The behavior of groundwater with dispersion in coastal aquifers

Abstract

A three-dimensional steady-state hydrodynamic dispersion model is used to simulate seawater encroachment in the confined aquifers in the estuaries of the Naka and Kiki Rivers in Japan. Two expressions of the dispersion coefficient are considered; one is constant over the entire region of the aquifer and the other is dependent on the flow velocity of the groundwater. The magnitudes of the constant dispersion coefficients in the horizontal and vertical directions, D xx and D zz , as well as the longitudinal and lateral dispersivities, a L and a T, are determined so as to reproduce the regional distributions of salt concentration in the confined aquifers in both estuaries. It is found that D xx = 5 cm 2s −1, D zz = 5-0.5 cm 2s −1 and a L = 1000–1250 m, a T = 100–125 m in the estuary of the Naka River; and D xx = 0.2 cm 2s −1, D zz = 0.2–0.02 cm 2s −1 and a L = 200 m, a T = 200-20 m in the estuary of the Kiki River. Examining the local distributions of the dispersion coefficient computed from the dispersivity and velocity fields of groundwater in both estuaries, the same value as estimated in the analysis with the constant dispersion coefficient is located in the middle layer of the aquifer. In the estuary of the Naka River, the piezometric surface predicted with the dispersion model with the velocity-dependent dispersion coefficient is almost the same as that predicted with the dispersion model with the constant dispersion coefficient and they are 5 10% lower than that predicted with the interface model (Kakinuma et al., 1984). They are, however, about 1.3 times the observed one.