| 基于Shepard和OI方法对雨量计逐时资料的分析 |
| |
| 引用本文： | 李春晖,梁建茵.基于Shepard和OI方法对雨量计逐时资料的分析[J].应用气象学报,2010,21(4):416-422. |
| |
| 作者姓名： | 李春晖 梁建茵 |
| |
| 作者单位： | 中国气象局广州热带海洋气象研究所，中国气象局热带季风重点实验室，广州 510080 |
| |
| 基金项目： | 广东省科技厅项目，中国气象局气象新技术推广项目 |
| |
| 摘 要： | 引入地形影响效应的日降水量气候分析场,分别运用Shepard和最优插值(OI)两种插值方法对广东和广西2007年5月20日—8月30日汛期小时雨量计降水量进行插值,得到0.125°×0.125°分辨率的规范化网格资料。结果表明:无论是直接插值还是用降水比率(地形影响效应的日降水量气候分析场)插值,两种方法均能很好地体现广东和广西雨量计站点观测降水的季内变化和日变化特征。虽然用直接插值方法比用降水比率插值方法得到的降水空间分布更为平滑,但估值精度没有用降水比率插值方法高。通过交叉检验进一步表明,总体上OI方法优于Shepard方法,而考虑地形影响效应的降水比率OI方法为最优,能有效提高相关性,并减少均方根误差和系统误差。
|
| 关 键 词： | Shepard插值 最优插值 降水比率 交叉检验 |
| 收稿时间： | 6/2/2009 12:00:00 AM |
Gauge Hourly Observations by Shepard Interpolation Method and Optimum Interpolation Method |
| |
| Li Chunhui and Liang Jianyin.Gauge Hourly Observations by Shepard Interpolation Method and Optimum Interpolation Method[J].Quarterly Journal of Applied Meteorology,2010,21(4):416-422. |
| |
| Authors: | Li Chunhui and Liang Jianyin |
| |
| Affiliation: | Tropical Monsoon Key Laboratory, Institute of Tropical and Marine Meteorology, CMA, Guangzhou 510080 |
| |
| Abstract: | The precipitation is one of the most important meteorological factors that impact on human activities. The high resolution precipitation products are useful on weather and climate monitoring, calibration of numerical model and hydrologic prediction. Due to the asymmetry and discontinuity of precipitation in spatial and temporal distribution, the excessive and absent asymmetric rainfall resulted in the meteorological disasters such as flood and drought. In order to avoid and reduce these disasters, it is important to know the amount and distribution of precipitation exactly. Using the rainfall recorder observation is the most direct, effective and common method to obtain precipitation information. But the rainfall recorder data only represent several isolated stations information, not the continuous grid data. In order to obtain the standard grid data, it is necessary to process the original rainfall data with math methods. A lot of interpolation methods have been proposed, such as Cressman, Barnes, Gandin, Shepard, Hulme. Bussieres and Hogg as well as Chen have compared and verified the interpolation methods in application of the daily and monthly rainfall analysis. It has been concluded that Gandin optimal interpolation possesses much more exactly, effectively and stably. But for smaller time scales, what math interpolation can be used to obtain a better results and how to set the correlative parameters needs further researches. Using the Parameter elevation Regressions on Slope Model (PRISM), an objective analysis of gauge hourly regional observation precipitation in Guangdong and Guangxi from 20 May to 30 August 2007 with the spatial resolution of 0.125°×0.125° by Shepard interpolation method are implemented, the results are compared with optimum interpolation method too. The comparison analysis shows that both the results can represent the inter seasonal and diurnal variation of gauge regional mean precipitation well between direct interpolation and PRISM interpolation. Although the spatial distribution of precipitation gained by the direct interpolation is smoother than that by PRISM interpolation, but the estimated precision of direct interpolation is lower than that of PRISM interpolation. The cross validation indicates that the accuracy of precipitation products by the optimum interpolation is better than that by Shepard interpolation. The PRISM optimum interpolation is an optimal method that can improve correlation and reduce root mean square error as well as system error. |
| |
| Keywords: | Shepard interpolation optimum interpolation precipitation ratio cross validation |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《应用气象学报》浏览原始摘要信息 |
|
点击此处可从《应用气象学报》下载全文 |
|
