Reports

Report Number: 20
Year: 1981
 

A One-Dimensional, Finite Element Salt Water Intrusion Model

No abstract was published. A summary of the Introduction and Results follows.

This report describes a one-dimensional, finite element salt water intrusion model. Such a model can be very useful in understanding local features of an extensive aquifer system when constrained by data acquisition, computer time and cost. This computer program has been written in FORTRAN. The nodes are numbered consecutively from left to right, and the elements are numbered so that they are the same as the node number at the left-hand end of the element. A pump can be located at any node in the system. The lower boundary can be of any general shape, with its elevation specified at each node. Aquifer properties are read in as a function of the element number. Both confined and unconfined flows can be handled in the program, as well as time-dependent situations. Three types of boundary conditions can be specified at either end. Type I specifies the variation of the head at the boundary as a function of time. Type II specifies the discharge at the boundary as a function of time. Type III is a mixed boundary condition, where q = K(Φfs).

The program code uses a time interval of Δt. The matrix of equations can be solved in a fully implicit condition (θ = 1.0), using the Crank- Nicholson approximation (θ = 0.5), or for any value of θ in between. A subroutine that solves a banded matrix using the Gauss Elimination technique is also provided in the code. Two variables – the freshwater head and the saltwater head – are solved for at each node. If the aquifer is such that either a saltwater toe develops along the lower boundary, then in the region where only freshwater is supposed to exist, a very thin (0.25 ft) layer of saltwater is also assumed to exist. The same is true in the case of freshwater. The location of the toe is obtained by extrapolating the interface from the adjacent element and determining where it intersects the impervious boundary.

The results of the program were compared with two analytical results – the steady interface flow in (1) an unconfined aquifer having a phreatic surface with precipitation, and (2) a confined aquifer with constant freshwater discharge. The computed results were within 1% of both analytical models' results.

Author(s):
Dinshaw N. Contractor