Uncertainty Aware Probabilistic Modelling of Terrestrial Free Space Optical Links under South African Continental Fog
This paper presents an uncertainty-aware, reproducible framework for modeling terrestrial free-space optical link attenuation under South African fog conditions, which validates its implementation against published data to demonstrate that reducing link length is more critical than wavelength selection for mitigating severe outages, while highlighting the need for station-level validation in future site acceptance.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the author. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a world where data travels not through buried cables, but as invisible beams of light shooting through the air between two towers. This technology, known as free-space optical communication, promises the speed of fiber optics without the cost of digging trenches. However, unlike a cable buried safely underground, these light beams are exposed to the sky. They can be scattered by dust, bent by heat, or blocked entirely by the weather. Among all weather conditions, fog is the most dangerous enemy. Fog is made of tiny water droplets that are roughly the same size as the light waves used for communication. When these droplets meet the beam, they act like a thick wall, absorbing and scattering the signal until the connection breaks. In many parts of the world, engineers have learned to predict how often this happens, but in South Africa, the picture has been unclear. The region contains a mix of coastal mists, highland fog, and arid dryness, yet previous studies often reported only average numbers without explaining the rare, catastrophic moments when the fog is so thick that no amount of planning can save the link.
A researcher at the Tshwane University of Technology has now stepped in to bring clarity to this uncertainty. Instead of claiming to have access to raw, minute-by-minute weather records from every station, the study works with published summaries of visibility data collected by the South African Weather Service. The goal was not to discover new weather patterns, but to build a rigorous, transparent framework that explains exactly how those existing numbers translate into real-world signal loss. The researcher treated the problem like a physics puzzle, taking the standard international rules for how light behaves in fog and applying them to South African conditions. The work involved checking whether the numbers previously published for seven different South African cities were actually calculated correctly, and then using those verified numbers to simulate what happens when the fog gets dangerously dense.
The investigation began by looking at a table of data that had been published for seven locations, including Cape Town, Pretoria, and Polokwane. This table listed the average loss of signal strength for three different colors of light, measured at 650, 850, and 1550 nanometers. A closer, mathematical audit revealed a critical misunderstanding in how these numbers were being read. The values in the table were not showing how much signal was lost per kilometer of distance, as many engineers had assumed. Instead, the numbers represented the total loss over a very short, fixed distance of about 230 meters. This distinction is vital. If an engineer mistakes a total loss for a loss per kilometer, they might design a system that fails immediately because they underestimated the true power of the fog by a factor of more than four. By correcting this interpretation, the study confirmed that the original calculations were mathematically sound, but they had been mislabeled. The researcher then used this corrected understanding to build a model that could predict how often a link would fail under different conditions.
The study then turned its attention to the most dangerous part of the weather: the thick fog that occurs rarely but causes total blackouts. In clear air or light mist, using a longer wavelength of light, such as 1550 nanometers, offers a distinct advantage, allowing the signal to travel further with less loss. However, the model showed that this advantage disappears completely when the fog becomes very dense. Once visibility drops below half a kilometer, the water droplets become so large and numerous that they block all wavelengths of light equally. At this point, choosing a specific color of light no longer helps; the fog is simply too thick. The only reliable way to keep a connection alive during these extreme events is to shorten the distance between the towers. The analysis demonstrated that reducing the span of the link is far more effective than trying to select a better wavelength when the fog is at its worst.
To understand the real-world impact, the researcher looked at the frequency of severe fog events in three major cities: Mbombela, Cape Town, and Polokwane. By treating the published probabilities of heavy fog as if they were steady, year-round patterns, the study calculated how many hours per year a link would be exposed to signal loss so severe that it would break a standard connection. The results were sobering. In Mbombela, the exposure to such severe conditions amounts to roughly 33.3 hours a year. In Cape Town, it is about 22.8 hours, and in Polokwane, around 10.5 hours. These numbers are significant because they exceed the total amount of downtime allowed for a system that promises to be available 99.9 percent of the time. In other words, if a network relies solely on a single optical link in these cities, the fog alone will cause it to miss its reliability target, regardless of how well the equipment is built.
The paper concludes by emphasizing that while the mathematical model is now clear and the calculations are verified, the work is not finished. The current study relies on published summaries and cannot yet confirm how the model performs against real-time, second-by-second measurements of light and fog at a specific site. The researcher argues that before any company builds a permanent link in South Africa, they must pair their optical equipment with a local sensor that measures visibility and logged signal strength directly on the path. This would allow for a final check to ensure the predictions match reality. Until that step is taken, the framework serves as a powerful tool for planning and risk assessment, showing that in the face of South African fog, the best defense is not a clever choice of light, but a shorter distance and a backup plan.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.