Environmental Engineering Reference
In-Depth Information
A more recent example with a gypsum-related issue is Caspe Dam in Spain. Cordova
and Franco (1997) cite dissolution of gypsum in the foundation as the cause of a large
leakage and damage to this 55 m high embankment dam in 1989 during its second filling.
Another is Tarbela Dam in Pakistan, where the right abutment rocks are intensely frac-
tured and include gypsum and “sugary” limestone which is friable and erodible. Grouting
has been only partially effective in controlling flows which have high contents of dissolved
solids. Additional drainage adits have been required (Amjad Agha, 1980).
Reports on the failure of Quail Creek Dike (James et al., 1989; O'Neill and Gourley,
1991) provide useful details on the weathered profile developed on the interbedded
dolomite, siltstone and “silty gypsum” which formed the foundation of that 24 m high
embankment dam. Its failure was not related to any post-construction solution effects, but
was due largely to the failure during design to recognize and take into account the gaping
joints in the near-surface rock produced by the past solution of gypsum.
3.8.2
Guidelines for dam construction at sites which contain evaporites
Predictive models derived by James and Lupton (1978) for the rates of dissolution of gyp-
sum and anhydrite have been discussed briefly above in Section 3.7.4. James and Lupton
also provide guidelines for the investigation of dam or reservoir sites where these miner-
als are present and indicate the kinds and levels of risks associated with construction and
operation at such sites. They point out that it is important to know the chemical compo-
sitions of the waters involved, because the solution rates of both gypsum and anhydrite
are increased by the presence of sodium chloride, carbonate and carbon dioxide. They
note that conglomerates which are cemented by gypsum or anhydrite “can produce a
material which is potentially very dangerous” if small proportions of the cement are
removed by solution. They conclude that the risks of accelerating solution effects are
higher with anhydrite than with gypsum and that an “efficient cutoff” is the only practi-
cal method to reduce seepage velocities to values low enough to provide “complete
safety” against solution effects in “massive anhydrite”. They also warn about the possi-
bility of anhydrite converting to gypsum and expanding, with the potential for heave.
James and Kirkpatrick (1980) confirm the conclusions of James and Lupton and pro-
vide further predictive data for gypsum and anhydrite and comparable data for halite (see
Tables 3.5
and
3.6
)
. They stress the need for chemical tests on drilling water and cores,
when halite is suspected, and special sampling procedures if halite is found. They con-
clude “It would be most unwise to build a dam on massive halite. Unconsolidated strata
containing sodium chloride may be unavoidable and control measures are feasible, if
costly”.
The authors acknowledge the valuable work carried out by James and Lupton (1978),
James and Kirkpatrick (1980) and James (1992) and endorse their views, subject to the
following:
-We would recommend extreme caution to those considering dam construction at sites
where thick beds of anhydrite occur at depths shallower than 150 m;
- As for sites on carbonate rocks (see
Section 3.7.6
)
we would have some reservations
about the effectiveness of cement grouting at sites containing thick beds of evaporites
severely affected by past weathering and solution. Special attention would be needed to
seepage monitoring and facilities for re-grouting during operation of the dam;
- The possibility that filter zones could become cemented with gypsum needs to be
considered;
- It is our view that engineers intending to build at sites containing evaporites would be
wise to get advice from persons with special knowledge of the evaporite minerals, and
from others with past experience of construction at such sites.
