Rifting and Heat Flow: Why the McKenzie Model is Only Part of the Story

Zhiyong He¹, Steven G. Crews², and Jeff Corrigan^1
1ZetaWare, Inc., Sugar Land, Texas
²Hess Corporation, Houston, Texas

Ever since McKenzie (1978) linked heat flow to the process of lithospheric thinning the “McKenzie Model” has been widely used by basin modelers to determine thermal history in basins that have undergone rifting. However this has led in many cases to an overestimation of synrift heat flow.
A central concept of the McKenzie model is that the increase in heat flow due to rifting is directly proportional to the beta factor (the ratio of lithosphere thickness before and after stretching/thinning). So for a beta factor of 2.0, the lithosphere would thin to half its original thickness and heat flow would double by the end of rifting. Jarvis and McKenzie (1980) amended the model to include the effects of transient heat flow, so the heat flow increase would lag slightly behind the actual mechanical rifting.
However, neither version of the McKenzie crust model incorporates the fact that a significant portion of surface heat flow is generated within the crust itself as a result of radiogenic heat production (RHP) from the decay of radioactive isotopes of elements, especially in granitic rocks of the upper continental crust. Thinning the crust during rifting therefore results in a loss of RHP that partially offsets the increase in heat flow caused by mantle upwelling.
RHP from the crust accounts for around 40%-50% of the surface heat flow in a typical sedimentary basin. This leads to a significant overestimate of total heat flow using the standard McKenzie model. For example, if beta factor was 2 and the heat flow was 60 mW/m² before rifting and the crust was contributing 30 mW/m² (50%) of the pre-rift heat flow, only 30 mW/m² could have come from the mantle. When rifting occurs, the 30 mW/m² from the mantle would double to 60 mW/m² but on average the amount of RHP from crust would decrease to 15 mW/m² as a result of thinning the crust by 50%. The net effect of the rifting would be an increase of total heat flow to 75 mW/m², or only a 25% increase in total heat flow as supposed to the 100% increase according to McKenzie model (an overestimate of 60%).
We have applied a more correct rifting model accounting for crustal RHP in several basins, using the 1D basin model Genesis. The synrift heat flow in our models is much lower than that reported in previously published work based on the standard McKenzie model. In areas such as the Campos basin of Brazil we have found that McKenzie models overestimate paleo heat flow significantly. These overestimates of paleo heat flow lead to pessimistic assessments of the timing of petroleum generation in which hydrocarbons are generated before traps form. In contrast, our model, with a lower paleo-heat flow, predicts a more favorable timing of generation (during or after containment).
Thinning of the upper crust during rifting is accommodated mainly by extensional faults, which are typically not distributed evenly across the rift and may be offset from the axis of lower crust and mantle lid stretching. Therefore loss of RHP may also vary across the rift. On rift shoulders and horsts, the crustal RHP may be relatively unaffected by rifting, while in the graben axes, the loss of RHP may be severe.
Other factors also tend to diminish the disparity between synrift and present day total heat flow. For example, RHP from sediments, which should also be considered when trying to reconstruct the heat flow history of rift basins. In some cases up to 30% of the present-day total surface heat flow can come from the syn- and post-rift sedimentary fill. This sedimentary RHP is not constant through time because it is a function of the thickness of the sedimentary column. At the culmination of the synrift period the sedimentary section was relatively thin; therefore the sedimentary RHP was lower. Finally, transient effects due to rapid sedimentation can also dampen the thermal effects of rifting, especially when stretching occurs over a longer time period.
A realistic assessment of synrift heat flow and its possible impact on timing of generation requires not only an estimate of the amount of stretching but also consideration of losses of RHP due to extension, gains in RHP due to sedimentation and transient effects related to sedimentation rate. In many cases the difference between present-day and synrift heat flow is not as high as a standard McKenzie model would suggest, and this is usually good news for the prospectivity of the basin.

AAPG Search and Discover Article #90066©2007 AAPG Hedberg Conference, The Hague, The Netherlands