Analysis and Mapping of an Updated Terrestrial Heat Flow Data Set

Abstract The number of heat flow measurements at the Earth surface has significantly increased since the last global analysis (Pollack et al., 1993, https://doi.org/10.1029/93RG01249 ), and the most recent of them provide insights into key locations. This paper presents a new compilation, which includes approximately 70,000 measurements. Continental heat flow (67 mWm −2 ) does not change significantly, but the differences are more important for oceanic heat flow. The divergence with conductive cooling models is reduced significantly for young ages of the seafloor, since the most recent measurements (92 mWm −2 ) are significantly higher on average than the older ones (79 mWm −2 ). This is related to a better quality and a better sampling of measurements in regions affected by hydrothermal circulation. The total Earth heat loss derived from these most recent measurements is estimated to ∼40–42 TW and represents only 3–5 TW less than with a conductive cooling model (∼45–47 TW). Hydrothermal heat loss in the oceanic domain is estimated with a new method based on the ruggedness of the seafloor and represents ∼1.5 TW more than previous estimates. The heat flow variability on continents is so large that defining a trend with stratigraphic or tectono‐thermal age is difficult and makes extrapolation from age a poor predictor. On the other hand, additional geological and geophysical information can be combined with age for better predictions. A generalized similarity method was used here to predict heat flow on a global 0.5° × 0.5° grid. The agreement with local measurements is generally good and increases with the number and quality of proxies.