hide
Free keywords:
-
Abstract:
In this study, rainfall-runoff simulation in the Mekong River basin was conducted using the SWAT model with satellite precipitation products (TRMM, GSMaP, PERSIANN-CDR) and gauge-based precipitation products (APHRODITE, GPCC). Four water level stations, Luang Prabang, Pakse, Stung Treng, and Kratie, which are significant outlets of the main Mekong river, were selected for rainfall-runoff simulation. The hydrologic model parameters were calibrated using APHRODITE as true observations for the period from 2001 to 2011 and runoff simulations were verified for the period from 2012 to 2013. In addition, using the ConvAE, a convolutional neural network model, spatiotemporal correction of original satellite precipitation products was performed, and rainfall-runoff performances were compared before and after the correction of satellite precipitation products. The actual satellite precipitation products and GPCC showed a quantitatively under or over-estimated or spatially very different pattern compared to APHPRODITE, whereas, in the case of satellite precipitation products corrected using ConvAE, the spatial correlation was dramatically improved. Likewise, the runoff simulation results using the bias-corrected satellite precipitation products for all the outlets have significantly improved accuracy than the runoff results using the original satellite precipitation products. Therefore, the ConvAE technique presented in this study can be used to generate spatiotemporal varying grid precipitation data sets for large basin hydrologic modeling.
