Last Updated on December 1, 2025 by John Berry
You might ask, “why make this web site?” Simple. To understand radiowave propagation in our atmosphere. But I’d like to go further too. I’d like to develop a “universal master algorithm for the prediction of signals of interest to the radio amateur”.
There is much known about radio wave propagation. There’s a whole ITU ‘sector’ dedicated to ‘radiocommunications’ and much of that is directly or indirectly associated with propagation. And there’s a huge body of knowledge set out in various papers and books by authors like Les Barclay, G3HTF (SK).
But the ITU and those authors have generally focussed on signal receipt for large percentages of time. For them, the signals received for small percentages of time are typically a nuisance – interference that would degrade otherwise healthy communications in the Mobile, Fixed, Broadcast and Satellite Services.
For radio hams (in the Amateur Service), it’s that low chance of reception that makes signals worth chasing. Something that’s only there for 0.1% of time or less is DX indeed.
My interest is in propagation that supports communications for small percentages of time.
Signals for high percentages of time (and to a lesser extent signals that would be interferers for small percentages) are described by algorithms and implemented as computer code in radio planning and modelling tools. A good example of a paper is Recommendation ITU-R Rec. P.526 Propagation By Diffraction, currently on its 14th edition. Any engineer wanting to model propagation in the troposphere over obstructed paths will most likely use the algorithms cited there.
So, signals for low percentages of time remains an interesting and seldom researched topic.
A place for AI
As machine learning evolves, many would be right to say, ‘just throw data at the problem and let an algorithm for small signals emerge’. That’s a valid thought. There are two issues with that.
- There needs to be something there in the first place to prime the pump. Even Bayesian techniques exploit initial ideas. So there’s still a place for a master outline.
- It’s unlikely that any institution is going to put money into predicting ham radio signals any time soon – unless they have economic value. Models will emerge for niche commercialisations like ultra-narrow band small signal propagation for smart metering and street lighting control. But no one is going to do a master algorithm for small signals.
About algorithms
Recommendation ITU-R P.526 is just one of many publications containing algorithms. If, for example, one station is ‘in the clutter’, moving through a town, a different algorithm would be used to describe propagation in this multi-path environment. This second algorithm would modify the first. And a third algorithm would be needed to describe how the received signal would likely vary in the long term.
Ultimately, one would need a clutch of algorithms to best model the path and resulting received signal. And those vary depending on the path geometry, propagation mode and technology. This algorithm amalgamation is the classic approach used in propagation modelling tools.
Having developed pseudocode for many of the models in HTZ Warfare and ICS Telecom, I’m most familiar with those tools from ATDI, but there are many. There are several in the amateur domain based on the professional tools and algorithms such as Radio Mobile by Roger Coudé, VE2DBE.
So, for propagation modelling – and hence propagation prediction – one requires a clutch of algorithms. Recommendation ITU-R P.1144-9 describes many algorithms or methods describing the high probability end of the science.
Decision maps
Professional tools use decision maps to determine which algorithms to use. Here’s an example taken from Rec. ITU-R P526-10.
The algorithms are annotated by the paragraph character (§). Each END mark shows the recommended algorithm. If the path suggests diffraction over the horizon, then §3.1 should be applied with its ten pages of formulae and nomographs. The §3.1 algorithms would then be added to those from other publications such as Recommendation ITU-R P.525-3 Calculation of Free-Space Attenuation to define path loss due to free-space plus diffraction – and so on. One can model every technology and propagation mode using this approach. Each algorithm is what I term on this website a propagation primitive.
But there are no such primitives targeting amateur radio.
Secondary research
There’s plenty myths about things like sporadic E propagation and propagation via the auroral oval. And there’s F-region propagation using Recommendation ITU-R P.533-9 (with parallels in VOACAP and others). But there’s no definitive decision map describing a master algorithm for signals received for low percentages of time.
My aim is to define a universal master algorithm for prediction of signals of interest to the radio amateur. It’s a long project that started in 2021! The target completion is 2026. Or maybe 2027… or…
And my focus is the use of academic and other validated publications: secondary research.
Guiding principles
I have two guiding principles.
Occams Razor
Given equal usefulness and descriptive power, I will always prefer an algorithm or model that requires the fewest assumptions and fewest model elements. This is a useful tenet to cut though the apparent complexity and lack of clarity in describing various effects. It’s especially useful in avoiding the disinclination of others to settle on cause. Occam’s razor is also known as the principle of parsimony.
Weight of evidence
Whilst I’m interested in new and novel approaches and descriptions in science, I need in this site to settle on descriptions that are useful. I also need them to be as true as possible. In determining what’s likely right in descriptions and models, I’ll weigh the evidence. It’s most likely that the option (in description or model) that is nearest the truth is the one with the greatest evidential support. Evidence will be pointed to in the bibliography in each topic.
And just in case you wonder, no Large Language Models (like ChatGPT) have been used in this site. If something is wrong, it’s through my ignorance and not through use of machines suggesting (un)truths.
It’s the truth that sets us free.