Sub-bottom temperature perturbations due to temperature variations at the boundary of inhomogeneous lake or oceanic sediments

For many years continental heat flow data were obtained from temperature gradients measured in boreholes and conductivity measurements made on representative cores. Because of the difficulty of obtaining deep holes in suitable regions, increasing use is being made of measurements made in the top few meters of the sediments at the bottom of continental lakes. For reliable heat flow results two conditions must be met. First, an accurate record of the temperature history at the top of the sediment should be available, preferably over a time span of 4 or 5 years; second, if the thermal properties are not constant with depth, then the depth-dependence should be known.Errors that may arise if either or both of these conditions cannot be met can be serious. Using various surface thermal histories we compare results from models assuming constant diffusivity, a diffusivity changing exponentially with depth and a layered medium, all being structures based upon observed variations near the sediment-water interface where conductivity may change by a factor of two over 2 or 3 m. In some situations, errors as large as several hundred percents may occur but in such cases it should be possible to deduce from the data that something has occurred to make the data unreliable; more realistically, there are occasions where a temperature-depth plot which is linear within normally acceptable limits of accuracy and with a gradient that is within the range of expectations may be as much as 100% in error.