Abstract The turbulent heat fluxes at the soil surface are not observed (or poorly observed) by existing observational systems. This affects the ability to reliably predict the consequences of climate changes on the hydrologic cycle. Therefore, an approach to estimating sensible surface heat fluxes based on combination of the K and similarity theories, and using routine meteorological observations available in Russia, was developed. This was possible for the former Soviet Union territory and some other countries, where the standard practice of hourly observations includes temperature measurements at the atmosphere–land surface boundary and codes of the surface conditions (wet, dry, snow covered, etc.). The approach is designed for use in climate change and/or climate feedback studies. A similar approach to estimating latent heat fluxes is developed, but only for saturated surfaces (wet and/or snow covered). The method has been tested on several observational datasets.