Estimation of the height of the turbulent mixing layer from data of Doppler lidar measurements using conical scanning by a probe beam

Abstract

Abstract. A method is proposed for determining the height of the turbulent mixing layer on the basis of the vertical profiles of the dissipation rate of turbulent energy, which is estimated from lidar measurements of the radial wind velocity using conical scanning by a probe beam around the vertical axis. The accuracy of the proposed method is discussed in detail. It is shown that for the estimation of the mixing layer height (MLH) with the acceptable relative error not exceeding 20 %, the signal-to-noise ratio should be no less than −16 dB, when the relative error of lidar estimation of the dissipation rate does not exceed 30 %. The method was tested in a 6 d experiment in which the wind velocity turbulence was estimated in smog conditions due to forest fires in Siberia in summer 2019. The results of the experiment reveal that the relative error of determination of the MLH time series obtained by this method does not exceed 10 % in the period of turbulence development. The estimates of the turbulent mixing layer height by the proposed method are in a qualitative agreement with the MLH estimated from the distributions of the Richardson number in height and time obtained during the comparison experiment in spring 2020.

On the estimation of boundary layer heights: a machine learning approach

“…Boundary layer height estimation algorithms (Tucker et al, 2009;Berg et al, 2017;Bonin et al, 2018) purely using the Doppler lidars are limited by the range of the Doppler lidars at the ARM SGP facility, which may or may not reach the top of the boundary layer. The data availability of the ARM Doppler lidar SGP C1 systems is typically less than 1-2 km (Newsom and Krishnamurthy, 2020), and the data availability of the Doppler lidar vertical stares and velocity azimuth display scans for the study period are shown in Appendix A.…”

“…The data availability of the ARM Doppler lidar SGP C1 systems is typically less than 1-2 km (Newsom and Krishnamurthy, 2020), and the data availability of the Doppler lidar vertical stares and velocity azimuth display scans for the study period are shown in Appendix A. The Tucker method z i is used for inter-comparison with the RF model in this study, primarily due to its ease in application to the multi-year dataset and known performance, including its established usage in studies using the ARM Doppler lidar and other locations and instruments (Träumner et al, 2011;Shukla et al, 2014;Schween et al, 2014;Berg et al, 2017;Lareau et al, 2018;Lareau, 2020). Vertical velocity variance profiles provided in the sgpdlprofwstats4newsC1.c1 VAP and a variance threshold of 0.04 m 2 s −2 are used to determine z i , following Tucker et al (2009).…”

“…Vertical velocity variance profiles provided in the sgpdlprofwstats4newsC1.c1 VAP and a variance threshold of 0.04 m 2 s −2 are used to determine z i , following Tucker et al (2009). The results are somewhat sensitive to the choice of threshold such that the z i estimates decrease as the threshold is increased (Berg et al, 2017).…”

“…Ground-based lidar systems, such as the Raman lidars (Turner et al, 2014), ceilometer, micro-pulse lidar (Campbell et al, 2002), atmospheric emitted radiance interferometer (Knuteson et al, 2004;Sawyer and Li, 2013), Doppler lidar (Tucker et al, 2009;Berg et al, 2017;Bonin et al, 2018;Vakkari et al, 2015;Banakh et al, 2021), and high-spectralresolution lidar, have been commonly used to provide PBL height estimates (Quan et al, 2013;McNicholas and Turner, 2014). For elastic lidar systems such as the micro-pulse lidar, z i is estimated by locating the height where the rangecorrected signal-to-noise ratio (SNR) or attenuated backscatter profile experiences a strong decrease with height (Emeis et al, 2008).…”

“…One method for estimating the convective boundary layer (CBL) depth using Doppler lidars is to find the height where the vertical velocity variance profile falls below some prescribed threshold, which in some cases can vary with time (Lenschow et al, 2000;Tucker et al, 2009;Lenschow et al, 2012;Berg et al, 2017). This method, which we refer to as the Tucker method, is simple to implement and provides a direct measure of z i .…”