Fog-top height (fog thickness) is very useful information for aircraft manoeuvres, data assimilation/validation of Numerical Weather Prediction models or nowcasting of fog dissipation. This variable is usually difficult to determine, since the fog-layer top cannot be observed from the surface. In some cases, satellite data, ground remote-sensing instruments or atmospheric soundings are used to provide approximations of fog-top height. These instruments are expensive and their data not always available. In this work, two different methods for the estimation of fog-top height from field measurements are evaluated from the statistical analysis of several radiation-fog events at two research facilities. Firstly, surface friction velocity and buoyancy flux are here presented as potential indicators of fog thickness, since a linear correlation between fog thickness and surface turbulence is found at both sites. An operational application of this method can provide a continuous estimation of fog-top height with the deployment of a unique sonic anemometer at surface. Secondly, the fog-top height estimation based on the turbulent homogenisation within well-mixed fog (an adiabatic temperature profile) is evaluated. The latter method provides a high percentage of correctly-estimated fog-top heights for well-mixed radiation fog, considering the temperature difference between different levels of the fog. However, it is not valid for shallow fog (∼ less than 50 m depth), since in this case, the weaker turbulence within the fog is not able to erode the surface-based temperature inversion and to homogenise the fog layer. 1 tem (ILS). Most airports have regulatory meteorological 2 instrumentation composed by surface visibilimeters, a 3 ceilometer (measuring cloud base and cloud cover) and 4 standard meteorological instrumentation, but all these 5 data are not enough to provide information about fog-6 top height. 7 Despite the numerous potential applications of this 8 variable, its numerical value is not always clear. Many 9 studies cannot provide information about observed fog-10 top height due to the lack of measurements in the ver-11 tical. In many cases, temperature and humidity data 12 from atmospheric soundings are used to estimate fog 13 thickness (e.g., Koračin et al., 2001; Liu et al., 2011; 14 Boers et al., 2013; Bari et al., 2015). However, these 15 soundings are not always available, or their temporal 16 frequency is not sufficient to cover the whole fog cy-17 cle. In other cases, remote sensing instruments are used 18 to estimate the fog top. Dabas et al. (2012) studied the 19 ability of using reflectivity measurements from sodar to 20 estimate fog-top height, while Boers et al. (2013) de-21 rived visibility from radar reflectivity for a case study of 22 radiation fog. Ceilometers detect cloud-base height of 23 low clouds (e.g., Dupont et al., 2012), but they are not 24 useful to provide information about fog-top height. All 25 these instruments are usually expensive and sometimes 26 their vertical resolution is not a...