人工神经网络反演算法反演北京地区湿度廓线

Retrieval of Humidity Profiles of Beijing Using Ack-propagation Neural Network

微波湿度计(MWHS)是风云三号卫星的主要有效载荷之一,其频率为150GHz(双极化)和183.3lGHz(三通道),分别用于探测大气窗区和水汽廓线。采用垂直于飞行方向的交轨扫描方式,科学目标是探测大气湿度的垂直分布。本文简要介绍微波大气湿度探测的基本原理,阐述了湿度廓线反演的各种方法,着重运用人工神经网络算法,反演北京地区的大气湿度的垂直分布。结果表明,运用风云三号微波湿度计数据,大气湿度垂直分布反演值与已知的探空数据相比,虽有误差,但在可接受范围之内,反演结果可应用于气象,通信等领域,并具有重要应用价值。

Microwave humidity sounder（MWHS） is one of primary instruments aboard on FY-3 satellite.It consists of a five channel microwave radiometer.The channels are centered on 150GHz,150GHz and 183.31±1GHz,183.31±3GHz,183.31±7GHz,known as channels 1 to 5 respectively.The first two of these channels fall in an atmospheric window（affected mainly by the water vapor continuum）,the three high frequency channels are centered on a water vapor absorption line.The instrument＇s antenna scans in cross-track type,sampling 98 earth-views within of nadir,each separated by.According the principle of atmospheric attenuation and radioactive transfer equation,the humidity profiles of Beijing can be retrieved using neural network.Known the information of FY-3 MWHS,the simulation results show that it has a small error which can be accepted.Also,the humidity profiles can be used in meteorological,agricultural and other fields.