The mesoscale Weather Research and Forecasting model（WRF） was used to dynamically downscale the data provided by the second version of Climate Forecast System（CFSv2） of the American Environmental Forecasting Center.Based on the hindcast of three typical years and the real-time monthly precipitation forecast in 2017, combined with the real-time precipitation data from hydrological stations and meteorological stations and the precipitation analysis products of 56 years, a two-step modification method of systematic error correction and historical extreme value adjustment for the monthly precipitation forecast in summer in Southwest China with complex terrain is constructed, and the correction effect is tested by errors and root mean square error（RMSE）.The results show that:（1） The spatial distribution of monthly precipitation forecast after two-step error modification was close to the observation, that is, there were more precipitation in the regions of the lower reaches than in the upper and middle reaches of Yalong River.（2） The uncorrected precipitation forecast obviously overestimated the rainfall in the southern part of upper reaches, the middle reaches and the lower reaches of the river, but after the systematic error correction and historical extreme adjustment, the overestimated precipitation forecast for the lower reaches of Yalong River was improved greatly.（3） Having been corrected, the errors of accumulated precipitation error prediction in June was reduced by 124.1 mm and the RMSE was reduced by 148.4 mm.Then, the errors and RMSE of accumulated precipitation forecast in July were reduced by 69.1 mm and 68.9 mm, respectively.（4） After being corrected twice, both the errors and RMSE in upper, middle and lower reaches of Yalong River basin were reduced, which reflected that the two-step modification method has significant correction effects.

基于欧洲中期天气预报中心全球高分辨率预报模式ECMWF、中国自主研发的新一代业务化区域数值模式GRAPES_Meso和WRF中国全国区域预报模式的降水预报结果进行未来3 d降水集成预报.以中国地面—卫星—雷达三源融合逐时降水格点产品(CMPA-Hourly,V2.0)作为"观测值"进行建模,采用消除偏差多模式平均法和基于无偏平均绝对误差集成法对中国大陆地区进行降水集成预报,同时采用2800个国家自动气象站降水观测数据对降水集成预报效果进行检验.结果表明:基于无偏平均绝对误差的降水集成方法能够综合每个模式成员降水预报场的优势,提供一种更为稳定可靠且具有更高分辨率的优质精细化降水预报产品;其在试验期间对中国大陆地区汛期的降水预报ETS评分,优于消除偏差多模式平均降水集成预报和最优单模式降水预报,BIAS评分更接近于1,与"实况"的距平相关系数也明显提高,是对降水大值捕捉能力较优的一种集成方法,尤其对中雨到暴雨量级预报的改进较好.

结合气象和水文观测资料,检验分析了ECMWF-IFS、NCEP-GFS、GRAPES-GFS三种全球模式在雅砻江流域2018年汛期的日降水预报.总体上,三天内三种降水预报与站点观测的一致性较好,可以较为准确地预报晴雨和小雨,漏报现象少,但空报偏多;对中到大雨的预报能力有限,空报或漏报都很普遍;在雅砻江流域,全球模式对于东移高原槽配合西南暖湿气流背景下的降水预报效果最优,低层切变影响下的降水预报效果最差,而当西伸副热带高压占据主导时,全球模式的降水预报表现不稳定.对比三种降水预报发现:整体预报效果的优劣排序为ECMWF-IFS、GRAPES-GFS、NCEP-GFS,但各模式对不同量级降水的预报效果存在差异:对于弱降水,GRAPES-GFS的预报最优,NCEP-GFS次之,ECMWF-IFS的预报效果较差,空报现象偏多;而当降水达中雨及以上量级时,ECMWF-IFS的检验评分最高,GRAPES-GFS稍逊之,但仍优于NCEP-GFS,后者易出现漏报;此外,ECMWF-IFS预报的降水量偏多,而其余两者的降水量偏少.

为了实现复杂地形下高分辨率风场的数值模拟及特征分析,采用中尺度气象模式WRF(Weather Research and Forecasting Model)结合牛顿松弛逼近Nudging资料同化技术,实现哈密地区水平分辨率1km的近地层风场数值模拟计算.基于模拟区域测风塔实测数据的对比检验发现,同化观测资料后风速风向的模拟结果均与实测更加接近,70 m高度风速模拟结果的绝对误差降低0.25m· s-1,同化后的模拟结果可以较好的修正风速较小时模拟值偏高和风速较大时模拟值偏小的问题,同时风廓线的模拟结果也与实测更加吻合.通过分析哈密复杂地形下水平分辨率1km逐10 min风场输出结果发现:(1)哈密地区地形比较复杂,风速平面分布差异很大,4月份风速较大区域主要分布在山北地区和西部山南垭口附近,而7月份风速较大区域则位于西部的山坳南部和北部地区;(2)复杂地形下风速较小时风速为负切变,且平均风速越小负切变值越大,地形越复杂负切变值越大;风速较大即使是复杂地形下同样为正切变,但是正切变值比平坦地区的值要小,平坦地形下风速越大正切变值越大;(3)哈密地区复杂地形下,风速12～25 m·s-1的风速占比在时间和空间上分布差异较大,风速较大的4月份,大部分地区占比达到20％以上,尤其是山北和西部垭口附近,占比甚至达到了50％以上,风速为12～25 m·s-1的情况下80m高度平均风速比60m高0.60～0.80m·s-1,比月平均风速的垂直变化值要大;(4)风速较大时,风向10 min变化不明显,风速较小时,风向变化值较大,且地形较平坦地区风向变化值较大,地形复杂地区变化值较小;(5)风向的垂直变化与风速大小关系比较明显,风速越小,其垂直变化越大,风向垂直变化的区域分布与地形复杂程度相关,地形越复杂风向的垂直变化值越大.

Using the tropical cyclone best track dataset from the Joint Typhoon Warning Center (JTWC) and the ERA-Interim reanalysis data on global 0.5°×0.5° grids from 1979 to 2015, characteristics of low-level winds around tropical cyclones (TC) over Hainan (15.5°N–23.5°N,106°E–116°E) are statistically analyzed. Results are as follows. (1)On average, annually there are 5.5 TCs entering the study region during April to December. Most of the TCs occur over the ocean to the southeast of Hainan and their intensities are stronger than those occurred to the northwest of Hainan. (2) When the TC centers are located at different quadrants of the island, the low-level wind fields in the TC periphery have different asymmetric features and high wind occurrence frequencies. The strong wind occurrence frequency is the highest when TC centers are located at the island, followed by that when the TC centers are located to the south of the island, and the occurrence frequency is the lowest when the TC centers are located to the north of the island. (3) Radiuses of the maximum wind (RMW) for TCs around Hainan are within the range of 9–185 km and the averaged value is 58.3 km. On the average, the RMWs of TCs located at the island (to the southwest of the island) are the biggest (smallest). (4) The maximum wind speeds near the TC centers sharply decrease from the ocean to the land for all TCs. And high wind speed regions are located at oceanic areas to the west and east of Leizhou Peninsula. (5) The 34-kts wind radiuses on the eastern side of the TCs are larger than that on the western side. And the stronger the TC is, the larger and more symmetry its specific wind radius will be. (6) Different regions have different degrees of wind circle deformation. On average, when the TC is located to the southwest (southeast) of the island, its wind circle experiences the largest (least) deformation.

应用WRF模式,以ECMWF全球环流预报场为模式背景场,对辽宁和吉林区域进行了4个典型月的10和3.33km分辨率的预报试验,并基于两个省的47座测风塔同期观测资料,对预报结果进行了常规统计检验、关键区间检验,以及预报误差的分解,探讨不同分辨率对东北地区近地层风场预报性能的影响.结果表明:1)模式水平分辨率对风速频率曲线的影响不大,但在辽宁的预报效果优于吉林;2)全风速段检验结果表明70m高度预报性能优于10m高度;3)模式分辨率的变化对均方根误差分解后各项的贡献率影响不大,但存在模式分辨率越高,位相差略微增加,系统差和幅度差的贡献略微降低的现象;4)分辨率的变化对不同风机工况的预测准确性有一定影响,在冬、春季表现为分辨率越高,无功工况下的预报准确性越高;而在夏、秋季分辨率越高满发工况准确性越高;5)不论哪个季节,分辨率的提高均导致发电工况下的TS评分降低,均方根误差增大,相对误差增大.可见在我国东北地区,并非模式分辨率越高风速预报准确性越高.

The fine data at 10min and lkm temporal and spatial resolutions,respectively,were used in the weather research and forecasting model assimilating synchronous 10min average wind-speed data from four masts and surrounding weather stations.The fluctuation characteristics of regional wind parameters were analyzed,using the wind capacity factor (CF),wind capacity credit factor (CCF),and wind peak-valley factor (PVF) and compared the differences in these parameters in three sizes of area.The major findings are as follows.1) Increasing the area where the grid interconnection of wind farms yielded a good smoothing effect on the fluctuation characteristics of wind power,which is apparent as decreases in the hourly and monthly variations of CF,CCF,and PVF with increasing area.2) Regional wind power had positive or opposite effects on the peak-load regulation of the receiving-end power grid,with varying effects with time and in different areas.3) CCF involvement in balancing the electrical power load was equal to 9％ and 5％ in the study region with a guaranteed rate of 85％ and 95％,respectively,and 29％ during peak-load hours.These methods and results are valuable for planning of large-scale wind power transmission and relevant policy making.

为了实现时间跨度较长(20-30年)的风能资源数值模拟评估,建立了一种新的分型方案,采用地面和探空观测资料,以风速、风向和最大混合层高度作为分型因子,针对每种类型随机抽取5％的日数进行真实算例的数值模拟.根据中国所有地区典型日的平均风能参数和30年观测平均结果的对比分析,得出以下结论:(1)中国所有分型站点挑选典型日的平均风速与30年平均风速的绝对误差均小于0.10 m/s,相对误差均低于6.50％;挑选典型日各风向频率值与30年结果的绝对误差平均值为0.28％-0.48％,风速频率绝对误差的平均值为0.09％-0.54％.(2)通过模拟区域分型站点以外探空气象站的风向频率和风速频率的对比检验发现:沿海地区风向频率绝对误差为0.27％-0.63％,风速频率绝对误差为0.14％-0.49％;内陆复杂山地风向频率绝对误差基本在0.57％以下,风速频率绝对误差为0.22 ％-0.60％;结果表明选取一个分型站点能够代表整个模拟区域内的风能资源特性.(3)根据沿海和内陆山地模拟区域重合范围内的探空站分型结果对比分析发现:对于模拟区域重合范围内的探空站,采用所有模拟区域分型站典型日结果加权平均后的风能参数对比误差大大低于各自模拟区域分型站点的对比误差.

采用数值模拟手段可以获取风能资源开发利用高度处资源量的区域分布状况,海面的动力学特性对近海面风速的数值模拟结果影响较大,论文采用WRF模式来研究不同海面动力学粗糙度参数化方案对近海及沿海地区风能资源的数值模拟评估结果影响以及方案的适用性.采用不考虑粗糙度变化(P 0)、WRF模式中参数化公式系数为定值的方案(PWRF)、Yelland和Taylor的分风速对参数化系数进行取值的方案(P YT)对杭州湾近海及沿海地区风场进行模拟研究.通过与观测资料的对比分析可以得出以下结论:通过参数化方法改变海面动力学粗糙度可以改进近海和沿海地区风速的模拟效果,且P YT方案的模拟效果要优于P WRF方案;当观测点为海风时,考虑粗糙度变化方案的风速模拟结果改进更加明显;观测点风速低于约8 m/s时,所有方案模拟结果的相对误差均随风速增加而降低,P YT方案的模拟误差最小,P0的误差最大;而当风速大于8m/s时,相对误差会随风速的增加而增加,P YT方案模拟误差最大,P0方案误差最小.三种方案模拟风能资源差异主要体现在陆地沿海和近海区域,离海岸线较远的陆上地区模拟结果相差不大,在沿海和近海区域方案PWRF平均风速模拟结果比方案P YT的偏高约为0.5 m/s,平均风功率密度偏高约为50W/m2.

Weather Research Forecast(WRF) model was used in experimental forecast of hourly solar radiation in four typical months in Beijing area. The solar radiation observational data of Nanjiao weather observatory were used to evaluate and correct the forecasting result. Conclusions were as follow s: the forecasting result of solar radiation with 5 km horizontal resolution was very similar to that with 1 km horizontal resolution in condition of the existing model. The forecast effect of solar radiation in sunny days was the best with WRF model,the secondary was in cloudy days,and the worst was in fully cloudy days or rainy days. Through analysis of error decomposition,the contribution of phase bias,system bias and amplitude bias was different obviously in each month. The ordinary linear correction method was useful to remove most of system bias and amplitude bias obviously,and in general,the correcting effect of double bias correction method was slightly better than that by linear regression method. The phase bias was difficult to eliminate only by ordinary linear correction method,so a new correction method should be adopted.

为了使检验的预报风速具有可比性,首先利用WRF模式对3种常用全球环流预报场(ECMWF、GFS、T639)在中国区域进行4个典型月的10 km水平分辨率的降尺度计算,再利用中国400座测风塔同期观测资料对24h内70 m高度的风速预报性能进行了对比检验.结果表明:(1)从4个典型月的全风速检验情况来看,ECMWF预报效果略好于GFS,T639稍差,但三者做简单的集合平均可获得最优的预报结果;(2)3种全球预报场降尺度后的预报风速误差平面分布比较相似,整体来看,内蒙古、东北和沿海部分区域误差较小,内陆地区误差较大,尤其在高原和内陆复杂地形下预报效果不佳;(3)从预报风速在4个风速区间的平均检验情况来看,在(0,3] m·s-1和(3,15]m·s-1区间内,ECMWF的TS评分略好一些,在(15,25]m·s-1区间GFS的TS评分最高,在25 m·s-1以上ECMWF具有明显优势;(4)在(3,15] m· s-1区间,3种全球环流预报场的风速预报误差平均在35％左右,ECMWF略好于GFS,GFS略好于T639;(5)从各测风塔点位最优预报效果的全球场统计情况来看,在(3,15]m·s-1区间内,有55.5％的点是ECMWF效果最优,24.8％是GFS最优,19.7％是T639最优.

In view of the high cost and sparse spatial resolution of the offshore meteorological observation,wind maps retrieved from the satellite synthetic aperture radar (SAR)data are valuable in offshore wind energy planning as a supplement to on site measurements,which is crucial in offshore wind resource assessment.The advanced SAR maps monitored by Envisat in 2002 have been widely used in off-shore wind resource assessment in Europe.The technical methods of off-shore wind resources as-sessment at Hangzhou Bay based on the near 1 km×1 km SAR wind retrieval data are investigated.Comparison of the in situ wind speeds at Hangzhou Bay and the SAR wind retrieval shows that (1 )the relative errors of most off-shore observation sta-tions (13/14)are less than 20%,of which 7 stations less than 10%,and the standard deviation is 2.29 m/s;(2)high consis-tency has been found on the wind energy parameters (shape parameter and scale parameter)calculated by the SAR wind retriev-al and the in situ wind;(3)compared with the control experiment,when the SAR wind retrieval assimilated into the numerical model (WRF),it shows a significant improvement on the wind speed correlation coefficient at most of the observation stations and some decrease in standard deviation as well as relative error,suggesting that high resolution SAR wind retrieval is able to be applied to offshore wind resource assessment in China.

Using the 100 m tower atmospheric turbulence observation experiment data during 2009-2010 in Inner Mongolia of-fered by the National Climate Center,the variation characteristics of vertical profile of wind speed under the different stability conditions are analysed,and the wind profile expression is established based on the local similarity theory.The results show that:(1)In the Inner Mongolia grassland near surface atmosphere,the probability of unstable stratification and stable stratifi-cation occurs much higher than the neutral and weakly unstable and weakly stable stratification.(2)Under stable stratifica-tion,the characteristics of the atmosphere change obviously in the vertical direction,which means that the hierarchical consid-eration is needed.According to the atmospheric turbulence observation experiment data,under stable stratification,the rela-tionship between the dimensionless velocity gradient and the stabilityζ can be expressed as 1 +βmζ,in whichβm changes with the height:βm takes 4.3-5.4 under 30 m height;βm takes 2.9-3.4 between 30-50 m;andβm takes 2.0-2.9 over 50 m.

Wind speed forecasting mostly adopts global circulation background as the initial fields of meso-scale model to downscaling, in order to obtain surface layer wind speed at a fine resolution. In China, ECMWF(EC in brief), GFS and T639 data are common circulation field data released by European centers and CMA. This study collected and collated the three sets of data from June 2010 to May 2011. To ensure data comparability, we adopted EC with a resolution of 0.25° grid(lat., Ion.), GFS with 0.5° grid(lat., Ion.)and T639 data with 0.28125° grid(lat., Ion.). We contrasted 677 weather stations and 400 wind masts over the same time period in China. We found that when the wind speed is between(0~3)m/s, the performance of the data sets is T639 > GFS > EC. When the wind speed is between(3~15)m/s the performance is EC > GFS > T639 when verified by masts and GFS > EC > T639 when verified by weather stations. When the wind speed is between(15~25)m/s, the performance is EC > GFS > T639. Based on the(3~15)m/s wind speed segment in northeast China, north China, northwest China and southwest China the performance is EC > GFS > T639, but the three sets of data for southwestern China are poor in Autumn and Winter. In east China during spring and autumn, EC and GFS are almost the same and both better than T639. In summer the performance is EC > GFS > T639 and in winter the three sets are almost the same. In south China during spring, autumn and winter, EC is the best and GFS and T639 are identical. In summer, the three sets of data are nearly identical.

The wind energy resources of 1 km×1 km resolution in Jiuquan region in Gansu Province from June 2009 to May 2010 was simulated using the mesoscale model MM5 combined with wind diagnosis field model CALMET and 1°×1° reanalysis data of NCEP.The comparison of the simulated and observed wind speeds shows that the relative errors of flat site and away from the mountain are smaller,the complicated topography is bigger.The relative errors in seven wind towers of ten wind towers are no more than 10%,the values of other three towers are within 20% too.The seasonal changes of wind speed and wind power density display that the data of spring is the biggest,then autumn and winter,and the summer is the smallest.This seasonal variation has a close relation with terrain,physiognomy and affected weather system.The wind speed increases with the increasing of height in Jiuquan region,especially in low latitude area,namely the most west part of Hexi corridor and north part of Jiuquan.Through the analysis of wind speed representativeness in the simulated year,there is stronger wind in Jiuquan region in the simulated year.If the average wind state is further analyzed,some correction of the simulated results should be done.

In order to develop and utilize the wind energy resource efficiently, and realize the target of stepping from a big wind energy country into a strong wind energy country, this paper summarizes the application status, problems and impacts on project benefits in various aspects of meteorological technology application in wind energy exploiting based on a large number of past projects practice, data analysis and test results. Furthermore, based on the analysis of atmospheric science principles and the special needs of wind power to meteorological technology, it is pointed out the key technical issues and direction should be noted in the applying process of wind energy assessment, numerical simulation and numerical prediction technology.

In this paper WRFV2.2 mesoscale numerical model is used to simulate meteorological elements of different heights over Poyang Lake area within 200km×200 km range.We got the wind,temperature,humidity fields and the profile distribution of Poyang Lake area atmospheric boundary layer with the high-resolution of 1 km,from 00:00 Nov 5,2009 to 12:00 Nov 6, 2009.The results show that:(1) The cool island effect exists between the lake-land during daytime accompanied by lake wind,and the heat island effect at night accompanied by land breeze, the maximum temperature difference between lake and land is up to 6℃.(2) Meanwhile,the terrain and surface types on the Poyang Lake area has a great influence on the distribution of the wind, mainly there exists a southwest-to-northeast low-level convergence zone at night,which disappears during the day.(3) In addition,the role of wind and topography makes the lake's moisture distribution non-uniform.The humidity thickness becomes lower during the day while on the contrary at night.On the right of Poyang Lake the humidity value is greater than the left. The results relatively well reflect the distribution of meteorological elements within the boundary layer in Poyang Lake area,which can help us to understand the characteristics of regional climate,and the distribution of wind,temperature and humidity at different levels.It provides a scientific basis for the Poyang Lake area local weather forecast,wind energy resource development and environmental protection.

文中将建筑物动力冠层方案引入到城市边界层精细模式中,该方案描叙了城市地Ⅸ建筑物对气流的拖曳作用以及建筑物形态对城市地区湍流活动动量输送的影响.采用建筑物拖曳法和建筑物动力冠层方案两种不同建筑物动力学处理方法对北京地区东南部进行模拟,通过城市地区水平风速的模拟结果与自动气象站实测资料对比发现模拟结果与实测相当吻合,建筑物动力冠层方案的引入能够更好地实施对城市地区的水平风速分布的数值模拟.分别采用建筑物拖曳法与建筑物动力冠层方案对一个实际小规模城市进行模拟试验,分析表明引入建筑物动力冠层方案可以模拟出小规模城市地区水平风速偏低的现象;还可以模拟出城市地区建筑物动力作用对湍流活动的影响;对湍流动能模拟结果表明比较符合实际分布情形.采用建筑物动力冠层方案对小规模城市地区建筑物高度变化对城市地区的流场及湍流活动的影响进行模拟分析,结果表明:城市地区建筑物高度增加,风速变小,但是高大建筑物底部风速略高于低矮建筑物底部;城市地区大气湍能增加,高大建筑物底部湍能较小,冠层高度内湍能随高度增加而增加;同时城市地Ⅸ的垂直扩散系数也随建筑物高度增加而增加;城市地区污染物排放高度处污染物浓度较低,下游乡村地区地面污染物浓度较低,但高空污染物浓度却较高.

Site selection in development zones and quantification of the technical exploitation amount of wind power through integration of accurate numerical simulation and objective assessment play a fundamental role in exploitation and utilization of wind energy resources.The distribution of wind energy resources at 50 m height with high resolution of 1 km×1 km at the planning area of wind power in Jiuquan City,Gansu Province was simulated using the wind energy numerical simulation model developed by the China Meteorological Administration.The model uses dynamic and statistical method,consisting of the large scale weather background classified model,meso-scale simulation model and statistical model.First,30-year climate datasets were classified as 448 clusters in terms of weather types.Second,each type was simulated by the meso-scale model.Third,the weighted average method was used to compute the distribution of wind energy resources in terms of the probability of each type in the 30-years datasets.Simulated results showed that the annual wind power density and annual effective hours in the south of Changma Town,Yumen City,Liuyuan Town,Guazhou County,the west of Dunhuang City and the surrounding of Mazongshan Town are relatively high.This suggests the large potential of wind energy resources and the feasibility of large scale exploitation of wind energy resources.The simulations of wind speed and wind power density are accurate through comparing with observations at three wind towers.Except one tower showing relatively large error in wind speed,error at the other towers was generally less than 5%.Major constraints on development and utilization of wind energy resources in the planning regions were carefully analyzed based on geographic information datasets consisting of topography,geomorphology,and land use data from the State Bureau of Surveying and Mapping using spatial analysis techniques and functionalities in ArcGIS software packages.Results suggested that the areas of desert and terrain with slope larger than 10% account for 90.91% of the total planning area,an explicit indication of limitations in wind energy resources exploitation.The area and technical exploitation amount of wind power on different grades in the planning regions excluding the limited development areas were computed.It was found that the rich zones for wind energy resources exploitation in Jiuquan City are Yumen,Guazhou and Subei counties.The technical exploitation amount of wind power on the third grade was found to be 117 million kW,whereas that on the fourth grade is only 3.6 million kW.As such,Jiuquan City seems is more suited for being taken as the third-grade wind farms in wind energy resources exploitation.This study provides a basis and reference for assessment and analysis of constraints on utilization of wind energy resources.

Lamb-Jenkinson atmospheric circulation classification scheme was used to classify the circulation type of daily average sea level pressure over Beijing in 1948-2008. The following conclusions are drawn that among 27 circulation types,primary types are A,C and SW,and their occurrence frequencies are 23.8%,15.4% and 7.0%,respectively. Frequencies of types A and C tend to increase,but the increasing rate of type A is much greater than that of type C. The seasonal primary circulation types are A and C in spring,C and SW in summer,and A in autumn and winter,but type CSW can not be ignored in summer. It is fine and cool in winter Beijing under type A,mostly rainy in summer under type C and rainless under type SW,and dry and hot under type CSW,especially after 2000.