Search

We present a new gravity model for the lithosphere beneath the Kenya Rift Valley, the Mozambique Belt, and the Tanzania Craton in western Kenya and northeastern Tanzania. The Kenya Rift lies within the eastern branch of the extensive Cenozoic East African Rift System and has developed almost entirely in the Pan-African Mozambique Belt about 50 to 150 km east of the exposed margin of the Archean Tanzania Craton. The gravity field over western Kenya and northeastern Tanzania is characterized by a long-wavelength Bouguer anomaly. We propose that this anomaly has two components: 1. (1) a "rift" signature, deriving from a shallow rift basin, a lower crustal intrusion and a low-density zone in the mantle lithosphere localized beneath the rift axis2. (2) a "suture" signature, arising from a crustal root along the boundary between the Mozambique Belt and Tanzania Craton and higher density crust in the mobile belt above part of the crustal root. Two lines of reasoning support our interpretation: 1. (1) Recent geological studies of the Mozambique Belt in Kenya and Tanzania suggest that it is a continent-continent collision zone, and continent-continent collision zones worldwide commonly exhibit a characteristic gravity anomaly.2. (2) The long-wavelength Bouguer anomaly has at least two minima, one over the craton-mobile belt boundary, and one or more over the rift valley. Corroborative evidence for our interpretation of the gravity field is provided by recent seismic investigations.

Estimates of subsurface temperatures in the Archean craton of southern Africa during the Archean derived from diamond thermobarometry studies are remarkably similar to temperatures estimated for the same depths today, even though heat production in the earth and the mean global heat flow were probably substantially higher in the Archean. We present multi-dimensional numerical models for the thermal environment of the Archean craton in southern Africa during the Archean in which deep mantle heat is diverted away from the craton toward the surrounding oceanic lithosphere by a lithospheric root beneath the craton. Extrapolation of present-day models to thermal conditions appropriate for the Archean is inadequate to explain the similarity of present-day and Archean temperatures in the cratonic root. Reconciliation of the modern and ancient temperature estimates requires either relaxation of the constraints that the cratonic crustal heat production and/or the earth's mean mantle temperature were higher in the Archean than they are today, or that substantial "erosion" of the lithosphere comprising the cratonic root has occurred since the Archean. The latter possibility could perhaps result from revolatilization of the cratonic root in association with thermal perturbations in the mantle, for which there is evidence in southern Africa in the form of post-Archean tectonic and igneous activity.

The surface heat flow in the interior of Archean cratons is typically about 40 mW m-2 while that in Proterozoic and younger terrains surrounding them is generally considerably higher. The eighty-four heat flow observations from southern Africa provide an excellent example of this contrast in surface heat flow, showing a difference of some 25 mW m-2 between the Archean craton and younger peripheral units. We investigate two possible contributions to this contrast: (1) a shallow mechanism, essentially geochemical, comprising a difference in crustal heat production between the two terrains, and (2) a deeper mechanism, essentially geodynamical, arising from the existence of a lithospheric root beneath the Archean craton which diverts heat away from the craton into the thinner surrounding lithosphere. A finite element numerical model which explores the interplay between these two mechanisms suggests that a range of combinations of differences in crustal heat production and lithospheric thickness can lead to the contrast in surface heat flow observed in southern Africa. Additional constraints derived from seismological observations of cratonic roots, the correlation of surface heat flow and surface heat production, petrological estimates of the mean heat production in continental crust and constraints on upper mantle temperatures help narrow the range of acceptable models. Successful models suggest that a cratonic root beneath southern Africa extends to depths of 200-400 km. A root in this thickness range can divert enough heat to account for 50-100% of the observed contrast in surface heat flow, the remainder being due to a difference in crustal heat production between the craton and the surrounding mobile belts in the range of zero to 0.35 [mu]W m-3.

A map of the heat flux out of the earth's mantle has been prepared by subtracting the heat flow arising in the earth's crust from the surface heat flow. In continental areas the crustal contribution of the enriched zone is determined from the parameters of the linear heat flow-heat production relationship q0 = q* + bA0 in areas where such data exist. Where heat flow-heat production data are not available, a new empiricism relating reduced heat flow q* to the mean heat flow of a province, is used to estimate the reduced heat flow, and the depth parameter b is assigned an average value of 8.5 km. The oceanic crustal heat flow contribution includes both heat liberated by cooling, which is a function of the age of the ocean floor, and a small radiogenic component. A spherical harmonic analysis to degree 18 of the computed 5[deg] x 5[deg] mantle heat flow values yields a mean of 48 mW m-2; the degree variance spectrum has prominent strength at degrees 1,4 and 5. Continent-ocean differences are more apparent in the mantle heat flow than in the surface heat flow. However, contrasting surface heat flow patterns within continents, such as in central and western Australia, which arise from different surficial radioactivity distributions, do not appear in the mantle heat flow. Mantle heat flow is positively correlated with the geopotential and negatively correlated with the topography of the earth. However, correlation of heat flow with a "normalized" topography, in which the base line elevation difference between oceans and continents is suppressed is significantly positive, in accordance with observation that ocean floor topography is thermally controlled.

Heat-flow results from eleven widely spaced sites in central and western regions of the Republic of Zambia range between 54 and 76 mW m-2. Ten of the sites are located in late Precambrian (Katangan) metasediments or Kibaran age basement, while one site is located in Karroo age sandstone. Compared to the global mean of 39 +/- 7 (sd) mW m-2 for Precambrian provinces elsewhere, these heat-flow results are anomalously high by some 25 mW m-2. Heat-production measurements on borehole core samples indicate that enhanced radioactivity of an enriched surface zone can account for only half of the observed anomaly. The remaining anomalous heat flow must have a deeper source, and can be interpreted as a flux from the asthenosphere, providing the overlying lithosphere has been thinned to less than 60 km. Such an interpretation supports the existence of an incipient arm of the East African rift system trending southwest from Lake Tanganyika into the central African plateau.

A global heat flow map has been derived from existing observations supplemented in areas without data by an empirical predictor based on tectonic setting and age. In continental areas the predictor is based on the observed correlation of heat flow with age of last tectono-thermal event, and in oceanic regions on the observed relation of heat flow to age of ocean floor. The predictor was used to assign mean heat flow values to 5[deg] x 5[deg] grid areas on the globe, weighted according to the relative area of tectonic provinces represented. A spherical harmonic analysis to degree 12 of the heat flow field yields a mean value of 59 mW m-2, a rms residual of 13 mW m-2, and an amplitude spectrum which decreases gradually and almost monotonically from n = 1. The spherical harmonic representation of the heat flow field is free of the unreal distortions which have characterized earlier analyses based on a geographically sparse data set. Areas with residuals greater than 15 mW m-2 comprise less than 19% of the area of the globe, thus indicating that most heat flow provinces have characteristic dimensions adequately represented in a 12-degree analysis.