It's COLD! Are you too LOW?

We learn about the effects of temperature on our altimeters in Private Pilot training, but it's usually an aside, or something memorized for the test - "High to Low or Hot to Cold, look out below..."  But if it gets really cold like it has this winter in much of the country, there can be some really big, and dangerous, effects!

I got an e-mail from a listener to my interview on the Stuck Mic AvCast.  He is a pilot in Vermont, and asked the following question (I have edited for length and content):

"What kind of temperature spread does TERPS use to calculate required obstacle clearance for altitudes on stepdown fixes for an ILS? (i.e -10C to +20C etc...)  At BTV [Burlington, VT] we have an ILS 33... East of the airport is the green mountain range and so the glideslope is at 3.2 degree with stepdown fixes leading up to the FAF.

The crossing altitude at NIDUQ is 5500' but if you notice the highest prominent obstruction in the area is a mountain named "Camel's Hump" about a half mile north of NIDUQ. Camel's Hump peak is 4088'.

As you'd figure, in the winter time it gets very cold here in Vermont. The question is does 5500' crossing at NIDUQ meet required obstacle clearance when its say -20C?"

This is a GREAT question.  We learn in our pilot training that temperatures below standard (15 C at sea level) cause the altimeter to read higher than we really are - or to turn it around, the plane is lower than the altimeter reads.  This is discussed some in the Aeronautical Information Manual, section 7-2-3, and in the Instrument Flying Handbook,starting on page 5-5.  It is typically described as the pressure levels "bunching up" in cold weather.  See Figure 5-6 from the IFH below:

So the airplane is lower than we think.  But how much?  The error is increased by two factors - 1) the current temperature at the "reporting point" (usually the airport) and 2) the height above that reporting point.  There's even a chart!  Figure 5-7 from the IFH:

Important - notice again that the temperature is that recorded at the airport, not at your current altitude.

You can see that for mildly cold temperatures and altitudes close to the station, there is only a little difference.  But for even colder temperatures and higher altitudes above the station, the error can be in the hundreds of feet!  This, of course, is only an estimate, and relies on certain assumption about lapse rate - nonstandard lapse rate or temperature inversions will affect this calculation.

Let's go back to the approach in question, KBTV ILS RWY 33.  This approach has a segment that passes over the Camel's Hump mountain.  Now, the writer asked if 5500 MSL to NIDUQ offered sufficient clearance, but the real question is does the 4800 MSL to HONIB clear the obstacle?  Let's work it out!

Camel's Hump is 4088 MSL and is shown on the chart between NUDIQ and HONIG.  Minimum altitude on this segment is 4800.  Let's say the temperature at BTV is -20C.  Does it ever get to -20C (-4F) in Burlington, VT?  Absolutely, at least several times a year (source - aviationweather.gov). 

KBTV is at 335 MSL.  The segment altitude, therefore, is 4800-335 = 4465 ft above the station elevation.  Using the chart above, find where -20C intersects 4465 ft.  There is no entry for 4465 ft, but that's about halfway between 4000 and 5000 ft, so I'd say the altitude error is about halfway between them, or 640 ft.  And remember, this is in the WRONG direction - so you and your altimeter think you're at 4800 MSL, but you're really at 4800-640=4160 MSL.  That 4088 MSL mountain is looking pretty close, isn't it?  How accurate is your altimeter?  The general rule on preflight is within 75 feet, so you might be at 4160-75=4085.  Now you're 3 feet below the peak of that mountain.  And that's only if you're maintaining altitude right on 4800 on the altimeter - have you ever been 20 or 40 feet below an assigned altitude?  Of course, everyone has.

So what do you do?

Back to one of the original questions - what consideration does TERPS (the approach procedure design rules) make for low temperatures?  The answer, in the vast majority of cases (except RNAV (RNP) procedures and Baro-VNAV minimums on an RNAV (GPS) procedure) is "none"!

This correction is left up to the pilot.  Some more advanced avionics systems may compensate for this, but that is on a case-by-case basis, and you'd have to read your specific flight manual.  Back to the IFH:

Under extremely cold conditions, pilots may need to add an appropriate temperature correction determined from the chart in Figure 5-7 to charted IFR altitudes to ensure terrain and obstacle clearance with the following restrictions:

* Altitudes specifically assigned by Air Traffic Control (ATC), such as "maintain 5,000 feet" shall not be corrected. Assigned altitudes may be rejected if the pilot decides that low temperatures pose a risk of inadequate terrain or obstacle clearance.

* If temperature corrections are applied to charted IFR altitudes (such as procedure turn altitudes, final approach fix crossing altitudes, etc.), the pilot must advise ATC of the applied correction.

For me, any time the airport is reporting around -10C or colder, I'll be paying much closer attention to that correction!

Approach procedure design considerations - example

During my interview with the Stuck Mic AvCast, we discussed a few of the approaches I designed while at the FAA.  I was asked several questions about what kind of considerations go into approach procedure design, so I wanted to go into more detail using an example.  For this, I'll discuss the St George, Utah, KSGU (formerly KDXZ) LDA/DME RWY 19 approach.  

This airport was one of my favorite projects.  The complexity of the terrain, the fact that it was at a brand-new airport, and the unique orientation of the localizer posed some interesting challenges.

This is an LDA approach, not a "localizer", that also has a glideslope installed.  Note the position of the antenna, at the approach end of the runway (normally a localizer is oriented at the far end of the runway, to provide signal guidance all the way down to rollout in some cases).  Also, the antenna itself is physically aimed about 7 degrees to the east instead of directly in line with the runway.  This siting was necessary in order to aim the final approach course between two mountains.  Had it been a standard straight-in alignment, the final would have come too close to the 10024 MSL peak.  So it's essentially an offset ILS.

This procedure has quite a few stepdown fixes, especially in the intermediate segment before the Final Approach Fix of PAYLR.  Determining the location and number of these fixes took a lot of manual evaluation of the terrain contours starting from HOPEB inbound.  You can see the rising terrain on the east side of final, but one of the most troublesome areas was actually the medium tan terrain on the northwest abeam WATLA and IPPOD.  It's higher and rises steeper on that side.  Fix placement is a balancing act - move the fix further south ("down" the mountain) and maybe you can get a lower altitude, but then the distance to the next fix is reduced, so that might require an excessive descent.  Move it further north ("up" the mountain), and you get more distance to descend but also a higher altitude to descend from.  Depending on the exact geometry of the situation, you may have to try several times to get just the right combination, as I did in this case.

One last thing - notice the note in the profile view "Loc unusable inside 0.7 DME".  What does that mean?  On any localizer, the signal gets narrower as you get closer to the antenna.  Usually, the localizer antenna is at the far side of the runway, so it's at least a mile away once you get to the missed approach point - getting too close to the antenna is not a factor.  On this procedure, however, the antenna is right there at the approach end of the runway.  So, at 0.67 nm from the antenna, flight inspection deemed the signal was too narrow to follow anymore, hence the restriction at 0.7 DME.  Notice this is also the location of the MAP when not using the glideslope.  Also, the DA for the LDA/GS line of minima is calculated to put you right at that 0.7 DME point when you reach 3170 MSL (286 feet above the touchdown zone elevation).

Those are just a few of the various types of considerations that go into an approach procedure like this. Generally, the more complicated the terrain, the more factors must be considered. Sometimes it can seem like a great big puzzle!

Interview with the Stuck Mic AvCast!

I was recently interviewed by the Stuck Mic AvCast for their most recent aviation podcast!  We mostly talked about my previous works designing instrument approach and departure procedures for the FAA. There was a lot of great discussion and lots questions (that I hopefully answered reasonably coherently), and it was a whole lot of fun.  These folks air a new podcast on the 1st and 15th of every month at the usual podcast locations or on their website.  Mine was just released today, check it out!

Stuck Mic AvCast

Direct link to the page with the podcast

Aspen EFD1000

At my flight school, we have the Aspen EFD 1000 Pro Pilot system installed in three of our aircraft.  The more I fly with this system, the more amazed I am with its capabilities.  The Aspen was designed as a replacement/upgrade for aircraft with standard 6-pack instruments, but it does so much more.  The Pro Pilot adds an HSI, bearing pointers, ground track icon, wind direction indicator, airspeed and altitude tapes with bugs, TAS, basically all the things you'd have with a glass cockpit, but in a package that just replaces the center two instruments - the AI and DG.

On a lesson tonight, my student flew the HAO ILS RWY 29.

I like this approach for two reasons.  One, it has intersections made up of NDB bearings (as well as DME fixes and VOR intersections), which is somewhat unusual but good practice with NDBs.  Second, the missed approach procedure takes you back to the Final Approach Fix, where you can fly the approach again.  We were able to fly the approach three times in rapid succession, helping to really drill on flying an ILS.

The Aspen, though, almost makes it too easy!  Our 15-20 degree crab angle at altitude was readily apparent and easily adjusted for.  At one point I unselected the Localizer from the HSI display so the student had to fly using raw data on a regular CDI.  Here's a pic of the Aspen before glideslope intercept (yes, we're at 3000 when GS intercept is at 2600).  Sorry for the slightly blurry picture, it was at night.  The HSI (green arrow) is set to display the Localizer, while the number 1 bearing pointer (blue arrow) is tuned to the RID VOR to assist in identifying HOLGR.  Unfortunately, the aircraft we were flying did not have an ADF installed, so we couldn't practice NDB work.

Instrument - descent rates and back courses!

I recently conducted a PCATD ("simulator") lesson with one of my instrument students.  When I "fly" on any type of simulator, I like to choose approaches which are interesting, appropriately challenging depending on their stage of training, and most definitely not local!  Why would I want to practice local approaches in the sim?  It's my goal to train the student to be a good instrument pilot, not a good "Southwest Ohio" instrument pilot.  (Insert your local area as appropriate.)  So I search out approaches which meet the training needs of the student and the syllabus.

Okay, enough philosophy.  Two of the approaches we flew recently were in Oregon, and they had some interesting challenges and educational value.

One was the KMFR VOR-A:

I like this approach (remember, I have a weird definition of fun).  It has three different radials involved all off the same VOR, so setting up the radios is important.  It has a little bend at the FAF, which could be easily overlooked if you're rushed as you begin your descent.  But what really caught my (and my student's) attention was the descent within the Procedure Turn.  Notice it requires 1700 feet of descent.  The maximum for a 10nm PT is 2000, so this is up there.  Now picture a normal 1-minute PT.  By the time you get turned around and lined upon the R-342 inbound, even if you do it perfectly, you now have barely more than a minute to lose 1700 feet.  That is a pretty excessive descent rate, especially in something like a Cessna 172. 

So how do you fly it safely?  Really the only way is to not fly a 1-minute PT, but to extend that outbound leg to 2 minutes.  Or even 3.  But you have to be careful, remember you need to stay within 10 nm of the VOR.  Naturally, the place to figure this out is not once you're inbound on the PT, by then it's too late - proper review of the procedure beforehand and a thorough approach briefing are important.

This procedure also brings up the question of "why is it a circling-only procedure? It's lined right up with the runway."  You do see this sometimes.  When the altitude to lose from the FAF to the runway is too steep for a normal approach (defined in the TERPS as 3.77 deg for Cats A-C, and 3.5 degrees for Cats D/E), the approach is labeled as circling only.  I calculate about a 5.3 deg glidepath for this one.  This means some maneuvering may be required once in visual conditions to land.  Of course, if you get the runway in sight in time to land safely straight ahead, there is no requirement to perform a circling maneuver.

Fun!

The other approach we "flew" and I wanted to discuss is the KSLE LOC BC RWY 13:

One of the topics of the lesson was to introduce LOC Back-course approaches.  This is where the localizer needle reads "backwards" from normal - to get it to center, you have to fly away from it!  This is referred to as "reverse sensing", and to fly it, you need to "pull the needle" back to center.  It's really not that difficult once you flip that little switch in your brain, but if you're maneuvering to get lined up on final it can be a bit of a mental exercise.  Once on final everything usually goes pretty smoothly, unless of course you forget it's a back course!

For this approach, I let the student enter the procedure from the UBG VOR.  Even I thought it would be cruel and unusual punishment to require him to fly the route from CVO, with a Procedure Turn, on his first exposure to back courses (there's always next time though...)

This procedure has an unusual note, notice the third line in the notes box "ARTTY INT not authorized for final approach fix."  What on earth does that mean?  Simply, it means that when you're inbound on final and identifying ARTTY as the FAF, you HAVE to use either DME, or the outer marker to do so.  Notice the route from CVO also proceeds to ARTTY as the IAF - for this route, ARTTY can be identified as the intersection of the CVO R-359 and the localizer course.  But once inbound on final, you must use DME or the marker beacon.  Why?  Look at the distance from CVO to ARTTY - 31.5 nm!  Since VOR radials spread out as you get farther from the transmitter, this means that the signal accuracy is okay to use when you're just entering the procedure (and at a higher altitude).  But once you get lower and are inbound approaching the FAF, the signal accuracy is no longer acceptable.  Logically, FAFs require somewhat tighter signal tolerances than IAFs and other fixes further out.

The missed approach also brings up a useful discussion - tracking inbound on the back course is the same as tracking outbound on the front course, so as you pass over the runway and fly the missed approach, nothing changes - you're still pulling the needle.  This is also true for the first half of the entry for the missed approach hold.  But once you get turned around, you're flying inbound on the front course, in which case the needle now reads normally!  Once in the hold, the needle remains reading normally, so once you flip the switch in your head back to "normal", leave it there.

It was an interesting and educational day.  (And yes, the student did great.)