Page 179 - The Petroleum System From Source to Trap

9. Overburden Rock, T e mperature, and Heat Flow 171

tion per unit volume per unit time. The extension of ture and thermal conductivity. The quantity that is
equation 4 to two or three dimensions is straightforward. � s�a�ed is the c nductive heat flow; there is usually an
�
In most geologic settings, however, the variation of ImpliCit assumption that advective heat transport by
temperature with respect to depth (dT I ()z) is much groundwater flow is absent. Because the holes used in
greater than lateral variations, and a one-dimensional traditional heat flow studies tend to be relatively shallow
approximation is justified. (-10�00 m), it is necessary to obtain continuous, high­
precisiOn temperature logs to derive accurate estimates
of thermal gradients. However, opportunities to log
SOURCES OF HEAT boreholes in sedimentary basins are scarce. Most wells
Roughly 40% of surface heat flow on the continents are drilled for the purpose of petroleum exploration and
comes from a superficial layer of radioactively enriched are either producing or cemented shut. The temperature
crystalline rocks about 1 0 km thick (Pollack and data usually available for analysis are bottom-hole
Chapman, 1977). The remaining 60%, or reduced heat temperatures (BHTs) measured during the geophysical
flow, comes from a combination of radioactive sources in logging of oil and gas wells. BHTs represent direct
the lower crust and upper mantle, as well as a convective measurements of temperature at depths (1--6 km) much
flux into the base of the thermal lithosphere. The half-life greater than those normally associated with traditional
of common heat-generating elements (K, U, and Th) is on heat flow studies in shallow (100-600 m) holes. In this
the order of 109 yr or greater; thus, the radioactive sense, use of BHTs obviates the need to infer tempera­
component of heat flow has not changed appreciably tures at depth indirectly from shallow measurements
since the Precambrian (assuming no loss or gain of and avoids the complications due to near-surface effects
mass). In contrast, heat flow into the base of the lithos­ (terrain corrections) that plague traditional heat flow
phere can vary markedly, as shown by the passage of the studies (Birch, 1950; Jaeger, 1965; Lachenbruch, 1968;
lithosphere over hot spots with resultant isostatic uplift, 1969; Blackwell et al., 1980; Powell et al., 1988; Lee, 1991).
enhanced heat flow, and volcanism (Crough, 1979). Unfortunately, BHTs are noisy. They tend to be lower
�ow long do thermal transients in the lithosphere than true formation temperatures due to the cooling
persist? A useful rule of thumb is that the time (t) taken effect of drilling fluid circulating at the bottom of
for a thermal disturbance to propagate a distance (y) boreholes. Corrections can be made for drilling distur­
through a material of thermal diffusivity (a) is (Lachen­ bances, but the information needed to make accurate
bruch and Sass, 1977) correcons is usually not available (Deming, 1989). It is
therefore necessary to take the presence of noise into
t = y2/4a (5) account when working with BHTs. The process of recon­
structing basin temperature from analysis of BHT data
For the lithosphere, a is approximately 32 km2/m.y., and can be generalized into three steps. First, raw data are
for an average lithospheric thickness of 100 km, transient extracted from well log headers and screened. Inconsis­
thermal events typically have lifetimes on the order of tent or implausible data are excluded at this stage.
50-100 m.y. Thus, the lithosphere has a relatively high Second, raw BHTs are corrected for drilling disturbances.
thermal inertia; background thermal states tend to persist A number of correction schemes are available (e.g.,
for periods of time that are comparable to the lifetime of Bullard, 1947; Lachenbruch and Brewer, 1959; Luheshi,
:
a petroleum system. In our studies of temperature­ 1983 Shen and Beck, 1 9 86); t�ese are reviewed by
dependent source rock maturation, it is therefore usually Dermng (1989), who offers practical recommendations.
relevant to determine the present-day thermal state as a Rarely, however, does sufficient information exist to
starting point for extrapolation back to the likely thermal allow accurate corrections to be made; the best that can
state at the time oil and/ or gas were formed. usually be hoped for is to reduce the systematic bias
introduced by the drilling disturbance. A third stage of
interpretation is essential. In the third stage, the corrected
ESTIMATING TEMPERATURE AND data are averaged through some interpretive model that
HEAT FLOW IN SEDIMENTARY BASINS reduces the random error in individual measurements.
By necessity, when the data are averaged, resolution
T e mperature suffers, although noise is reduced.
Some aspects of interpreting BHT data can be illus­
Traditionally, geothermal studies have been trated through the following examples. Figure 9.4 shows
concerned with estimating heat flow, although tempera­ BHTs collected from a small area a few kilometers square
ture is the actual quantity of interest. The utility of heat near the Iberia salt dome in south Louisiana. These data
flow studies lies in the ability to infer thermal conditions were collected from original well log headers and
at great depth from measurements in shallow boreholes. corrected for drilling disturbances using the AAPG
Measuring the geothermal gradient alone is not nearly as depth-based correction (Kehle, 1971, 1972; Deming,
revealing because the geothermal gradient may change 1 9 89). Following correction, an average geothermal
markedly with depth due to changes in thermal conduc­ gradient of 21.6°C/km was derived by a least-squares
tivity. regression. The implicit supposition is that the regression
Heat flow is never measured but instead estimated line (Figure 9.4) is a more accurate description of temper­
from equation 2 by making measurements of tempera- ature than the corrected BHTs from which it was derived