Radiofax: DDH3/DDK3/DDK6 (Deutscher Wetterdienst, Germany)

The triplet of stations from Germany (DDH3/DDK3/DDK6) are distinguished in the radiofax field. They operate on a +/-425Hz shift (whereas others operate on +/-400Hz), and they have the highest transmission powers – 10kW for DDH3 and 20kW for DDK3/DDK6. This makes the station relatively “loud” – for comparison, VMC/VMW operate at just 1kW. In fact, I was able to receive them from home relatively weakly in the past, although this was the exception rather than the norm.

While I started monitoring Deutscher Wetterdienst’s transmissions again recently, I went through the UTwente WebSDR using Fldigi decoding. I found that propagation was not maintained well enough to receive all of the fax variants. Additionally, processing the data would be difficult, so I switched over to the use of Initially, I used DL/JO43UG at Tostedt, Germany which provided better results but then I discovered DF0KL’s KiwiSDR which was practically right next to the transmitting station. As a result, a solid signal was received around the clock allowing for practically perfect copy!

The faxes presented here represent a snapshot in time and are not updated. They should not be used for navigation purposes. Copyright in the faxed material remains property of the transmitting agency, Deutscher Wetterdienst of Germany. Faxes are reproduced for documentary purposes covering the technical aspects of the transmission, formatting of the content and as proof of activity and reception on the given dates.


Owing to the advantageous situation, I am able to present faxes in schedule transmitted order, omitting charts using the same formatting and repetitions. However, due to connection unreliability resulting in a loss of decoding from the SDR, there is a mixture of charts received on 29th and 30th January 2019 (and not necessarily in order either). All three frequencies were found to be active and were able to be received successfully, however, the best chart is presented which usually corresponds to the higher frequency except in cases of QRM.

The schedule for DWD begins at 0430UTC, with the first chart (received 30th January 2019) being a surface analysis. The chart has a fairly standard length of phasing and moderately long start tone. Margins are white in all faxes, and they come in various lengths with short-charts being relatively common. The faxes often also have a text product identifier string which contains a number of underscores – this is placed near the caption which is in the bottom left.

The above chart shows extremely fine detail for a radiofax – note the small observations figures labelled throughout the chart. If it were not for the perfect signal from DF0KL’s receiver, it may have been impossible to read these numbers. Even the legend and ID are relatively small on this fax.

The next is a 36hr MSL pressure forecast at 0512UTC, with “COR” in the legend which I suspect means it has been corrected/amended. Less fine detail in this chart and I suspect “Icon” is the name of their weather model/observation system.

At 0525UTC, we receive our first tall chart which is intended to be read rotated. This chart shows a surface analysis with iceberg limits and tropical storms (if any). Note the lines spill into the double-line border section.

The 0638UTC chart was not transmitted as it is a storms chart during season only, so we move to the 0651UTC chart which is the first “tiled” chart, consisting of four sub-charts arranged so as to be read rotated. This includes Temperature 500hPa, Geopotential height, Pressure Reduced to MSL.

This is followed by 0704UTC chart which has Temperature 850hPa and Relative Humidity 700hPa. Despite being tiled in a similar way, the label text is much bigger on this chart.

At 0730UTC, the surface pressure forecast for 48-hours is broadcast in the same format as the earlier 36h version.

This is followed by +60h at 0743UTC.

This is followed by +84h at 0804UTC.

Finally, the series concludes with the +108h chart at 0817UTC, all adhering to the same format that you could actually animate them.

The next type of chart is the wind and swell forecasts. At this time, QRM was affecting the 13882.5kHz frequency, so some of the following charts were received from 7880kHz instead. The first chart at 0830UTC is the prediction for +24h, a short chart with a graphic that fills nearly the whole image area and includes the DWD logo.

This is followed by the +48h chart at 0842UTC.

Which is followed by the +72h chart at 0854UTC.

The series concludes with the +96h chart at 0906UTC. The 0930UTC chart that follows adheres to the same format as the 0704UTC chart with different validity times in the caption and is not included.

The next chart is the Sea Surface Temperatures for the North Sea. This is a tall chart with very bold lines, including the transmitter’s callsigns and the logo for Bundesamt Fur Seeschiffahrt Und Hydrographie. Looks like the sea is a chilly 4-11 degrees C.

The 1007UTC chart is a repeat of the 0525UTC chart and is omitted. Another tall chart follows at 1029UTC which indicates sea state and wave height. Of note is a lot of blank area on the last two charts.

At 1050UTC, there is a surface analysis transmitted again – this one for 06UTC following the same format as the 0430UTC. It seems their charts are so systematically generated that it’s hard to tell the difference without looking closely at the captions – hence why I omit quite a few charts because they only differ very subtly. In this vein, all charts from 1111UTC through to the 1356UTC chart are either repeats or variants of the charts already shown and are omitted.

Unfortunately the link did drop-out on the 30th January 2019, so charts from here are received the day before on 29th January 2019. At 1425UTC, the first part of the schedule is sent – a tall chart but sent sideways. Large text, bold writing and lots of white space make it easy to read.

The schedule is broken into two with the second part sent at 1445UTC. Of note is that it clearly states the transmission to have +/-425Hz shift (850Hz) rather than the more common +/-400Hz (800Hz), so when tuning with a regular decoder without a change of configuration, blacks can be “below black” and white can be “above white”. This is of no great consequence, as all program matter appears to be mostly black and white with very little greyscale usage.

At 1508UTC, we get an iceberg analysis courtesy of the Canadian Ice Service Ottawa being sent via DWD’s transmitters. This chart is sent sideways as a short chart, but it’s good to see the collaborative effort.

The effort extends as at 1520UTC, a tall ice chart from SMHI Sweden is sent. This chart is especially fine, but also has a “waviness” to it that suggests it was printed on paper at one stage and scanned back into the system.

In what is truly a collaborative effort, this is followed at 1540UTC by an ice chart from the Norwegian Meterological Institute. This chart is a square-ish chart which seems to have one edge cut off to the right, and attempts to use various shading and greyscales to convey the data. Unfortunately, some of that seems to have been lost in transmission.

Charts at 1636UTC through to 1831UTC follow the same format as charts shown above with different validity times and are excluded.

This chart at 1834UTC was supposed to be a +24h surface forecast, but instead was a +36h forecast. I suppose maybe the +24h product is no longer offered. All charts from 1847UTC through to the final chart of the day at 2200UTC have been omitted as they are either repeats of charts above or follow the same format.


I’m rather pleased to have been able to receive the faxes with practically perfect clarity thanks to the advantageous location of DF0KL’s KiwiSDR and rather excellent antenna. As a result, I was able to document the full range of charts sent by Deutscher Wetterdienst which includes charts produced by other agencies, most of which I haven’t seen before in this clarity. The charts contain surprisingly fine detail for radiofax and a mixture of sizes. Some charts show evidence that they have been “scanned in” to the system, revealing a chain of “analog” processes to deliver the charts to users.

It looks like with a little more careful planning and choice of receiver, receiving preservation-grade images is possible. Unfortunately, for many other radiofax stations, their weaker transmissions and the location of receivers don’t allow for this to happen so “noisy” images are the result.

P.S. On the upside, the NWS/NOAA have updated the radiofax schedule, but on the downside, the formatting is a little disrupted and one of my amendments didn’t make it through correctly. I’ve sent them another e-mail so they can hopefully remedy the issue.

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in Radio and tagged , , , . Bookmark the permalink.

Error: Comment is Missing!