March, 1982                     -29-                    786085
The seepage transport algorithm contains dispersion
and retardation parameters which may be calibrated by com-
paring predicted seepage transport to observed.      This is
complicated, however, by the scale dependency of the co-
efficient of dispersivity.     Also, such observed data is
seldom available requiring the use of values from the lit-
erature.
In cases where calibration is nonexistent or weak,
sensitivity studies should be made to determine the impact
of the parameter values selected on the model results.
Quite often some parameters are found to have very little
effect over the range of required results and typical val-
ues may be used. Also, inferred parameters should be cho-
sen in such a manner to insure conservative results.
The model calibration for the Crandon site was based
on a detailed geologic    investigation(1), pump test per-
formed in June, 1980(2) and Dames and Moores environmental
baseline study(10).   This site specific data provided the
initial set of data from which the calibration effort be-
gan. Simulated contours were compared to observed contours
to determine the accuracy of the calibration. The observed
groundwater contours used are those presented in refer-
ence 11 which reflects all data in the area gathered to
date and a comprehensive set of piezometer readings in
September, 1980.
Several parameters were based upon the above listed
sources of information and were not varied during the cali-
bration process. These were the average horizontal hydrau-
lic conductivity of the coarse grained stratified drift of
35 feet/day and storage coefficients of the coarse grained

Golder Associates