Integration of remote sensing and meteorological data to predict flooding time using deep learning algorithm

Accurate flood forecasting is a vital need to reduce its risks. Due to the complicated structure of flood and river flow, it is somehow difficult to solve this problem. Artificial neural networks, such as frequent neural networks, offer good performance in time series data. In recent years, the use of Long Short Term Memory networks hase attracted much attention due to the faults of frequent neural networks. In this study, it is tried to predict the daily discharge of the Aqqala station in Golestan provice, for the next three days, using Long Short Term Memory network. This network is very suitable for time series predictions, due to its special structure and ability to learn long-term dependencies. On the other hand, the desired network is stable and contains the maximum default parameters, which indicates its usability for other regions. Also, this algorithm has the ability to use topography and flow data from other stations in the region. To predict the discharge at the target station, several data combinations; the discharge data of Aqqala station alone and together with its upstream stations, the elevation model of Aqqala city and Golestan province were used as network inputs. The present research outcome was compared with simple regression network, support vector machine-regression, and frequent neural network. The results show that Long Short Term Memory network is superior to other networks with Nash-Sutcliffe Efficiency values above 91%. In future study, authors are going to to use other influential data on flood occurrence as well as network development into fully automated network.