资源科学 ›› 2018, Vol. 40 ›› Issue (8): 1622-1633.doi: 10.18402/resci.2018.08.12

• • 上一篇    下一篇

基于统计检验的降雨侵蚀力简易计算模型比较

马小晴1,2(), 郑明国1()   

  1. 1. 中国科学院地理科学与资源研究所,北京100101
    2. 中国科学院大学,北京 100049
  • 收稿日期:2017-11-15 修回日期:2018-02-27 出版日期:2018-08-25 发布日期:2018-08-10
  • 作者简介:

    作者简介:马小晴,女,吉林白山人,硕士生,主要从事降雨侵蚀力研究。E-mail:mxq.15s@igsnrr.ac.cn

  • 基金资助:
    国家自然科学基金资助项目(41671278);水利部黄土高原水土流失过程与控制重点实验室开放课题基金项目(2016002)

Statistical evaluation of proxies for the R factor of the Universal Soil Loss Equation

Xiaoqing MA1,2(), Mingguo ZHENG1()   

  1. 1. Institute of Geographic Sciences and Nature Resources Research, Chinese Academy of Sciences, Beijing 100101, China
    2. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2017-11-15 Revised:2018-02-27 Online:2018-08-25 Published:2018-08-10

摘要:

降雨侵蚀力因子(R)的准确估算对提高侵蚀预测的精度具有重要作用,但R的计算需要高时间分辨率的降雨过程资料。基于日、月、年降雨资料,目前已开发了众多的R值简易计算模型(PR)。本文通过Meng’s检验比较了不同PR模型结果与R的相关性是否存在显著性差异,并以此为基础评估了国内外15种PR模型在中国东部地区的适用性。结果表明,15种PR模型计算结果与R均显著相关(r >0.59,p<0.03),表明这些方法均适用于中国东部地区,但由于存在系统偏差,许多模型需要按文中建立的回归关系式校正后得到更优的估算结果。相关系数的统计学比较表明,日降雨PR模型的计算结果与R相关性明显优于年和月降雨模型,而年和月降雨模型差异不明显。所有模型中,3种基于中国数据(包括国内最常用的章文波模型和第一次全国水利普查采用的模型)和1种基于澳大利亚数据建立的日降雨PR模型的相关性并列处于最高等级(r >0.99),但该澳大利亚模型的计算偏差随纬度增高而变大,章文波模型也存在约30%的系统误差,两者校正后使用能获得更好的效果;第一次全国水利普查采用的模型则可以直接用于R的估算,不需要进行校正处理。论文最终给出了获取不同时间分辨率降雨资料时建议使用的PR模型。

关键词: 土壤侵蚀, 降雨侵蚀力, 相关系数, Meng’s检验, 统计检验, 简易计算模型

Abstract:

The accurate estimation of rainfall erosivity (R) is significant for soil erosion prediction. However, the algorithm for the R factor of the universal soil loss equation (USLE), the most widely used algorithm to calculate R, needs pluviograph data with high temporal resolution. Thus, studies have developed a large number of proxies for the R factor (PR) using daily, monthly, yearly rainfall data instead of pluviograph data. It is noted that the statistical evolutions of the PRs is desirable, which is currently lacking. This study aimed to evaluate 15 widely used PRs, including two yearly, six monthly, and seven daily ones, and further rank them in terms of their correlation with R using a stepwise Meng’s test procedure and data collected at 14 rainfall stations in the eastern China. The results indicated that the 15 PRs under examination were all significantly correlated with R (r > 0.59, p < 0.03), implying all of them can be reasonably used as an R predictor. The stepwise Meng’s test illustrated that the daily PRs were better correlated with R than the monthly and yearly PRs with four daily PRs, which included three developed in China and one developed in Australia, ranking the first (r > 0.99) between the 15 PRs. Additional examination did show that the Australia PR exhibited a relative error that increased with the latitude and thus should be calibrated by means of the regression equation we developed. One of the three Chinese PRs did not show a systematic error and can be directly used to estimate R without calibration. The two other PRs showed approximately 30% overestimation and also be calibrated by the developed regression equations. We also recommended the PRs to be used if only the monthly or yearly rainfall data is available.

Key words: soil erosion, rainfall erosivity, correlation coefficient, Meng’s test, statistical text, simplified estimating model