Observations of the Early Morning Boundary-Layer Transition with Small Remotely-Piloted Aircraft

Abstract

A remotely-piloted aircraft (RPA), equipped with a high resolution thermodynamic sensor package, was used to investigate physical processes during the morning transition of the atmospheric boundary layer over land. Experiments were conducted at a test site in heterogeneous terrain in south-west Germany on 5 days from June to September 2013 in an evolving shallow convective boundary layer, which then developed into a well-mixed layer later in the day. A combination of vertical profiling and constant-altitude profiling (CAP) at 100 m height above ground level was chosen as the measuring strategy throughout the experiment. The combination of flight strategies allows the application of mixed-layer scaling using the boundary-layer height $z_i$, convective velocity scale $w_*$ and convective temperature scale $\theta _*$. The hypothesis that mixed-layer theory is valid during the whole transition was not confirmed for all parameters. A good agreement is found for temperature variances, especially in the upper half of the boundary layer, and the normalized heat-flux profile. The results were compared to a previous study with the helicopter-borne turbulence probe Helipod, and it was found that similar data quality can be achieved with the RPA. On all days, the CAP flight level was within the entrainment zone for a short time, and the horizontal variability of temperature and water vapour along the flight path is presented as an example of the inhomogeneity of layer interfaces in the boundary layer. The study serves as a case study of the possibilities and limitations with state-of-the-art RPA technology in micrometeorology.

Appendix: Uncertainty Estimation of Surface Heat Fluxes

The uncertainty of the estimated surface heat flux $H_0$ can be evaluated theoretically. We consider the uncertainties $\sigma _B = 0.1$, $\sigma _{z_i} = 20\,\text{ m }$, $\sigma _{\dot{\theta }} = 1\times 10^{-4}\,\hbox {K\,s}^{-1}$, and apply Gaussian error propagation to Eq. 8,

$$\begin{aligned} \sigma _{H} = \sqrt{\left( \frac{{\partial }{H}}{{\partial } {B}}\sigma _B\right) ^2+ \left( \frac{{\partial }{H}}{\partial z_i}\sigma _{z_i}\right) ^2+ \left( \frac{{\partial }{H}}{\partial {\dot{\theta }}}\sigma _{\dot{\theta }}\right) ^2} \end{aligned}$$

(13)

with the boundary conditions $B = 0.68$, $z_i = 200$ m, $\dot{\theta } = 1\times 10^{-3}\,\hbox {K\,s}^{-1}$, $\rho = 1.18\,\hbox {kg\,m}^{-3}$ and $c_\mathrm{p} = 1006\,\hbox {J\,kg}^{-1}\,\text{ K }^{-1}$. The result is an uncertainty of $\sigma _H=$ 33 W m$^{-2}$ for a surface heat flux of 161 W m$^{-2}$, which is approximately 20 %.