Adjoint inversion of the thermal structure of Southeastern Australia

The variation of temperature in the crust is difficult to quantify due to the sparsity of surface heat flow observations and lack of measurements on the thermal properties of rocks at depth. We examine the degree to which the thermal structure of the crust can be constrained from the Curie depth and surface heat flow data in Southeastern Australia. We cast the inverse problem of heat conduction within a Bayesian framework and derive its adjoint so that we can efficiently find the optimal model that best reproduces the data and prior information on the thermal properties of the crust. Efficiency gains obtained from the adjoint method facilitate a detailed exploration of thermal structure in SE Australia, where we predict high temperatures within Precambrian rocks of 650 degrees C due to relatively high rates of heat production (0.9-1.4 mu W m(-3)). In contrast, temperatures within dominantly Phanerozoic crust reach only 520 degrees C at the Moho due to the low rates of heat production in Cambrian mafic volcanics. A combination of the Curie depth and heat flow data is required to constrain the uncertainty of lower crustal temperatures to +/- 73 degrees C. We also show that parts of the crust are unconstrained if either data set is omitted from the inversion.