Abstract: The spectral estimation method based on the Fourier analysis（DFT） has been widely used in Doppler weather radar to measure rainfall velocity, and has proved to be almost optimal for small normalized spectral widths[(σN)]. However, the approach limites to small [σN][(σN<0.1)], which ground and space based Doppler weather radar satisfies. Developing a Doppler weather radar installed in the geostationary orbit satellite for the global vertical precipitation rate has become necessary. This paper extends a spectral moments estimation approach for spaceborne radar with larger values of [σN][(σN～0.2)]. By comparing with several other forms of DFT spectrum estimation methods and the Pulse Pair Process（PPP）, the simulation results show that for the same system parameters, the proposed new method has the optimal performance and better than 0.5 m/s Doppler velocity accuracy.

Key words: spaceborne weather radar, spectral moments estimation, characteristics of weather echo signal, Discrete Fourier Transform（DFT）, Pulse Pair Process（PPP）

摘要： 基于傅里叶分析（DFT）的谱参数估计方法已被广泛用于测量多普勒天气雷达降水速度，已经证明对于较小的归一化谱宽[(σN)]而言该方法几乎最优。然而该方法局限于地基和空基多普勒天气雷达中较小的[σN][(σN<0.1)]。研究安装在静止轨道卫星上的多普勒天气雷达获取全球的垂直降水速度很有必要。针对星载天气雷达较大的[σN]（[σN～0.2]），扩展了一种适用于星载场合的谱参数估计方法。对该方法性能进行了分析，与其他几种形式的DFT谱估计方法及脉冲对处理方法（PPP）比较可看出：在相同的系统参数下，提出的新方法性能最优，能够获得优于0.5 m/s的多普勒测速精度。

关键词: 星载天气雷达, 谱参数估计, 气象回波特征, 离散傅里叶变换, 脉冲对处理