中国农业气象 ›› 2013, Vol. 34 ›› Issue (01): 30-36.doi: 10.3969/j.issn.1000-6362.2013.01.005

• 论文 • 上一篇    下一篇

流域离散化对分布式水文模型径流模拟的影响

史晓亮1,2,李颖,赵凯1,张丹3,朱红雷1,2   

  1. 1中国科学院东北地理与农业生态研究所,长春130102;2中国科学院大学,北京100049;3黑龙江省齐齐哈尔水文局,齐齐哈尔161005
  • 收稿日期:2012-06-08 出版日期:2013-02-20 发布日期:2013-04-17
  • 作者简介:史晓亮(1985-),陕西宝鸡人,博士生,主要从事遥感与地理信息系统在水文学中的应用研究。Email:shixiaoliang@neigae.ac.cn
  • 基金资助:

    中国科学院重大碳专项项目(XDA05060400)

Effect of Watershed Delineation on Flow Simulation of Distributed Hydrological Model

SHI Xiao liang1,2, LI Ying1,ZHAO Kai1,ZHANG Dan3,ZHU Hong lei1,2   

  1. 1Northeast Institute of Geography and Agroecology, CAS, Changchun130102, China;2University of Chinese Academy of Sciences,Beijing100049;3Hydrology Bureau of Qiqihaer, Heilongjiang Province, Qiqihaer161005
  • Received:2012-06-08 Online:2013-02-20 Published:2013-04-17

摘要: 以嫩江上游诺敏河古城子水文站以上集水区为研究对象,利用分布式水文模型SWAT(Soil and Water Assessment Tool),对研究区进行6种子流域划分,研究不同时间尺度上径流模拟随流域离散化水平的变化,以了解输入数据空间集总程度对分布式水文模拟的影响。结果表明,在年和月尺度上,不同子流域划分对径流模拟精度影响较小,但在日尺度上对径流模拟精度有影响;不同时间尺度上流域离散化水平对出口流量的影响均较小,但地表径流随子流域数量的增加而增大,相对误差达到354%;径流模拟结果对降水量较敏感,湿润年份模拟精度更高,而子流域划分对径流模拟结果的影响在干旱年份以及夏季汛期更显著;当流域离散化达到一定水平即最小集水面积阈值为流域总面积的2%时,增加子流域划分对径流模拟结果的影响较小,但减少子流域划分则会导致模拟结果不稳定,且难以满足流域空间格局分析的要求。

关键词: 分布式水文模型, SWAT, 流域离散化, 径流, 降水

Abstract: The effect of input date aggregation on flow simulation was studied at different time scale in the catchment upstream Guchengzi hydrological station of Nuomin river, upper reaches of Nenjiang river, using the Soil and Water Assessment Tool(SWAT) with six watershed delineations, each with a different number of subwatersheds. The results showed that overall accuracy of flow output over a monthly or annual time scale was not affected from watershed delineation, but the daily output that captures hydrologic events was affected by watershed delineation. Average flow at different time scale were not affected by changes in sub watershed size, but the runoff yield increased as the increase of subwatershed amount, and the relative error reached up to 35.4%. The model performance was sensitive to precipitation, and the better simulation accuracy for wet years (above average rainfall) compared with dry years (below average rainfall) for the study area. The watershed delineation had a more profound effect on flow simulation for dry years and summer. There was a reasonable watershed delineation level for distributed hydrological model, and the critical source area accounted for 2% of drainage area. Increasing the number of sub watersheds over the level of watershed delineation couldn't significantly affect the runoff yield. Besides, reducing the number of subwatersheds would create instability of runoff prediction, and it might not fulfill the demand of spatial pattern analysis.

Key words: Distributed hydrological model, SWAT, Watershed delineation, Runoff, Precipitation