Abstract
The technique developed by Gal-Chen in 1978 is used to derive vertical velocities, buoyancy, and pressure perturbations from dual-Doppler radar observations of the planetary boundary layer (PBL). Several approaches to verification are pursued. They include: (a) scan-to-scan temporal continuity of the derived fields; (b) an objective test to find out how well the derived pressure perturbations balance the dynamical equations; (c) comparison of dual-Doppler derived, horizontally averaged fluxes of heat versus in situ measurements and other data sets; and (d) a noteworthy improvement in the quality of the retrieved pressure when tendencies are included.
Previous studies indicate that in order for the method to be viable the radars have to resolve the PBL with at least ten vertical levels. One such event occurred on 27 September 1978 during project PHOENIX, conducted at the Boulder Atmospheric Observatory (BAO) 300 m tower. An inversion above a shallow boundary layer of height around 800 m was eroded, and the PBL grew to a height of 2.4 km in less than half an hour. During that period, the vertical profiles of potential temperature and pressure variance derived from the two NOAA/Wave Propagation Laboratory X-band (3 cm wavelength) Doppler radars suggest the existence of two inversions. Two inversions are also indicated by the aircraft data.
Some aspects of the derived heat flux profiles, such as negative heat flux at the top of the mixed layer, are classical and constitute further evidence of the plausibility of the results. Some other aspects such as positive vertical gradient of the heat flux profile near the first inversion (where the heat flux is still positive) are not commonly observed. Based on the available data, it is speculated that this latter feature is transient, indicative of the mixing (during the growth of the PBL) of the potentially warmer upper layer with the potentially colder lower layer.
Several closure approximations for three-dimensional PBL models are tested. Nonlinear eddy viscosities are derived from the observed second moments of the Doppler spectrum and are used to estimate the frictional dissipation in a three-dimensional numerical model of the PBL. Except near the ground, the derived temperature and pressure are only slightly sensitive to factor-of-two variation in the value of the eddy viscosity. Furthermore, it is found that adding frictional dissipation does not reduce the imbalance between the horizontal pressure gradient and the horizontal accelerations. Recalling that in a “perfect” three-dimensional model exact balance must prevail, one concludes that this particular subgrid parameterization could be merely a device to prevent excessive accumulation of energy in the smallest resolvable scale of a numerical model.
