Lidar determination of the entrainment zone thickness at the top of the unstable marine atmospheric boundary layer

The thickness of the entrainment zone at the top of the marine atmospheric boundary layer (MABL) has been documented by an airborne lidar on two consecutive days during a cold-air outbreak episode over the Mediterranean. In addition to the lidar observations, in situ turbulent flux measurements at three levels in the MABL were made by a second aircraft. The flights' tracks are broken down in segments 25-30 km long and the data are filtered for the parametrization of turbulent entrainment in the MABL at scales smaller than a few kilometres. The structural parameters of the entrainment zone are determined by lidar from the distributions of the instantaneous MABL top height, The average values P-h0 and P-h2 of the cumulative probability distributions are used to define the bottom and top heights of the entrainment zone h(0) and h(2), respectively. The parameters h(0) and h(2) are calculated by reference to a linear vertical buoyancy flux profile in the framework of a first-order jump model. The model is constrained by both lidar and in situ data to determine P-h0 and P-h2 and so h(0) and h(2). In unstable conditions the average fraction P-h0 is estimated to be 6.0 +/- 1%. It is shown to be slightly sensitive to the presence of cloud at small cloud fractions. The mean value of the ratio of the inversion level buoyancy flux to the surface buoyancy flux A(R upsilon) is found to range from 0.15 to 0.30 depending on the shear in the MABL. The average value is 0.22 +/- 0.05. Our results are in good agreement with previous analysis at comparable spatial scales. In purely convective conditions, the value of A(R upsilon) given by the parametrizations fitted to our results is about 0.10-0.12, a value smaller than the commonly accepted value of 0.2. When compared to previous parametrization results, our proportionality constant for the mechanical production of turbulent kinetic energy is also found to be scaled down, in good agreement with large-eddy simulation results. It is suggested that mesoscale organized motions in the MABL is the source of this difference.