The physics of PMSE
Radar echoes occur if the radar refractive index (which is directly proportional to electron density in the mesosphere) possesses variations at spatial scales of half the radar wavelength. This condition is often called the Bragg condition for radar scattering. Most PMSE observations are conducted with VHF radars operating at frequencies around 50 MHz with corresponding Bragg scales of about 3 m. In the upper mesosphere, however, such small spatial scales should be efficiently destroyed by molecular diffusion, such that structures at the Bragg scale should actually not exist. The existence of charged ice particles, however, affects the mobility of the free electrons by ambipolar forces.
Owing to our today’s understanding of reduced electron diffusivity in the presence of charged ice particles, we can today also understand why previous rocket sounding revealed the absence of turbulence at some PMSE altitudes, even though the standard assumption has long been that turbulence actually is the physical process which creates the small scale structures in the first place. Our analysis of the lifetime of such small scale structures in the presence of charged ice particles showed that this lifetime strongly depends on the ice particle radius r, i.e., it varies with the square of r.
The figure sketches the time development of a PMSE signal after the decay of some turbulent activity (indicated as a rectangular pulse) for the cases of different ice particle radii. These calculations demonstrate, that PMSE might `survive’ the decay of neutral air turbulence by minutes to hours depending on the ice particle properties involved. Hence, turbulence and PMSE become decoupled in the time domain and it is not at all surprising that one phenomenon was observed in the absence of the other.
This microphysical picture further led to our proposal of a microphysical proxy for PMSE, i. e., the product of the ice particle charge number density and the square of the particle radius. Using CARMA to calculate height profiles of this proxy demonstrates that this simple model reproduces the main observed features like the altitude range of PMSE and the altitude of its maximum sufficiently well. In this context it is further interesting to note that the same CARMA calculations predict the altitude range of the simulated NLC at the bottom of the PMSE/proxy-layer in full agreement with all available observations.