There are three likely refractive modes which can give rise to NDT:
If you can think of others, particularly if you can give experimental evidence for them, please let me know.
You can estimate the height of convective cells responsible for the first mode of propagation, according to the time of day at which it first occurs. Because convective cells normally start near to the ground and ascend at about 1 km every 3 hours, on a long summer's day you are likely to see refractive tropo as follows:
local time
height (km)
0900
1
1200
2
1500
3
It is much easier to model refractive modes than scatter. AREPS 1.01 is the US Navy's advanced refractive propagation modelling tool, which uses a combination of ray tracing and the parabolic equation method of predicting signal strengths. The following images produced by AREPS illustrate some important points about refractive modes. In each, the software has been used to predict signal loss (in dB, at 144.3 MHz) over a ground path of 100 nautical miles, and up to 4000 feet into the air (i.e. the lowest part of the lower troposphere). Regions with different losses are represented by colours, as shown in the key at the bottom of each (red, yellow = low loss, to dark green and blue = high loss). Although set up to model conditions at my home QTH, the numbers are relative rather than absolute.
Signal loss over a sea path - standard conditions, no ducting or refractive NDT modes
Signal loss over a sea path - surface ducting (at 1000 feet) present, showing its effect on both ducted and non-ducted radio waves
Signal loss over an obstructed path - standard conditions, no ducting or refractive NDT modes, illustrating how a poor take-off increases losses at elevation, but not at the greatest distances
Signal loss over an obstructed path - surface ducting (at 1000 feet) present, showing the increased losses resulting from a poor take-off
Although two of these examples are for surface ducts, the effect of convective cells should be similar, although the pseudo-duct produced by a series of cells would of course be considerably more leaky, resulting in higher losses.
Last updated 27 Dec 1998
Howard Oakley
Mail howard@quercus.demon.co.uk