Normally, the path taken by the radio waves is assumed to follow refraction over a 4/3 radius earth, to a common volume situated between the earth and the tropopause.
However, over shorter paths (less than about 500 km), Geoff G3NAQ has suggested that the path may be refractive through a series of convective cells, and this has been confirmed (but not yet proved) by recent studies for the Propagation Studies Committee of the RSGB, by Ray Cracknell G2AHU, Chris Deacon G4IFX, and Ian Brotherton G2BDV (watch here for a link to this report when it becomes available on the Web). These convective cells are mainly situated in the lower troposphere, below 3 km above the earth's surface, and can be viewed as forming a train of lenses, or a broken duct effect, and fits well with the pattern of NDT at lower VHF frequencies.
There are some problems with low-level refraction in convective cells, though. To achieve 400 km in a single 'reflection' hop would require that 'reflection' (in fact, refraction in the convective cell) to occur at an altitude of about 4 km, which is well above the height of convective cells in normal temperate shallow convection. Given the ease with which 400 km QSOs can be achieved on lower VHF bands, convective cells do not appear to account for such everyday propagation, although they may still account for improved conditions at times.
Alternatively, over the same 400 km path, if the 'reflections' are occurring in cells at altitudes of around 2 km, quite a few cells would need to be aligned along the middle section of the path to support the QSO. The first and last 135 km of the path would be accounted for in the waves attaining the 2 km cell height, leaving a gap of about 130 km which would need to be bridged by a succession of convective cells. If those cells could be 20 km apart, then a total of 7-8 cells could provide the whole path. If the cells would need to be 5 km apart, then that rises to a total of about 27 cells. The probability of the path being complete is not too strong: assuming a 50% chance of each cell being correctly placed, the chance of the whole path being complete is 1 in 128 (for 7 cells). Extend the path to 600 km, and probability would range from 1 in 2 to the power of 18 to 1 in 2 to the power of 67 - clearly nowhere near as 'everyday' as experience suggests!
The details of refraction by convective cells are also somewhat speculative. Papers by Burrows (Burrows WG, The effects of atmospheric convective circulation on a narrow radio beam, J Inst Electronic Radio Eng 58:S185-S194, 1988) draw on old data and the idealised model of the Benard cell, which are not borne out by more recent radar results. Whether Burrows' proposal, that the rising warm air can actually form a cylindrical biconcave lens, is feasible remains open to question.
NDT at higher frequencies does not seem to follow this pattern though, at least over longer paths, being more sensitive to events in the middle and upper troposphere (such as folding of the tropopause). There is thus stronger evidence for conventional paths to an elevated common volume.
Some lifts may be largely refractive, for instance through an occlusion (G3NAQ and M1BWR are currently analysing two such events over the UK), when the tropopause 'bubbles up' and there are fronts high up in the troposphere to account for an elevated path.
Last updated 5 Apr 1999
Howard Oakley
Mail howard@quercus.demon.co.uk