IMPROVING RAINFALL PERFORMANCE BY DECAYING AVERAGE BIAS CORRECTION VIA LYAPUNOV THEORY

Abstract

The bias correction is the main tool for improving the rainfall simulation from the model to
improve performance and increasing accuracy with observation. Since, if the good estimate and more accuracy
of rainfall simulation are crucial to helping the risk assessment policies for increasing demands from
agricultural, industrial and domestic sectors for many countries. So, the aim of this study, to improving
decaying average bias correction by using the Lyapunov theorem for simulating rainfall over Indochina
Peninsular. The time period for exampling the results were in Mar, April, and May 2015. The results were
shown a comparison between standalone model simulation results and bias correction results (Theorem 2 and
Theorem 3) as shown in time series and statistical method value. The times series results were shown the results
from bias correction improving by Lyapunov (Theorem 2 and Theorem 3) that show good estimates than the
standalone model simulation. In statistical analysis, the bias correction improving by the Lyapunov theorem
(Theorem 2 and Theorem 3) were shown the highest accuracy (MAE and RMSE) than standalone model
simulation. However, the results from the time sires and statistical analysis were guaranteed the bias correction
improving by the Lyapunov theorem that can improve the results of the model and increase more accuracy
when compared with reanalysis grid observation data.