摘要
用反距离加权法、普通克里格法、薄盘光滑样条函数法三种空间插值方法,对西南地区复杂山地92个气象站点30 a(1971-2000年)的温度、降水和相对湿度月平均资料进行分辨率为1.0km×1.0km的空间插值,并进行了交叉验证和随机站点验证。结果表明:薄盘光滑样条函数法的插值精度最优,特别是温度和降水的效果尤其明显;对于三种插值方法,相对湿度的插值精度在夏季差异不明显。在样本数选取上,样本数的增加使温度和湿度的插值精度有较大提高,但对降水插值精度的影响不明显。
Three kinds of interpolation methods including Inverse Distance Weight (IDW) , Ordinary Kriging (OK) and Thin Plate Smoothing Spline (TPS) were used to interpolate monthly average air temperature, precipitation and relative humidity at the spatial resolution by 1.0 km × 1.0km in 1971 - 2000 in 92 meteorological stations, which covered the mountainous areas of the Southwest China. The results were verified by the cross validation and the random validation. The results showed that TPS was of the best accuracy of three methods for temperature, precipitation and relative humidity, significantly for air temperature and precipitation. The difference of the accuracy of three interpolation methods for the aestival relative humidity was not significant. However, the interpolation accuracy for the temperature and relative humidity were greatly improved as the number of samples increased, but it was not obviously for the precipitation.
出处
《中国农业气象》
CSCD
2008年第4期458-462,共5页
Chinese Journal of Agrometeorology
基金
国家自然科学基金(40765004)
中国气象局科技发展专项基金(CMATG2008S10)
中国气象局成都高原气象开放实验室基金(LPM2007012)
关键词
山地气候
空间插值
薄盘光滑样条函数
农业气象
西南地区
Mountain climate
Spatial interpolation
Thin plate smoothing spline (TPS)
Agri-meteorology
Southwest China