In an article comparing constant-angle, nonprecision, straight-in instrument approach procedures (IAPs) with “dive-and-drive” (stepdown) IAP designs (“Continuous Descent,” ASW, 7/13), the article’s source makes a case against airline flight crews’ continued use of dive-and-drive IAPs.
In dive-and-drive, the pilot flying passes the depicted final approach fix (FAF) — or, if not depicted, the point where the aircraft is established inbound on the final approach course — at a specified altitude. He or she then typically descends, possibly with one or more stepdown fixes, to arrive at the minimum descent altitude (MDA) prior to reaching the missed approach point (MAP). On many IAPs today, a vertical descent point (VDP) provides a defined point on the final approach course from which a “normal” (usually 3.0-degree glidepath) descent from the MDA to the runway touchdown point may be commenced, provided the pilot has established visual reference.
But, as importantly, the VDP is almost universally considered now to be the last point from which it is acceptable to be in your descent from the MDA to the runway. Beyond the VDP, you’ll be too close and, hence, too steep. Aviation regulations generally specify that the pilot cannot descend below the MDA unless in a normal position to continue the approach to landing. This is not a trivial point, as a pilot very likely would run into trouble with a pilot examiner during a practical test if he flew past the VDP, and then decided to descend for landing. Either he would be steeper on approach than is deemed safe, or he would touch down beyond the touchdown zone.
The ASW article’s source first argued that dive-and-drive IAPs inherently are prone to becoming unstabilized because of the fairly steep rate of descent — usually between 1,000 and 1,500 fpm (the “dive”), which operational experience and research consider unacceptable below 1,000 ft above ground level — required to level off at the MDA sufficiently prior to the VDP. Second, even if prior to reaching the VDP, the pilot has sufficient visual reference to continue the approach, he must fly level until reaching the VDP (the “drive”), and then re-establish a descent. The source then said that an added problem is that once you re-start the descent, you may also have to reconfigure the airplane. In his words, “You’re configuring the airplane late, quite often reading checklists right down to the last moment.” Aside from all other considerations, this was deemed contrary to today’s stabilized-approach safety philosophy. I agree it would be contrary to stabilized-approach philosophy, but I don’t agree that it is usually, or even commonly, called for.
A better method, according to the article, is the constant-descent approach (CDA). Here the pilot flying conducts an IAP designed or adapted for a continuous rate of descent from the FAF/FAF altitude to arrive at the MDA and VDP at the same time. The pilot then just continues the descent, assuming sufficient visual reference has been established, to landing. Visual reference is a complex concept beyond the scope of this article, so I’ll just say “runway in sight” for simplicity from now on.
I disagree with the CDA-only philosophy as presented for a few reasons. First, as pointed out in the article, not all airplanes have the capability to generate the guidance that many airlines require for a CDA. The source noted that some nonprecision, straight-in IAPs provide a table showing distances vs. altitudes which, if followed, approximate the proper descent rate. For IAPs that do not, the pilot would have to calculate these numbers as the airplane descends or refer to another tool if available.
Yet doing this mental calculation accurately, which is critical this close to the ground, is not assured, as pointed out in the article itself. The pilot must accurately calculate while flying the airplane, looking for the runway and maintaining situational awareness. Although letting an autopilot do the flying can help, if you get behind — or, worse, err in your calculations — you can put yourself in a bad situation. Even when provided with a table giving you distance vs. altitude numbers, having to constantly refer to it by looking down at the approach plate for the number, then to the flight instruments to see if “you’re there,” and also looking out the windscreen, is not easy or, perhaps, safe. Humans are not great multitaskers, and flying an approach in bad weather already involves lots of tasks, without adding math — and math that must be accurate.
Second, as mentioned, I disagree with the article’s quote about level of risk while changing configuration and doing checklists after passing the FAF. There may be some airlines that specify this, but I rather doubt it is anywhere near common. I can’t provide substantiation, but I know from the three or four operators with which I am familiar, that their standard operating procedures specify that the airplane be configured in the landing configuration by the FAF, with all checklists complete. This is required specifically so that the pilots do not have to concentrate on anything other than safely flying the last phase of the approach, and fully supports the stabilized-approach safety concept. This would apply in both CDA and dive-and-drive cases.
Third, the MDA of a nonprecision IAP is the lowest altitude to which a pilot can fly prior to having the runway in sight, as opposed to reaching the decision height (DH)/decision altitude (DA) on a precision IAP. The DH/DA allows continuing the descent for the brief period while the pilot decides whether there is sufficient visual reference to continue to land, or executes a missed approach. Not so with the MDA.
When you fly a CDA, the goal is to reach the MDA at the VDP. If you don’t have the runway in sight at that time, you must go around. But the problem is that with a CDA, by definition, you’re still in a descent as you reach MDA and make the go-around decision. If you see the runway, no problem; if you decide that you must miss the approach, you are below MDA in violation of regulations, and you may also be unsafe. If, for example, it takes 150 ft to transition from your descent to level flight at the MDA, then you must start to level off 150 ft prior to reaching the MDA. You are limiting yourself to having to either see the runway 150 ft above minimums, or leveling off at minimums where, if you see the runway, you’ll be in the same position as if you’d done a dive-and-drive to begin with.
There’s also the chance that the descent angle required from the FAF/FAF altitude to arrive at the MDA and VDP at the same moment will not be 3 degrees; this is not something the pilot has a choice in. I assume the goal in establishing the location and altitude of the FAF is to locate it on a nominal 3.0-degree glidepath to the touchdown point. But if anything (obstacles or perhaps other altitude restrictions) dictates that the FAF altitude be higher or lower than that provided by a 3-degree glidepath, then the descent path during the CDA will get you to the VDP via a steeper or shallower gradient than you want from the VDP to touchdown.
So if the idea of a CDA is to look out the windscreen as you approach the MDA-VDP, see the runway, and then continue the same rate of descent to landing, you may end up short or long. I understand that judging the final descent is what we get paid to do, but using a CDA may set you up to start off at the wrong descent rate.
Will the latest versions of dive-and-drive IAP solve any of these problems? No and yes.
No, because in a dive-and-drive you will, in fact, have to transition from level flight at the VDP to a stable descent. This will make you slightly steeper than the nominal 3 degrees. A mitigating factor is that, at least in the United States, pilots are allowed to begin a descent up to 0.2 nm (0.4 km) prior to a charted descent point. But as the ceiling is just as likely to be the reason you can’t see the runway; if you level earlier at the MDA, you may see the runway earlier and then be able to start the descent by the VDP.
No, because the airplane will be level at a more nose-up attitude than if it reached the VDP in a descent. This will likely restrict the pilot’s forward and down vision to some extent, depending on airplane design.
Yes, because the pilot is not giving up the last 150 ft or so of altitude approaching the MDA, where he might see the runway. If the ceiling is very near the MDA, then a CDA is likely to end in a missed approach for that reason.
Yes, because the pilot does not have to do math or continuously refer to the approach plate, checking altitude to go vs. distance while performing all other required tasks.
Yes, because if an autopilot can be used for vertical navigation, the pilot can let it level off and fly at the MDA while devoting more crosscheck capacity to looking for the runway prior to reaching the VDP. Even if this saves only a few seconds, it provides the pilot with a chance, when the weather is marginal, to see the runway before reaching the VDP.
If an autopilot is used to fly a CDA, which is a requirement at some airlines, then the autopilot will allow you to do the same search for visual reference, but it’s going to transition to level-off mode before you reach the VDP, as mentioned above. And if you have to do the math or check the altitude/distance table, there is not much time to be looking outside.
While considering a response to this article, I had the fortune to change seats at my airline. My new airplane, a large commercial transport jet, has flight management/autoflight systems that are fully capable of not only flying the airplane via either dive-and-drive or CDA IAPs, but they also can calculate the constant descent angle required and display this, very much like an instrument landing system glideslope. No math or table required.
And still, I’ve experienced exactly the dilemma mentioned: On each of the CDAs I fly (actually, our policy is that the automation must fly CDAs), when the pilot monitoring calls ”approaching minimums“ 200 ft above the MDA, I know that if I don’t have the runway visually in the next 40 ft, the autopilot will start to level the aircraft at the MDA. Once that begins, I have no choice but to commence a missed approach, as the airplane is already past the point where I can cross the VDP at the MDA in a descent, which are the criteria my company (rightly) requires: That’s the entire point of the CDA.
While I’ve always looked with suspicion on the concept of a pilot-derived CDA, it was mainly because of the added mental workload needed to constantly monitor the “distance to go vs. altitude” progress, and also because of how often I’ve made errors doing this calculation in line operations. I’ve often gotten behind, or paid too much attention to the math, and lost situational awareness in one way or another.
But by practicing CDAs in training with an airplane that does all the calculating and flying for me, I’ve become more concerned about the other issue: If I’d been able to descend to the MDA even a few seconds prior to the VDP, assuming the ceiling was right at minimums, I could’ve gained sufficient visual reference to descend below the MDA. True, I’d have had to start down very slightly before reaching the VDP to get on a 3-degree descent angle, or accept a slightly steeper descent by leaving the MDA at the VDP. I feel those are acceptable choices, and certainly no worse than the possibility of not seeing the runway until the airplane is already leveling off, and then having to miss the approach.
Even if I have the runway in sight as the autopilot levels the airplane at the VDP, I’m not supposed to disconnect the autopilot and dive back onto the glide path. So what do I do? Set up for another approach? What if that’s the only IAP available? If I assume, because I just managed to see the runway as I passed the VDP after the autopilot leveled me off on the last approach, that I’ll see the runway next time, do I disconnect the automation on the next try and hand fly the descent the last 150 ft to the MDA, so the airplane doesn’t again start to level off? What if the weather has changed and I don’t see the runway this next time? Then I bust the MDA. Is this so unsafe as to be a real issue? I can’t answer that.
But I know that I could do that nasty old dive-and-drive IAP on my second approach, make sure I get to the MDA soon enough to ascertain I can still see the runway at minimums, and then start the descent just prior to, or upon, reaching the VDP. Yes, it might take a higher rate of descent from the FAF to get to the MDA early enough, but I don’t feel it would be such a high rate as to be a greater threat than not seeing the runway and having to again miss the approach. And if I had done the dive-and-drive the first time around, I might already be on the ground, with more fuel. Perhaps, while planning for the approach, if I know it’s going to take an unacceptably high rate of descent to reach the MDA prior to the VDP, then it’s time to acknowledge that that approach, and today’s weather, just aren’t a safe combination.
All of this is not to say that I don’t like the CDA concept. I just don’t like CDAs when the ceiling is close to the MDA. And how can I be sure of where I’ll break out in changing, or marginal, conditions?
There is one other possibility: Have the regulatory authorities recognize the problem of flying approaches to MDAs. Instead, convert all MDAs to DHs or DAs. That will likely mean the DH/DA will have to be somewhat higher than the old MDA for the same approach, to allow for the transition from a descent to a climb in the event of a missed approach. But at least then the pilot could reach DH/DA at the VDP and decide what to do. If there’s no runway, miss the approach knowing that the airplane will (legally) descend (safely) below the DH/DA while transitioning to the missed approach climb or, if the runway is in sight, then the descent path can be continued. Of course, this doesn’t address my concern about today’s airline pilots calculating their own CDA flight paths. That is their individual decision.
In the past few months, I’ve learned that this appears to be exactly what another airplane model in my company’s fleet does. Some of the IAPs have been modified to take into account the amount of altitude used in leveling off during the CDA. An altitude that is not an MDA is used as the missed approach decision point, prior to the autoflight system beginning the level-off. I look forward to the day when all nonprecision IAPs make use of this new idea, with an aircraft-generated CDA.
Alan Gurevich is a captain for an airline operating wide-body aircraft in both U.S. domestic and international operations. His background includes flying military fighters and transports, and engineering work in both the aerospace and systems safety fields.