기계학습을 활용한 Digital Recursive Filter의 보완 및 WAPLE과의 비교

Baseflow is important component for understanding streamflow and water resource management. Baseflow was proven as most unpredictable component of streamflow throughout various researches. However recent method for estimating baseflow are owing to development of theoretical and computer techniques. Digital recursive filter by Eckhardt (2005) is newest technique for separating baseflow from streamflow. This paper attempted to optimize two parameters (a parameter and BFImax)of this technique and derived baseflow estimation with better understandings about watershed’s characteristic. This study proposed reinforced Digital recursive filter combining with machine learning techniques. Result is comparison between baseflow separation from pre-defined values of parameters and proposed method. And proposed showed stable and realistic results compare to result from pre-defined parameters and WAPLE. This study will let researchers to reinforce their study with accurate material and policies as well.