我国西南山区降雨侵蚀力时空变化趋势研究

降水是一种常见的自然现象，在气候调节中起着极为重要的作用. 同时，由于降水在空间、时间上的不均匀性，会发生诸如洪涝、干旱等自然灾害，给人类的经济、生活等造成严重影响，为了防治这类自然灾害，需要我们进行准确而及时的降水监测. 本文使用人工智能的手段，将无线电在传播过程中由于降水引起的衰减作为研究对象，尝试从不同的衰减中找到降水量信息，进而达到降水监测的目的. 实测结果表明，使用通信基站作为信号源，以数据采集时的标准的气象降水信息作为标签，神经网络成功地学习到了衰减特征，且实验准确率在95%左右. 这说明，使用无线电传播链路结合人工智能的方法进行降雨监测是一种新型的降雨测量方法，具有一定的研究意义.

Temporal and spatial variations of rainfall erosivity in southwest China from 1960 to 2009

Precipitation is a common natural phenomenon that plays an important role in climate regulation. At the same time, due to the spatial and temporal heterogeneity of precipitation, natural disasters such as floods and droughts often occur, which can seriously damage the economy and livelihood of humans. We use AI technology to study the attenuation caused by precipitation during the propagation of radio wave, trying to find the precipitation information from different attenuation, and then achieve the purpose of precipitation monitoring. The experimental results show that the neural network successfully learns the attenuation characteristics using the communication base station as the signal source and the standard meteorological precipitation information at the time of data collection as the label, and the experimental accuracy is around 95%. This suggests that the use of radio links combined with AI for rainfall monitoring is a novel approach to rainfall measurement and has some research implications.