This article wants to be some kind of introduction to the technique first discovered and exposed by Harel Dan in his Medium post from 2018 (highly recommended read) which involves using the publicly available SENTINEL-1 satellites data to detect and locate powerful sources of radio frequencies in (some of) the C band.
We will also use this space to act as sort of an hub for content related to this technique, it’s workings, the hardware detectable and in general anything new that pushes our knowledge about the method forward. (So, check this space for updates)
Radars, according to needs and design specs can operate on completely different radio frequencies. For clarity, the vast spectrum of frequencies available has been divided into frequency ranges called “bands”. There are multiple standards to define them, but two are way more common than others when discussing radars. Unfortunately they have overlapping nomenclature for the various bands which can be rather confusing.
- IEEE (Institute of Electrical and Electronics Engineers) standard, covers most civilian and military use frequencies.
- NATO, EU, US ECM frequency designation is a standard agreed upon NATO countries to better define the frequencies used by military systems such as radars (which to make things worst, has undergone a complete naming change for the bands, so old and new nomenclature for the bands, under the same standard, can be used according to sources).
The satellites of the SENTINEL-1 constellation use a radar which operates in the IEEE defined C band hence its name “C-SAR” (C band Synthetic Aperture Radar). Specifically it operates at 5.405 Ghz .
IEEE C band (SENTINEL-1) covers the frequencies ranging from 4.0 to 8.0 gigahertz (GHz), same range is instead divided into two bands under the NATO standard (new nomenclature), the G band (4.0 to 6.0 Ghz) and the H band (6.0 to 8.0 Ghz).
You can read more about commonly used radio frequencies, their nomenclature and the various standards following this link.
If operating on similar frequencies and based on their relative distance and power, radio sources and receivers can disturb each others for multiple reasons, this is what we see in SENTINEL-1 data when the typical “interference lines” show up. A powerful radio emitter, operating in a frequency close to that of the satellite sensor (5.405 Ghz), is disturbing the reflected radio waves that reach back to the satellite radar.
The “interferences lines”
The interferences we are discussing about manifest themselves on SENTINEL-1 data as sort of lines which are always perpendicular to the satellite orbit trajectory.
SENTINEL-1 mission is composed of two satellites traveling both in near polar orbits but in opposite directions and with a slight angular offset. Their orbits are called Ascending and Descending.
The “interference lines” we are talking about are clearly visible in the two pictures posted above (Fig1 and Fig2). Also quite clear the different angle the lines have if the satellite that caught them was in ascending or descending orbit (this due to the satellites having slightly different orbit angles and the “lines” appearing always perpendicular to the satellite orbit).
The lines can be hundreds of kilometres long and tell us that somewhere along those lines a powerful radio source close to 5.405 Ghz (such as some radars) is active and emitting.
If the interferences we are analysing are pictured in both ascending and descending satellite passes (which can have days of offset unfortunately), we can then overlay them to create a composite image (Fig3). The interference source will be exactly at the intersection of the lines. This works very well for fixed installations that are kept active over long period of times.
In the case pictured above, at the centre of the “X” there is a Patriot battery protecting Dubai that has been practically constantly active for the last 2 years.
In many cases sources will be active just for short period of times, meaning that they will either never be captured by a SENTINEL-1 pass or be captured just in an Ascending or Descending one. We can (in some cases) overcome this by generating overlaying data of passes stretching over very long time spans increasing the chances of having the interference caught by at least one pass for each satellite (hence allowing us to generate a “X” to pinpoint the source).
The interference sources
It appears that to disturb, at least in a recognisable way, the SENTINEL-1 sensors the radio source has to be very powerful, this is generally reserved to military grade radars.
In his first exposure Harel Dan isolated 2 rather popular radar systems responsible for at least some of the characteristic “interference lines”. Specifically the AN/MPQ-53/65 that powers Raytheon’s Patriot systems and the STRIL array in Sweden which is composed of many radars all along the Swedish coast.
More systems are being found to cause the characteristic artefacts on SENTINEL-1 data, and as they will be exposed we will list and link them here.
The systems :
Most recent discoveries:
- EMPAR (SPY-790) – Designed for maritime operations by SELEX (later purchased by Finmeccanica / Leonardo)
- KNONOS Grand – The evolution of EMPAR produced by Leonardo (destined to Navy use)
- FCS-3 – Main Japanese radar employed in most modern vessels of Japanese Maritime Self-Defence Force.
- Sea Giraffe – the sea version of SAAB Giraffe radars operates on frequencies that should be detectable by this technique. We are still investigating this.
- Multiple Chinese Navy radars such as : Type 382 Radar, Type 381 Radar , Type 346 radar and other older models.
- Harder to detect but also some EW tools, especially Turkish ones have been detected both in Syria and Libya.
- Some older Soviet era radars are also detectable, visible in North Korea and and rarely in Libya
The known unknowns
There is still a lot that can be researched, discovered and refined in regards to this technique, I’m working on some, but hope that others will chime in and share their knowledge or researches on the subject.
- Exactly why the interferences manifest themselves as lines always perpendicular to the satellite trajectory (we have theories about this, but still not enough to be sure about it… hope to update the post soon with more info)
- The exact frequency range that can be revealed by this technique
- A possible relation between “lines” width, patterns and shapes and the power output or frequency of the interference source
- The possibility to isolate “fingerprints” in signals so to better point to source system
There are surely more but for the moment this seem as good starting points.
The known knows
What we know is altho enough for multiple types of analysis.
- The technique can in many case pinpoint the interference source or provide a general area for it.
- The frequency with which we receive new data from the satellite allows us to study deployments over time with often 2 or 3 days precision.
- We can also infer the level of “alertness” of certain sites given how often their radar is on and in case of radars mounted on vessels we can track their operations across the oceans and seas of the planet.
All this data then, coupled with other OSINT sources such optical satellites, social media pictures, posts and videos etc, can provide a more precise outlook on what is being observed.
Often even a single pass capturing an interference is enough to isolate a research area which via other means can then be narrowed down to obtain both an exact location and identification of the system being examined.
Other times location and precise id of a system aren’t necessary, in general the technique provides unprecedented access to (some) previously unaccessible military deployment data.
The unknown unknowns
SENTINEL-1 data is transmitted daily to ESA servers, from there it reaches various online platforms that can be used for free.
- Copernicus SCI-HUB: from this portal you can access raw data, as soon as it is available on ESA servers, from all SENTINEL missions. The data will be huge and not very user friendly but for those willing there multiple open source tools that make the task of analysing it more bearable.
- EO-Browser is a great online portal to access most recent open source satellite data visually. It receives new data to visualise with a very short delay compared to Copernicus and it comes with a growing list of visualisation options and some simple tools to perform some initial analysis.
- 5 GIT (5 Ghz Interference Tracker) is a tool developed in Google Earth Engine (GEE) specifically to facilitate researches around the technique described in this article. GEE updates their SENTINEL-1 satellite with approximately 24 hours delay compared to the other tools described above (currently at least, situation might improve in the future) but it allows for way more efficient researches thanks to its ability to overlay SENTINEL-1 data from both ascending and descending trajectories over large time ranges and impressively vast areas. It will also be updated to include eventual suggestions and discoveries. (Here you can find some sort of tutorial for it and a demo research using it)
- Google Earth Engine is an impressive web based engine offered by google to develop applications that require to process immense amounts of data to study our planet. It requires some scripting knowledge but it also has access to a vast array of databases both of raw satellite data (such as SENTINEL-1) and processed models for environmental studies. In general, well worth a look at if you aren’t scared by the coding requirement.
- C-SAR Eye – all new tool developed in GEE (Google Earth Engine) that allows to analyse and isolate differences in C-SAR data so easily detect terrain changes, new buildings, radars etc.
Want to help or have questions ?
Great ! Join the OSINT Editor discord server where (as the community grows at least) you will find users able to answer your questions, the author of this terribly written article, and in general users with which you can coordinate your research efforts.
Contact the author via email