Journal of Arid Meteorology

Previous Articles     Next Articles

Error Statistics and Correction Method of QPE Product in SWAN System

LIU Weicheng 1, 2WANG Yong 1ZHOU Xiaojun 3
  

  1. 1. Lanzhou Central Meteorological Observatory,Lanzhou 730020,China;
    2. College of Atmospheric Sciences,Lanzhou University,Lanzhou 730000,China;
    3. Lanzhou Meteorological Bureau of Gansu Province,Lanzhou 730020,China
  • Online:2014-12-31 Published:2014-12-31

SWAN系统QPE产品的误差统计及订正方法研究

刘维成1,2,王 勇1周晓军3   

  1. 1.兰州中心气象台,甘肃 兰州 730020;
    2.兰州大学大气科学学院,甘肃 兰州 730000;
    3.甘肃省兰州市气象局,甘肃 兰州 730020
  • 通讯作者: 刘维成(1984-),男,汉族,甘肃镇原人,工程师,硕士,主要从事雷达资料应用及数值预报工作.E-mail:gsliuwc@163.com
  • 作者简介:刘维成(1984-),男,汉族,甘肃镇原人,工程师,硕士,主要从事雷达资料应用及数值预报工作.E-mail:gsliuwc@163.com
  • 基金资助:

    中国气象局气象关键技术集成与应用重点项目(CMAGJ2013Z09)、公益性行业(气象)科研专项(GYHY201206029)及中国气象局预报员专项(CMAYBY2012-063)共同资助

Abstract:

Based on QPE product from SWAN system and precipitation data from automatic stations in Gansu Province,the effect of QPE products during the flood season from 2010 to 2012 was tested. The results showed that the estimated values of QPE products for 0.1 ~ 0.9 mm precipitation were slightly bigger than the observation,while to other grades precipitation those were smaller,and the under valuation was obvious as the magnitude of precipitation increase. For less than 50 mm precipitation,the error of QPE products was small,and the average error was less 2.0 mm. But it appeared a sharp increase for more than 5.0 mm precipitation,the error increased with the intensity of precipitation,and the effect of decreasing with the increase of rainfall intensity was more noticeable. The hit rates of estimation values when precipitation was less than 2.5 mm were high,but decreased rapidly with the increase of precipitation.And the responses of the hit rate with precipitation estimation absolute error to less than 10.0 mm precipitation was good,similarly reduced quickly with the increase of rainfall.The altitude,longitude and latitude affected greatly the QPE product error in Gansu.A simple and multiple variables weighted regression equation to correct the QPE product were respectively established based on three effect factors.Through application test in 2012,the effect of the method on QPE product correction was obvious.

Key words: QPE, error statistics, correction method

摘要:

利用SWAN系统QPE产品资料和自动站降水量,对甘肃省2010~2012年汛期QPE产品的应用效果进行检验。结果表明:QPE产品在0.1~0.9  mm降水量级上存在较小的高估,而在其它降水量级上均表现为低估,且随着降水量级的增加低估的越为明显;对于5.0  mm以下降水,QPE产品估测误差较小,平均误差<2.0 mm,但对于>5.0 mm降水会出现估测误差跃增现象,且随着降水强度的增加误差增大,估测产品的效果明显减弱;降水估测值命中率POD1在降水量<2.5 mm区间效果较好,但随着降水量的增加其命中率迅速减小,而降水估测绝对误差命中率POD2对于<10.0 mm降水的响应较好,同样地随着降水量的增加其命中率快速减小;海拔高度经度和纬度是影响甘肃地区QPE产品误差的主要因子,利用影响降水分布较大的3个因子分别建立一元多元加权回归方程,在此基础上对QPE产品进行订正,通过2012年的应用检验表明该方法对QPE产品的订正效果明。

关键词: QPE, 误差统计, 订正方法

CLC Number: