摘要
利用重庆气象局CINRAD/SA气象雷达降水回波资料和相应地区的地面雨量站资料,基于径向基函数神经网络,建立雷达定量估测降水模型,将其用于地面降水估测。作为比较,同时以变分法得到的Z-R关系式估测所得降水。经二者对比试验结果表明:建立的雷达定量估测降水模型的估测精度和稳定性要明显优于Z-R关系式,能较好地反映降雨的真实情况。
Based on RBF neural network, the radar quantitative precipitation estimation model is established through using CINRAD/SA radar echo rainfall data from Chongqing meteorological bureau and the corresponding ground station data for estimating surface rainfall. As a comparison, Z-R relation is used to estimate rainfall by the variational method at the same time. The comparison experiment results show that the established radar quantitative precipitation estimation model is much better than Z-R rela- tion both in accuracy and stability, and it can reflect the real situation of rainfall better.
出处
《气象科学》
北大核心
2015年第2期199-203,共5页
Journal of the Meteorological Sciences
基金
国家自然科学基金资助项目(61202138)
关键词
大气探测
雷达估测降水
RBF神经网络
Z-R关系
估测精度
Atmospheric sounding
Radar rainfall estimation
RBF neural network
Z-R rela-tion
Estimation accuracy
