中国农业气象 ›› 2017, Vol. 38 ›› Issue (06): 342-352.doi: 10.3969/j.issn.1000-6362.2017.06.002

• 论文 • 上一篇    下一篇

不同订正方法下CFSR再分析数据在天山西部山区径流模拟中的应用

高瑞,穆振侠,彭亮,周育琳,尹梓渊,汤瑞   

  1. 新疆农业大学水利与土木工程学院,乌鲁木齐 830052
  • 收稿日期:2016-10-18 出版日期:2017-06-20 发布日期:2017-06-17
  • 作者简介:高瑞(1992-),硕士生,主要从事水文水资源研究。E-mail:grui92@163.com
  • 基金资助:
    国家自然科学基金项目(51469034;51209181;51569031);新疆自治区地方公派出国留学成组配套项目(XJDF201307);新疆水文学及水资源重点学科基金项目(xjswszyzdxk20101202)

Application of CFSR Data under Different Correction Methods in Runoff Simulation in Western Tianshan Mountains

GAO Rui, MU Zhen-xia, PENG Liang, ZHOU Yu-lin, YIN Zi-yuan, TANG Rui   

  1. College of Water Conservancy and Civil Engineering,Xinjiang Agricultural University,Urumqi 830052, China
  • Received:2016-10-18 Online:2017-06-20 Published:2017-06-17

摘要: 以天山西部山区喀什河流域为典型研究区,基于1990-2000年的水文气象站点实测资料、2005年喀什河流域遥测降水数据及各时间段对应的CFSR(Climate Forecast System Reanalysis)再分析数据,通过实测站与其最近CFSR站多年月平均降水的比值,CFSR站点处2005年的年降水量与同期遥测年降水量的比值,以及基于实测站点降水数据和各CFSR站点与实测站点最近的CFSR站点的空间关系3种方法对CFSR降水数据进行订正。借助VIC模型,分析不同订正方法下研究区径流的模拟效果。降水时间序列分析结果表明,CFSR降水数据能够较好地在时间和空间上反应喀什河流域内月降水的变化情况。径流模拟结果表明,(1)CFSR降水数据的不同订正方法对径流结果影响较为明显;(2)基于实测站点降水数据以及各CFSR站点与实测站点最近的CFSR站点的空间关系订正的所有CFSR降水,在径流模拟中表现最好,模型率定期和验证期的模型效率系数均在0.80以上,多年径流相对误差在15.0%以内。该订正方法不仅考虑了降水在时间维度上的变化,还考虑其在地理空间维度的变化,较其它两种方法在径流模拟中具有更好的应用前景。

关键词: CFSR降水数据, 降水订正方法, VIC模型, 稀缺资料地区

Abstract: Kashi river basin, the west part of Tianshan Mountain, was chosen for this study. Based on the data of hydro-meteorological station from 1990 to 2000, telemetry station in 2005 and CFSR station from 1990 to 2000 and 2005, three correction methods were used to correct the precipitation of CFSR stations: the ratio between the mean monthly precipitation of Jilintai station and the mean monthly precipitation of No.11 CFSR station from 1990 to 2000 was the first method to correct the CFSR precipitation data; the ratio between the amount of telemetry precipitation and the amount of CFSR precipitation in 2005 at each CFSR station was the second method; the correlation between No.10 CFSR station and each CFSR station and the precipitation data of Wulasitai station was the third method. Then the effects of runoff simulation were analyzed based on the VIC model. After the analyzing of time series of precipitation, the features of CFSR precipitation in time scale and spatial scale are same to the features of real precipitation. The results of runoff simulation showed that, (1) different methods to correct CFSR precipitation have obviously different effects on runoff simulation. (2) Compared with other methods to correct CFSR precipitation, the way which used the correlation between No.10 CFSR station and each CFSR station and the precipitation data of Wulasitai station is the best one among other methods from the effects of runoff simulation. The efficiency coefficients of this model based on the best correction method in the calibration and verification period are both above 0.80, the relative errors of runoff are both under 15.0%. The best correction method in this study takes two factors - the change of precipitation in time scale and spatial scale - into consideration, and it has a better application in the numerical simulation of runoff than other methods, especially in scare data area.

Key words: CFSR precipitation data, Precipitation correction method, VIC model, Scarce data area