DECEMBER 16, 1993
.LETTER FROM PETER F. ANDERSON, P.E., GEOTRANS, INC.
46050 Manekin Plaza Suite 100 Sterling, Virginia a 20166
703 444 7000
December 16, 1993 .... ----.-..-
Ms. Barbara Vergara ....-... ........
St. Johns River Water Management District I
P.O. Box 1429 .-
Palatka, Florida 32178
Once again, I enjoyed meeting with you, your staff, and the
other consultants of the Needs and Sources panel two weeks ago. I
have several comments I would like offer as closure to my
involvement in this effort.
1. I must restate my opinion that the currently proposed
methodology relies too heavily on the numerical models for
determining critical areas. Although the models are the best
available technology, they are not the only technology available
and they do have their shortcomings. The sharp lines on the wall-
size color figures of model results certainly were impressive, but
how representative of actual conditions (or future conditions) are
they? They were derived from a conceptual model that has
significant uncertainty with fundamental aspects of the problem:
Where is the lower concentration boundary? What is the thickness
and conductivity of the Middle semi-confining unit? What are the
concentrations in the Lower Floridan and Middle semi-confining
unit? Are current conditions at a stable equilibrium as
represented in the model? I believe that SJRWMD should use the
models; however, another methodology should be used for
verification or support.
In my last correspondence, I described the SWISS method,
which was a ranking system based on causal factors, geographic
factors, and historical data. Whether SJRWMD uses this method,
some variation of it, or another method, it should be able to
support the modeling based on semi-independent computations.
2. I am concerned that the quantification of sensitivity
(uncertainty) was not carried through to the predictions. My
impression is that a sensitivity analysis on various parameters
has been performed with the steady-state calibration. This is
useful in understanding which parameters are most influential -on
model results, but it does not quantify how the prediction is
affected. For instance, we know that the Middle semi-confining
unit leakance is an influential parameter that causes the
equilibrium position of the saltwater front to shift upwards if
Ms. Barbara Vergara 2 December 16, 1993
increased. However, if the leakance really is on the high end,
how does this affect the movement of the interface in a stressed
condition? Even though time consuming, requiring several
recalibrations, this type of sensitivity analysis should be
conducted on the influential parameters and boundary conditions in
order to have any confidence in statements regarding future
interface movement. It may not be necessary or desirable to
perform the analysis on individual parameters. Perhaps worst case
and best case analyses can be conducted.
3. I do agree that the 20-year results may be used for
management decisions as long as the 50-year results are presented
and factored into the decisions and for planning/update purposes.
This is, of course, contingent upon alleviation of my two concerns
enumerated above. I also recommend running the models to a steady
state condition in addition to the 20 and 50-year simulations.
This will determine how close the 20 and 50-year results are to
the ultimate magnitude of the problem.
4. I do not understand the controversy over whether SJRWMD
should concern themselves with a rate of interface movement or
absolute distances of movement. Since distance moved is merely
rate multiplied by time, the two seem equivalent to me. It should
be recognized that effective porosity is directly proportional to
rate and amount of movement over a time increment. Because the
uncertainty associated with effective porosity is on the order of
100% (between 0.15 and 0.3), rate and amount of movement of the
saltwater interface may have an uncertainty of 100% associated
with them (e.g., 30 ft/yr or 15 ft/yr).
5. The fundamental question that was posed to the panel -
what constitutes an unacceptable condition? was never resolved.
I do not believe that a rate or amount of movement can be judged
unacceptable until it is determined what damage to the system is
unacceptable. A 500 mg/l increase in chloride concentration or 50
ft upward movement of the interface may be unacceptable in areas
near public or domestic wells, but may have little consequence in
undeveloped areas. To effectively manage, SJRWMD must make
decisions on which areas are necessary to maintain at current
conditions and which are not critical. One way I would suggest is
to develop histograms for the number of public or domestic wells
(or alternatively, quantity of withdrawn water) subject to certain
increases in chloride content. The attached histograms illustrate
that in case 1, only a limited number of wells are affected by
future increase in pumpage; in another area (case 2), a
significant number of wells are affected. Even though the rates
or ultimate movement of the interface may be identical for the two
cases, case 1 may be acceptable while case 2 is not. This type of
analysis can be conducted by simply noting the closest node in the
model for each production well and determining the increment of
Ms. Barbara Vergara
concentration change that each node falls into. The number of
affected wells for each concentration change interval may then be
Having made these five points,-I would like to close my
discussion. Thank you for the opportunity to be a part of this
interesting project. Please contact me if I can be of further
assistance to the SJRWMD on this or other problems. Best wishes
for a happy holiday season.
Peter F. Andersen, P.E.
december 16, 1993
BY DATE 1___ 6 PROJECT S J -A N C +,
CHKD. BY DATE So-5C- \
SHEET NO. OF
0-to blc-o 160-so50o 560
500 1 6-
0-10 to-0too 00oo-0Sto 500ooo
I I ff I I I ,