Integration of Satellite Precipitation Data and Deep Learning for Improving Flash Flood Simulation in a Poor-Gauged Mountainous Catchment

Xuan Tang,Hongxia Li,Zhaorui Yin,Guanghua Qin,Li Guo

doi:10.3390/rs13245083

Abstract

Satellite remote sensing precipitation is useful for many hydrological and meteorological applications such as rainfall-runoff forecasting. However, most studies have focused on the use of satellite precipitation on daily, monthly, or larger time scales. This study focused on flash flood simulation using satellite precipitation products (IMERG) on an hourly scale in a poorly gauged mountainous catchment in southwestern China. Deep learning (long short-term memory, LSTM) was used, merging satellite precipitation and gauge observations, and the merged precipitation data were used as inputs for flood simulation based on the HEC-HMS model, compared with the gauged precipitation data and original IMERG data. The results showed that the application of original IMERG data used directly in the HEC-HMS hydrological model had much lower accuracy than that of gauged data and merged data. The simulation using the merged precipitation in HEC-HMS exhibited much better performances than gauged data. The mean NSE improved from 0.84 to 0.87 for calibration and 0.80 to 0.84 for verification, while the lower NSE improved from 0.81 to 0.84 for calibration and 0.73 to 0.86 for verification, which showed that accuracy and robustness were both significantly improved. Results of this study indicate the advances of remote sensing precipitation with deep learning for flash flood forecasting in mountainous regions. It is likely that more significant improvements can be made in flash flood forecasting by employing multi-source remote sensing products and deep learning merging methods considering the impact of complex terrain.

Highlights

Published: 14 December 2021Flash flood is one of the most serious natural disasters around the world
The performance of the IMERG product for a poor-gauged mountainous catchment in China was assessed, and deep learning was used for precipitation data merging
The merged precipitation data, compared with the gauged data and original IMERG data, were used as inputs for flood simulation based on the HEC-HMS model

Summary

Introduction

Published: 14 December 2021Flash flood is one of the most serious natural disasters around the world. Many studies indicated that precipitation data were one of the essential inputs for hydrological modeling, and about 70–80% of the uncertainties of hydrological simulations were due to the uncertainties in precipitation data [2,6,7,8]. The commonly used precipitation data are (1) rain gauge data, the advantage of which is providing accurate point precipitation information. Many small catchments with complex topography are poorly gauged, so their spatial representativeness is deficient, which impacts the accuracy of flash flood forecasting [9,10,11]. (2) Satellite precipitation products (SPPs)—these satellite remote sensing technologies provide new ways for precipitation monitoring, which has wide spatial coverage making up for the inadequate and uneven distribution of ground precipitation observation, especially for ungauged basins [12,13,14,15]. SPPs have become potential precipitation data sources for hydrometeorological studies [16,17,18,19,20]

Objectives

Methods

Results

Discussion

Conclusion