Estimating tidal effects in spring discharge: A multiscale method using correlated phenomena

Modeling groundwater flow within a coastal karst aquifer is a difficult task because of the influences of rainfall, sinking streams, lunisolar tides, and storm tides. A primary goal of modeling this flow is the accurate simulation of discharge due to rainfall and sinking streams. A necessary preliminary step to accomplish this is to remove from the discharge signal that portion resulting from the influence of tides. This paper presents a new wavelet method for separating these two signals. The statistical concept of correlation is used as a threshold parameter to accomplish the separation.