JOURNAL OF NATURAL RESOURCES ›› 2013, Vol. 28 ›› Issue (2): 321-327.doi: 10.11849/zrzyxb.2013.02.013

Previous Articles     Next Articles

The Confidence Coefficient of Mean Annual Rainfall Erosivity Influenced by Record Length of Rainfall Datasets

YAN Ye-chao1, YUE Shu-ping1, ZHANG Shu-wen2   

  1. 1. College of Remote Sensing, Nanjing University of Information Science & Technology, Nanjing 210044, China;
    2. Northeast Institute of Geography and Agroecology, CAS, Changchun 130102, China
  • Received:2011-06-16 Revised:2012-06-24 Online:2013-02-20 Published:2013-01-30

Abstract:

Record length of rainfall datasets is an important element which should be taken into account in the process of computing the mean annual rainfall erosivity. Based on the monthly rainfall datasets for the period 1980-2009, the annual rainfall erosivity for 601 weather stations of China was calculated using a simplified method originally proposed by Wischmeier and Smith. According to the theory of statistics, by drawing simple random samples of 30 years, 20 years, 10 years and 5 years of the datasets, the confidence coefficients of the mean annual rainfall erosivity were calculated based on the percent sampling errors of 10% and 25%. The results suggest that: 1) record length of datasets does have an effect on the confidence level for the mean annual rainfall erosivity; 2) the confidence coefficients for the mean annual rainfall erosivity vary greatly across China, and higher confidence level lies in south of the Yangtze River, eastern Tibetan Plateau and the mountainous area in southern part of Hexi Corridor; 3) when modeling soil erosion with limited rainfall data, it’s necessary to analyze the confidence level of sampling error by setting the allowable errors of the mean annual rainfall erosivity according to the research goal and predefined accuracy.

Key words: GIS, rainfall erosivity, probability and mathematical statistics, record length of rainfall data sets, uncertainty

CLC Number: 

  • S157.1