Removing the contribution of mesoscale fluctuations to surface sensible and latent heat fluxes over a heterogeneous surface

Point measurements of temperature, moisture, and vertical velocity in the surface layer are commonly used to compute local surface kinematic heat and moisture fluxes. In this study, we examine the dependence of estimated surface heat and moisture fluxes on the time scale of the mean that is used to estimate deviations from the mean (Φ_T=Φ-Φ_M, where Φ is a scalar and the subscripts T and M represent turbulence and mesoscale). Since 2006, point measurements have been archived at Howard University Beltsville, MD Research Campus (HUBC). The campus, where a 31-m flux tower is installed, is covered by mixed deciduous and coniferous forests, but is also influenced by urban infrastructure less than 2 km to the west and north of the site. The eddy covariance method is applied to data collected under fair-weather conditions for summers 2006 and 2007. The surface fluxes estimated with the typical 30 min averaging period are compared with those whose averaging periods are determined by the cospectral method. The cospectral method is applied by using consecutive positive cospectral values as an indicator of turbulence, since typically, during the daytime, turbulent heat and moisture fluctuations are dominantly transported by turbulent vertical velocities from the surface to the atmosphere. The range of the averaging periods to filter out mesoscale fluctuations in estimating turbulent surface fluxes at HUBC will be reported. Also the contribution of mesoscale fluctuations to surface heat and moisture fluxes will be investigated from the aspect of the four different terms contributing to total covariance: w_TΦ_T, w_TΦ_M, w_MΦ _T, and w_MΦ_M (where w is vertical velocity).