Page 111 - Geothermal Energy Renewable Energy and The Environment

Exploring for Geothermal Systems 97

1998). Also, fluids migrating to the surface exsolve a certain proportion of their dissolved gases,
including CO , as discussed in Chapter 5. As a result, the carbonate–bicarbonate equilibria that can
2
have a strong influence on pH are likely to be representative of conditions during fluid ascent, not of
bicarbonate and pH values in the reservoir. Finally, if steam separation has taken place, partitioning
of gases such as CO into the steam phase will influence pH. Silica, too, is somewhat partitioned
2
into the steam phase. This can result in low apparent silica concentrations in the collected water
(Truesdell 1984).
Geochemical thermometers, or geothermometers, can nevertheless be constructed from empiri-
cal data that are conceptually grounded in the approach delineated above. As an example, if there
is a silica polymorph in the reservoir with which the fluid has equilibrated, the silica concentration
will be directly controlled by that polymorph. That, in turn, will exert influence on the equilibration
of other minerals with the fluid. The end result is a cascading, coupled relationship of interdepen-
dencies that will be expressed as a temperature-dependent evolution of the concentration ratios of
various solutes.
Since the 1960s, formulation of a variety of geothermometers has progressively developed a suite
of applicable relationships useful for exploring potential geothermal reservoirs. Currently more
than 35 different formulations of geothermometers have been published (see Verma, Pandarinath,
and Santoyo 2008 for a recent compilation). Each geothermometer has been formulated by numeri-
cally fitting a functional form to data sets that have known or well-constrained temperatures and
compositional data. Thus far, equations have been published for concentrations as a function of
temperature for the following systems: SiO , Na–K, Na–Ca–K, K–Mg, Na–K–Mg, Na–K–Ca–Mg,
2
Na–Li, and Mg–Li. Shown in Figures 6.6 and 6.7 are examples of computed results from different
formulations for the SiO and Na/K geothermometers using five natural geothermal systems and
2
one experimental study. Consideration of these two geothermometers allows consideration of the
utility and limitations of the geothermometer approach to exploration for geothermal resources.
In Figure 6.6, agreement between observed and computed temperatures varies significantly
among the different formulations for the SiO geothermometer. For the experimental study, quartz
2
was the controlling silica polymorph (Pang and Reed 1998), and the computed temperatures for the
quartz geothermometers are all well within + /−20°C of the actual value. The chalcedony geother-
mometer significantly underestimates the temperature, as would be expected because of the higher
solubility of chalcedony (Chapter 5). At temperatures below 150°C, the quartz geothermometer
significantly overestimates the temperature, as does the chalcedony geothermometer although the
latter does so with a smaller error. At higher temperatures, the empirical geothermometers show
a broad array of results, differing from the measured values by as much as 120°C and as little
as 5°C.
The discrepancies between measured and computed temperatures provide insight into the cave-
ats associated with empirical geothermometers, some of which are similar to the caveats associ-
ated with the Q/K approach described above. Empirical geothermometers represent a strategy for
overcoming the inherent limitations of laboratory experimental systems. Geothermal waters are,
as previously noted, compositionally very diverse. Sodium-chloride-, bicarbonate-, sulfate-, and
calcium-chloride-dominated solutions characterize water types that have been reported. These nat-
ural waters are invariably compositionally complex, requiring detailed and accurate representation
of activity-composition relationships in the mass action expressions for the solubility equations, in
order to accurately compute a temperature from a silica concentration. Since activity–composition
relationships are not generally available in a simple computational scheme but, rather, require accu-
rate and thorough compositional data for use in sophisticated computer models (discussed later),
the strategy has been developed to fit curves of an appropriate form to observed data to allow an
approximate temperature to be computed. The majority of waters (but by no means all) used to
fit curves for geothermometers are of the sodium chloride-type at near-neutral pH. As a result,
the fitted curves that are generated will have varying degrees of validity for waters for which the
compositions differ from those used in the fitting database. As well, because different authors use