If the
well were only open to one layer, the discharge, Q, would be specified in that
layer.
Given that this well is defined as open to all the layers,
a common approach is to divide the total Q with the Q in each layer proportional
to the transmissivity, T, of each layer.
An alternative approach is to define a small
grid cell to represent the wellbore. It is assigned a K that is numerous orders
of magnitude (without sacrificing model stability and mass balance) greater than
the formations, and storage parameters of 1.0. The discharge is then withdrawn
from the lowest cell in which the well is open. Of course, the adjacent grid cells
must gradually increase in size away from the well
Worse yet, if the top layer were to "go dry" then that portion of the discharge
would be omitted from the simulation.