一种2米温度误差订正方法在复杂地形区数值预报中的应用
投稿时间: 2017-08-21  最后修改时间: 2017-10-27  点此下载全文
引用本文:
摘要点击次数: 465
全文下载次数: 
作者单位E-mail
赵滨 国家气象中心 zhaob@cma.gov.cn 
张博 国家气象中心  
中文摘要:模式地形与实况地形间固有差异严重影响了2米温度的评估效果,传统二维插值方法仅满足预报要素与实况观测间经纬度二维空间上的一致性,而忽略了垂直方向的一致性,使得预报与观测不在同一三维空间上进行检验,从而引起严重的评估误导。日变化是2米温度的重要特征,受辐射等物理过程表达缺陷的限制,模式在日变化预报上常存在较大偏差。本文利用模式三维预报变量,结合近地面要素预报产品,采用2米温度三维插值方法进行地形订正,以确保预报与观测在三维空间上的一致性,在地形订正基础上,利用历史月均预报误差为参考误差,剔除模式系统性误差,获取具备日变化特征的预报产品。基于陕西地区复杂地形条件下的典型观测站点,利用2016年8月28日48小时预报个例进行对比分析发现,三维插值方法有效改善了地形误差引起的评估误导问题,但无法改进模式预报的日变化趋势,进一步采用系统性误差订正后,日变化特征明显改善,特别是前24小时预报效果体现出与实况良好的一致性及更佳的预报技巧。通过2016年夏季统计评估表明,误差订正后的2米温度预报产品有效改善了周期性误差振荡,均方根误差稳定在2 K左右,显示出明显的改进优势。
中文关键词:2米温度,三维插值方法,复杂地形区,误差订正,日变化
 
Application of a Bias Correction Scheme for 2 Meter Temperature over Complex Terrain
Abstract:The inherent difference between the observation topography and model terrain has seriously affected the 2 m temperature verification accuracy. The traditional two-dimensional interpolation scheme can only ensure forecast element and observation consistency in latitude and longitude location of the two-dimensional space, while ignoring the vertical direction consistency, which makes verification result of forecast and observation is not from the same position, thereby causing an evaluation misleading. Diurnal cycle is an important feature of the 2 m temperature, Due to the limitation of physical process such as radiation, large bias always appears in the diurnal cycle forecast. In this paper, using three-dimensional forecast variables, combined with near-surface elements of forecast products, an advanced three-dimensional interpolation scheme is developed to ensure consistency with the observed three-dimensional space forecasting. Based on the topography correction methods, monthly forecast error is used as reference bias products to eliminate systematic error and obtain forecast products with characteristics of diurnal cycle. Based on typical observation gauges in Shanxi province complex terrain, 48 hours forecast products on August 2016 are used to compare and it is found that the three-dimensional interpolation scheme effectively solve the evaluation misleading caused by the height bias between the model terrain and observation topography, however it can hardly effectively improve the diurnal cycle trend. After the systematic error correction is adopted, the diurnal cycle forecasting features are improved obviously. In particular, it shows better consistency with observation and the higher skill score, especially in the first 24 hours. By statistical evaluation on summer 2016, it indicates that 2 m temperature after the bias correction can effectively improve the oscillation of periodic errors, the RMSE keeps at about 2 K which shows the obvious advantage of improvement.
keywords:2 meter temperature, three-dimensional interpolation scheme, complex terrain, bias correction,diurnal cycle
查看全文  查看/发表评论  下载PDF阅读器