Turbulent flux estimation in the atmospheric surface layer through flux-gradient similarity during the ALEX17 field campaign

Turbulent flux measurements require high frequency sampling in order to characterize appropriately all the variability scales of the atmosphere. A 3D sonic anemometer coupled with a gas detector allows for applying the eddy-covariance method which has become the standard. However, the high cost of this system often implies to look for alternative methods, specially when multiple stations are required. Turbulent fluxes can also be estimated through the flux-gradient similarity theory, requiring observations of mean quantities of (at least) air temperature and humidity at two levels and wind at one height. This approach is more sensitive to the disturbing influence of heterogeneous and complex surfaces and a comparison between methodologies is required under these conditions.

The data used in this study is part of the ALaiz EXperiment 2017-2018 (ALEX17). This campaign was the last within the New European Altas project. It had a duration of over a year with measurements in complex terrain. The location of the experiment is a valley bounded by two mountain ranges that rise 150 m north and over 600 m south. A central site in the centre of the valley was instrumented with a sodar-RASS, an 80-m tower, a surface energy balance (SEB) station with an eddy-covariance system and a surface-layer station (SLS) with the necessary measurements to estimate the turbulent fluxes. In addition, eight supplementary SLS were deployed along the longitudinal and transverse valley axes to characterize the surface layer variability within the valley.

This communication will present a comparison of the friction velocity and sensible heat flux obtained from both the eddy-covariance system and the flux-gradient method at the central site for a time series of 8 months. Friction velocity is highly comparable between methodologies with a correlation of 0.92 and a standard deviation of 0.05. The performance of the sensible heat flux estimation differs between stable and unstable cases, with a correlation of 0.70 and 0.89, respectively, after applying a quality control procedure. The poorer results obtained under stable conditions points out the need for alternative estimations of the sensible heat flux for these cases.