An investigation of flux-variance methods and universal functions applied to three land-use types in unstable conditions

Fluxes calculated from three flux-variance methods, which depend upon three different forms of the normalized standard deviation functions (referred to as universal functions) of the surface atmospheric stability have been tested and compared with measurements for temperature and water vapour. The flux measurements were made over a fully leafed deciduous forest, a leafless deciduous forest and over a wetland region during the summer. The first method (referred to as the variance method) allows for certain constants, which are associated with the universal functions, to vary with land-use type and the scalar for which the flux is computed; the second method uses the form of the universal function developed by Tillman, which depends upon two constants; and the third method, known as the Wyngaard method, is the simplest and depends upon one constant only. Flux estimates from the variance method yield the best agreement with observations over the three hand-use types and the Wyngaard method appears to yield estimates that are quite comparable. The measurements for the univeral functions agree better with the Wyngaard function for temperature and better with the Tillman function for water vapour, although both show some scatter. The simplest formula developed by Wyngaard may be considered adequate for computing fluxes of temperature and water vapour from their variances. The main advantage of the formulae is to help define an upper limit for actual fluxes.