摘要
以洪湖2007-2010年间水质因子为研究对象,利用地统计学和常用的5种插值方法(径向基函数、反距离权重、全局多项式、局部多项式和普通克里格插值法)对洪湖水质的时空状况进行分析,并用误差评价因子(MAE、MRE和RMSE)对其拟合情况进行评价。结果表明,各水质因子均具有较高的时空异质性,其中,总氮质量浓度受到温度、水深以及水循环的影响;在5种插值方法中,径向基函数取得了较好的模拟预测结果,洪湖总氮质量浓度具有不同的空间分布特征。
The water quality of Honghu Lake between 2007 and 2010 was taken as an example and the geostatistics and five commonly used interpolation methods （radial basis function, inverse distance weighted,global polynomial, local polynomial and ordinary Kriging） were used to analyze the temporal and spatial conditions of water quality of Honghu Lake. Error evaluation factors （MAE, MRE and RMSE） were employed to evaluate the fitting situation. It concluded that all factors of the water quality had high spatial-temporal heterogeneity and the concentration of total nitrogen was influenced by temperature,water depth and water cycle. Among the five kinds of interpolation methods, the radial basis function had a better predict result. The total nitrogen concentration of Honghu Lake had different characteristics of spatial distribution.
出处
《华中农业大学学报》
CAS
CSCD
北大核心
2013年第3期82-87,共6页
Journal of Huazhong Agricultural University
基金
国家科技重大水专项项目(2009ZX07631)
关键词
水质因子
地统计学
时空特征
插值模型
误差评价因子
洪湖
water quality factor
geostatistics
space-time characteristics
interpolation model
error evaluation factors
Honghu Lake
