I raised an issue about AIS altitude data on the ship card for AIS aircraft (SAR) where the altitude did not make sense when compared with ADSB reported altitude. Jasper checked it was OK AIS-catcher wise and likely originated in the original AIS data.

Background: a container vessel lost some containers in the deep water channel 5NM east of the Isle of Wight (near Portsmouth UK) which triggered a search by coastguard aircraft (fixed wing and rotor) for several days.

The issue is closed: number 482 for reference and background.

Then Dan over at the PlanePlotter group https://planeplotter.groups.io (who does not have a github account) sent me a post with some useful background information there.

Dan has given me permission to post that here (thanks Dan) on his behalf. I think it may be useful in showing the many variables involved in trying to work out why there is a difference between AIS reported altitude and that reported by ADSB.

I reproduce his posts below for information as a discussion, in case it is useful in the future and does not get lost in a closed issue.

Post 1

Hi Roger,

This is reference your discussion over on the AIS-catcher github. You were discussing differences you see between the SAR aircraft below on ADS-B and AIS. I don't have a github account, so I put this here. I know you also follow the SP AIS group, but we are both much more active here. This is just additional info that might point you in the right direction.

reg = G-HMGF
hex = 407f58
type = B350

Short answer: I suspect the AIS altitude you are seeing for the SAR aircraft is completely GPS based, so it is a true altitude (not barometric). As a true altitude, it will not vary based on weather and atmospheric pressure.

The ADS-B altitude you normally see is barometric, and based on a standard day 2992/1013 altimeter setting.

AIS-catcher github discussion:
#482

Looking at that area, the local altimeter setting is around 29.65/1004. Normally, with a lower pressure like that, I would expect to see an aircraft showing a lower altitude by about 250 feet or so. Your aircraft doesn't match that, so there is something else going on.

If you look at the track for that aircraft, adsbexchange reports both the barometric altitude and the Geometric WGS84 (GPS true) altitude. It also reports the QNH setting that the aircraft actually has dialed in to the altimeter. Baro reporting by ADS-B is always based on standard day, but it helps to see what altimeter setting the crew has dialed in.

Here's the slightly strange part. Here are the QNH altimeter settings that the aircraft had dialed in today.

1003 = When they first arrived in the search area. This is good.
992 = The entire time they performed the search pattern.
1001 = When they left the search pattern. This is good.

The 1003 and 1001 above closely matches the current local altimeter. The 992 is a bit different, and would account for about a 90 or 100 foot difference.

For reference. wx info for EGKA just north of the search area today. The "Q1004" is the altimeter setting.

METAR EGKA 091750Z 20015KT 9999 SCT021 13/12 Q1004
METAR EGKA 091720Z 20018G28KT 9999 SCT016 13/12 Q1004
METAR EGKA 091650Z 19020G32KT 8000 -RA BKN011 13/12 Q1004
METAR EGKA 091620Z 19018KT 9999 BKN022 13/11 Q1003
METAR EGKA 091550Z 19019KT 9000 -RA BKN027 13/11 Q1003
METAR EGKA 091520Z 18019KT 9000 -RA BKN038 13/11 Q1003
METAR EGKA 091450Z 18020KT 9999 BKN038 13/11 Q1003
METAR EGKA 091420Z 18019KT 9999 SCT048 13/11 Q1003
METAR EGKA 091350Z 18019KT CAVOK 13/11 Q1003
METAR EGKA 091320Z 18016KT CAVOK 13/11 Q1003
METAR EGKA 091250Z 17013KT CAVOK 12/11 Q1003
METAR EGKA 091220Z 16013KT CAVOK 13/11 Q1003

So, I think you are seeing a combination of things. AIS altitude for this aircraft is probably completely independent of the aircraft avionics. It is most likely completely GPS based, and will always be a true altitude.

Aircraft ADS-B altitude is likely barometric, and based on a standard day altimeter setting. I'm not sure why the aircraft used a different altimeter setting during the search, but I would not expect that to make a difference in the reported altitude.

If you watch the difference between altitudes on different days, I suspect you will notice a change based on local altimeter setting and local barometric pressure. You will also notice that the difference varies greatly at higher altitudes.

Have fun watching that traffic. I rarely see my local SAR aircraft broadcasting AIS.

Regards,
-Dan

Post 2

On Tue, Dec 9, 2025 at 12:31 PM, Dan Henry wrote:

Here's the slightly strange part.  Here are the QNH altimeter settings that the aircraft had dialed in today.
 
1003 = When they first arrived in the search area.  This is good.
992 = The entire time they performed the search pattern.
1001 = When they left the search pattern.  This is good.
 
The 1003 and 1001 above closely matches the current local altimeter.  The 992 is a bit different, and would account for about a 90 or 100 foot difference.

Hi Roger,

Mainly trivia.

Reference their lower 992 altimeter setting during the search pattern, here are a few possible ideas. Remember that crew set QNH setting will not change the broadcast ADs-B altitude, but the setting is broadcast for ATC to see.

I wouldn't consider their 1,600 foot altitude as "low altitude", but they might use several different procedures during search, especially over water, on how they determine their actual local altimeter setting.

With multiple search/military aircraft close together, it is common to have a common altimeter setting. That lower 992 might have come from another site than the nearby airport at 1004.

For safety, a lower altimeter setting will always keep you a bit higher above terrain or the water. They might have used a forecast "min altimeter" setting for the area of 992.

Normal altimeters usually have an acceptable accuracy of +/- 75 feet when on the ground at a known test elevation with a local current altimeter. This means that worst cast, the pilot and co-pilot altimeters could be as much as 150 feet different. One could be max low, and the other max high, but both technically within limits. They might use a procedure over water to adjust to the "safest" of the two systems. This way you can sync the two altimeters to show the same displayed altitude.

Guaranteed that SAR aircraft has a radio altimeter. They might possibly have used a procedure to manually change the baro altimeter setting so that their altimeter matches the radar altimeter over water. I wouldn't expect them to do this at 1,600 feet, but it might be a very valid procedure for extremely low altitudes like a few hundred feet.

In a previous life, we flew for hours at 250 feet AGL doing terrain following in mountainous terrain. We also had a procedure where we flew visually at 100 feet AGL over water or flatter terrain. Flying 100 feet over water, at night, will definitely keep you alert! For that we relied on the radar altimeter, and the "low light" setting was set to 90 feet.

Regards,
-Dan

Post 3 (mine)

On Tue, Dec 9, 2025 at 07:19 PM, Dan Henry wrote:

That lower 992 might have come from another site than the nearby airport at 1004.

Dan

The plane was based at Humberside Airfport EGNJ, 200mi NNE of the search area, so that might explain that difference. That was where it started and returned. Low pressure further north than that location.

Roger

Post 4

On Wed, Dec 10, 2025 at 03:25 AM, Roger G7RUH wrote:

Yet another rabbit hole to explore....

Hi Roger,

When you explore that rabbit hole, also take a look at the link below. It does a good job of introducing ellipsoids, datums, and geoids. The GPS altitude is WGS84 based, and some systems can display the raw value, the corrected value, or both. I'm not sure which is going out in that SAR AIS altitude. The UK altitude correction is a bit larger than the US correction, around 100-150 or so feet I believe.

Mean Sea Level, GPS, and the Geoid
https://www.esri.com/about/newsroom/arcuser/mean-sea-level-gps-geoid

Things get really fun when dealing with different charts and all the datum/ellipsoid combinations available. Then you toss in dealing with different coordinate systems, English/metric units, and display formats, and things can get really strange.

One interesting paragraph from the link above describing how much goes in to "GPS altitude" and how complicated it gets.

However, most users expect accurate elevation readings that are related to MSL. Consequently, newer GPS devices output orthometric (geoid) height measurements as a product of “behind the scenes” calculations based on a combination of formulas, tables, and matrices that use geographic coordinates as inputs. The appropriate height for the geographic location taken from a coarse or fine DEM matrix is provided instead of a real measurement of the z value (or height). Some receivers use approximations of the geoid height to estimate the orthometric height from the ellipsoid height. Still other units, based on an older technology, provide direct readings of the z value based on the ellipsoid.

Regards,
-Dan

OT: the reference to rabbit hole came from my comment in another post on that forum where I keep on finding rabbit holes to explore, for clarity:

"OT: I figure that, as long as I do not meet a hookah smoking caterpillar** on my travels, I am still doing OK.
** From Alice's Adventures in Wonderland by Lewis Caroll, (for those who are interested in yet another rabbit hole)."