<![CDATA[Papua, an equatorial region in Indonesia, faces unique geographical and natural challenges, including heavy annual rainfall. This heavy rainfall increases flooding risks and impacts infrastructure, the economy, and daily life. Despite the importance of rain gauges for monitoring floods and climate change, Papua's difficult geography and limited transportation infrastructure hinder their installation and maintenance. In this work, we investigate a deep learning one-dimensional convolution neural network (1DCNN) model to estimate rainfall intensity using energy per symbol to noise power density ratio (Es/No) of the signals received from a communication satellite signal coupled with additional data representing satellite daily movement. The findings of this study demonstrate that the performance of the proposed model has a higher accuracy for moderate to heavy rainfall than for light rainfall. The NRMSE values for light rain, moderate rain, and heavy rain are 47.09, 31.78, and 33.58%, respectively. These results show that this method is promising for monitoring heavy rainfall as a flood mitigation effort. However, there is still room to improve the accuracy of the estimation such as using other secondary data that is highly correlated with rain at the satellite transceiver location.]]>
Copyrights © 2024
