Random modelling of the lithospheric thermal regime: forward simulations applied in uncertainty analysis

Abstract

A random modelling technique was applied in uncertainty analysis of forward geothermal modelling of the lithospheric thermal regime. We present results for estimating the effects of uncertainties in thermal conductivity, heat production rate, model basal temperature and basal heat flow density on calculated lithospheric temperature and heat flow density (HFD). We analysed two models: first a 4-layer synthetic model representative of typical shield conditions with thick crust and lithosphere, and second a two-dimensional case history from the Fennoscandian (Baltic) Shield. Thermal conductivity (normally distributed) and heat production (log-normally distributed) as well as temperature or heat flow density (normally distributed) used as the lower boundary condition in the mantle were randomly varied in the simulations. Calculations based on 1500 independent cases of the layered model indicate, for instance, that a standard deviation (STD) of 50 K in calculated Moho temperature results with uncertainties either in thermal conductivity of about ±0.5 W m −1K −1, in heat production rate of ±0.2 log A ( A in μW m −3) or ±115 K in basal temperature, but only ±2 mW m −2 in basal heat flow density. Respectively, the same values result in an uncertainty of ±2–10 mW m −2 in calculated surface heat flow density. If conductivity and heat production rate are varied simultaneously, the resulting uncertainty in calculated Moho temperature increases to about ±70 K. Adding also basal temperature variation increases the Moho temperature variation to about ±85 K. Results calculated with the two-dimensional transect in the Baltic Shield indicate analogously that uncertainty in temperature at 50 km depth (approximately at Moho) is ±35–60 K (using temperature as the lower boundary condition) and ±50–85 K (using heat flow as lower boundary condition). The corresponding variations in surface heat flow density are ±6–15 mW m −2. The choice of the lower boundary condition has an essential effect on the variation of mantle temperatures, and models using heat flow density as a lower boundary condition yield higher uncertainties in calculated mantle temperatures than those based on temperature as the lower boundary condition.