Comparison of Sensible Heat Flux as Measured by Surface Layer Scintillometer and Eddy Covariance Methods Under Different Atmospheric Stability Conditions

Abstract

Surface fluxes of sensible heat and latent energy are important in many atmospheric processes and can be measured with reasonable accuracy over homogeneous surfaces. Direct measurements of turbulent fluxes such as sensible heat and latent energy, which are components of the shortened energy balance, are usually achieved by the eddy covariance (EC) method, which is considered the standard method for sensible heat and latent energy flux measurement and basically involves the use of single-point measurements from EC instruments mounted on a mast. However, the application of the EC method is often problematic. In this paper, results of the EC and surface layer scintillometer (SLS) estimations of sensible heat flux for different atmospheric stability conditions, namely, unstable, and near-neutral atmospheric stability conditions are presented. The aim of the study was to assess sensible heat flux measurements obtained by SLS, which relies on Monin-Obukhov Similarity Theory (MOST) and therefore assumes stationarity and homogeneity of the surface where measurements are taken with those obtained using the EC method for different stability conditions and different times (seasons) of the year. Statistical analysis of the data reveals a seasonal trend in the sensible heat flux, F h comparisons between the two measurement methods. There seems to be better agreement in the measurements obtained by the two methods, as noted by higher correlation coefficients and t-values, obtained in warm summer period from November to December during unstable atmospheric conditions while lower agreement in the values are recorded in the cold months of June and August. Also noted is a slight bias in the SLS measurement of F h compared to the EC measurements. The bias in SLS F h measurements is noticed for unstable atmospheric conditions whereas the EC method seems to record slightly greater values when the atmospheric condition is near-neutral. However the agreement between the F h values measured by the two measurement methods is still good. The agreement is even more remarkable considering that the EC method is a point measurement method depending on the covariance between w and sonic temperature T whereas the SLS method is an areal-averaging method that depends on MOST and therefore also on z−d. The inner scale length l o values measured by the SLS method are larger in the evening and night-time when the atmospheric condition is stable than during the daytime when the atmosphere is mainly unstable. This can be attributed to greater turbulent mixing during the unstable atmospheric condition compared to low turbulent mixing when the atmosphere is mainly stable. As for the agreement in the sensible heat flux values measured by the EC and SLS methods, there is no distinct or consistent pattern. Vegetation height also seems to slightly influence the agreement in the F h measurements obtained by the two methods (EC and SLS) as is noticed from the slope values. In general, the slope values approach 1 with increasing vegetation height for both unstable and near-neutral atmospheric conditions, although there are exceptions especially for the month of November.