全球和区域数值模式在甘肃2020年汛期降水预报中的检验评估

2020年汛期6—8月甘肃降水日数多、持续时间长、范围广、强度大,对该时间段内3种全球模式(ECMWF、GRAPES_GFS和NCEP_GFS模式)和4种区域模式[GRAPES区域数值预报业务系统(GRAPES_3 km)、西北区域区域模式(GRAPES_LZ10 km)、西北区域快速更新循环预报系统(GRAPES_LZ3 km)和华东区域模式(SMS-WARMS)]24 h累计降水预报性能进行检验评估。结果表明:(1)全球模式中ECMWF模式的预报性能优于其余2个模式,而区域模式中GRAPES_3 km和SMS-WARMS模式预报性能相对较好,且SMS-WARMS模式预报性能更稳定。(2)区域模式晴雨准确率及小雨和中雨的TS评分、ETS评分、命中率低于全球模式,暴雨优于全球模式;大雨和暴雨的空报率和预报偏差均高于全球模式。(3)根据500 hPa环流形势可将甘肃汛期降水划分为副高边缘型和低槽型2种类型,针对2020年4次副高边缘型和3次低槽型降水进行分类检验评估。全球模式和区域模式均对前者的各个量级降水预报性能优于后者;ECMWF模式和区域模式对2种类型大雨和暴雨预报效果优于NCEP_GFS和GRAPES_GFS模式;全球模式中ECMWF模式、区域模式中SMS-WARMS模式对2种类型降水预报效果最好。(4)7种模式对2种类型中雨和大雨雨带走向预报较好,对副高边缘型降水过程降水落区的预报能力优于低槽型降水过程,但预报降水强度较观测偏强,尤其是降水中心区域。     

位温方程拉格朗日平流计算方案的改进及其在GRAPES_GFS全球模式的应用

GRAPES全球数值预报系统(GRAPES_GFS)个例试验中,模式高层垂直速度在南极大陆上空会出现较大的计算噪音,进而影响模式的积分稳定性,甚至积分中断。对GRAPES_GFS的动力框架以及动力、物理耦合方案进行诊断分析,发现位势温度在拉格朗日上游点上的垂直插值是垂直速度噪音产生的主要原因。2013年7月实际资料试验表明,在位温方程中引入垂直非插值的拉格朗日平流方案,可以减少或消除该计算噪音。使用新的方案后,北半球和热带地区模式预报的整体性能得到了明显提高,8 d预报的全球质量损失较为严重的问题也得到了缓解。     

GRAPES_GFS在东亚地区中期预报技巧极端下降现象的初步研究

利用2016年4月1日-2018年12月31日GRAPES_GFS模式的业务预报数据,将GRAPES_GFS模式在东亚地区144小时预报500 hPa高度场的距平相关系数小于0.4,均方根误差大于60 gpm的事件定义为模式在东亚地区的中期预报技巧极端下降事件,据此标准筛选出29个个例进行了研究.结果表明,GRAPES...     

GRAPES_GFS全球中期预报系统的研发和业务化

该文回顾了中国气象局全球中期数值天气预报系统GRAPES_GFS的研发历程,重点介绍了近年来在GRAPES_GFS研发过程中的重要进展,概要阐述了这些进展对GRAPES_GFS业务:化的贡献。动力框架方面的改进主要包括位温垂直平流的算法、极区滤波方案、标量平流方案、垂直速度衰减（damping）算法、提高模式分辨率等,改善了模式框架的稳定性、计算精度以及质量守恒性。物理过程方面的改进主要包括RRTMG辐射方案、CoLM陆面过程方案、积云对流、边界层过程、双参数云物理方案,以及物理过程的调用计算等,全面提升了模式物理过程的预报能力。全球三维变分同化方面,研发了模式空间三维变分（3DVar）系统、资料质量控制和偏差订正技术、卫星资料同化方面的相关技术等。同时,对目前GRAPES_GFS2.0的预报能力进行了评估,总体来说,该系统各项预报指标全面超越GRAPES_GFS1.0,与T639相比等压面要素预报在对流层也有明显优势,降水、2 m温度等预报也优势明显。     

GRAPES全球/区域集合预报系统10年发展回顾及展望

较系统地概述了中国气象局全球/区域集合预报系统及描述模式初值和模式自身不确定性的集合预报扰动技术发展历程,回顾了GRAPES(Global/Regional Assimilation Pr Ediction System)全球集合预报的奇异向量初值扰动方法、GRAPES区域集合预报的集合变换卡尔曼滤波初值扰动方法和多尺度混合初值扰动方法、GRAPES全球/区域集合预报模式不确定性的随机物理过程倾向项扰动方法和动能后向散射随机补偿方法等研究成果,介绍了GRAPES全球/区域集合预报业务系统构建参数设置和预报性能,最后分析了GRAPES全球/区域集合预报中存在的问题,展望了未来发展方向。     

GRAPES全球系统的研发进展

2007年7月始,中国气象科学研究院数值预报研究中心在“十五”攻关成果的基础上研发了运行稳定、结果合理的GRAPES全球预报系统1．0版本（简称GRAPES_GFS1．0）。2009年3月,该系统通过了中国气象局预测减灾司组织的准业务化运行专家评审,已批准投入准业务运行。这是我国完全自主开发的全球预报系统,标志着我周全球数值预报迈向了自主创新、可持续发展的道路。     

GRAPES与日本模式对东莞单站预报效果的对比

利用统计检验方法,对2011年前汛期(4～6月)GRAPES区域中尺度模式与日本GSM全球谱模式东莞单站的逐日降水和气温预报进行定量检验和对比。结果表明:降水预报GRAPES区域中尺度模式略优于日本模式;对于前汛期由锋面低槽和季风槽引起的明显降水过程,GRAPES区域中尺度模式预报好于日本模式,且两个模式都对锋面低槽降水预报优于各自对季风槽降水预报;对日最高气温的预报,日本模式比GRAPES区域中尺度模式表现的更加稳定,对日最低气温的预报,日本模式具有较好的参考性。     

2012年汛期华南中尺度模式与日本模式的降水预报评估分析

采用累加降水量级检验、降水空间分布和时间演变以及天气系统分类等方法,对广东2012年汛期GRAPES华南中尺度模式和日本气象厅全球谱模式(JMA GSM,以下简称JMA)降水预报产品进行分析评估,结果表明:两个模式对小雨预报能力相当,对于中雨、大雨和暴雨GRAPES预报效果好于JMA;两个模式对广东雨带分布预报相反,GRAPES降水值随着预报时效的增加而增加;两个模式对广东较大降水过程的发展趋势预报基本准确, GRAPES降水预报以偏大为主,JMA以偏小为主;将汛期较大的降水过程按照其主导天气系统分为6类,GRAPES对其中5种类型降水预报效果好于JMA,两个模式对台风暴雨预报效果最好,对回流暴雨预报相对较差。      

GRAPES_GFS在西南地区的预报稳定性 及其误差与地形的关系

根据地形特征,将西南地区划分为高原区、边坡区和盆地区,引入统计学"不稳定度"定量描述模式预报稳定性,对2016年6月—2017年9月全球中期天气预报(GRAPES_GFS)和欧洲中期天气预报中心(EC)在西南地区的高层形势场、主要的天气影响系统和地面要素预报性能进行了主客观检验,一定程度揭示了GRAPES_GFS和EC在西南地区的预报稳定性、地形的影响以及二模式预报性能的异同。结果显示:GRAPES_GFS高空高度场、温度场预报不稳定度分布呈北高南低型,相对湿度、风速预报不稳定度大值区在高原边缘;各要素预报不稳定度季节性周期最为显著,其位相和振幅因要素不同而有所不同;地形主要影响温度和风向预报误差值,但对相对湿度和风速预报的影响则体现在误差随时效的增长速率差异上;"漏报"是模式对西南地区天气系统的主要预报误差源,"低报"则是模式对西南地区2 m温度预报误差的最大来源;模式对西南地区降水落区预报有效率大约为50%,但强度预报通常偏低。EC与GRAPES_GFS的误差特征没有本质区别,但EC误差更小,稳定性更高。     

质量守恒的订正算法在GRAPES_GFS中的应用

如何在长时间积分的过程中保证模式大气的质量守恒是数值模式动力框架面临的基本问题之一,对于半拉格朗日动力框架来说, 在理论上要满足质量守恒面临着诸多的困难,质量订正方案不失为一种简单可行的选择。中国气象局数值预报中心的GRAPES_GFS (Global-Regional Assimilation and PrEdiction System, Global Forecast System) 在长时间积分过程中质量损失问题较为明显,该研究工作借鉴C-CAM (Climate-Community Atmosphere Model) 中对地面气压进行订正控制模式质量守恒的思路,在GRAPES_GFS中开发了一种对每个网格内的质量按照不同权重系数进行调整、控制模式大气总质量守恒的算法。经过一系列的试验,验证了该方法在GRAPES_GFS中的可行性,在保证模式质量守恒的情况下可有效减小高度场的预报偏差,缓解模式预报结果中天气系统偏弱的问题,在实际业务预报中有一定应用价值。     

A comparative study of the observations of high clouds and simulations by an atmospheric general circulation model

The importance of clouds in the upper troposphere (cirrus) for the sensitivity of the Earth's climate e.g., requires that these clouds be modeled accurately in general circulation model (GCM) studies of the atmosphere. Bearing in mind the lack of unambiguous quantitative information on the geographical distribution and properties of high clouds, the simulated distribution of upper tropospheric clouds in a spectral GCM is compared with several satellite-derived data-sets that pertain to high clouds only, for both winter and summer seasons. In the model, clouds are assumed to occupy an entire gridbox whenever the relative humidity exceeds 99%: otherwise the grid box is assumed to be free of cloud. Despite the simplicity of the cloud prediction scheme, the geographical distribution of the maxima in the model's upper tropospheric cloud cover coincides approximately with the regions of the observed maxima in the high cloud amount and their frequency of occurrence (e.g., intertropical convergence zone and the monsoon areas). These areas exhibit a minimum in the outgoing longwave radiation (OLR; Nimbus-7) and are also coincident with regions of heavy precipitation. The model, with its relatively simple cloud formation scheme, appears to capture the principal large-scale features of the tropical convective processes that are evident in the satellite and precipitation datasets, wherein the intense, upward motion is accompanied by condensation and the spreading of thick upper tropospheric layers of high relative humidity and cloudiness in the vicinity of the tropical rainbelt regions.This paper was presented at the International Conference on Modelling of Global Climate Change and Variability, held in Hamburg 11–15 September 1989 under the auspices of the Meteorological Institute of the University of Hamburg and the Max Planck Institute for Meteorology. Guest Editor for these papers is Dr. L. Dümenil     

云分析系统在台风莫拉菲数值模拟中的应用

借鉴美国CAPS开发的中尺度ARPS模式的资料分析系统ADAS,设计了基于GRAPES_Meso的云分析系统,实现了地面云观测资料、卫星云图、多普勒雷达反射率资料在GRAPRS模式中的综合融合应用。通过对登陆台风莫拉菲 (0906) 数值模拟检验云分析系统的性能。结果表明:云分析系统通过地面云观测资料、卫星红外云图、可见光云图、多普勒雷达反射率的同化,能够反演出合理的三维云覆盖状况;在三维云覆盖的基础上结合云底云顶高度,进而反演出云水、云冰、雨水、雪、霰等云微物理量,并显著改善模式初始湿度场;在台风登陆过程的模拟中,虽然对台风路径的预报有所偏差,但对比控制试验,在台风登陆地点、台风强度、路径的预报中云分析发挥了正作用;通过对云分析后初始场反演的雷达回波进行检验,验证了云分析的可行性与正确性,对整体的降水预报效果也有明显改善。     

基于CERES资料的新疆云物理参数时空分布特征

利用2011—2020年ERA5再分析降水资料、CERES云物理参数产品，分析新疆云参数的时空变化分布特征，归纳总结云物理参数与降水的相关性，结果表明：1）云水路径（冰相）值、云粒子有效半径（冰相）、云光学厚度与降水量的空间分布一致，均为山区最大，北疆次之，南疆最小。2）夏季（6—8月）在南、北疆、山区云水路径（液、冰相）、云顶（底）温度、云光学厚度与降水量呈同位相变化；云粒子有效半径（液、冰相）、云顶气压与降水量呈反位相变化。3）夏季（6—8月）北疆、山区的云水路径（液、冰相）值、云顶（底）温度、云光学厚度，南疆云光学厚度与降水量呈正相关；北疆云粒子有效半径（冰相），南疆云粒子有效半径（液相）、云顶气压，山区云粒子有效半径（液、冰相）、云顶气压与降水量呈负相关。     

CONVECTIVE ANOMALIES IN TROPICAL OCEAN AREAS AND LONG-LEAD FORECAST OF SUMMER RAINFALL IN SHANDONG

This study focuses on deep convection anomalies in tropical regions in winter-spring period and their possible influence on the following summer rainfall in Shandong province. On the basis of monthly precipitation wet and dry summers in Shandong are defined according to a precipitation index. Then monthly OLR data, observed by NOAA satellites, are used to diagnose the features of deep convection for both wet and dry summers. It is found that negative anomalies seem dominant prior to wet summers, while large areas of positive anomalies appear prior to dry summers in tropical oceans. The differences are remarkable especially in the western, middle and eastern tropical Pacific as well as in the tropical Indian Ocean. Correlative analysis confirms the relations between OLR and precipitation. Subtropical High, which plays an essential role in summer rainfall, is also connected with the deep conviction. Altogether eight EOF-CCA forecast models are established on the basis of the above study. The assessment of the models relies on the gauge observing precipitation in 1997 and 1998. The results show that models using spring OLR data appear to be more practicable than those using winter OLR data, and the models established with OLR in western Pacific and the Indian Ocean perform better than the others.     

相变修正方案在GRAPES模式标量平流中的应用

如何更好地模拟水物质的空间分布和小尺度变化,对于数值天气预报效果的改进,特别是对于更好地模拟降水过程,具有重要的意义.计算机的飞速发展使数值模式的分辨率不断提高,云的显式计算成为可能,这样就要求水物质在平流的过程中必须要做到高精度、守恒、保形.水物质场是正定标量的场,具有空间和时间变化幅度大、存在强梯度甚至不连续的特点,水物质场的合理模拟一直是数值预报中的一个难题.GRAPES模式中的标量平流方案采用PRM分段有理函数方法,比较好地解决了该半拉格朗日模式中水物质平流的高精度、守恒、保形问题,但是当有凝结潜热发生时,由于半拉格朗日平流方案求解上游点时的插值,在云边缘区域会造成虚假的云水,进而导致不合理的相变过程.为了解决以上问题,本研究在GRAPES模式中PRM平流方案的案础上,加入了非线性半拉格朗日相变潜热的修正方案,旨在改进GRAPES模式对水物质平流问题的模拟,提高降水的预报效果.该研究通过理想试验,验证了非线性半拉格朗口相变修正方案可以有效地限制云边缘由于半拉格朗日平流方案插值产生的虚假柑变;然后将该方案加入GRAPES模式的PRM水物质平流方案中,通过实际个例模拟验证了加入非线性半拉格朗日方案以后,模式可以更好地模拟水物质的平流过程,且对云中热力场及水物质分布地模拟更加合理,同时预报出的雨带中心区与实况更加符合.     

GRAPES-Meso模式浅对流云辐射效应的改进试验

在万子为等（2015）对GRAPES-Meso模式浅对流参数化改进的基础上,进一步引入了浅对流云量诊断计算,并设计旨在完善浅对流云辐射效应的浅云云量和云中水凝物的补偿方案,以改进模式低层云量偏少和浅对流云辐射效应不足的问题。通过对数值试验结果的诊断和对比分析以及与观测的比较,重点考察了浅对流云量计算与浅对流激发的协调性、浅对流云对低云补偿后所产生的辐射效应以及对模式地面要素预报的影响等,验证了改进方案的合理性与有效性。结果表明:（1）浅对流云量诊断计算合理,其云覆盖区与浅对流激发区相吻合,引入浅对流云量的计算可减小模式云量的计算偏差、使其向观测结果靠近;（2）改进方案在浅对流发生区低层0.5-4 km高度范围内,对影响模式云辐射过程的浅云云量和云中水凝物形成有效补偿,最明显的浅云补偿在1-1.5 km高度处,浅对流活跃时期浅对流过程对浅云水凝物（云水和雨水之和）的补偿量可达20%-55%;（3）云光学厚度对浅云水凝物的补偿响应合理,即水凝物的补偿引起云光学厚度增大,两者的变化特征在时空分布上十分相似,且云光学厚度之变化受云水补偿的影响比受雨水补偿的影响更明显;（4）在白天时段,浅云补偿所产生的辐射效应使模式地表太阳总辐射有所下降,缩小了与观测的偏差,进而使地表温度和地面2 m气温模拟偏差减小。改进方案在缓解模式云量偏少、地表太阳总辐射偏强和地面2 m气温偏高等方面的作用,在批量试验中得到了验证。      

The WRF 3DVar System Combined with Physical Initialization for Assimilation of Doppler Radar Data

The three-dimensional variational data assimilation (3DVar) system of the Weather Research and Forecasting (WRF) model (WRF-Var) is further developed with a physical initialization (PI) procedure to assimilate Doppler radar radial velocity and reflectivity observations. In this updated 3DVar system, specific humidity, cloud water content, and vertical velocity are first derived from reflectivity with PI, then the model fields of specific humidity and cloud water content are replaced with the modified ones, and finally, the estimated vertical velocity is added to the cost-function of the existing WRF-Var (version 2.0) as a new observation type, and radial velocity observations are assimilated directly by the method afforded by WRF-Var. The new assimilation scheme is tested with a heavy convective precipitation event in the middle reaches of Yangtze River on 19 June 2002 and a Meiyu front torrential rain event in the Huaihe River Basin on 5 July 2003. Assimilation results show that the increments of analyzed variables correspond well with the horizontal distribution of the observed reflectivity. There are positive increments of cloud water content, specific humidity, and vertical velocity in echo region and negative increments of vertical velocity in echo-free region where the increments of horizontal winds present a clockwise transition. Results of forecast experiments show that the effects of adjusting cloud water content or vertical velocity directly with PI on forecast are not obvious. Adjusting specific humidity shows better performance in forecasting the precipitation than directly adjusting cloud water content or vertical velocity. Significant improvement in predicting precipitation as well as in reducing the model's spin-up time are achieved when radial velocity and reflectivity observations are assimilated with the new scheme.     

Cloud cover over the Tibetan Plateau and eastern China: a comparison of ERA5 and ERA-Interim with satellite observations

This study examines the progress made by reanalyses and satellite products in the estimation of cloud cover over China: the ECMWF reanalyses ERA5 and ERA-Interim, geostationary satellite observation Himawari-8 (H8) and the International Satellite Cloud Climatology Project H-series (ISCCP) product. There is great similarity in spatial patterns of cloud cover in reanalyses and satellite observations, especially between ERA5 and H8. Distinct characteristics of the seasonal evolution of cloud cover are shown over the Tibetan Plateau (TP), the southeast (SE) and northeast (NE) of China. Differences in magnitudes of cloud cover exist. Overestimations are about 10% for reanalyses and about 20% for ISCCP in compared with certain cloud cover in H8. When probable cloud (about 10%) in H8 is included in the estimation, biases reduce the most in ERA5. The cloud hit rate (CHR) and false alarm rate (FAR) in against H8 and ISCCP reveal that simulated clouds in ERA5 have been improved especially over eastern China, but with limited improvement over TP in compared with ERA-Interim. Diurnal variations of cloud cover are characterized by increases during daytime over those three regions. Amplifications of diurnal variation vary over different regions and months. Satellite observations and ERA5 indicate distinguished diurnal cycle of cloud cover over TP, while further investigation based on ERA5 reveals coherent diurnal cycle in meteorological environment. Long-term changes of cloud cover highlight decreasing trends over TP and particular during March in past decades based on ISCCP and ERA5, which require further investigation in future.     

Study of tropical cyclogenesis using satellite data

Summary Satellite data are used to study the cloud development and water vapor supply during the genesis of Typhoon Nina (November 1987). Using satellite microwave and infrared data, the following physical parameters are retrieved and analyzed: water vapor path, surface wind, sea surface evaporation, precipitation, and cloud type.During the week prior to the genesis of Nina, several cloud clusters were observed in the region of the subsequent genesis (near 5° N, 170° E). Cloud type studies showed that several clusters had similar structures. By examining the sea surface evaporation and precipitation in the cloud clusters, we found that the precipitation exceeded evaporation by several times in the precipitating areas of the cluster that evolved into a tropical storm, indicating that local evaporation alone could not supply enough water vapor, and that horizontal transfer of water vapor from surrounding areas is required for the tropical cyclogenesis. Surface wind fields indicated that there was a constant increase of cyclonic wind in the area of the cloud cluster that finally led to the tropical storm, while no apparent increase of wind was found in the other cloud clusters. In addition, water vapor amount did not decrease for several days until the disturbance was upgraded to a tropical storm, while it was found to decrease after the mature stage for the other cloud clusters that did not evolve into tropical storms.From consideration of the water vapor balance, the cyclogenesis can be interpreted as a transition from an unbalanced cluster to a balanced cluster. Horizontal transfer of water vapor in a water vapor-unbalanced cloud cluster is not large enough to overcome the deficit caused by precipitation over evaporation. The shortage of water vapor in the unbalanced cluster results in a short-lived cloud cluster. When the sum of evaporation and horizontal transfer can provide enough water vapor supply to balance the removal by precipitation (balanced cluster), the precipitation does not dry up the atmosphere. This is the necessary condition for the cyclogenesis. The increase in horizontal transfer of water vapor is found in this study to be associated with the increase of the surface cyclonic wind.With 5 Figures