摘要
基于是否耦合城市冠层模式的3个对比试验(无城市冠层模式:W-NUR;单层冠层模式:W-UCM;多层冠层模式:W-BEP),应用WRF模式模拟了珠江三角洲(简称珠三角)地区发生于2011年6月21日午后的一次强降水过程,发现W-UCM、W-BEP试验的模拟结果均优于W-NUR试验,但降水落区与实际观测相比仍有差异。为此应用前期10 d(2011年6月11—20日)的模拟结果,通过对珠三角城市群6个站点基本气象要素模拟效果的评估,重点考察了W-NUR及W-UCM试验结果的差异,并对模拟降水存在差异形成的原因进行分析。评估结果表明:无论是W-NUR还是W-UCM,模拟的10 m高度风速普遍偏强,其中W-NUR模拟的平均风速比观测偏强1.61 m/s,W-UCM偏强1.58 m/s;W-NUR模拟的2 m高度温度及温度露点差均较观测偏低,温度平均偏差-1.28℃,温度露点差平均偏差-1.39℃,而W-UCM模拟的温度及温度露点差较观测偏高,温度平均偏差略高0.14℃,温度露点差平均偏高1.12℃。平均偏差及均方根误差分析反映出,温度模拟的误差最小,其次为温度露点差,风速误差最大,而且模拟温度和温度露点差与观测相关性更好,相关系数分别大于0.60和0.50,通过0.001显著性水平检验,而风速的相关性则相对较弱。总体来看,尽管耦合冠层模式后WRF对地面气象要素的预报有所改善,但模拟的10 m高度风速仍然偏强。就21日强降水过程的模拟来说,由于偏北风偏强造成切变线南压可能是模拟降水落区偏南的一个原因。
Based on three comparing experiments in which the WRF model is coupled without any urban canopy model, denoted as W-NUR, or is coupled with a single-layer urban canopy model, W-UCM, or is coupled with a multilayer urban canopy model, W-BEP, a process of afternoon heavy precipitation on June 21,2011 over Pearl River Delta metropolitan area is simulated through the WRF model and the results show that the W-UCM simulation is better than the other experiments, though its simulated locations of precipitation are deviated from the observation. Therefore, the simulations of the early ten days （from June 11 to 20 in 2011） are applied to investigate the difference of simulated experiments between the W-NUR and the W-UCM through estimating the essential factors of six meteorological stations over Pearl River Delta metropolitan area. The assessment results indicate that wind speeds simulated by both the W-NUR and the W-UCM at 2 m height are stronger than the observed but the mean bias is reduced using the W-UCM （1.61 m/s for the W-NUR versus 1.58 m/s for the W-UCM）. The simulated temperature at 2 m height with the W-NUR is colder while that of the W-UCM is warmer significantly reduced using the W-UCM （-1.28 ℃ for the than the observed. However, the mean bias is W-NUR versus 0.14 ℃ for the W-UCM）. From the simulated dew-point depression at 2m height of both experiments, the mean bias of the W-NUR is smaller than that of the W-UCM （-1.39 ℃ for the W-NUR versus 1.12 ℃ for the W-UCM）. Comparisons of the mean biases and correlation coefficients of wind speed, temperature and the dew-point temperature between the simulation and the observation show that the mean biases and the correlation coefficients of the 2 m temperature are the smallest and largest of the three elements, respectively, while those of the wind speed are the largest and smallest, respectively. Overall, although the forecast of surface elements is much improved after using the WRF coupled with the urban canopy model, the simulated wind speed at 10 m height is stronger than the observed. For the simulation of heavy precipitation on June 21, a strong northward wind induces a shear line to move southward, which is one of the reasons resulting in the simulated precipitation area shifting southward.
出处
《热带气象学报》
CSCD
北大核心
2013年第6期935-946,共12页
Journal of Tropical Meteorology
基金
行业专项资助(GYHY200906026)
国家自然科学基金(41105006)共同资助
关键词
城市冠层模式
地面气象要素
数值模拟
影响评估
rban canopy model, surface meteorological elements
numerical simulation
impact assessment
