ABSTRACT
A common assumption of ground water models formulated using a block-centered finite-difference method is that a well is located at the center of a cell regardless of its actual location. Due to this assumption, errors are introduced in the spatial distribution of simulated heads. This paper presents an alternative approach for assigning the pumping rates of wells that are located off cell centers. This approach consists of assigning the pumping rate not only to the cell in which the well is located but also to adjacent cells, taking into account the length of the well screen, the hydraulic conductivity, and the distance from the well to the center of its cell. The advantage of this alternative approach over the conventional one is illustrated with a test problem of a synthetic aquifer. Statistical measures of error indicate a much better model fit when pumping rates of wells are distributed over several cells.
