for simulating atmospheric CO2 can generally be divided into global and regional models by the simulation domains. Global models such as TM5 (Krol et al., 2005), GEOS-Chem (Nassar et al., 2010), and CarbonTracker (Cheng et al., 2013) cannot identify mesoscale effects generated by land surface heterogeneity and complex topography (Ahmadov et al., 2009). In contrast, it is possible for regional models to simulate atmospheric CO2 concentrations at high spatial and temporal resolutions. Regional models like WRF-VPRM (Ahmadov et al., 2007), RAMS-CMAQ (Kou et al., 2013), and REMO (Chevillard et al., 2002) are able to capture sharp changes in diurnal variability in CO2 concentrations within a much shorter spatial distance than global models. For example, Ballav et al. (2012) simulated the regional transport of atmospheric CO2 using 5 different CO2 fluxes from ocean, fossil fuel, and terrestrial biospheres at the horizontal resolution of 27 × 27 km; Kou et al.(2015, 2017) assessed the impacts of biosphere–atmosphere CO2 exchange on seasonal variations of atmospheric CO2 concentrations over East Asia, using a regional model.

      East Asia has a broad terrain with complex topography and diverse climates (Lau et al., 2007). Along with the rapid expansion of energy consumption, anthropogenic CO2 emissions from human activities have significantly increased in recent years (Guan et al., 2009). Moreover, regional economic development imbalance still persists in the region, which is a main factor contributing to CO2 inhomogeneity (Wang and Liu, 2008). Previous studies have found that the CO2 concentrations change sharply at Shangdianzi (from 377.8 to 391.7 ppmv) and Lin’an (from 383.2 to 393.2 ppmv) due to the great influence of urbanization in the Beijing–Tianjin– Hebei region and the Yangtze River Delta region (Liu et al., 2009). Measurements from national background stations indicate that the atmospheric CO2 concentration at Waliguan station is 383.5 ppmv, while ambient CO2 mixing ratio over urban Nanjing reaches 406.5 ppmv (Huang et al., 2015). Given such spatial variations in CO2 concentrations, it is of scientific significance to explore the spatiotemporal distribution of atmospheric CO2 concentrations and the temperature effects of non-homogeneous CO2 in East Asia.

      Previous studies have employed global transport models to assess spatiotemporal variability of CO2 concentrations (Feng et al., 2011; Ballav et al., 2016). Fewer studies have considered the effect of CO2 spatial variations on regional temperature (Yang et al., 2012). Regional models with high spatial resolutions can be powerful in simulating the processes that influence atmospheric CO2 concentrations, such as transport, emissions, and consumption. The primary purpose of this study is to investigate the spatiotemporal characteristics of atmospheric CO2 by using the regional climate model RegCM4 and to assess the impacts of the dynamic non-homogeneous CO2 concentrations on temperature in East Asia. Sources and sinks of CO2 are treated as prescribed surface fluxes in the modified RegCM4 model, while CO2 is transported as a tracer in this model. The model can retrieve signals from the anthropogenic emissions and the biosphere–atmosphere exchanges, and thus capture the spatiotemporal variations of ambient CO2 concentrations. To identify and quantify the effect of non-homogeneous CO2 concentrations on regional temperature, the radiative transfer scheme in RegCM4 is revised to consider the spatial and temporal distributions of atmospheric CO2, by substituting the prescribed constant CO2 concentrations with the simulated non-homogeneous CO2 concentrations. The model and input data are described in detail in Section 2, followed by a presentation and discussion of the results in Section 3. Summary and conclusions are provided in Section 4.

    2.   Methods and data
    • The regional climate model RegCM4 is an evolution of its previous version, RegCM3, developed by the International Center for Theoretical Physics (ICTP) in Italy (Giorgi et al., 2012). The model dynamics is the same as that of the hydrostatic version of MM5 (Grell et al., 1994), which has been used for a wide range of applications in the world (Giorgi et al., 2006). RegCM4 includes a mixed convection and tropical band configuration, and modifications to the pre-existing boundary layer schemes. The radiative transfer calculations in RegCM4 are carried out with the radiative transfer scheme of the global model CCM3 (Kiehl et al., 1996), which includes contributions from all the main greenhouse gases, i.e. H2O, CO2, O3, CH4, N2O, and CFCs. The land surface module of RegCM4 is optional, with the Biosphere–Atmosphere Transfer Scheme version 1e (BATS1e) (Dickinson et al., 1993) and a new option to use the Community Land Model, version CLM3.5 (Oleson et al., 2008).

      In this study, the Holtslag PBL scheme (Holtslag et al., 1990), the Grell cumulus convection scheme (Grell, 1993), and the CLM3.5 land surface module are employed. In order to reproduce the real-time atmospheric CO2 distribution, we have revised RegCM4 to include CO2 simulations. CO2 is treated as an inert chemical species, whose concentrations are modulated by atmospheric transport, and source and sink processes in the model. Atmospheric transport includes horizontal/vertical advection and diffusion. Sources and sinks of CO2 are prescribed as surface fluxes that include fossil fuel emissions, biomass burning, air–sea CO2 exchanges, and terrestrial biosphere CO2 fluxes.

      Advection and diffusion are parameterized by using Giorgi’s method (Giorgi et al., 2006). Anthropogenic CO2 emissions are from the Emission Database for Global Atmospheric Research version 4.2 (EDGAR4.2). This dataset provides greenhouse gas emissions per country and on a 0.1° × 0.1° grid for all anthropogenic sources identified by the IPCC for the 1970–2012 period (Olivier, 2015). The CO2 emissions have been aggregated into five main source sectors, including fossil-fuel combustion, fugitive emissions from fuels, cement production and other carbonate uses, feedstock and other non-energy uses of fossil fuels, waste incineration and fuel fires. Biomass-burning emissions from forest wildfires, biosphere–atmosphere exchange, and ocean flux of CO2 are adopted from the NOAA CarbonTracker CT2011 (data available at http://carbontracker.noaa.gov). CarbonTracker uses atmospheric CO2 observations from a global observational network and simulates atmospheric transport to provide optimized biospheric and oceanic CO2 fluxes (Peters et al., 2007). This dataset is available on the global 3° (longitude) × 2° (latitude) grids at a 3-h time resolution.

    • To reveal the temporal and spatial variations of CO2 in East Asia, revised RegCM4 with prescribed surface CO2 fluxes is run from 1 December 2008 to 31 December 2009 with the first month as a spin-up period. Only the results in 2009 are used for analyses. The study domain is 7260 km × 4620 km, with a spatial resolution of 60 km × 60 km on a Lambert conformal map projection centered at 34°N, 117°E (Fig. 1). This domain, covering most of East Asia, has large variations in topography and land cover types. In the vertical, 18 levels are set, with a top at 50 hPa. The lateral meteorological boundary conditions are updated every 6 h using the ERA-Interim reanalysis data, which can provide an improved representation of the hydrological cycle compared to previous products, especially over the tropical regions (Uppala et al., 2008). The Optimum Interpolation Sea Surface Temperature (OISST) of the NOAA Climate Analysis Center (CAC) Weekly Optimal Interpolation dataset in the original NetCDF format is used as the SST boundary forcing. CO2 concentrations obtained from CT2011 optimized estimation (CT2011_oi; 3° longitude × 2° latitude) are interpolated into the model grids as the initial and boundary conditions.

      Figure 1.  Model domain with topography (shaded, m) and locations of the observation stations (red dots). See Table 1 for station full names.

      Two numerical experiments are carried out to further assess the effects of non-homogeneous CO2 concentrations on longwave radiation flux and temperature. One (named EXP_CTRL) adopts the primitive radiation scheme of the RegCM4 model, which uses constant CO2 concentrations, i.e., 385.9 ppmv for the experiment period 2009. The other experiment (named EXP_REL) is conducted by using simulated CO2 mixing ratios, which are closer to the real-time atmospheric CO2 distributions. The column averaged CO2 concentrations of EXP_REL is 387.6 ppmv, which is slightly higher than that in EXP_CTRL (385.9 ppmv), although the difference between them can bring about some uncertainties in our assessment. Nevertheless, the difference is much lower than the CO2 spatial inhomogeneity, which varies from 379.1 to 411.8 ppmv in EXP_REL. Therefore, the difference in the simulations between the two experiments is mainly caused by non-homogeneous CO2 concentrations. Comparing these two experiments, we can assess the impact of non-homogeneous CO2 on longwave radiation flux and regional temperature.

    • The observed CO2 concentration data is taken from GLOBALVIEW-CO2 (2013), which is provided by the Cooperative Atmospheric Data Integration Project. This dataset provides pseudo-weekly CO2 for each site by fitting the CO2 flask or in-situ observations with a smooth curve (Masarie and Tans, 1995). Besides, the project is coordinated and maintained by the Carbon Cycle Greenhouse Gases Group of the NOAA Earth System Research Laboratory (ESRL), which is a cooperative effort among many organizations and institutions making high-quality atmospheric CO2 measurements available.

      The locations of seven observation stations in East Asia are given in Fig. 1. Among the seven stations, Guam (GMI) station is located in the western Pacific island, surrounded by ocean, far from anthropogenic sources and terrestrial ecosystems. The Mt. Waliguan (WLG) station is the global background station of WMO/GAW (Global Atmosphere Watch) in the central Eurasian continent (Liu et al., 2009). Lulin (LLN) station and Tae-ahn Peninsula (TAP) station are located in populous areas, but being isolated and far from megacities. Detailed geographical information of these stations is listed in Table 1.

      StationAbbreviationLongitude (°E)Latitude (°N)Altitude (m)
      LulinLLN120.8623.462867
      Mariana Islands, GuamGMI144.6613.39 5
      Mt. WaliguanWLG100.9036.293815
      Sary Taukum, KazakhstanKZD 75.5744.45 412
      Plateau Assy, KazakhstanKZM 77.8843.252524
      Tae-ahn PeninsulaTAP126.1236.74 21
      Ulaan Uul, MongoliaUUM111.1044.451012

      Table 1.  Geographic information of the observation stations

    3.   Results and discussion
    • Figure 2 compares the monthly mean CO2 mixing ratios at the seven stations between the model simulations and measurements, with the detailed statistical information given in Table 2. The simulated CO2 concentrations are generally in good agreement with the observations, particularly at UUM and GMI. CO2 mixing ratios show strong seasonality at the inland stations (UUM, WLG, KZM, and KZD), typically with a winter or spring maximum and a summer minimum. This seasonality is mainly driven by the CO2 flux variation between the atmosphere and terrestrial vegetation and the seasonal patterns of atmospheric mixing and transport. The annual mean CO2 mixing ratios at TAP and KZD stations are higher than the other stations (Table 2). It is predominantly related to the strong influence of local emissions. KZM, which is geographically adjacent to KZD, shows lower CO2 all year round due to a higher terrain elevation (Table 2). Apparently, the model performs less well in urban and coastal areas than inland areas. For example, the model generally overestimates CO2 at LLN station where it is in urban areas of the Taiwan Island and sometimes underestimates CO2 at TAP station located in a coastal area of Korean Peninsula. Because the input anthropogenic emissions are adopted from yearly mean EDGAR4.2 dataset, it is difficult for model to capture the local emission fluctuations in urban regions. The effects of complex topography and microscale circulation also make it hard for the model to well reproduce the seasonality of CO2 concentrations (Pillai et al., 2011).

      Figure 2.  Comparison of simulated (blue) and observed (red) monthly mean CO2 mixing ratios (in ppmv) at the seven stations.

      StationMeanAnnual amplitudeStandard deviation
      Sim (ppmv)Obs (ppmv)Bias (%)Sim (ppmv)Obs (ppmv)Sim (ppmv)Obs (ppmv)
      LLN387.12382.51 1.20 6.4611.332.123.84
      GMI385.75385.31 0.11 4.35 6.781.732.46
      WLG386.82386.12–0.18 8.41 9.692.863.11
      KZD387.57389.63–0.5315.3015.795.275.59
      KZM386.55385.67 0.2312.1512.674.204.44
      TAP390.90392.05–0.2911.9212.883.913.62
      UUM386.67386.90–0.0613.6415.254.905.00
      Note: Bias = (simulation – observation)/observation × 100%.

      Table 2.  Statistical characteristics of simulated (Sim) and observed (Obs) monthly CO2 mixing ratios at the seven stations

      For these seven stations, the biases between the simulated and observed monthly mean CO2 mixing ratios range from –0.53% at KZD to 1.20% at LLN (Table 2). The difference between the simulated and observed annual amplitude of monthly mean CO2 concentrations is less than 2.5 ppmv, indicating the seasonal variation of atmospheric CO2 levels is well captured by RegCM4. The remote ocean station (GMI) shows the lowest seasonal variation, with the annual amplitude of 6.78 ppmv and standard deviation of 2.46 ppmv in the observations, respectively. At KZD and UUM, the standard deviations of the monthly mean CO2 mixing ratios are higher than at the other stations, indicating remarkable impacts of terrestrial biosphere. The difference between simulated and observed standard deviations of monthly mean CO2 mixing ratios is less than 0.4 ppmv, suggesting that RegCM4 can reasonably reproduce the monthly variations of CO2 mixing ratio. At the mountain station (LLN), the simulated annual mean CO2 mixing ratio (387.12 ppmv) is much higher than the observation (382.51 ppmv), while the simulated annual amplitude is much weaker. The discrepancy is probably due to the inability of the model to resolve the complex local topography and small-scale system effects.

    • Figure 3 depicts the simulated spatial variations of atmospheric CO2 mixing ratios in four seasons. At the surface layer, atmospheric CO2 varies with region. The CO2 mixing ratio in urban areas is evidently higher than that in remote regions with a maximum discrepancy of approximately 30 ppmv. Discernible gradients of surface CO2 concentrations can be found between southwestern and eastern China, especially in winter and spring. Surface CO2 mixing ratios can reach 410 ppmv in parts of East China and Korea, especially in urban areas.

      Figure 3.  Spatial distributions of atmospheric CO2 mixing ratio (ppmv) at surface, 800 hPa, and 600 hPa in four seasons: (a) winter, (b) spring, (c) summer, and (d) autumn.

      It is also seen from Fig. 3 that the simulated surface CO2 mixing ratios in East Asia exhibit notable seasonal variations, which agree with numerous ground-based in situ measurements (e.g., Dlugokencky and Tans, 2017). CO2 concentrations are generally higher in winter, particularly near the surface, similar to the observations (Dlugokencky and Tans, 2017). Moreover, the seasonality of surface CO2 concentrations varies with region, with weaker seasonal variation over oceans than over continental areas. In winter, high CO2 mixing ratios mainly appear in the Sichuan basin, North China, Japan, and Korea, where considerable anthropogenic CO2 is released from intensive human activities. Due to the strong influence of the biosphere, CO2 concentrations are much lower in summer, with the maximum of 400 ppmv in the North China Plain.

      Due to the influences of human activities and the ecosystem, the CO2 concentrations in the lower troposphere (the surface to 800 hPa) are significantly affected by the underlying surface. Anthropogenic emissions make more contributions to CO2 concentrations in urban areas than in remote regions, where CO2 concentrations are primarily influenced by the terrestrial ecosystem (Kou et al., 2013). Atmospheric CO2 mixing ratios in the Sichuan basin are higher than neighboring areas, which is related to the inverse temperature structure (Ge et al., 2011), and local emissions.

      Different from in the lower troposphere, the CO2 concentrations in the middle troposphere (600 hPa) tend to have smaller longitudinal variations with weaker land–sea CO2 contrasts, resulting from lesser influence from the underlying surface (Fig. 3). The result suggests that air masses from the middle to upper troposphere are well mixed, with similar spatial distribution patterns and small CO2 horizontal variations.

    • By conducting two numerical experiments, the impacts of non-homogeneous CO2 concentrations on longwave radiation flux are assessed. Figure 4 demonstrates the difference in annual mean net upward longwave radiation under clear sky and the total sky conditions between EXP_CTRL and EXP_REL at top-of-atmosphere (TOA) and the surface. Generally, the regionally averaged difference is negative for all the four conditions, ranging from –0.31 to –0.03W m–2. There is more longwave radiation that has reached the earth surface and atmosphere after using the simulated spatially explicit CO2 concentrations in the radiative transfer scheme. Comparing Figs. 4a, b with Figs. 4c, d, we find that the difference of upward longwave flux at the earth surface is evidently larger than that at TOA, because the CO2 spatial variation is larger in the lower troposphere than in the upper troposphere. The impact of non-homogeneous CO2 concentrations on radiation flux for the total sky condition is more notable than for the clear sky condition. The altered CO2 spatial distribution affects the energy balance of the atmosphere, changing the amount of water vapor content and clouds, and resulting in a feedback on the radiation flux.

      Figure 4.  Differences of the annual mean upward longwave flux (W m–2) under (a, c) clear sky and (b, d) total sky conditions between EXP_CTRL and EXP_REL at (a, b) TOA and (c, d) the surface.

      The differences between EXP_CTRL and EXP_REL vary spatially. From Fig. 4a, it is seen that the change of clear sky TOA upward longwave flux is positive in most areas of East Asia except for eastern Sichuan and Hebei Provinces, and Japan as well. This difference is more evident under the total sky condition (Fig. 4b). The spatial distributions with positive or negative differences in Fig. 4b are similar to those in Fig. 4a, but with larger absolute values, ranging from 1 to 2.5 W m–2. This indicates the indirect impact of cloud variation induced by non-homogeneous CO2 concentrations on radiation flux. In Fig. 4c, the difference of surface net upward longwave radiation flux between the two experiments is relatively homogeneous over ocean areas under the clear sky condition, similar to the distribution of atmospheric CO2 concentrations (Fig. 3). The difference in surface net upward longwave radiation flux under the total sky condition ranges from 3 to 5 W m–2, the largest among all four conditions, which is related to the sharp CO2 horizontal gradient near the surface (Fig. 3). Evidently, non-homogeneous distribution of CO2 concentrations near the surface and cloud forcing are crucial factors affecting the longwave radiation flux at TOA and the surface.

      To further understand the impact of non-homogeneous CO2 on longwave radiation flux, the zonal averaged annual mean energy balance is explored. Figure 5 shows the net upward longwave radiation under the total sky and clear sky conditions at TOA and the surface, and longwave cloud forcing at TOA and the surface. The difference between the clear-sky and total-sky net upward longwave radiation is defined as cloud forcing. The blue curves represent the results of EXP_REL, and the green curves show the difference between the two experiments. Figure 5 shows that the TOA net upward longwave radiation reduces with latitude, mainly due to the colder surface and atmosphere in the north of the study domain than in the south of it. Due to the accommodation of non-homogeneous CO2 concentrations, the TOA net upward longwave radiation under the total sky condition decreases in southern East Asia over 15°–45°N, but increases in northern East Asia over 45°–50°N (Fig. 5a). The differences between the two experiments range from –1.25 to 0.5 W m–2, with the maximum difference appearing near 37°N. Different from the total sky condition, the discrepancy of the TOA net upward longwave radiation under the clear sky condition between the two experiments is quite small, ranging from –0.25 to 0.25 W m–2 in East Asia (Fig. 5b). As seen from Fig. 5c, TOA longwave cloud forcing increases with latitude and reaches the maximum near 32°N, which is rather unlike the total- and clear-sky TOA net upward longwave radiation. Opposite to Fig. 5a, the discrepancies of TOA longwave cloud forcing between the two experiments are positive in the regions of 15°–46°N, but negative in the regions of 46°–50°N. The maximum (1.1 W m–2) appears near 37°N while the minimum (–0.4 W m–2) is found at 48°N. As mentioned, the spatial variation of atmospheric CO2 has slight impact on the clear-sky TOA upward longwave radiation. However, a non-homogeneous CO2 distribution notably affects the spatial and temporal distributions of cloud cover, contributing to the alteration of the total-sky TOA outgoing longwave radiation.

      Figure 5.  The zonal averaged annual mean net upward longwave radiation under the total sky condition at (a) TOA and (d) the surface, net upward longwave radiation under the clear sky condition at (b) TOA and (e) the surface, and longwave cloud forcing at (c) TOA and (f) the surface for EXP_REL (blue, scales on the left) and for the differences between EXP_CTRL and EXP_REL (green, scales on the right). Units are W m–2.

      In contrast to the situations at TOA, the energy balance at the earth surface is more complicated. The zonal averaged surface net upward longwave radiation varies with latitude, and the relatively high longwave radiation flux appears over 30°–40°N. This is possibly related to the integrated effects of the large water vapor gradients, cloud variation, and complicated topography in these regions. Comparing the two experiments finds that the difference of the total sky surface net upward longwave radiation in East Asia is negative over 21°–40°N, where atmospheric CO2 concentrations are much higher than in other areas due to intensive human activities (Fig. 3), with a minimum of –1.1 W m–2 (Fig. 5d). The maximum difference between adjacent latitudes even reaches 0.5 W m–2, probably due to the substantial CO2 spatial variations. In Fig. 5e, the discrepancy between the two experiments under the clear sky condition ranges from –0.5 to 0.5 W m–2, smaller than that under the total sky condition, but similar to that under the total sky condition in the spatial pattern of positive and negative areas. By considering the non-homogeneous CO2 concentrations, surface longwave cloud forcing is enhanced across 22°–43°N, which is in agreement with net upward longwave radiation flux. Non-homogeneous CO2 concentrations affect the longwave radiation balance at the surface, resulting in a decline in net upward longwave radiation flux in the subtropics. Furthermore, cloud forcing derived by CO2 variation strengthens this trend.

    • Figure 6 indicates the absolute and the relative (defined as the ratio of heating rate differences to the heating rate from EXP_CTRL) differences of annual mean zonal longwave radiation heating rate between the two experiments at multiple altitudes. Broadly speaking, the application of non-homogeneous CO2 concentrations in the radiative scheme exhibits an evident effect on the heating rate of longwave radiation at different altitudes, and the range of heating rate difference is from –0.13 to 0.13 K day–1.

      Figure 6.  (a) Difference (K day–1) and (b) relative difference (%) in the annual mean zonal heating rate between EXP_CTRL and EXP_REL.

      In Fig. 6a, positive heating rate differences mainly appear at the altitudes of 970–830 hPa, and over the regions between 32°–41°N (about 0.08 to 0.13 K day–1) and 44°–48°N (about 0.07 to 0.09 K day–1). It is probably due to high atmospheric CO2 levels in these regions (Fig. 3), causing more absorption of longwave radiation. The lower troposphere over tropical areas (15°–21°N) tends to be colder with the application of non-homogeneous CO2 concentrations in the radiative scheme, with a decline of –0.08 to –0.05 K day–1 in the heating rate. It is consistent with the increase in the net surface upward longwave radiation in these areas (Fig. 5). The relative differences vary between –5% and 5% for most altitudes (Fig. 6b). In the surface layer, longwave radiation heating rate changed by non-homogeneous CO2 concentrations is comparatively larger than at other altitudes. It is supposed to be a result of substantial variations in CO2 concentrations near the surface (Fig. 3), which is induced by the combined effects of local emissions, terrestrial biosphere, and mesoscale transport.

      Previous studies have found that longwave radiation heating is a primary mechanism. The thermodynamic structure is affected by the heating rate parameterization in the radiative scheme, which, in turn, can change the atmospheric temperature and even the dynamic structure of the atmosphere in climate models (Iacono et al., 2000). The change of longwave radiation heating rate contributes to increases in surface and lower tropospheric annual mean zonal temperature (Fig. 7). The maximum increase (approximately 0.18 K) is found near the surface in the temperate zone, i.e., between the altitude of 1000–900 hPa and over the latitudes of 42°–50°N. Temperature is enhanced even from the surface to 800 hPa over the tropical zone (15°–24°N), with a range of 0.05–0.1 K. Due to atmospheric CO2 heterogeneity, the difference of temperature between the two experiments varies spatially. The upper troposphere tends to be cold due to less longwave radiation. The cold center appears between the altitudes of 300–150 hPa and over 32°–45°N, with a maximum of –0.05 K. It is evident that non-homogeneous distribution of CO2 concentrations affects the atmospheric temperature, and this impact varies with altitude and latitude. Considering the influence on longwave radiation, the application of simulated CO2 concentrations in the radiative scheme enhances downward longwave radiation, especially in the lower troposphere, and then warms the earth surface and the lower troposphere, resulting in an increase in temperature near the surface.

      Figure 7.  Difference of annual mean zonal average temperature (K) between EXP_CTRL and EXP_REL.

      Figure 8 indicates the surface temperature difference between the two experiments for different periods. In general, surface temperature differences vary by region, especially in summer. Surface temperature increases for most parts of mainland in East Asia from a one year simulation, while the maximum increment appears in central South Korea (approximately 1.3 K) (Fig. 8a). However, for eastern Sichuan and Hebei provinces, the earth surface tends to be colder with a maximum of about –0.15 K. Seasonally, the surface temperature difference for summer (averaged for July) and winter (averaged for January) is given in Figs. 8b, c, respectively. The surface tends to be warmer in summer in northeastern China, the Yangtze River Delta, and Korean Peninsula. The maximum increase even reaches 3.5 K in Liaoning Province. In central Inner Mongolia, northern China, and Sichuan Province, surface temperature decreases by 1–4 K. However, regionally, the wintertime surface temperature difference is smaller. Most regions in East Asia exhibit a warming trend, but the temperature increment is usually less than 0.5 K except for central South Korea and the Pearl River Delta of China. For areas such as central Japan and Shandong Province, a cooling trend is shown with a range of –0.6 to –0.2 K. The maximum absolute values of the annual, summer, and winter surface temperature differences are 1.27, 4.18, and 1.10 K, respectively. Overall, the difference in summer is obviously larger than that in winter, and varies more regionally. Summer is a season when plants grow vigorously. Strong photosynthetic absorptions of atmospheric CO2 combined with enhanced transpiration induce large decrease in temperature in vegetated areas. While in urban areas, high CO2 concentrations and more intense summer radiation result in a warmer surface than in other seasons. The annual, summer, and winter temperature differences are all positive in the entire simulation domain, indicating that the application of non-homogeneous CO2 concentrations in the model generally warms the earth surface.

      Figure 8.  Surface temperature difference (K) between EXP_CTRL and EXP_REL for (a) the annual mean, (b) the July mean, and (c) the January mean.

    4.   Summary and conclusions
    • In this study, regional climate model RegCM4 is revised to simulate the spatial and temporal distributions of atmospheric CO2 in East Asia. The comparison with the ground observations suggests that the modified RegCM4 can reasonably reproduce the spatial inhomogeneity and seasonal variation of CO2 concentrations, despite of uncertainties resulting from inconsideration of the dynamic influence of anthropogenic emissions in the model. Surface CO2 mixing ratios are relatively higher in winter and spring. It is probably related to the source/sink transition during different phases of plant growth and the seasonal patterns of meteorological conditions. The difference in the surface CO2 between urban and remote regions is ascribed to anthropogenic CO2 emissions and biogenic CO2 flux.

      Furthermore, in order to assess the effect of non-homogeneous CO2 concentrations on regional longwave flux and temperature, two simulation experiments are conducted, one with a prescribed constant CO2 concentrations and the other with simulated CO2 concentrations that vary spatially in the radiative transfer scheme in the model. Comparing the two experiments finds that the surface energy balance is altered when considering the spatial distribution of atmospheric CO2 in the radiative scheme. The implementation of non-homogeneous CO2 results in a maximum of –1.1 W m–2 decrease in net surface upward longwave flux in subtropical regions (21°–40°N). Accumulated atmospheric CO2 in the lower troposphere engendered by local emissions traps more longwave radiation, and consequentially enhances the downward longwave flux at the surface. With the altered earth energy balance, the lower tropospheric atmosphere tends to be warmer, with a maximum temperature increment of 0.18 K.

      Our results suggest that the surface temperature simulated by current regional climate models may be biased, especially in summer, because of the hypothesis of well mixed CO2 concentrations in the radiative scheme. The discrepancies in regional longwave flux and temperature due to non-homogeneous CO2 concentrations found in this study indicate that the spatial and temporal distributions of CO2 concentrations should be taken into account in future climate modeling.

      Acknowledgments. The authors thank Jingxian Liu for help with the language check. We also thank the anonymous reviewers for their constructive and valuable comments on this paper.

Reference (42)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return