Impacts of Different Monin–Obukhov Parameterizations on Seasonal Turbulent Heat Flux Estimates Over Snow Surfaces

Abstract

Abstract Modeling of turbulent heat fluxes over snow surfaces almost exclusively depends on the Monin–Obukhov similarity theory. Within this framework, this study evaluates the impact of different parameterizations, including roughness lengths and stability correction functions, on half‐hourly estimates of sensible heat flux ( H ) and latent heat flux ( λE ) and their seasonal cumulative values over snow surfaces based on eddy covariance data for four winter seasons. The representative roughness lengths for momentum ( z 0 ), temperature ( z T ), and water vapor ( z q ) obtained by optimizing the agreement between modeled and observed H and λE during periods of sufficient snow cover were 7.19 × 10 −3 m, 5.88 × 10 −6 m, and 1.26 × 10 −8 m, respectively. Furthermore, using dual representative roughness lengths for high‐ and low‐albedo snow conditions improved the agreement between modeled and observed latent heat fluxes. Stable conditions accounted for 37% of the quality‐controlled flux data, and the choice of stability correction functions affected seasonal cumulative H estimates by up to 19%, highlighting the importance of accurately considering stability effects. We propose a simple yet physically interpretable stability correction function ( Ψ ) under stable conditions in the form of Ψ = α log(1 + βζ ), where ζ is the stability parameter and α and β are empirical constants. While the values of α and β are site dependent, the functional form is expected to be applicable across a wide range of sites. Despite the improvement achieved using dual representative roughness lengths, temporal variability in z 0 remains a key challenge for further improving seasonal cumulative flux estimates.