LAUSR

Data from 88 Oklahoma Mesonet stations and one micrometeorological tower in southern Brazil show a consistent relationship between mean wind speed V$$ V $$ and net radiation Rn$$ {R}_{\mathrm{n}} $$ in the stable boundary layer (SBL): for large radiative loss (typically |Rn|>40W·m−2$$ \mid {R}_{\mathrm{n}}\mid >40\kern0.3em \mathrm{W}\cdotp {\mathrm{m}}^{-2} $$ ), the average V$$ V $$ increases linearly with |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ at the vast majority of the sites considered. In most of them, such a linear relationship may be extrapolated all the way to the origin of the V×Rn$$ V\times {R}_{\mathrm{n}} $$ diagram. We present a detailed analysis of these observations that supports the hypothesis that such an extrapolated line is a good indicator of Vr$$ {V}_{\mathrm{r}} $$ , the wind‐speed threshold for the turbulence regime transition in the SBL. The physical concept behind this is that a weakly stable regime of sustained turbulence in the SBL occurs when the turbulent heat flux is capable of transferring back to the surface much of the energy it loses by radiation, explaining the increase of Vr$$ {V}_{\mathrm{r}} $$ with |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ . Conditions with large |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ occur almost exclusively under this weakly stable regime, which favors a warmer surface with increased long‐wave emission, making the average V$$ V $$ a good proxy for Vr$$ {V}_{\mathrm{r}} $$ when |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ is large. The rate of increase of V$$ V $$ with Rn$$ {R}_{\mathrm{n}} $$ for large |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ is highly variable among the sites, being larger in places where the occurrence of the weakly stable regime demands larger wind speed. We propose to use the rate of increase of V$$ V $$ with |Rn|$$ \mid {R}_{\mathrm{n}}\mid $$ as a quantifier of the surface coupling strength ( CS$$ CS $$ ), the definition of which is the amount of net radiative loss that can be sustained by a unit of mean wind speed at any given site. A semi‐empirical expression for CS$$ CS $$ , sustained by the observational analysis, is proposed.