Measuring Amazon rainfall intensity with sound recorders: data and code

Many regions still lack a network of ground weather observations, hampering effective climate monitoring and disaster management. In the Amazon basin, this occurs due to its remoteness and the challenging measurement of rainfall within the forest. Innovative rainfall estimation methods are thus requested to fulfill this gap. Our approach allows to estimate rainfall based on a sound recorder fixed on a tree trunk and supervised machine learning. The method is very promising for future weather monitoring of remote tropical areas. This dataset provides all data used in the paper “Measuring Amazon rainfall intensity with sound recorders”, as well as three sample audio recordings, used for the plotting of Figure 2. These files can also be used to extract the frequency range associated with the power spectrum density values. Contents - SMM00894_20230510_224500.wav represents a light rainfall recording - SMM00894_20230510_225500.wav represents a heavy rainfall recording - SMM00894_20230510_223500.wav represents an audio without rainfall - data used for training the models - Jupyter Notebook (code).

众多地区仍缺乏地面气象观测网络，这阻碍了有效的气候监测和灾害管理。在亚马逊流域，这一问题源于其偏远性以及森林内部降雨测量的挑战。因此，迫切需要创新的降雨估算方法以填补这一空白。本研究采用的方法基于固定在树干上的声音记录仪和监督式机器学习来估算降雨量，该方法对于未来偏远热带地区的天气预报具有极大的应用前景。本数据集提供了论文《使用声音记录仪测量亚马逊降雨强度》中使用的所有数据，以及用于绘制图2的三个样本音频录音。这些文件还可用于提取与功率谱密度值相关的频率范围。数据内容包含：- SMM00894_20230510_224500.wav 代表轻度降雨录音；- SMM00894_20230510_225500.wav 代表重度降雨录音；- SMM00894_20230510_223500.wav 代表无降雨音频；- 用于训练模型的训练数据；- Jupyter Notebook（代码）。