Study on Clouds in the Hinterland of the Taklimakan Desert Based on Consecutive MMCR Observations

基于地基毫米波雷达连续探测数据对塔克拉玛干沙漠腹地云的研究

+ Author Affiliations + Find other works by these authors
  • Corresponding author: Minzhong WANG, wangmz@idm.cn
  • Funds:

    Supported by the National Natural Science Foundation of China (41775030) and National Science Foundation for Young Scientists of China (41805006)

  • doi: 10.1007/s13351-021-1023-0

PDF

  • This study mainly focused on the structure and evolution of desert clouds, and was the first to conduct high-resolution consecutive detection of clouds over the hinterland of the Taklimakan Desert (TD) from April to June 2018 using the ground-based Ka-band Millimeter Wave Cloud Radar (MMCR). The reflectivity factor, cloud boundary, and liquid water content (LWC) were calculated based on the power spectrum data observed by the MMCR, which were verified by comparing the detection data of the cloud profile radar (CPR) on the CloudSat board. The results showed that: clouds over the TD were dominated by medium and high clouds, with their thicknesses generally being less than 2 km; moreover, the mean LWCs of the medium and high clouds were less than 0.01 g m−3, which implied that cirrus and stratiform clouds were predominant. However, for the low clouds, the average thickness was 3166 m and the drizzles were concentrated within 2.5 to 4.5 km, which indicated that precipitation would more likely occur in the low clouds. The mean LWC in the clouds over the TD was 0.0196 g m−3, which was less than that of clean clouds. Compared to the other periods, the average durations and LWCs in clouds increased significantly around noon owing to the obvious surface heating by sensible heat flux. The average duration for the evolution of high to low clouds was approximately 2 h, and the average maximum LWC increased from 0.008 to 0.139 g m−3. These results provide a key database for further studies on the desert cloud structure and the evolution characteristics of clouds over TD.
    为了揭示塔克拉玛干沙漠(TD)云的结构和演变特征,本研究首次利用地基毫米波雷达(MMCR)于2018年4–6月对TD腹地的云进行了连续探测。首先利用MMCR探测的功率谱数据,计算得到了TD云的反射率因子、云层边界和液水含量(LWC);然后与CPR(安装在Cloudsat的云雷达)探测的数据进行了对比验证。通过分析得到:TD的云主要是中云和高云,云体厚度一般小于2 km。通过分析得到:TD的云主要是中云和高云,云体厚度一般小于2 km。对中云和高云而言,平均LWC小于0.01 gm−3,主要由卷云和层云组成;而对低云而言,平均云体厚度为3166 m,并且在2.5–4.5 km 含有大量水滴,这表明TD的降水多产生于低云。TD云的平均LWC为0.0196 gm−3, 小于其他地区净云的平均LWC。由于地面的加热,在中午时段云的持续时间和LWC都比其他时段更高。当高云发展到低云时,平均时长为2 h,云中的最大LWC由0.008 gm−3增大到0.139 gm−3。这些结论对未来更深入地研究TD云的结构和演变提供了宝贵的数据基础。
  • 加载中
  • Fig. 1.  (a) The geographic position of Tazhong and (b) Ka-band millimeter cloud wave radar (MMCR).

    Fig. 2.  The identification of the cloud boundaries during the periods from (a) 1500 to 1700 BT on 22 May and (b) 1000 to 1700 BT on 27 June. The red and black asterisks represent the CTHs and CBHs derived from the MMCR, respectively. The black round circles denote the vertical position and thickness of clouds derived from the CPR at 1540 BT on 22 May and 1510 BT on 27 June.

    Fig. 3.  The empirical relationships between Z and LWC (or IWC) proposed by different researchers: (a) drizzle-free clouds and (b) clouds with drizzle.

    Fig. 4.  The variations of LWCs at different heights during the periods from (a) 1500 to 1700 BT on 22 May and (c) 1100 to 1700 on 27 June. Both the mean LWCs derived from the MMCR and CPR during the above two periods are shown (b) in panels and (d) on the right, respectively.(请作者确认b和d图的横纵左边名称及单位).

    Fig. 5.  The temporal variations of the vertical distribution of low, medium, and high clouds derived from the reflectivity factors in (a) April, (b) May, and (c) June.

    Fig. 6.  As in Fig. 5, but for the cloud thickness.

    Fig. 7.  As in Fig. 5, but for the average IWCs (retrieved by Zong) and LWCs.

    Fig. 8.  The proportion of the LWCs at different-order magnitudes for (a) all, (b) low, (c) medium, and (d) high clouds.

    Fig. 9.  The diurnal variation of the frequency of hourly occurrence of (a) all clouds and (b) low, medium, and high clouds.

    Fig. 10.  The diurnal variations of the monthly mean (a) Z and (b) LWC at different heights from April to June in 2018.

    Fig. 11.  The vertical distributions of the mean Zs for (a) low, (b) medium, and (c) high clouds. As in (a−c), but for (d−f) the mean LWCs.

    Fig. 12.  The proportions of the durations of (a) high, (b) medium, and (c) low clouds.

    Fig. 13.  The conceptual models for the evolution of (a) high to medium clouds and (b) medium to low clouds. The red and blue lines represent the temporal variation of the CTH and CBH, respectively.

    Fig. 14.  The vertical profiles of the monthly mean temperature over the TD in (a) April, (b) May, and (c) June 2018.

    Fig. 15.  The (a) visibility, (b) Z, and (c) average Z at different heights of the dust high clouds at 1400−1600 BT on 20 May. As in (a−c), but for (d−f) the clean high clouds at 0600−1000 BT on 26 June.

    Fig. 16.  As in Fig. 15, but for (a−c) the dust medium cloud at 2000−2300 BT on 20 May and (d−f) the clean medium cloud at 1300−1900 BT on 23 June.

    Table 1.  The parameters of the Ka-band millimeter cloud wave radar (MMCR)

    Parameter name Parameter value Parameter name Parameter value
    Wavelength 8.26 mm Pulse width 2560 μs
    Distance resolution 10 m Transmit power 10 w
    Center frequency 35 GHz FFT points 256
    Horizontal beam width $ {1.2}^{0} $ Feeder loss 1.2 dB
    Vertical beam width $ {1.2}^{0} $ Antenna gain 40 dB
    Download: Download as CSV

    Table 2.  Regression parameters of the Z-LWC/IWC

    Author Equation With drizzle or drizzle-free
    Matrosov et al., 2004 $\mathrm{L}\mathrm{W}\mathrm{C}=2.4{{Z} }^{0.5}$ Stratiform cloud (drizzle-free)
    Zhao et al., 2017 $\mathrm{L}\mathrm{W}\mathrm{C}=6.3{{Z} }^{0.5}$ Drizzle-free (the cloud height is < 1.5 km)
    Baedi et al., 2000 $\mathrm{L}\mathrm{W}\mathrm{C}=0.46{{Z} }^{0.19}$ With drizzle
    Krasnov and Russchenberg, 2005 $\mathrm{L}\mathrm{W}\mathrm{C}=0.026{{Z} }^{0.63}$ With drizzle
    Zong et al., 2013 $\mathrm{L}\mathrm{W}\mathrm{C}=0.09{{Z} }^{0.63}$ With drizzle
    $\mathrm{L}\mathrm{W}\mathrm{C}=0.34{{Z} }^{0.82}$ Drizzle-free
    ARM (Dunn et al., 2011) $ \mathrm{I}\mathrm{W}\mathrm{C}=0.097{Z}^{0.59} $ Ice cloud
    Download: Download as CSV

    Table 3.  The number and proportion of different cloud base heights from April to June 2018

    Month Low cloud Medium cloud High cloud
    Number Proportion (%) Number Proportion (%) Number Proportion (%)
    4 277 8.3 1599 48.1 1450 43.6
    5 81 2.5 1676 51.8 1477 45.7
    6 280 6.5 2111 49.0 1918 44.5
    Download: Download as CSV

    Table 4.  The distribution of different cloud thicknesses of the low, medium, and high clouds

    Type Average CBH Average cloud thickness Proportion of different thicknesses (%)
    $ \mathrm{H} < 1 $ $ 1\leqslant \mathrm{H} < 2 $ $ 2\leqslant \mathrm{H} < 4 $ $ \mathrm{H}\geqslant 4 $
    Low cloud 1416 3166 22.1 10.3 31.3 36.2
    Medium cloud 3704 1934 32.4 25.8 29.8 11.4
    High cloud 6396 1100 54.0 28.8 16.4 0.3
    All cloud / / 41.38 26.29 23.91 8.42
    Download: Download as CSV

    Table 5.  The mean LWCs in clouds over different locations in China

    Location Average
    LWC (g m−3)
    Author
    Taklimakan Desert (TD) 0.0196 This paper
    Tibetan Plateau 0.05 Li and De, 2001
    Hebei Province in China 0.04 You et al., 1994
    Shandong Province in China 0.06 Zhang et al., 2011
    Shanxi Province in China 0.03 Sun et al., 2014
    Download: Download as CSV
  • [1]

    Baedi, R. J. P., J. J. M. de Wit, H. W. J. Russechenberg, et al., 2000: Estimating effective radius and liquid water content from radar and lidar based on the CLARE98 data-set. Phys. Chem. Earth, Part B: Hydrol., Oceans Atmos., 25, 1057–1062. doi: 10.1016/s1464-1909(00)00152-0.
    [2]

    Chen, X., Q. Tang, S. C. Xie, et al., 2015: A variance-based decomposition and global sensitivity index method for uncertainty quantification: Application to retrieved ice cloud properties. J. Geophys. Res. Atmos., 120, 4234–4247. doi: 10.1002/2014jd022750.
    [3]

    Chin, H.-N. S., D. J. Rodriguez, R. T. Cederwall, et al., 2000: A microphysical retrieval scheme for continental low-level stratiform clouds: Impacts of the subadiabatic character on microphysical properties and radiation budgets. Mon. Wea. Rev., 128, 2511–2527. doi: 10.1175/1520-0493(2000)128<2511:amrsfc>2.0.co;2.
    [4]

    Choobari, O. A., P. Zawar-Reza, and A. Sturman, 2014: The global distribution of mineral dust and its impacts on the climate system: A review. Atmos. Res., 138, 152–165. doi: 10.1016/j.atmosres.2013.11.007.
    [5]

    Clothiaux, E. E., M. A. Miller, B. A. Albrecht, et al., 1995: An evaluation of a 94-Ghz radar for remote sensing of cloud properties. J. Atmos. Ocean. Technol., 12, 201–229. doi: 10.1175/1520-0426(1995)012<0201:aeoagr>2.0.co;2.
    [6]

    Dee, D. P., S. M. Uppala, A. J. Simmons, et al., 2011: The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597. doi: 10.1002/qj.828.
    [7]

    Dong, X. Q., and G. G. Mace, 2003: Profiles of low-level stratus cloud microphysics deduced from ground-based measurements. J. Atmos. Ocean. Technol., 20, 42–53. doi: 10.1175/1520-0426(2003)020<0042:pollsc>2.0.co;2.
    [8]

    Dunn, M., K. Johnson, and M. Jensen, 2011: The Microbase Value-Added Product: A Baseline Retrieval of Cloud Microphysical Properties. DOE/SC-ARM/TR-095, U.S. Department of Energy, doi: 10.2172/1015189. (查阅所有网上资料, 未找到对应的出版地和页码信息, 请联系作者确认).
    [9]

    Fu, P. J., J. P. Huang, C. W. Li, et al., 2008: The properties of dust aerosol and reducing tendency of the dust storms in northwest China. Atmos. Environ., 42, 5896–5904. doi: 10.1016/j.atmosenv.2008.03.041.
    [10]

    Ge, J. M., J. P. Huang, C. P. Xu, et al., 2014: Characteristics of Taklimakan dust emission and distribution: A satellite and reanalysis field perspective. J. Geophys. Res. Atmos., 119, 11,772–11,783. doi: 10.1002/2014jd022280.
    [11]

    Ge, J. M., H. Y. Liu, J. P. Huang, et al., 2016: Taklimakan Desert nocturnal low-level jet: climatology and dust activity. Atmos. Chem. Phys., 16, 7773–7783. doi: 10.5194/acp-16-7773-2016.
    [12]

    Guo, J. P., M. Y. Lou, Y. C. Miao, et al., 2017: Trans-Pacific transport of dust aerosols from East Asia: Insights gained from multiple observations and modeling. Environ. Pollut., 230, 1030–1039. doi: 10.1016/j.envpol.2017.07.062.
    [13]

    Huang, J., Q. Fu, J. Su, et al., 2009: Taklimakan dust aerosol radiative heating derived from CALIPSO observations using the Fu-Liou radiation model with CERES constraints. Atmos. Chem. Phys., 9, 4011–4021. doi: 10.5194/acp-9-4011-2009.
    [14]

    Huang, J. P., T. H. Wang, W. C. Wang, et al., 2014: Climate effects of dust aerosols over East Asian arid and semiarid regions. J. Geophys. Res. Atmos., 119, 11,398–11,416. doi: 10.1002/2014jd021796.
    [15]

    Illingworth, A. J., R. J. Hogan, E. J. O’Connor, et al., 2007: Cloudnet: Continuous evaluation of cloud profiles in seven operational models using ground-based observations. Bull. Amer. Meteor. Soc., 88, 883–898. doi: 10.1175/bams-88-6-883.
    [16]

    Kogan, Z. N., D. B. Mechem, and Y. L. Kogan, 2005: Assessment of variability in continental low stratiform clouds based on observations of radar reflectivity. J. Geophys. Res. Atmos., 110, D18205. doi: 10.1029/2005jd006158.
    [17]

    Kollias, P., E. E. Clothiaux, M. A. Miller, et al., 2007: Millimeter-wavelength radars: New frontier in atmospheric cloud and precipitation research. Bull. Amer. Meteor. Soc., 88, 1608–1624. doi: 10.1175/bams-88-10-1608.
    [18]

    Krasnov, O. A., and H. W. J. Russchenberg, 2005: A synergetic radar-lidar technique for the LWC retrieval in water clouds: Description and application to the Cloudnet data. Proceedings of the 32nd Conference of Radar Meteorology, Albuquerque, NM, American Meteorology Society, Albuquerque, New Mexico, 1–13. (查阅所有网上资料, 未找到对应的出版信息, 请联系作者确认).
    [19]

    Li, L. G., and L. De, 2001: Analyses of microphysical features for spring precipitation cloud layers in east of Qinghai. Plateau Meteor., 20, 191–196. doi: 10.3321/j.issn:1000-0534.2001.02.013. (in Chinese)
    [20]

    Lin, Y., Q. A. Wang, S. S. Gu, et al., 1995: Atmospheric Sounding Course. China Meteorological Press, Beijing, 20–55. (查阅所有网上资料, 未找到本条文献信息, 请联系作者确认). (in Chinese)
    [21]

    Ling, X. L., W. D. Guo, and C. B. Fu, 2011: Composite analysis of impacts of dust aerosols on surface atmospheric variables and energy budgets in a semiarid region of China. J. Geophys. Res. Atmos., 119, 3107–3123. doi: 10.1002/2013jd020274.
    [22]

    Liu, L. X., X. Huang, A. J. Ding, et al., 2016: Dust-induced radiative feedbacks in north China: A dust storm episode modeling study using WRF-Chem. Atmos. Environ., 129, 43–54. doi: 10.1016/j.atmosenv.2016.01.019.
    [23]

    Mace, G. G., E. E. Clothiaux, and T. P. Ackerman, 2001: The composite characteristics of cirrus clouds: Bulk properties revealed by one year of continuous cloud radar data. J. Climate., 14, 2185–2203. doi: 10.1175/1520-0442(2001)014<2185:tccocc>2.0.co;2.
    [24]

    Matrosov, S. Y., T. Uttal, and D. A. Hazen, 2004: Evaluation of radar reflectivity-based estimates of water content in stratiform marine clouds. J. Appl. Meteor. Climatol., 43, 405–419. doi: 10.1175/1520-0450(2004)043<0405:eorreo>2.0.co;2.
    [25]

    Mehta, M., N. Singh, and Anshumali, 2018: Global trends of columnar and vertically distributed properties of aerosols with emphasis on dust, polluted dust and smoke-inferences from 10-year long CALIOP observations. Rem. Sens. Environ., 208, 120–132. doi: 10.1016/j.rse.2018.02.017.
    [26]

    Moran, K. P., B. E. Martner, M. J. Post, et al., 1998: An unattended cloud-profiling radar for use in climate research. Bull. Amer. Meteor. Soc., 79, 443–456. doi: 10.1175/1520-0477(1998)079<0443:aucprf>2.0.co;2.
    [27]

    Pan, H. L., M. Z. Wang, K. R. Kumar, et al., 2019: Seasonal and vertical distributions of aerosol type extinction coefficients with an emphasis on the impact of dust aerosol on the microphysical properties of cirrus over the Taklimakan Desert in Northwest China. Atmos. Environ., 203, 216–227. doi: 10.1016/j.atmosenv.2019.02.004.
    [28]

    Pan, H. L., W. Huo, M. Z. Wang, et al., 2020: Insight into the climatology of different sand-dust aerosol types over the Taklimakan Desert based on the observations from radiosonde and A-train satellites. Atmos. Environ., 238, 117705. doi: 10.1016/j.atmosenv.2020.117705.
    [29]

    Protat, A., D. Bouniol, J. Delanoe, et al., 2009: Assessment of cloudsat reflectivity measurements and ice cloud properties using ground-based and airborne cloud radar observations. J. Atmos. Ocean. Technol., 26, 1717–1741. doi: 10.1175/2009jtecha1246.1.
    [30]

    Sassen, K., and Z. E. Wang, 2008: Classifying clouds around the globe with the CloudSat radar: 1-year of results. Geophys. Res. Lett., 35, L04805. doi: 10.1029/2007gl032591.
    [31]

    Sauvageot, H., and J. Omar, 1987: Radar reflectivity of cumulus clouds. J. Atmos. Ocean. Technol., 4, 264–272. doi: 10.1175/1520-0426(1987)004<0264:rrocc>2.0.co;2.
    [32]

    Sheng, P. X., J. T. Mao, J. G. Li, et al., 2013: Atmospheric Physics. 2nd ed. Peking University Press, Beijing, 296–315. (查阅所有网上资料, 未找到对应的英文翻译, 请联系作者确认). (in Chinese)
    [33]

    Shikwambana, L., and V. Sivakumar, 2018: Global distribution of aerosol optical depth in 2015 using CALIPSO level 3 data. J. Atmos. Sol. -Terr. Phys., 173, 150–159. doi: 10.1016/j.jastp.2018.04.003.
    [34]

    Stephens, G. L., D. G. Vane, R. J. Boain, et al., 2002: The Cloudsat mission and the A-train: A new dimension of space-based observations of clouds and precipitation. Bull. Amer. Meteor. Soc., 83, 1771–1790. doi: 10.1175/bams-83-12-1771.
    [35]

    Su, J., J. P. Huang, Q. Fu, et al., 2008: Estimation of Asian dust aerosol effect on cloud radiation forcing using Fu-Liou radiative model and CERES measurements. Atmos. Chem. Phys., 8, 2763–2771. doi: 10.5194/acp-8-2763-2008.
    [36]

    Sun, H. P., P. R. Li, S. M. Yan, et al., 2014: Characteristics of cloud microphysical structure based on aircraft data in 2008–2010 in Shanxi Province. Meteor. Sci. Technol., 42, 682–689. doi: 10.3969/j.issn.1671-6345.2014.04.027. (in Chinese)
    [37]

    Twomey, S., 1977: The influence of pollution on the shortwave albedo of clouds. J. Atmos. Soc., 34, 1149–1152. doi: 10.1175/1520-0469(1977)034<1149:tiopot>2.0.co;2.
    [38]

    Wang, J. H., J. X. Ge, M. Wei, et al., 2016: Study of the relationship between IWC and Z for nonspherical ice particles at millimeter wavelength. J. Trop. Meteor., 22, 78–88. 本条文献在正文中未被引用 doi: 10.16555/j.1006-8775.2016.s1.008.
    [39]

    Wang, M. Z., and H. Ming, 2018: Study on radar detection of one stratiform cloud precipitation process in the central part of the Tianshan Mountains in China. Asia-Pac. J. Atmos. Sci., 54, 511–524. doi: 10.1007/s13143-018-0059-2.
    [40]

    Wang, M. Z., H. Ming, Z. Ruan, et al., 2018: Quantitative detection of mass concentration of sand-dust storms via wind-profiling radar and analysis of Z-M relationship. Theor. Appl. Climatol., 131, 927–935. doi: 10.1007/s00704-016-2012-6.
    [41]

    Wang, Z., Z. H. Wang, and X. Z. Cao, 2016: Consistency analysis for cloud vertical structure derived from millimeter cloud radar and radiosonde profiles(DOI链接打不开,请作者核查). Acta Meteor. Sinica., 74, 815–826. doi: 10.11676/qxxb2016.057. (in Chinese)
    [42]

    Wang, Z., Z. H. Wang, X. Z. Gao, et al., 2018: Comparison of cloud top heights derived from FY-2 meteorological satellites with heights derived from ground-based millimeter wavelength cloud radar. Atmos. Res., 199, 113–127. doi: 10.1016/j.atmosres.2017.09.009.
    [43]

    Xu, X. F., H. Wu, X. Y. Yang, et al., 2020: Distribution and transport characteristics of dust aerosol over Tibetan Plateau and Taklimakan Desert in China using MERRA-2 and CALIPSO data. Atmos. Environ., 237, 117670. doi: 10.1016/j.atmosenv.2020.117670.
    [44]

    You, J. Y., Y. Duan, and L. G. You, 1994: Study on Cloud Precipitation Physics and Artificial Precipitation Enhancement Technology. China Meteorological Press, Beijing, 83–88. (查阅所有网上资料, 未找到对应的英文翻译, 请联系作者确认). (in Chinese)
    [45]

    Yumimoto, K., K. Eguchi, I. Uno, et al., 2009: An elevated large-scale dust veil from the Taklimakan Desert: Intercontinental transport and three-dimensional structure as captured by CALIPSO and regional and global models. Atmos. Chem. Phys., 9, 8545–8558. doi: 10.5194/acp-9-8545-2009.
    [46]

    Zhang, D. G., X. L. Guo, D. L. Gong, et al., 2011: The observational results of the clouds microphysical structure based on the data obtained by 23 sorties between 1989 and 2008 in Shandong Province. Acta Meteor. Sinica., 69, 195–207. doi: 10.11676/qxxb2011.017. (in Chinese)
    [47]

    Zhang, P. C., B. Y. Du, and T. P. Dai, 2000: Radar Meteorology. China Meteorological Press, Beijing, 77–87. 本条文献在正文中未被引用(查阅所有网上资料, 未找到对应的英文翻译, 请联系作者确认). (in Chinese)
    [48]

    Zhang, Y., Q. Zhou, S. S. Lyu, et al., 2019: Elucidating cloud vertical structures based on three-year Ka-band cloud radar observations from Beijing, China. Atmos. Res., 222, 88–89. doi: 10.1016/j.atmosres.2019.02.007.
    [49]

    Zhao, C. F., S. C. Xie, X. Chen, et al., 2014: Quantifying uncertainties of cloud microphysical property retrievals with a perturbation method. J. Geophys. Res. Atmos., 119, 5375–5385. doi: 10.1002/2013jd021112.
    [50]

    Zhao, C. F., L. P. Liu, Q. Q. Wang, et al., 2017: MMCR-based characteristic properties of non-precipitating cloud liquid droplets at Naqu site over Tibetan Plateau in July 2014. Atmos. Res., 190, 68–76. doi: 10.1016/j.atmosres.2017.02.002.
    [51]

    Zhou, Q., Y. Zhang, B. Li, et al., 2019: Cloud-base and cloud-top heights determined from a ground-based cloud radar in Beijing, China. Atmos. Environ., 201, 381–390. doi: 10.1016/j.atmosenv.2019.01.012.
    [52]

    Zong, R., L. P. Liu, and Y. Yin, 2013: Relationship between cloud characteristics and radar reflectivity based on aircraft and cloud radar co-observations. Adv. Atmos. Sci., 30, 1275–1286. doi: 10.1007/s00376-013-2090-7.
  • 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

Study on Clouds in the Hinterland of the Taklimakan Desert Based on Consecutive MMCR Observations

    Corresponding author: Minzhong WANG, wangmz@idm.cn
  • 1. School of Electrical & Electronic Engineering, Shandong University of Technology, Zibo 255000
  • 2. Taklimakan Desert Meteorology Field Experiment Station of CMA, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002
  • 3. Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science and Technology, Nanjing 210044
  • 4. State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081
Funds: Supported by the National Natural Science Foundation of China (41775030) and National Science Foundation for Young Scientists of China (41805006)

Abstract: This study mainly focused on the structure and evolution of desert clouds, and was the first to conduct high-resolution consecutive detection of clouds over the hinterland of the Taklimakan Desert (TD) from April to June 2018 using the ground-based Ka-band Millimeter Wave Cloud Radar (MMCR). The reflectivity factor, cloud boundary, and liquid water content (LWC) were calculated based on the power spectrum data observed by the MMCR, which were verified by comparing the detection data of the cloud profile radar (CPR) on the CloudSat board. The results showed that: clouds over the TD were dominated by medium and high clouds, with their thicknesses generally being less than 2 km; moreover, the mean LWCs of the medium and high clouds were less than 0.01 g m−3, which implied that cirrus and stratiform clouds were predominant. However, for the low clouds, the average thickness was 3166 m and the drizzles were concentrated within 2.5 to 4.5 km, which indicated that precipitation would more likely occur in the low clouds. The mean LWC in the clouds over the TD was 0.0196 g m−3, which was less than that of clean clouds. Compared to the other periods, the average durations and LWCs in clouds increased significantly around noon owing to the obvious surface heating by sensible heat flux. The average duration for the evolution of high to low clouds was approximately 2 h, and the average maximum LWC increased from 0.008 to 0.139 g m−3. These results provide a key database for further studies on the desert cloud structure and the evolution characteristics of clouds over TD.

基于地基毫米波雷达连续探测数据对塔克拉玛干沙漠腹地云的研究

为了揭示塔克拉玛干沙漠(TD)云的结构和演变特征,本研究首次利用地基毫米波雷达(MMCR)于2018年4–6月对TD腹地的云进行了连续探测。首先利用MMCR探测的功率谱数据,计算得到了TD云的反射率因子、云层边界和液水含量(LWC);然后与CPR(安装在Cloudsat的云雷达)探测的数据进行了对比验证。通过分析得到:TD的云主要是中云和高云,云体厚度一般小于2 km。通过分析得到:TD的云主要是中云和高云,云体厚度一般小于2 km。对中云和高云而言,平均LWC小于0.01 gm−3,主要由卷云和层云组成;而对低云而言,平均云体厚度为3166 m,并且在2.5–4.5 km 含有大量水滴,这表明TD的降水多产生于低云。TD云的平均LWC为0.0196 gm−3, 小于其他地区净云的平均LWC。由于地面的加热,在中午时段云的持续时间和LWC都比其他时段更高。当高云发展到低云时,平均时长为2 h,云中的最大LWC由0.008 gm−3增大到0.139 gm−3。这些结论对未来更深入地研究TD云的结构和演变提供了宝贵的数据基础。
    • The Taklimakan Desert (TD) is the largest dust source region in Asia and is located in northwest China (Yumimoto et al., 2009; Shikwambana and Sivakumar, 2018), and contributes significantly to global dust emissions (Mehta et al., 2018). As the Qinghai−Tibet Plateau and its surrounding mountainous areas block the transport of ocean water vapor, the TD region is an extremely arid climate area with little rain (Wang et al., 2017 本条文献在文后文献中未体现; Pan et al., 2020). Although many studies have shown that the frequency of dust storms over the TD has decreased in recent years (Fu et al., 2008; Guo et al., 2017), blowing sand and dust storms still occur often from April to June (Choobari et al., 2014; Liu et al., 2016). Moreover, dust aerosols from the TD can be blown at a height above 5 km into the upper troposphere, which has an indirect influence on the microphysical properties of clouds (Su et al., 2008; Ge et al., 2014). Subsequently, clouds with dust aerosols have a much greater potential impact on climate change and radiation balance (Ling et al., 2011; Pan et al., 2019). Therefore, accurate monitoring of dust clouds over the TD is imperative for reducing the risk of dust storm disasters along with climate change research (Huang et al., 2014; Ge et al., 2016).

      In recent decades, satellite remote sensing instruments have been used for studying dust aerosols and clouds over the TD, and these studies have mainly focused on the interactions between aerosols and clouds (Huang et al., 2009; Pan et al., 2020; Xu et al., 2020). As the A-train satellites pass through the same region every 16 days, it is difficult to conduct an effective detection of the life cycles and real-time vertical structure characteristics of clouds over the TD by using the Cloudsat product. Thus, the structure and evolution characteristics of clouds over the TD remain elusive. This study addresses these problems by utilizing the millimeter-wave cloud radar (MMCR) to conduct consecutive detection of clouds over the TD.

      As the wavelengths of MMCRs are shorter than those of weather radars, MMCRs can resolve the vertical structures of clouds for their occurrences and microphysical properties (Clothiaux et al., 1995; Mace et al., 2001; Stephens et al., 2002; Kollias et al., 2007). MMCRs operated at the US Department of Energy’s Atmospheric Radiation Program (ARM) provide consecutive long-term observations of precipitation and weakly precipitating clouds (Moran et al., 1998). In Europe, the cloud net program has utilized ground-based remote sensing instruments (ceilometers and MMCRs) to study clouds for more than 15 yr (Ιllingworth et al., 2007; Protat et al., 2009). In China, MMCRs have been widely used to study the vertical structure and microphysical properties of clouds over the last 10 yr; cloud top heights (CTHs) and cloud base heights (CBHs) were obtained by MMCRs and verified by comparison with observed satellite data (Wang et al., 2018 本条文献在文后文献中未体现; Zhang et al., 2019; Zhou et al., 2019). The LWCs and droplet radii were retrieved from the reflectivity factors observed by MMCRs (Zhao et al., 2017).

      As blowing sand and dust storms over the TD mainly occur from April to June, the Ka-band continuous MMCR was utilized to detect the non-precipitating clouds in the hinterland of the TD from April to June 2018. First, the reflectivity factors were calculated using the power spectrum data observed by MMCR. Second, using the reflectivity factors, the CTHs, CBHs, and LWCs were calculated and analyzed. Third, the basic and daily characteristics of clouds over the TD were analyzed. Finally, the vertical and temporal characteristics of the three cloud types were analyzed.

    2.   Detection equipment and data
    • The detection is conducted at the Tazhong Experiment Station ($ {39}^\circ {00}^{'}\mathrm{N}, {83}^\circ {40}^{'}\mathrm{E} $; 1009.3 m above sea level), which is located at the center of the TD (Fig. 1a). The climate of this region is drought-prone with little precipitation. The sparse vegetation was distributed around the observational field, which is representative of the climatic features of the TD.

      Figure 1.  (a) The geographic position of Tazhong and (b) Ka-band millimeter cloud wave radar (MMCR).

      The main detection equipment used in this study is the ground-based Ka-band continuous MMCR made by Anhui Sichuang Electronics Company (Fig. 1b), and its main parameters are presented in Table 1. The radar can work nonstop for 24 h and generate a set of data every 3 min. The consecutive detection of clouds over the TD was conducted from April to June 2018 using MMCR.

      Parameter name Parameter value Parameter name Parameter value
      Wavelength 8.26 mm Pulse width 2560 μs
      Distance resolution 10 m Transmit power 10 w
      Center frequency 35 GHz FFT points 256
      Horizontal beam width $ {1.2}^{0} $ Feeder loss 1.2 dB
      Vertical beam width $ {1.2}^{0} $ Antenna gain 40 dB

      Table 1.  The parameters of the Ka-band millimeter cloud wave radar (MMCR)

    • This study focuses on non-precipitating clouds in the hinterland of the TD. From April to June 2018, the MMCR detects 10,869 times of non-precipitating clouds over the TD, with a total duration of 32,607 min (3326 times in April, with a duration of 9978 min; 3234 times in May, with a duration of 9702 min; 4309 times in June, with a duration of 12,927 min). As the MMCR is affected by noise and the formation height of the radar beam, observed data with a height below 300 m are unreliable. Wang Z. et al. (2016) analyzed the consistency of the MMCR with radiosonde vertical structure observations and concluded that the MMCR could accurately characterize the vertical structure of clouds below 10 km. Hence, this study mainly uses power spectra data ranging from heights of 300 m to 10 km to study the non-precipitating clouds over the TD. Additionally, we also used the 2B-CLDCLASS and 2B-CWC-RD in CloudSat cloud products (Sassen and Wang, 2008), the monthly means of daily means of 0.75° × 0.75° isobaric surface field data from ERA-Interim reanalysis data from April to June in 2018 (Dee et al., 2011), and the visibility data derived from the visibility meter at the Tazhong Station.

    3.   Methods
    • Using the radar meteorological equation, the reflectivity factor (Z) was calculated using the following formula (Eq. (1)).

      $$ Z=\frac{1024\times {\rm{ln}}2\times {\lambda }^{2}\times {R}^{2}\times L\times {P}_{r}}{{\pi }^{3}\times {P}_{t}\times {\rm{c}}\times \text{τ} \times {\rm G}^{2}\times \text{φ} \times \text{θ} \times {10}^{-0.2{\int }_{0}^{R}k\mathrm{d}R}\times {\left|\dfrac{{m}^{2}-1}{{m}^{2}+2}\right|}^{2}}, $$ (1)

      where $ \mathrm{\lambda } $ is the wavelength (8.26 mm); R is the target distance (km); L is the feeder loss (1.2 dB); $ {P}_{t} $ is the radar transmitting power (10 W); $ \mathrm{c} $ is the transmission speed of the electromagnetic wave (3×108 m s−1); τ is the pulse width (2560 µs); G is the antenna gain (40 dB); φ is the horizontal beam width (0.02 rad); θ is the vertical beam width (0.02 rad); and $ {\left|\dfrac{{m}^{2}-1}{{m}^{2}+1}\right|}^{2} $ is the square of the complex refraction index (0.928).

      $ {P}_{r} $ represents the echo power, which can be calculated using the signal-to-noise ratio (SNR). The equation is as follows:

      $$\hspace{-68pt} {P}_{r}={\rm{SNR}}\times {\rm N}_{\rm f}\times {\rm{K}} \times {T}_{0}\times {\rm{B}}, $$ (2)

      where K is the Boltzmann constant ($ 1.38\times {10}^{-23}{\rm{J/K}} $), B is the receiver bandwidth (4 MHz), $ {T}_{0} $ is the radar antenna temperature expressed by absolute temperature (290 K), and Nf is the noise factor (4.5 dB).

    • By comparing the CBH and the CTH results from the MMCR and other atmospheric sounding data (e.g., radiosonde and meteorological satellites), previous studies have verified that −40 dBZ is a reliable threshold for MMCR to determine the CBHs and CTHs. (Wang Z. et al., 2016, 2018). For the MMCR used in this study, the Zs caused by noise ranges from −55 to −40 dBZ; additionally, Z values less than −45 dBZ account for more than 95%, while only less than 5% of Z values are within the range of −45 to −40 dBZ. The noise can be effectively distinguished from clouds using a threshold of −40 dBZ. To eliminate the radar detection errors and ensure the accuracy of the boundary, the values of the three successive vertical bins are all greater than −40 dBZ in this method. CBHs and CTHs were determined by the following criteria.

      Firstly, the bins at a certain height $ \left(H_{0}\right) $ with a reflectivity factor ($ Z_{H_0 }$) larger than $ -40\;\mathrm{d}\mathrm{B}\mathrm{Z} $ are selected, and the formula can be written as follows.

      $$\hspace{-137pt} Z_{H_0}\geqslant -40, $$ (3)

      Secondly, if $ Z_{H_0}\geqslant -40, $ we determine whether the reflectivity factors ($ {Z_{H_{0}-10}} $, $ Z_{H_{{0}}-20} $, and $ Z_{H_{0}-30} $) of the heights, which are 10, 20, and 30 m (the vertical bin of the MMCR is 10 m) lower than the current height, respectively, are all less than $ -40\;\mathrm{d}\mathrm{B}\mathrm{Z} $ simultaneously; the formula can be expressed as:

      $$ Z_{H_{0}-10} \leqslant -40 \; \& \; {Z}_{H_{0}-20} \leqslant -40 \; \& \;{Z}_{{H}_{0}-30} \leqslant -40, $$ (4)

      Finally, when Eqs. (2), (3) are satisfied, we determine whether the reflectivity factors ($ {Z}_{{H}_{0}+10} $, $ Z_{{H}_{0}+20} $, and $ Z_{{H}_{0}+30} $) of the heights, which are 10, 20, and 30 m higher than the $ {H}_{0} $, respectively, are all simultaneously larger than $ -40\;\mathrm{d}\mathrm{B}\mathrm{Z} $; the formula can be expressed as:

      $$\hspace{-16pt} {Z}_{{H}_{0}+10}\geqslant -40 \;\&\;{Z}_{{H}_{0}+20}\geqslant -40\;\&\;{Z}_{{H}_{0}+30}\geqslant -40, $$ (5)

      When Eqs. (3)–(5) are satisfied simultaneously, the current height is regarded as the CBH.

      Similarly, we can obtain the CTH when the reflectivity factors at a certain height $ {\mathrm{H}}_{0} $ satisfy Eqs. (6)–(8).

      $$ {Z}_{{H}_{0}}\geqslant -40, $$ (6)
      $$ {Z}_{{H}_{0}-10}\geqslant -40\;\&\;{Z}_{{H}_{0}-20}\geqslant -40\;\&\;{Z}_{{H}_{0}-30}\geqslant -40, $$ (7)
      $$ {Z}_{{H}_{0}+10} \leqslant -40\;\&\;{Z}_{{H}_{0}+20} \leqslant -40\;\&\;{Z}_{{H}_{0}+30} \leqslant -40, $$ (8)

      Based on this method, the CBHs and CTHs of clouds with thicknesses and intervals larger than 30 m were determined effectively. Furthermore, combined with the precipitation data from in situ measurements from April to June, the precipitation cloud data were deleted. Finally, the CBHs and CTHs of non-precipitation clouds were obtained.

      At 1540 BT, on 22 May and 1510 BT, on 27 June, the Tazhong station was within the scan range of the 94 GHz Cloud Profile radar (CPR); thus, we could obtain the CBHs and CTHs from CPR at two instances (Sassen and Wang, 2008). The reflectivity factors detected by MMCR between 1500 and 1700 BT of 22 May and between 1000 and 1700 BT of 27 June are utilized to determine and compare the CBHs and CTHs from the CPR and MMCR, as shown in Fig. 2. The cloud heights detected by the CPR agree with those by MMCR (the height difference between the two methods is less than 0.3 km). This indicates that the cloud heights acquired in this study, with a reflectivity factor −40 dBZ as the threshold, is consistent with the true values.

      Figure 2.  The identification of the cloud boundaries during the periods from (a) 1500 to 1700 BT on 22 May and (b) 1000 to 1700 BT on 27 June. The red and black asterisks represent the CTHs and CBHs derived from the MMCR, respectively. The black round circles denote the vertical position and thickness of clouds derived from the CPR at 1540 BT on 22 May and 1510 BT on 27 June.

    • According to Rayleigh’s study on a cloud droplet, the reflectivity factor Z can be defined by the cloud droplet particle spectrum n(r) as follows:

      $$ {Z}={2}^{6}{\int }_{0}^{{\infty }}{n}\left({r}\right){{r}}^{6}{d}{r}, $$ (9)

      where r is the particle radius. The effective liquid water content (LWC) of clouds (Dong and Mace, 2003; Zhao et al., 2014; Chen et al., 2015) can also be defined by the cloud droplet particle spectrum, as shown in Eq. (10).

      $$\hspace{-100pt} \mathrm{L}\mathrm{W}\mathrm{C}={\int }_{0}^{\mathrm{\infty }}\frac{4}{3}\mathrm{\pi }{{r}}^{3}\mathrm{\rho }{n}\left({r}\right){\rm d}{r}, $$ (10)

      Based on Eqs. (9), (10), the relationship between LWC and Z (Matrosov et al., 2004; Zhao et al., 2017; Atmospheric Radiation Measurement (ARM), 2010 本条文献在文后文献中未体现) can be expressed as follows.

      $$\hspace{-152pt} \mathrm{L}\mathrm{W}\mathrm{C}={a}{{Z}}^{{b}}, $$ (11)

      According to previous studies on cloud reflectivity factors (Sauvageot and Omar, 1987; Chin et al., 2000; Kogan et al., 2005; Zong et al., 2013), the reflectivity factors of ice clouds are generally less than −15 dBZ, whereas those of the drizzle cloud are normally more than −15 dBZ. Table 2 presents the relationship between IWC/LWC and Z, as summarized in previous studies.

      Author Equation With drizzle or drizzle-free
      Matrosov et al., 2004 $\mathrm{L}\mathrm{W}\mathrm{C}=2.4{{Z} }^{0.5}$ Stratiform cloud (drizzle-free)
      Zhao et al., 2017 $\mathrm{L}\mathrm{W}\mathrm{C}=6.3{{Z} }^{0.5}$ Drizzle-free (the cloud height is < 1.5 km)
      Baedi et al., 2000 $\mathrm{L}\mathrm{W}\mathrm{C}=0.46{{Z} }^{0.19}$ With drizzle
      Krasnov and Russchenberg, 2005 $\mathrm{L}\mathrm{W}\mathrm{C}=0.026{{Z} }^{0.63}$ With drizzle
      Zong et al., 2013 $\mathrm{L}\mathrm{W}\mathrm{C}=0.09{{Z} }^{0.63}$ With drizzle
      $\mathrm{L}\mathrm{W}\mathrm{C}=0.34{{Z} }^{0.82}$ Drizzle-free
      ARM (Dunn et al., 2011) $ \mathrm{I}\mathrm{W}\mathrm{C}=0.097{Z}^{0.59} $ Ice cloud

      Table 2.  Regression parameters of the Z-LWC/IWC

      Considering −15 dBZ as the critical threshold of the reflectivity factor for drizzle-free clouds and clouds with drizzle, the equations (the relationships between LWC/IWC and Z) presented in Table 2 are shown in Fig. 3. Owing to the following three reasons in this paper, Eqs. (12), (13) given by Zong et al. are adopted to retrieve the LWCs of clouds.

      Figure 3.  The empirical relationships between Z and LWC (or IWC) proposed by different researchers: (a) drizzle-free clouds and (b) clouds with drizzle.

      1. In Zong’s study, the wavelength of MMCR is 8.3 mm and the frequency is 35 GHz, which is consistent with the parameters of our study.

      2. The corresponding line between Z and LWC summarized by Zong is shown as the red line in Fig. 3b. It demonstrates that when the reflectivity factor is −5 to $ ~ $25 dBZ, the LWC is 10−2 to 100 g m−3, which is consistent with our result of the stratiform cloud over the TD (Wang and Ming, 2018).

      3. The results of Zong’s study were verified using aircraft remote sensing data, and the study was conducted in several places in China, instead of only one location. Thus, the results of Zong’s study are universal to an extent.

      $$ \mathrm{L}\mathrm{W}\mathrm{C}=0.34{{Z}}^{0.82}\left(\mathrm{d}\mathrm{r}\mathrm{i}\mathrm{z}\mathrm{z}\mathrm{l}\mathrm{e}{\text{-}}\mathrm{f}\mathrm{r}\mathrm{e}\mathrm{e}\right), $$ (12)
      $$ \mathrm{L}\mathrm{W}\mathrm{C}=0.09{{Z}}^{0.63}\left(\mathrm{w}\mathrm{i}\mathrm{t}\mathrm{h}\;\mathrm{d}\mathrm{r}\mathrm{i}\mathrm{z}\mathrm{z}\mathrm{l}\mathrm{e}\right), $$ (13)

      For instance, we analyzed the retrieval IWCs from Z values during the two periods (from 1500 to 1730 BT on 22 May and 1000 to 1700 BT on 27 June), as shown in Fig. 4. Figures 4b and 4d show the vertical distributions of LWCs at 1540 BT on 22 May and 1510 BT on 27 June, as detected by the CPR and MMCR, respectively. The two values are of the same order of magnitude, and the distribution shapes of their LWCs are generally consistent with each other. Therefore, the LWCs retrieved using the method proposed by Zong et al. are reliable.

      Figure 4.  The variations of LWCs at different heights during the periods from (a) 1500 to 1700 BT on 22 May and (c) 1100 to 1700 on 27 June. Both the mean LWCs derived from the MMCR and CPR during the above two periods are shown (b) in panels and (d) on the right, respectively.(请作者确认b和d图的横纵左边名称及单位).

    4.   The characteristics of the clouds
    • According to the criteria released by Lin et al. (1995) and Sheng et al. (2013), clouds are divided into three types (low, medium, and high clouds). Clouds with CBH less than 2 km, ranging from 2 to 5 km, and higher than 5 km, were classified as low, medium, and high clouds, respectively. On this basis, we conducted a statistical analysis on these three types of clouds.

    • Figure 5 shows the temporal distribution of the low, medium, and high clouds in April, May, and June, calculated by the reflectivity factors. As shown in Fig. 5, compared to the period from 1 April to 10 June, the frequency of occurrence of clouds was significantly less than that from 11 June to 30 June. The daily average durations of clouds during the former and latter periods were 313 and 516 min, respectively. Owing to the increase in thermal convection and water vapor due to the subtropical westerly jet, clouds were more frequent over the TD in June than in April and May.

      Figure 5.  The temporal variations of the vertical distribution of low, medium, and high clouds derived from the reflectivity factors in (a) April, (b) May, and (c) June.

      The numbers and proportions of the low, medium, and high clouds during the three months are presented in Table 3. As shown in Fig. 5 and Table 3, the clouds over the TD are dominated by medium and high clouds, which account for 94%. We identified that the lack of water vapor results in a higher lifting condensation level, thereby reducing the possibility of the occurrence of low clouds over the TD. Compared to June, this characteristic becomes more apparent in April and May due to drier air.

      Month Low cloud Medium cloud High cloud
      Number Proportion (%) Number Proportion (%) Number Proportion (%)
      4 277 8.3 1599 48.1 1450 43.6
      5 81 2.5 1676 51.8 1477 45.7
      6 280 6.5 2111 49.0 1918 44.5

      Table 3.  The number and proportion of different cloud base heights from April to June 2018

    • Figure 6 shows the temporal distribution of the cloud thickness in April, May, and June, calculated by the reflectivity factors. The thicknesses were mainly below 2 km (67.67%), while those exceeding 4 km were very less (8.42%). The frequency of cloud occurrence decreased with increasing cloud thickness. The distributions of the different cloud thicknesses (H) of the low, medium, and high clouds are presented in Table 4. As shown in Fig. 6 and Table 4, regarding the low clouds, clouds with thicknesses greater than 2 and 4 km account for approximately 67.5% and 36.2%, respectively. In contrast, regarding the medium and high clouds, their proportion decreased with increasing cloud thickness. The thicknesses of the high clouds were less than 1 km in most cases, and only 0.3% of them had thicknesses larger than 4 km. Overall, the low clouds were thicker, implying that the cumuliform clouds accounted for a large proportion of the low clouds. However, the medium and high clouds were thinner, indicating that the medium and high clouds were mainly cirrus and stratiform clouds.

      Type Average CBH Average cloud thickness Proportion of different thicknesses (%)
      $ \mathrm{H} < 1 $ $ 1\leqslant \mathrm{H} < 2 $ $ 2\leqslant \mathrm{H} < 4 $ $ \mathrm{H}\geqslant 4 $
      Low cloud 1416 3166 22.1 10.3 31.3 36.2
      Medium cloud 3704 1934 32.4 25.8 29.8 11.4
      High cloud 6396 1100 54.0 28.8 16.4 0.3
      All cloud / / 41.38 26.29 23.91 8.42

      Table 4.  The distribution of different cloud thicknesses of the low, medium, and high clouds

      Figure 6.  As in Fig. 5, but for the cloud thickness.

    • As shown in Figs. 6, 7, the temporal distribution of the average LWCs is consistent with that of the cloud thicknesses, which implies that thicker clouds usually have larger LWCs. Moreover, the average LWCs range from 10−4 to 100 g m−3, with a maximum of less than 2 g m−3 and a minimum greater than 10−4 g m−3. Figure 8a shows that 92% of the LWCs of clouds are within the range of 10−4 to 10−1 g m−3, while LWCs greater than 100 g m−3 accounted for only 1.1%. The LWCs of low clouds were generally larger than 10−2 g m−3, and 35% of LWCs were larger than 10−1 g m−3 (Fig. 8b). However, the LWCs of medium and high clouds were generally less than 10−1 g m−3, which accounted for 91% and 100%, respectively. In short, the low clouds have more drizzles; however, the medium and high clouds are dominated by ice clouds.

      Figure 7.  As in Fig. 5, but for the average IWCs (retrieved by Zong) and LWCs.

      Figure 8.  The proportion of the LWCs at different-order magnitudes for (a) all, (b) low, (c) medium, and (d) high clouds.

      The statistical results presented in Table 5 indicate that the mean LWC for all clouds over TD is approximately 0.0196 g m −3, which is significantly lesser than that over other locations in China. In particular, the mean LWC of clouds over TD is less than that in clean clouds, which is consistent with the “Twomey effect” (Twomey, 1977), which states that more dust aerosols correspond to lesser LWCs. The lower mean LWC in clouds over the TD can also be attributed to the dominant high and medium clouds with less drizzles; the special geographical location of the TD also leads to lower LWCs in clouds (Pan et al., 2020).

      Location Average
      LWC (g m−3)
      Author
      Taklimakan Desert (TD) 0.0196 This paper
      Tibetan Plateau 0.05 Li and De, 2001
      Hebei Province in China 0.04 You et al., 1994
      Shandong Province in China 0.06 Zhang et al., 2011
      Shanxi Province in China 0.03 Sun et al., 2014

      Table 5.  The mean LWCs in clouds over different locations in China

    • To reveal the diurnal characteristics of the clouds over the TD, we categorized the 24-h clouds on an hourly basis, and the numbers of the low, medium, and high clouds are shown in Fig. 9. As shown in Fig. 9a, the frequency of hourly clouds shows an increasing trend in the morning (0300–1100 BT), which reaches the maximum value around noon (1100–1600 BT), and decreases gradually in the afternoon and early night (1600–0200 BT). Figure 9b demonstrates that, regarding hourly proportions, the low cloud accounts for less than 20%, while both the medium and high clouds account for more than 40%. From 0200 to 0800 BT, the proportion of high clouds is greater than that of the medium clouds, while the situation reverses during other periods. Moreover, the medium clouds have more obvious diurnal variations, which is consistent with the characteristics shown in Fig. 9a. This is mainly attributed to the diurnal variation of the local boundary layer and the development of the fair-weather cumulus clouds. However, the low clouds do not have obvious diurnal variations, which indicates that the low clouds over the TD are mainly related to the large-scale weather system rather than the diurnal variation of the boundary layer.

      Figure 9.  The diurnal variation of the frequency of hourly occurrence of (a) all clouds and (b) low, medium, and high clouds.

      The relationship between the total detection days (ND), the hourly average number of clouds ($ {\mathrm{N}\mathrm{C}}_{{h}} $), and the total number of clouds detected per hour (NC) can be expressed as follows.

      $$ {\mathrm{N}\mathrm{C}}_{{h}}=\mathrm{N}\mathrm{C}\cdot {\mathrm{N}\mathrm{D}}^{-1}, $$ (14)

      As the total detection days by MMCR are 91 (ND = 91) and the total number of clouds detected per hour is shown in Fig. 9a (NC = Fig. 9a), the hourly average numbers of clouds ($ {\mathrm{N}\mathrm{C}}_{{h}} $) can be calculated using Eq. (14). The $ {\mathrm{N}\mathrm{C}}_{{h}} $ ranges from 2.7 to 5.3, with an average duration of 8–16 min. The maximum $ {\mathrm{N}\mathrm{C}}_{{h}} $ (5.3) occurs between 1100 and 1300 BT, with a duration of 16 min; however, the minimum $ {\mathrm{N}\mathrm{C}}_{{h}} $ (2.7) occurred between 0200 and 0300 BT, with a duration of 8 min.

      Overall, as the obvious sensible heat flux changes the vertical thermal structure of the atmosphere and increases the vertical thermal instability, the number of clouds and average durations during 1100 to 1600 BT are greater than those during other periods. After 1600 BT, as the probability of bubbles within the boundary layer lifting the water vapor to the condensation height decreases gradually with time, the number of clouds and average durations also decrease with time.

    • Reflectivity factors can indirectly reflect different components within the cloud. The reflectivity factor increases with an increase in the amount or radius of the drizzle within the cloud (Zong et al., 2013). The diurnal variations in the monthly mean reflectivity factor (Z) and LWC at different heights from April to June are shown in Fig. 10. When 5 km ≤ H < 8 km, the mean Zs are less than −20 dBZ with average LWCs less than 0.0078 g m −3; the mean Zs and average LWCs decrease with increasing heights. When H < 5 km, the mean Zs and LWCs increased with time from to 0600–0900 BT. However, during 1100–1600 BT, both the mean Zs and LWCs reached their maximum values of approximately −5 dBZ and 0.045 g m−3, respectively. These results indicate that these clouds consist of a large amount of drizzle during this period; thus, precipitation tends to occur if strong convections occur within the clouds. Furthermore, from 1600 to 2200 BT, when the height is located between 3.5 and 5 km, the mean Zs rapidly drop to −12 dBZ and the average LWCs drop from 0.045 to 0.015 g m−3; the mean Zs and LWCs change slightly below 3 km. From 2200 to 0300 BT, the mean Zs are mainly less than −15 dBZ, while the LWCs less than 0.015 g m−3. Additionally, when the height ranges from 3 to 5 km, the mean Zs and LWCs are greater than those at other heights in this period. However, from 0300 to 0600 BT, both the mean Zs and LWCs decrease to a relatively low level with little vertical variation. This implies less drizzle in clouds and little possibility of precipitation during this period.

      Figure 10.  The diurnal variations of the monthly mean (a) Z and (b) LWC at different heights from April to June in 2018.

    5.   The vertical and temporal characteristics of the three types of clouds
    • The vertical distributions of the mean Zs and LWCs of the three types of clouds were calculated and are shown in Fig. 11. Considering the low clouds (Figs. 11a, d), the mean Zs range from −23 to −5 dBZ, with average LWCs ranging from 0.003 to 0.045 g m−3. When H < 2.5 km, the average Z and LWC increased with increasing height and reached a maximum value of −5 dBZ and 0.045 g m−3 at a height of 2.5 km, respectively. However, when 2.5 ≤ H < 4.2 km,, the mean Z and LWC maintained high values, which are basically the same for varying the height. Moreover, when 4.2 ≤ H < 5 km, the mean Z and LWC decreased with increasing height. Considering the medium clouds (Fig. 11b, e), when the height increases from 4 to 5.8 km, the mean Z decreases from −15 to −20 dBZ, while the mean LWC decreases from 0.01 to 0.0078 g m−3. Considering the high clouds (Fig. 11c, f), when the height increases from 6.2 to 7.3 km, the mean Z decreases from −28 to −32 dBZ, and the mean LWC decreases from 0.0017 to 0.0008 g m−3.

      Figure 11.  The vertical distributions of the mean Zs for (a) low, (b) medium, and (c) high clouds. As in (a−c), but for (d−f) the mean LWCs.

      In general, the vertical profiles of LWC in low clouds can be roughly described as a unimodal curve, and the drizzles are mainly concentrated at heights of 2.5 to 4.5 km. As the clouds over the TD are dominated by medium and high clouds, the LWCs are generally less than 0.01 g m−3, which implies that the clouds are dominated by ice clouds with little rain.

    • In this study, we propose a cloud process concept. The start and end times of the cloud process are defined as the times at which cloud and cloudless intervals are identified by the vertical profiles of Z from the MMCR, respectively. From April to June 2018, the number of cloud processes derived from the MMCR for low, medium, and high clouds were 112, 114, and 54, respectively. Figure 12 demonstrates that the durations of both the high and low clouds are usually less than 2 h (81.25% of the high clouds and 77.8% of the low clouds), while durations longer than 3 h account for a very small proportion (12.55% of the high clouds and 18.5% of the low clouds). However, considering the medium clouds, the percentage of the duration less than 2 h accounted for 57.89%, while that longer than 3 h reached 28.95%. Overall, the average duration of clouds over the TD was less than 2 h, which indicates that the life cycles of clouds over the TD are relatively short.

      Figure 12.  The proportions of the durations of (a) high, (b) medium, and (c) low clouds.

    • The transformations of clouds from lower to higher heights usually occur at the end of the life cycle of the clouds, and the sample sizes are not sufficient for scientific analysis. Thus, this paper focuses on the analysis of transformations from higher to lower clouds.

      Among the 112 high cloud processes observed by MMCR, high clouds turned into medium clouds (26.8%) 30 times. These instances of cloud development were used as the samples for the analysis. The statistical results showed that when the cloud base ranged from 6500 to 5000 m, the durations of the evolution processes were generally within 30 to 90 min (the proportion was 78%), with an average of 60 min. Moreover, the average cloud thickness increased gradually from 2100 to 3400 m, and the average maximum liquid water content ($ {\mathrm{L}\mathrm{W}\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{x}} $) increased from 0.008 to 0.028 g m−3. Assuming that the height of the cloud varies linearly with time, the conceptual variation model for the cloud base from 6500 to 5000 m is shown in Fig. 13a.

      Figure 13.  The conceptual models for the evolution of (a) high to medium clouds and (b) medium to low clouds. The red and blue lines represent the temporal variation of the CTH and CBH, respectively.

      Among the 114 medium cloud processes detected by the MMCR, 26 of those medium clouds developed into low clouds (22.81%). These 26 cloud development processes were considered as the samples for the analysis. When the cloud base ranges from 3500 to 2000 m, the durations of the evolution processes mainly range between 15 to 45 min (the proportion is 76%) with an average of 38 min. The average cloud thickness gradually increases from 4500 to 5800 m with time, while the average $ {\mathrm{L}\mathrm{W}\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{x}} $ increases from 0.038 to 0.139 g m−3 with time. Assuming that the heights of clouds vary linearly with time, the conceptual model for the medium to low clouds (cloud base from 3500 to 2000 m) is shown in Fig. 13b.

      Among the 54 low cloud processes observed by the MMCR, low cloud processes generated precipitation (70.3%) 38 times; this generally occurred as intermittent stratiform cloud precipitation. The duration of each precipitation usually ranged from 20 to 90 min (82%), with a cloud thickness of more than 6000 m.

    6.   Discussions
    • 1) It is almost impossible for the CPR to conduct effective detection of the life cycles and real-time evolution characteristics of clouds over the TD because of the long scanning cycle. Compared to the CPR, the ground-based continuous MMCR has higher vertical (10 m) and temporal (a set of data is generated every three minutes) resolutions, thereby enabling the real-time observations of clouds. This study is the first to conduct continuous cloud detection in the hinterland of the TD using the ground-based continuous MMCR. From the perspective of instruments and data, our research is innovative and appropriate for the investigation of clouds over the TD.

      2) According to Dunn’s (ARM) research, if the ambient temperature is lower than $ -16^\circ {\rm{C}} $, all of the Z can be attributed to backscattered radiation from ice water particles (Dunn et al., 2011). As shown in Fig. 14, the approximate heights of ambient temperature equal to -16°C in different months are estimated using ERA-Interim reanalysis data, which are 4720, 5280, and 6250 m in April, May, and June, respectively. Therefore, clouds with heights higher than these can be regarded as ice-phase clouds. For ice-phase clouds, there are some deviations in the retrieval of the ice water content (IWC) in this study. An accurate estimate of the uncertainty in the MMCR retrieval is desired; therefore, we recalculated the IWCs using Eq. (15) (Dunn et al., 2011).

      Figure 14.  The vertical profiles of the monthly mean temperature over the TD in (a) April, (b) May, and (c) June 2018.

      $$ \mathrm{I}\mathrm{W}\mathrm{C}=0.097{Z}^{0.59}, $$ (15)

      No significant differences were found when comparing the total content (including both LWCs and IWCs) results from Zong’s and Dunn’s method, as shown in Fig. 8. For medium and high clouds, the results from Dunn’s method are approximately 4% greater than those from Zong’s method in the range of 10−4 to 10−3 g m−3. For all three types of clouds, there are subtle differences between the interior hydrometeors (including LWCs and IWCs) of clouds derived from these two methods (0.0194 vs. 0.0196 g m−3). In our subsequent research, we intend to utilize other sounding data (e.g., aircraft data) to study the LWC within the clouds over the TD and constantly improve the Z-LWC relationships.

      3) Every year from April to June, blowing sands and dust storms occur frequently over the TD. According to the data from the Tazhong Weather Station in 2018, the frequency of occurrence of blowing dust and dust storms was 24 and 4 days from April to June, respectively. Hence, the dust clouds, which occurred in April and June, embody the paradigmatic features of the TD dust clouds.

      To discuss the difference between dust and clean clouds, we selected some high and medium cloud cases with obvious visibility differences for comparative analysis. As shown in Figs. 15, 16, the average Zs of clean clouds are obviously greater than those of dust clouds, which indicates that the average LWCs in clean clouds are obviously greater than those in dust clouds. These features are associated with small drizzles in dust aerosol clouds. Further observational data analysis is required to elucidate the interaction between dust aerosols and clouds over the TD.

      Figure 15.  The (a) visibility, (b) Z, and (c) average Z at different heights of the dust high clouds at 1400−1600 BT on 20 May. As in (a−c), but for (d−f) the clean high clouds at 0600−1000 BT on 26 June.

      Figure 16.  As in Fig. 15, but for (a−c) the dust medium cloud at 2000−2300 BT on 20 May and (d−f) the clean medium cloud at 1300−1900 BT on 23 June.

    7.   Conclusions
    • Based on the data observed by the MMCR, the reflectivity factors, cloud boundaries, and LWCs were calculated and analyzed. The conclusions are as follows.

      1) During the period from April to June, the average proportions of low, medium, and high clouds were 5.8%, 49.6%, and 44.6%, respectively. The cloud top height in the range of 4 to 8 km accounted for 72.1%, and the cloud thicknesses were generally below 2 km (67.7%).

      2) In terms of the daily detection, the numbers of the observed cloud per hour range from 2.7 to 5.3, with durations between 8 and 16 min. Moreover, 5.3 clouds were observed between 1200 and 1300 BT, with a duration of 16 min, while 2.7 clouds with a duration of 8 min were observed between 0200 and 0300 BT.

      3) In general, the vertical profiles of LWC in low clouds can be roughly described as a unimodal curve, and the drizzles are mainly concentrated at a height of 2.5 to 4.5 km.

      4) The mean LWC for all clouds over the TD is approximately 0.0196 g m−3, which is significantly less than that over other locations in China.

      5) The average transition duration from high (medium) to medium (low) clouds was approximately 1 h (38 min), and the $ {\mathrm{L}\mathrm{W}\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{x}} $ increased from 0.008 (0.038) to 0.028 (0.139) g m−3.

Reference (52)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return