双参数微物理方案的冰相过程模拟及冰核数浓度的影响试验1

利用耦合Morrison 2-mon(MOR)双参数微物理方案的中尺度天气研究与预报模式(WRF)中的单气柱模式,对热带暖池国际云试验(TWP-ICE)期间的个例进行数值模拟。通过与观测资料和云分辨率模式的模拟结果进行对比,检验MOR方案对热带对流云系的微物理特征的模拟能力。模拟结果显示:MOR方案能够较好地模拟出热带云系中液相和冰相水凝物的垂直分布以及随时间的演变特征。地表向下长波辐射和大气顶向外长波辐射的量级和时间演变趋势同观测也非常接近。对与冰晶和雪有关的云微物理特征分析之后发现:季风活跃期,冰晶主要的源汇项有凝华增长过程、沉降过程、冰晶向雪的自动转化以及冰晶被雪碰并的过程。由于冰晶主体位于温度低于―20°C的高空,因而它对雨水的形成主要是间接贡献。同时期雪的主要源汇项中,凝华增长和沉降过程占据着主导地位。雪的凝华过程消耗了大量的水汽,可能抑制了冰晶的增长。另外雪的融化过程非常强盛,是产生降水的重要因子。季风抑制期,冰相的微物理过程变得相对简单且整体削弱,以凝华升华和沉降过程为主。凝华凝冻核的数浓度(Ndep)的气溶胶敏感性试验表明:季风抑制期,高空的冰晶云的宏观和微观性质对凝华凝冻核数浓度的响应情况呈现显著的线性特征。冰晶的含量随着Ndep的增加而增加,反之降低。该时期微物理过程主要同冰晶有关,水分的分配较为简单,Ndep增加时,高空冰云中小冰晶粒子数目增多且云顶升高,使得大气顶部向外长波辐射(OLR)值减小,反之冰云主体中冰晶有效半径增加,高空的冰云更加透明,云顶更低,对OLR值增加起促进作用。而季风活跃期,微物理过程复杂,冰晶云的宏微观特征对Ndep的响应表现出一定的不规律特征。     

双参数微物理方案在WRF单柱模式中的模拟检验和对比研究

利用耦合Milbrandt 2-mon(MY)和Morrison 2-mon(MOR)两种双参数微物理方案的WRF中的单柱模式,对TWP-ICE(Tropical Warm Pool International Cloud Experiment)试验期间的个例进行数值模拟。通过与观测资料和云分辨率模式的模拟结果进行对比发现:两种双参数微物理方案能较好地模拟出TWP-ICE期间热带云系的宏观和微观的特征。模拟的降水率、地表向下长波辐射和大气顶向外长波辐射的量级、时间演变趋势与观测相一致;总的液相和冰相水凝物的垂直分布以及随时间的演变特征总体与观测以及云分辨率模式的结果也较接近。在整个时期,两种方案水云中的雨滴宏观和微观特征差异较小,而云滴混合比在两种方案之间的差别显著;冰晶对冰云的贡献在MY方案中占据主导地位,而MOR方案中雪在冰云中扮演的角色相比于在MY方案中更为重要。微观上与MY方案相比,MOR方案中的云滴是由数量更大的小云滴构成,但冰晶却是由数量较少的大冰晶粒子构成。微物理过程转换率的区别是造成两种方案冰云宏观分布特征差异的主要原因。与冰晶和雪有关的微物理过程转换分析表明:活跃期两种方案中与冰晶有关的主要微物理转换项有冰晶的凝华增长、冰晶向雪的自动转化、冰晶被雪碰并以及冰晶的沉降过程。而雪主要的转换项包含沉降和凝华过程等,其中MY方案中雪的主要转换项更为丰富。该时期两种方案冰晶和雪的主要微物理转换项的垂直分布以及量级特征的差异同冰云的宏观分布相一致。季风抑制期,两种方案中冰晶主要的源汇项包括凝华增长和沉降过程。MY方案中凝华凝冻核化也是主要的源汇项之一。抑制期MOR方案中高空的雪发展较强,参与的微物理过程较MY方案更为丰富,主要转换项比MY方案高出约一个量级。     

GRAPES双参数云微物理方案的改进和云降水个例模拟研究：GRAPES_SCM对热带对流云个例的模拟研究

应用全球-区域同化预报系统单柱模式(GRAPES_SCM),对热带暖池国际云试验(TWP-ICE)个例进行数值模拟。通过和实际观测资料进行对比,诊断并改进了Liuma云微物理方案对热带对流云微物理特征的模拟能力。结果显示在GRAPES_SCM框架下,Liuma原方案和WSM6(WRF single moment)方案均能呈现出TWP-ICE期间热带云系的发展特征,并能够明显区分试验期间的季风活跃期和季风抑制期。活跃期Liuma原始方案和WSM6方案模拟的冰云组成结构差异显著,在Liuma原始方案所模拟的冰相水凝物分布中,存在冰雪含量过少、霰过多的现象。改进后的Liuma方案对程序中各微物理过程计算顺序进行了优化,改进后霰质量混合比明显减少,冰雪质量混合比明显增加,冰相水凝物分布较合理。     

台风螺旋雨带云结构和降水形成机制研究

应用数值模式结果,选择台风登陆后两个不同时次螺旋雨带中两个强降水中心,对台风螺旋雨带的云结构和降水形成机制进行诊断分析.结果发现螺旋雨带云结构和降水形成机制有如下特点:在9～13 km高空范围内冰晶的非均质核化非常活跃,冰晶转化率高于台风眼壁暴雨数倍,但是冰晶通过贝吉龙过程生长为雪、雪通过凝华增长生长为霰的过程相对台风眼壁很弱,螺旋雨带雨水形成微物理机制以霰粒子融化成雨水(pgmlt)为主,冰相粒子转化率大值区位于垂直上升气流大值区,8 km高度霰收集雪(dgacs)干增长是最主要的冰相粒子生长过程,与北方层状云比较,螺旋雨带暴雨冷云中的凝华过程和撞冻过程非常活跃.螺旋雨带云水凝结过程呈双峰型,位于7～8 km高度冷云区的云水凝结峰值较大,暖云区0.5～1.5 km高度云水凝结峰值次之.     

WSM6云微物理方案对华北地区一次降雪预报偏强的原因分析

国家气象中心GRAPES区域业务模式对2019年11月29—30日在华北地区降雪过程的预报出现显著高估现象,针对该模式中采用的WSM6云微物理方案进行了深入分析,并与Liu-Ma云微物理方案以及ERA5再分析数据进行比较,探究其可能存在的原因。主要结论如下:冰晶和雪的沉降是WSM6方案在本次地面降雪形成的最主要贡献,Liu-Ma方案则是以大粒子雪和霰的沉降为主,冰晶产生的贡献较少。WSM6方案严重低估了大气中的液态水含量,冰相粒子构成中以冰晶含量为最多,雪含量次之,这些特征都与ERA5资料和Liu-Ma方案有显著的不同,后两者具有较好的一致性。与Liu-Ma方案相比,WSM6方案在模式低层冰晶含量更高、冰晶平均落速更大,二者共同作用使冰晶沉降在本次降水形成中具有重要贡献;WSM6方案中雪的平均落速大于Liu-Ma方案,这是其雪的柱积分总量小而雪的沉降降水多于Liu-Ma方案的直接原因。在WSM6方案中冰晶的凝华/升华过程在冰相微物理过程中占据主导地位,致使雪和霰的凝华过程以及云水凝结过程都明显不足,这是该方案冰晶偏多、雪偏少、液水明显偏少的主要原因。针对冰晶凝华/升华过程(SVI)的敏感性试验发现,SVI转化率与地面降水呈正相关关系、与液水柱积分总量呈"跷跷板"关系,当降低SVI的转化率,地面降雪将显著减少,而柱积分液水总量则会明显增多。     

北京一次突发性降雪的云场结构特征分析

本文对利用三重嵌套的高分辨率中尺度模式对北京一次突发性降雪过程进行了数值模拟。重点分析伴随降雪过程的云各微物理含量构成和云场三维图像的演变特征,初步分析表明：此次降雪过程中云中的主要成份是冰晶粒子,雪其它粒子的含量较少,冰晶粒子含量随时间呈由高空向低空增加的趋势,它在降雪过程中起重要作用;云的三维结构图也清楚地反映云顶有不断降低和向下伸展过程,这与冰晶粒子的增长和沉降有关。     

区域冷锋降水微物理机制研究

20世纪利用一维层状云模式对2002年4月4~5日河南省冷锋降水过程进行了模拟。数值模拟结果显示,此次冷锋降水属于冷云降水过程,冷锋前后云中主要以冰相粒子为主,云中水质粒自上而下的空间分布依次为冰晶、雪、云水、霰、雨水。冷锋前后,各种水质粒有着不同的含量及数密度,但形成水质粒的主要微物理过程都表现为:冰晶数密度的增加主要依靠核化、繁生,大部分雪主要靠凝华、撞冻过冷云水和冰晶增长,霰的质量增加主要靠撞冻雪、过冷云水和雪自动转化而来,大部分的雨水是由霰融化而来,因而此次冷锋降水机制表现为“水汽—雪—霰—雨水”。     

SACOL站冰云粒子下降末速度的反演及其时空分布特征研究

冰云是影响气候变化最为重要的因子之一,其生命周期的变化在很大程度上决定了冰云的气候辐射效应。冰云粒子下降末速度是影响冰云生命周期的关键参数。为了开展对冰云粒子下降末速度的研究,利用兰州大学半干旱气候与环境监测站Ka波段毫米波云雷达2013年8月至2015年7月连续观测数据,反演了冰云粒子的下降末速度（V_t）,并根据雷达反射率因子（Z）与V_t的关系计算了拟合因子a、b的值;在此基础上应用聚类分析方法,对比分析了4种不同特性冰云Z、V_t和拟合因子a、b的时、空分布特征,进而尝试通过参数垂直分布特征识别研究云中不同位置上云微物理过程的变化。结果表明:冰云粒子下降末速度的分布与雷达反射率因子有很好的对应,最大频率都出现在距离地面约7 km高度处,且具有显著的季节变化,粒子下降末速度在暖季较冷季可增大25%,峰值出现在6月和9月;云层较厚且持续时间长的第一、三类冰云,其雷达反射率因子、粒子下降末速度及拟合因子a和b的平均值都显著大于云层较薄且持续时间短的第二、四类云。垂直方向上,Z、V_t和拟合因子b从云顶到云底随着高度的降低呈现先增大后减小的趋势,体现了云粒子在云顶区域成核和水汽凝华效应,随着粒子在下落过程中碰并增长,云滴粒子逐渐增大,水汽的凝华和粒子的聚合起主要作用,最后在云底部分,云粒子蒸发、升华减小消亡的过程。由此表明中纬度干旱半干旱地区冰云是从云顶到云底自上而下的形成过程。      

北京地区一次降雪过程的人工催化数值模拟研究

利用中尺度气象模式WRF的双参数显示云物理方案,开展冬季冷性层状云降水过程的数值模拟和人工增雨催化数值试验。模拟个例为2013年3月19日北京地区的一次典型降水过程,在分析模拟得到的云中水成物和上升速度分布的基础上设计不同催化试验,研究不同催化时刻(云体发展期、云体成熟期)和三种催化剂量对地面降水、云中水成物浓度、动力场和热力场以及微物理转化过程的影响。模拟试验结果表明:模拟的自然降水分布和实测结果较为一致;不同的催化试验都可以使地面雨量增加,在云体发展期以107个·kg-1剂量进行催化的效果最佳;引入人工冰晶后催化区域水汽和过冷云水含量明显减少、冰晶和雪的含量有所增加、催化区域上升气流明显增强,温度提高;催化后40 min时雪的增长主要依靠其凝华增长、冰晶向雪的自动转化、雪和云滴之间的碰冻以及冰晶和雪之间的碰并;催化后200 min,催化云中各种微物理过程对雪的贡献高于自然云,催化前期消耗了过冷云水,此时云中雪和云滴之间的碰冻对雪的贡献非常微弱,雪的增长主要依靠凝华增长以及雪和冰晶的相互作用。     

WRF单柱模式中单参数方法对热带对流模拟能力的影响

利用耦合Milbrandt 2-mon(MY)双参数微物理方案的WRF中的单柱模式,对TWP-ICE试验(Tropical Warm Pool International Cloud Experiment)期间的个例进行数值模拟和敏感性试验。通过与观测资料和云分辨率模式的模拟结果进行对比发现:MY方案默认的双参数版本和单参数版本均能够再现TWP-ICE期间的热带云系的总体宏观和微观特征。MY方案的双参数版本模拟的降水率的演变特征同观测十分吻合,冰相粒子的微观特征同观测事实较为一致。单参数默认版本的降水率、液态云的构成及冰相粒子微观特征方面同观测事实存在明显差距。然而实际业务应用中单参数方案由于计算量较小应用更为广泛,但模拟效果有待改善。为了使方案保持计算量较为合理的同时具有较好的模拟效果,参考双参数控制试验中的冰相物质的微观特征,尝试对单参数方案中冰相粒子的单参数方法进行改进。冰晶单参数改进试验中虽然对于冰晶数浓度采用两种不同的处理方法,但模拟效果均未明显改善。其中冰云总含量更加接近观测,且冰云构成发生显著变化,主要归因于冰晶有效半径的减小间接削弱了雪和霰的发展。云滴含量的异常增强导致液态水含量比观测偏高约一个量级,暖云异常增厚则与上升运动的增强直接相关。雨水含量明显增强及雨滴有效半径减小综合导致了降水率仅有微弱改善。雪的单参数改进试验中,雪的截距值增加及环境场过饱和条件改善促进了冰云的发展。通过适当调整雪的截距的经验诊断公式,雪的截距、液态水含量以及降水率均得到较好的改善;而指定雪截距为常数的处理方式使液态云更为偏厚,降水率演变细节同观测仍然差异显著。改进试验结果表明,单参数方案中采用适当的经验公式诊断雪的截距的处理方法对改善单参数方案的模拟能力具有一定的可行性。     

Numerical comparison study of cloud microphysical parameterization schemes for a moderate snowfall event in North China

Summary A moderate snowfall event in North China is simulated using the high-resolution mesoscale model MM5. A fourfold-nest experiment, with a minimum horizontal grid size of 2 km, is run. In order to study the cloud microphysics processes associated with the snowfall, two experiments were conducted in two inner domains, one using the Goddard scheme (Goddard experiment), and the other using the Reisner scheme (Reisner experiment). The analysis focused on the comparison of the cloud microphysics processes which occurred in the experiments. It is shown that there is no implicit precipitation of cumulus parameterization in the domain of grid scale 18 km. The snowfall distribution patterns in the experiments are slightly different, but the microphysical characteristics and processes may have considerable differences between the two experiments: (1) The water substances in the cloud have cloud water, cloud ice and snow, but no rainwater and graupel in the Goddard experiment. However, the water substances in the cloud have cloud ice, snow, and graupel, but no cloud water and rainwater in the Reisner experiment. (2) The cloud ice mixing ratios in the Goddard experiment are larger than those in the Reisner experiment. (3) In the Goddard experiment, the dominant cloud microphysical processes include the growth of cloud water by the condensation of supersaturated vapor, the depositional growth of cloud ice, the initiation of cloud ice, the accretion of cloud ice by snow, the accretion of cloud water by snow, the deposition growth of snow and the Bergeron process of cloud ice. In the Reisner experiment, the dominant cloud microphysical processes include the depositional growth of cloud ice, the conversion of cloud ice to snow, the deposition of snow, and the deposition growth of graupel. (4) There is only snowfall in the Goddard experiment. Meanwhile, there is ice fall, snow fall, and graupel in the Reisner experiment. But the ice fall and graupel in the Reisner experiment is very slight and can be ignored.     

The Cloud Processes of a Simulated Moderate Snowfall Event in North China

The understanding of the cloud processes of snowfall is essential to the artificial enhancement of snow and the numerical simulation of snowfall. The mesoscale model MM5 is used to simulate a moderate snowfall event in North China that occurred during 20–21 December 2002. Thirteen experiments are performed to test the sensitivity of the simulation to the cloud physics with different cumulus parameterization schemes and different options for the Goddard cloud microphysics parameterization schemes. It is shown that the cumulus parameterization scheme has little to do with the simulation result. The results also show that there are only four classes of water substances, namely the cloud water, cloud ice, snow, and vapor, in the simulation of the moderate snowfall event. The analysis of the cloud microphysics budgets in the explicit experiment shows that the condensation of supersaturated vapor, the depositional growth of cloud ice, the initiation of cloud ice, the accretion of cloud ice by snow, the accretion of cloud water by snow, the deposition growth of snow, and the Bergeron process of cloud ice are the dominant cloud microphysical processes in the simulation. The accretion of cloud water by snow and the deposition growth of the snow are equally important in the development of the snow.     

两类双参数云微物理方案对夏季强降水事件模拟能力的对比研究

运用中尺度WRF模式,分别采用Morrison(MOR)和Milbrandt-Yau(MY)双参数化云微物理方案,对2010年7月20—21日辽宁省的一次强降水过程进行模拟,通过对比分析两个方案所对应的地表累积降水量、降水强度、云中微物理量的模拟结果,评估两个双参数方案对强降水事件的模拟能力及主要微物理过程的差异。结果表明,在对雨带和强降水中心的位置上,MOR方案的模拟能力优于MY方案,但MY方案对强降水中心强度模拟能力则优于MOR方案;两方案对强降水宏观特征的模拟差异在一定程度上体现了它们在微物理具体方案上的差异,相比MY方案而言,MOR方案模拟降水发展期的垂直水汽通量高,使得雪晶的凝华增长、碰连增长增强,从而导致MOR方案的冰晶含量低,雪晶含量高,通过雪晶的凇附作用形成的霰含量也比MY方案高,霰的凇附增长消耗了大量过冷水,使冷云中云滴(过冷水)含量减少; MOR方案模拟得到的600 hPa到地表的雨滴直径均为1 mm,与实际雨滴直径的观测值不符,需要未来进一步开展研究,对原方案进行优化。     

云微物理过程对台风数值模拟的影响

将中国气象科学研究院（CAMS）混合双参数云微物理方案用于中尺度天气模式WRF,开展了对2013年超强台风天兔（1319）的模拟,通过与台风最佳路径、强度及热带降雨测量卫星（TRMM）资料对比,分析CAMS云微物理方案在模拟台风中的适用性及云微物理过程对模拟台风天兔的影响机制。设计了3组敏感性试验:修改雪粒子质量和落速系数（EXP1）,采用海洋性云滴参数（EXP2）,同时修改雪粒子质量和落速系数并采用海洋性云滴参数（EXP3）。结果表明:EXP1和EXP3由于霰碰并雪速率的增加及减小的雪下落通量,导致雪含量显著降低,同时也减少了整体冰相物的含量;EXP2和EXP3模拟的台风眼区对流有效位能快速减小,再现了前期台风的快速增强过程,路径偏差也最小;各试验模拟的小时降水率总体偏强,EXP3的降水空间分布与实况更接近,明显降低雪粒子含量,并一定程度上改善模拟的台风路径、强度及降水分布等。该结果不但可为改进适用于台风的云微物理参数化方案提供思路,也可加深云微物理过程对台风影响的认识。     

2014年南京青奥会开幕式日降水过程数值模拟研究

为评估2014年南京青奥会开幕式日的人工催化消减雨作业效果,利用中尺度数值模式WRF对当日的云降水过程和催化作业开展数值模拟。本文系第一部分工作。首先对常用的八种云微物理方案的降水模拟效果进行评估,进一步选取Thompson和Milbrandt-Yau两个微物理方案对此次降水过程的云系结构和降水形成机制进行对比分析。模拟结果表明,采用Thompson和Milbrandt-Yau两个方案模拟的云系结构和降水形成的微物理机制是一致的。开幕式当天影响奥体场馆的降水由弱的积层混合云系产生,降水过程以冰相微物理过程为主。雪的融化是雨水的主要源项,Thompson方案中雪的融化对雨水的贡献率为72%,Milbrandt-Yau方案为60%,蒸发则是雨水的主要汇项,Thompson方案中蒸发对雨水的消耗率达94%,Milbrandt-Yau方案为95.6%。     

Evaluation of cloud and precipitation parameterization using a single-column model: A TWP-ICE case study

Cloud and precipitation parameterization schemes are evaluated, and their sensitivity to the method and/or parameters used to determine cloud physical processes is examined using a singlecolumn version of the Unified Model (SCUM). In the experiment for TWP-ICE, cloud fraction is overestimated (underestimated) in the upper (lower) troposphere due to the wet (dry) bias. The precipitation rate is well simulated during the active monsoon period, but overestimated during the suppressed monsoon and clear skies periods. In the moist convection scheme, trigger condition and entrainment process affect the lower tropospheric humidity through the impact on convective occurrence frequency and intensity, respectively. Strengthening the trigger condition and using the adaptive entrainment method alleviate the low-level dry bias. In the microphysics scheme, more large-scale precipitation is produced with prognostic rain, due to rain sedimentation considering vertical velocity of rain drop, than with diagnostic rain. Less ice/snow deposition with the prognostic two-ice category results in lower ice water content and upper-level cloud fraction than with the diagnostic splitting method for the twoice category. In the cloud macrophysics scheme, the prognostic cloud fraction and cloud/ice water content scheme produces a larger cloud fraction and more cloud/ice water content than the diagnostic scheme, mainly due to detrainment from moist convection (cloud source) that surpasses the effect of convective heating and drying (cloud sink). This affects temperature by influencing the radiative, convective, and microphysical processes. The experiment with combined modifications in cloud and precipitation schemes shows that interaction between modified moist convection and cloud macrophysics schemes results in more alleviation of the cold bias not only at the lower levels but also at the upper levels.     

Numerical Simulation of Microphysics in Meso-β-Scale Convective Cloud System Associated with a Mesoscale Convective Complex

Numerical simulation of meso-β-scale convective cloud systems associated with a PRE-STORM MCC case has been carried out using a 2-D version of the CSU Regional Atmospheric Modeling System (RAMS) nonhydrostatic model with parameterized microphysics. It is found that the predicted meso-γ-scale convective phenomena are basically unsteady under the situation of strong shear at low-levels, white the meso-β-scale convective system is maintained up to 3 hours or more. The meso-β-scale cloud system exhibits characteristics of a multi-celled convective storm in which the meso-γ-scale convective cells have lifetime of about 30 min. Pressure perturbation depicts a meso-low after a half hour in the low levels. As the cloud system evolves, the meso-low inten-sifies and extends to the upshear side and covers the entire domain in the mid-lower levels with the peak values of 5-8 hPa. Temperature perturbation depicts a warm region in the middle levels through the entire simulation period. The meso-γ-scale warm cores with peak values of 4-8oC are associated with strong convective cells. The cloud top evapo-ration causes a stronger cold layer around the cloud top levels.Simulation of microphysics exhibits that graupel is primarily concentrated in the strong convective cells forming the main source of convective rainfall after one hour of simulation time. Aggregates are mainly located in the stratiform region and decaying convective cells which produce the stratiform rainfall. Riming of the ice crystals is the predominant precipitation formation mechanism in the convection region, whereas aggregation of ice crystals is the predominant one in the stratiform region, which is consistent with observations. Sensitivity experiments of ice-phase microphysical processes show that the microphysical structures of the convective cloud system can be simulated better with the diagnosed aggregation collection efficiencies.     

Tropical convective responses to microphysical and radiative processes: a sensitivity study with a 2-D cloud resolving model

Summary Prognostic cloud schemes are increasingly used in weather and climate models in order to better treat cloud-radiation processes. Simplifications are often made in such schemes for computational efficiency, like the scheme being used in the National Centers for Environment Prediction models that excludes some microphysical processes and precipitation-radiation interaction. In this study, sensitivity tests with a 2-D cloud resolving model are carried out to examine effects of the excluded microphysical processes and precipitation-radiation interaction on tropical thermodynamics and cloud properties. The model is integrated for 10 days with the imposed vertical velocity derived from the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment. The experiment excluding the depositional growth of snow from cloud ice shows anomalous growth of cloud ice and more than 20% increase of fractional cloud cover, indicating that the lack of the depositional snow growth causes unrealistically large mixing ratio of cloud ice. The experiment excluding the precipitation-radiation interaction displays a significant cooling and drying bias. The analysis of heat and moisture budgets shows that the simulation without the interaction produces more stable upper troposphere and more unstable mid and lower troposphere than does the simulation with the interaction. Thus, the suppressed growth of ice clouds in upper troposphere and stronger radiative cooling in mid and lower troposphere are responsible for the cooling bias, and less evaporation of rain associated with the large-scale subsidence induces the drying in mid and lower troposphere.