模式背景场及边界层参数化方案对近地层风场预报的影响

发布时间：2018-03-12 15:55

本文选题：WRF模式　切入点：风场预报　出处：《中国气象科学研究院》2017年硕士论文　论文类型：学位论文

【摘要】：近地层风场的预报是风能资源利用的重要关注点。背景场资料是区域模式的初边值来源,而模式近地层的湍流过程需要边界层参数化方案来描述。因此,两者对近地层风场的预报有着重要的影响。利用WRF中尺度模式和GSI同化系统分别研究了背景场资料和边界层参数化方案对蒙西地区的风场预报试验,并利用该地区的70个国家地面站和18座测风塔资料对试验结果进行了评估和分析。模式背景场使用GFS资料和T639资料的试验结果表明:(1)两种背景场资料的预报都能较好地再现蒙西地区模拟时段内的风场变化情况;两种背景场预报的10 m风速都较观测偏高,对风向的预报都有比较小的顺时针偏差;(2)多种统计指标分析表明GFS对10 m和70 m风速风向的预报效果稍好于T639,但两者差异较小;(3)模式的分辨率和预报时长对预报结果有比较显著的影响;分辨率为4 km的风场预报效果要优于分辨率为12 km的预报效果;第12~35时的风场预报效果优于第36~59时的预报效果,而第60~83时的预报效果最差;(4)风场预报误差较大的站点主要集中在阴山山脉南部的低海拔地区,表明复杂的地形分布会影响模式的预报效果。对比7种不同边界层参数化方案的试验表明:(1)各方案的预报都可以抓住实际风速风向的演变和位相变化;YSU方案对10 m风速的预报效果最好,ACM2对70 m风速预报效果最好;各方案对10 m风场的预报差异较70 m的更为明显,且模拟的风速偏大;模式对低风速的预报能力较差;(2)各方案对有效风速(3m/s 15 m/s)的预报效果最好,对满发风速(15 m/s 25 m/s)的预报效果次之,对无效风速(m/s)的预报效果最差;(3)各方案都能模拟出风速风向的日变化特征,表现为白天风速随时间增大和风向的顺转,傍晚到夜间风速的减小和风向的逆转;(4)TEMF、YSU以及ACM2方案对风切变指数的预报较好,大气层结的稳定性对低层风场的模拟有比较重要的影响。综合而言,模式背景场采用GFS资料和T639资料均能得到较好的预报效果。模式水平分辨率采用4 km时结果较为理想,预报时效在12~35小时优于其他时段。就边界层参数化方案而言,YSU方案10 m风速的预报效果明显优于其他方案,各方案对70 m风速的预报效果差异较小,其中ACM2表现相对较好。TEMF、YSU以及ACM2方案对风切变指数的预报较好。复杂地形是影响近地层风场预报效果的重要因素,大气层结的稳定性影响着低层风场的模拟效果。
[Abstract]:The prediction of near-ground wind field is an important concern in the utilization of wind energy resources. The background field data is the source of the initial and boundary values of the regional model, and the turbulence process of the model near-ground layer needs to be described by the boundary layer parameterization scheme. The WRF mesoscale model and the GSI assimilation system are used to study the background field data and the boundary layer parameterization scheme for the wind field prediction test in Mengxi area, respectively. The test results are evaluated and analyzed by using the data of 70 national earth stations and 18 wind towers in this area. The results of the model background field using GFS data and T639 data show that the prediction of the two kinds of background field data can be achieved by using the two kinds of background data. The variation of wind field in Mengxi region during the simulation period is well reproduced. The wind speed of 10 m predicted by both kinds of background fields is higher than that of observation. Statistical analysis shows that GFS is better than T639 in predicting wind direction of 10 m and 70 m, but the difference between the two models is small. The results were significant. The prediction effect of the wind field with a resolution of 4 km is better than that of the one with a resolution of 12 km, and the forecasting effect of the wind field at the time of 12 ~ 35 is better than that at the time of 36 ~ 59. However, the stations with the worst prediction effect at 6083rd hour are mainly concentrated in the low altitude areas in the southern part of the Yinshan Mountains where the wind field prediction errors are large. The comparison of 7 different boundary layer parameterization schemes shows that each scheme can grasp the evolution of actual wind speed and wind direction and phase change. The prediction effect of m wind speed is the best and that of ACM2 is the best for 70 m wind speed. The difference of prediction for 10 m wind field is more obvious than that of 70 m, and the simulated wind speed is larger than that of 10 m, and the prediction ability of the model for low wind speed is poor. (2) each scheme has the best prediction effect on the effective wind speed of 3 m / s 15 m / s. For the full wind speed of 15 m / s 25 m / s, the prediction effect is second, and for the ineffective wind speed of 1 m / s, the forecast effect is the worst. (3) each scheme can simulate the diurnal variation characteristics of wind speed and direction, which shows that the daytime wind speed increases with time and the direction of wind changes along with time. The decrease of wind speed from evening to night, the reversal of wind direction and the ACM2 scheme can predict the wind shear index better, and the stability of atmospheric stratification has an important effect on the simulation of wind field in the lower layer. Both GFS data and T639 data can be used to predict the model background field, and the results are satisfactory when the horizontal resolution of the model is 4 km. The prediction time of 10 m wind speed of YSU scheme is obviously better than that of other schemes, and the difference of prediction effect of each scheme for 70 m wind speed is small, in terms of boundary layer parameterization scheme, 10 m wind speed prediction effect is obviously superior to other schemes. The prediction of wind shear index by ACM2 and ACM2 scheme is relatively good. The complex terrain is an important factor affecting the prediction effect of wind field in the near strata, and the stability of the atmospheric stratification affects the simulation effect of the wind field in the lower layer.
【学位授予单位】：中国气象科学研究院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P456.7

【相似文献】