Temperature effects on altimeters: How is it handled IRL?

After flying the Phantom out of hot and high Nevada a lot recently, I noticed my altimeter consistently reporting my elevation 150~200ft below actual field elevation. After going down a rabbit hole of research I found this is normal and can be calculated, and the calculations are reflected in DCS with your altimeter being accurate on a standard day, understating on a hot day, and overstating on a cold day.

What I couldn’t find in that rabbit hole is how this is handled in real life. I always hear that pilots take the pressure reported by ATIS or tower, dial it in, and the altimeter should read field elevation, with no mention of any adjustment for high elevation or nonstandard temperatures.

So, do pilots just roll with the under/over-stated altimeter reading and mentally apply that difference to their pattern altitudes, etc., or does ATIS/tower give a pressure reading that’s somehow adjusted for that temperature to give correct airfield elevation?


It’s a big deal in cold climates and requires a bit of training in order to take advantage of the cold altimeter corrections that procedures can allow. My airline rarely flies to places where the lack of training and capability hurt us from a reliability perspective. So we take the hit of being forced to use rather high minimums for some non-precision approaches on cold days (< -15C). There is never any trick-fornication with the actual altimeter setting. Not trained to overcome these procedural limitations myself, I am not equipped to explain it fully. @Troll is probably the guy to chime in.


Yeah I found various documents talking about it in cold weather, which makes sense because the error could lead to flying into obstacles, but I can’t find anything about hot-and-high situations. FAA says altimeters should be accurate to within 75ft, but when hot-and-high the reading is going to be up to 200ft off unless somehow adjusted.


This pic illustrates it pretty well:

Better to be high than low, which is a the issue, and why some approach procedures require temperature compensation below a given temperature.

Our OPSSPECS allow us to fly temperature compensated approaches, but to be honest I’ve only done so a handful of times IRL. The FMS handles the magic (it’s literally a button press and compensated minimums post up), and I couldn’t tell you what our book says off hand without looking it up, which is what I have to go back and do whenever conditions are such that it might be necessary.

In the real airplane, at an airport without an AWOS, I’ll dial in the known field elevation and know that I have an altimeter setting that’s close enough to depart with. When you set the altimeter to what the AWOS reports, it’s rarely the exact field elevation, generally off by .01-3”hg.

Altimeter settings aren’t an exact science in real world aviation. Often center gives us one that differs slightly from what approach says which differs from what the ATIS says which differs from what tower has (but not by much). Whatever the last received (ATIS or tower) is controlling on the approach, IMO. We all have radar altimeters anyway, and can set minimums based on either radar or baro, although baro is normally used in our aircraft.

The main thing is that all the airplanes in a given airspace are flying at the same altimeter setting or close enough to maintain separation on their assigned altitudes.


So would that mean that QNH given by ATIS/tower is prioritized over whatever the altimeter says the pressure is if you set it to airfield elevation, even if it’s understating the altitude by 200ft or so?

What I’m still confused on is, say I’m at Nellis on a 45C day and ATIS (rather DCS) tells me QNH is 29.70. I dial in 29.70 and instead of getting 1,800ft (+/-75ft) field elevation I get something like 1,600ft. According to the temperature effects I researched this is normal, but according to a lot of other stuff I read/hear the altimeter should still read 1,800ft (+/-75ft) even with that 45C weather.

On one hand it doesn’t matter if a) everyone is on the same QNH, and b) I do mental math to avoid hitting mountains, but which is correct? The 1,600ft reading or the 1,800ft reading?

1 Like

Warm weather altimeter errors are never even mentioned. Honestly I’ve never given it a thought outside density altitude. As you’ve said we are all operating off the same error and those errors take us farther from the rocks, so, no problem!


Not temperature related, but where I’ve flown, mostly in the SE, runways often have a slope and vary from one end to the other. Or have a big dip or hump in the middle. They are almost never flat.

1 Like

True, and good point. Any airport could have several points where the actual elevation varied significantly from the published field elevation. Mostly what’s charted is TDZE for each runway, but any location on the airport could be higher or lower than the reference point.


That plate is a good example, a 70 foot difference between thresholds on 3R/21L.


Also: Oooof! I just got back from 3 back-to-back trips into LAS and saw temps of 41C, which was insufferably hot there. Walking across the ramp the pavement was nearly soft, lol. Anything above 40C in our jet and we’re having to leave fuel behind to make climb requirements, depending on the payload. Guess a Phantom probably doesn’t have that problem, but I bet it’s noticeably different than a standard day.

1 Like

FWIW, I’m pretty sure that most RNP approaches have a temp range for uncompensated altimeters, so that would address both positive and negative temperature extremes (by restricting the acceptable deviation) with systems not designed to address them.

Also, I think that temperatures were traditionally addressed (to a certain degree) by the broad brush approach. For instance, most enroute altitudes provide 2000ft clearance in the mountains, 1000 in the flats. So, if you are IMC and operating in these regions, you would most likely be well clear of any terrain by maintaining at least the appropriate minimum altitude, which would address most smaller temperature variations.

In the approach environment, remember that the point was, and arguably mostly still is, to get pilots low enough or close enough to the field to conduct a visual landing.

So, again in a sense, close enough is close enough. :wink:

Moreover, high temperatures (in addition to resulting in deviations in the non-crashy direction) are unusual in that, speaking anecdotally, you rarely have low ceilings/bad visibility and exceptionally hot temps at the same time. In the desert, rain will often never even make it to the ground in dead summer, evaporating into Virga before it does so. Arguably, you are more likely to encounter low ceilings/visibility from blowing sand than moisture out in the desert SW.

That’s why, again just an example, I don’t recall that either LAS or PHX have approaches lower than CAT I ILS. They statistically don’t need them often enough to justify the cost.

Incidentally, (since we are speaking of the Phantom and altimetry) if the F-4 altimeter is anything like the one in the A-4, it will lag with any appreciable climb or descent rate.

So, if you wanted to release your bombs at, say, 5000ft, you might need to pickle them off at 5500ft on the altimeter. IIRC, the correction used to be in the charts.

Just thought that worth a mention since it’s another example of just having to use TLAR at times when working with the tools we’re given…particularly in the old days! :slightly_smiling_face:


We do fly a lot in cold air, yes… :slight_smile:
Many years ago we used a rule of thumb that I think was add 4% to the minimum altitude for every 10°C below ISA (which is 13°C at MSL and drops 2° for every 1000ft altitude increase).
These days the vertical approach profile is coded in the FMC database and we just tell it what the temperature is at the airfield and the profile is automatically adjusted accordingly.


We’re a cheap airline I guess. No compensated baro for us; leastwise not on the 757/767 or 737.

It sounds like typical use is to set QNH as given by ATIS/tower and ignore the discrepancy from published airfield elevation, then? Doesn’t seem like anyone’s actually paid attention to what the altimeter reads after dialing in the QNH :joy:

1 Like

We notice the difference but ultimately it doesn’t make much difference, I guess. If I want to reference how close to the ground I am, the RA is my first glance. If it’s still caged then we’re 2500’+ and so a 100’ difference doesn’t register I guess.

It’s a different ballgame when you’re a tiny company than when you have tail #’s in the thousands. I hadn’t really considered it anything fancy, the previous operation I worked at had just under 30 and had it on about half the jets, depending on age and brand of avionics. I think it’s pretty standard on WAAS upgraded Collins suites, and every Garmin G3000/5000 has had it incorporated from the beginning as far as I’m aware. Then the OpSpec is just paperwork, I don’t think there’s anything too involved in getting authorization except maybe training burden. But also, 121 is always the last to adopt new technologies in the cockpit (how many are using LPV approaches now?). Mostly because there’s no need, virtually every runway they land on has a good ILS, for example. Synthetic vision is probably years from hitting an airliner cockpit because there’s no real benefit.

Cost optimization at large operations has always been rampant; see the classic olives in the in flight meals salad story.

I spoke to a crew the other day from the largest 91K/135 operator in the US, who told me their company just stopped subscribing to charts in the panel, and has reverted to only having them on the iPads, with the explanation that it will save them several million a year in costs.


Okay you can’t just leave that sitting there, with no explanation.


It’s one of those oft-repeated stories that continue to propagate through email fwd:fwd:fwd: chains and web forums, but probably has some truth to it.


On the ground with everyone experiencing the same pressure and temperature as the reporting agency (Tower/ATIS/AWOS), the altimeters read field elevation.

Ahh I think I get why this has you befuddled. You are wondering why we all just blindly accept that our altimeters are wrong when we are given the QNH by ATIS. This is not how QNH works. QNH is determined by the controlling agency by dialing their own calibrated altimeters located outside to the tower elevation, and reading the resulting pressure. This is why all altimeters ON THE GROUND are correct when set to QNH regardless of temperature.


Great posts @WarPig and @smokinhole!

My companies never had corrected baro (or LPV mins) either. The only aircraft that I flew that had it was the Cessna Citations that I flew in the Marine Reserves.

I agree about the large airlines being the last to change. I recall them stiff-arming the FAA about moving to RNAV for a long time and they will never abandon ILS until the GPS can get them to the same mins!

Good call on QNH as well. I have to admit I was trying to puzzle through that myself as, like Warpig, I couldn’t recall ever seeing the altimeter out of bounds, even in LAS, PHX, or RNO.


So if we’re talking DCS terms and let’s say there’s a functioning ATIS, the QNH they give us to show field elevation would be different than the QNH set in the misison editor due to the temperature difference?

Or alternatively, the pressure readout from their calibrated altimeters won’t necessarily match the pressure from a weather report?

1 Like