Toward quantifying turbulent vertical airflow and sensible heat flux in tall forest canopies using fiber-optic distributed temperature sensing

Abstract

Abstract. The paper presents a set of fiber-optic distributed temperature sensing (FODS) experiments to expand the existing microstructure approach for horizontal turbulent wind direction by adding measurements of turbulent vertical component, as well as turbulent sensible heat flux. We address the observational challenge to isolate and quantify the weaker vertical turbulent motions from the much stronger mean advective horizontal flow signals. In the first part of this study, we test the ability of a cylindrical shroud to reduce the horizontal wind speed while keeping the vertical wind speed unaltered. A white shroud with a rigid support structure and 0.6 m diameter was identified as the most promising setup in which the correlation of flow properties between shrouded and reference systems is maximized. The optimum shroud setup reduces the horizontal wind standard deviation by 35 %, has a coefficient of determination of 0.972 for vertical wind standard deviations, and a RMSE of less than 0.018 ms−1 when compared to the reference. Spectral analysis showed a fixed ratio of spectral energy reduction in the low frequencies, e.g., <0.5 Hz, for temperature and wind components, momentum, and sensible heat flux. Unlike low frequencies, the ratios decrease exponentially in the high frequencies, which means the shroud dampens the high-frequency eddies with a timescale <6 s, considering both spectra and cospectra together. In the second part, the optimum shroud configuration was installed around a heated fiber-optic cable with attached microstructures in a forest to validate our findings. While this setup failed to isolate the magnitude and sign of the vertical wind perturbations from FODS in the shrouded portion, concurrent observations from an unshrouded part of the FODS sensor in the weak-wind subcanopy of the forest (12–17 m above ground level) yielded physically meaningful measurements of the vertical motions associated with coherent structures. These organized turbulent motions have distinct sweep and ejection phases. These strong flow signals allow for detecting the turbulent vertical airflow at least 60 % of the time and 71 % when conditional sampling was applied. Comparison of the vertical wind perturbations against those from sonic anemometry yielded correlation coefficients of 0.35 and 0.36, which increased to 0.53 and 0.62 for conditional sampling. This setup enabled computation of eddy covariance-based direct sensible heat flux estimates solely from FODS, which are reported here as a methodological and computational novelty. Comparing them against those from eddy covariance using sonic anemometry yielded an encouraging agreement in both magnitude and temporal variability for selected periods.