Landing too fast (xhtml w3c 12/09)

Decreasing your exposure to risk

3. Don't land too fast in an emergency

Rev. 5 — page content was last changed 19 December 2009
consequent to editing by RA-Aus member Dave Gardiner
www.redlettuce.com.au

Content

3.1 Minimising touchdown velocity in a forced landing
3.2 Plan and control the potential crash!
3.3 Deceleration forces and energy absorption
3.4 Your final defence perimeter — a safe occupant zone

Forced landings in recreational aircraft — due to engine/propeller failure or fuel starvation, exhaustion or contamination — are certainly not uncommon; but our pilots cope well and, in terms of injury, recreational aviation forced landings are generally uneventful. But occasionally something goes wrong. Light aircraft accident statistics from the US indicate that the most prevalent cause of a forced landing gone wrong is because the approach is too fast, leading to a heavy impact perhaps followed by a bounce and capsize.

Could this happen to you?

While flying at 3000 feet near Rochester, Vic in the company of two other aircraft, the pilot of the Corby Starlet reported that his engine had stopped.

The landing area selected was a flat irrigation bay 40 metres wide extending 575 metres north-south and covered with short [50 mm] grass on a firm base. There was an east-west fence line across the bay approximately 175 metres in from the north boundary leaving 400 metres unrestricted for the landing. It would have been difficult to see this fence from a distance, however the aircraft had not made contact with it.

Landing was from the north on a heading of 180 degrees. At about 410 metres from the northern boundary of the bay or 235 metres into the available landing area (165 metres remaining) were the start of the fresh wheel markings on the ground. They continued south for about 48 metres. It appeared that the plane may have then lifted off the ground for a short distance and then touched again. The wheel marks then continued for about 54 metres to an impact point where the plane hit a mound beside an irrigation drain. It was flung into the air, striking a fence and landing inverted on an internal farm road.

The witnesses to the fatal accident, including two people on the ground, described the final approach of the aircraft's emergency landing, as "looking too fast" and again, the aircraft "flying and descending to about half a metre above the ground and maintaining that height above the ground". The touch down was made two thirds of the way along the available field . "His speed did not seem to drop in the length of the field." Ground level wind was reported to be nil.

3.1 Minimising touchdown velocity in a forced landing

The problem with always using the best glide speed for distance

Following power loss the importance of establishing the aircraft at the best glide speed for distance [Vbg], appearing in the aircraft flight manual or pilot's operating handbook [POH], is emphasised in training and in the text books. This emphasis is valid to the extent it provides a reasonably safe initial flight speed to attain and hold whilst ascertaining the situation, planning appropriate actions and subsequent manoeuvring into the final approach position. In simulated engine failure procedures Vbg is often used throughout the approach simply because it is safer to do so; but it may not be best practice for the real thing. See 'V-speeds' for an explanation of the various codes.

Note: the Vbg stated in the POH is for MTOW and should be decreased by half the percentage reduction in aircraft weight from MTOW — and of course the Vs1/Vso stall speeds decrease in the same way. For example when there is no passenger in a two-place aircraft gross weight might be 16% below MTOW thus Vs1/Vso and Vbg (and Vmp below) are all reduced by 8%. So if the POH states Vs1 is 40 knots and Vbg is 60 knots but actual operating weight is 16% below MTOW then adjusted Vs1 and Vbg are 37 knots and 55 knots respectively.

There is often an impression that in an emergency the pilot should peg Vbg and stay with it otherwise the consequences may be dire. (This concept pre-supposes that a reasonable landing site is always at extreme range and that Vbg is a fixed value.) What may not be mentioned is though Vbg provides the lowest glide angle (the flattest path and hence the longest air distance), it provides neither the lowest forward speed nor the lowest rate of sink. (The term 'rate of sink' is synonymous with 'negative rate of climb'.) Airspeeds lower than Vbg should generally be used in the final stages of an approach in a real forced landing, unless the landing site selected is familiar to the pilot and known to be dimensionally ample and quite safe.

speed/sink rate polar curve

This airspeed/sink rate curve is a generalised plot of the relationship between rate of sink and airspeed when gliding an erect light aircraft in still air with the propeller stationary (a windmilling propeller increases drag); it is essentially an inverted power curve. Vmp, the speed for lowest rate of sink or minimum power (i.e. drag × speed), is at the highest point of the curve. The best distance glide speed is ascertained by drawing the red line from the zero coordinate origin tangential to the curve (i.e. just touching); the point of contact is where the ratio of rate of sink to airspeed is at a minimum and Vbg is directly above the contact point. Also the angle between the red line and the horizontal is allied to the angle of descent and it is obvious that Vbg occurs at the smallest possible descent angle, though it can be seen that even in nil wind conditions Vbg is not a clearly defined point value; rather, it's the mid-point of a speed range for maximum glide distance.

However, aircraft don't often operate in still air conditions and Vbg should be increased or decreased so that the ratio of rate of sink to ground speed is at a minimum. When countering a headwind the ground speed will reduce so that the ratio of rate of sink to forward speed will be less favourable. It is usual to increase Vbg by one third to one half of the estimated wind speed, which improves penetration by increasing ground speed. Although the rate of sink then also increases because of the increased airspeed, the ratio will again approach the minimum. This could be represented on the curve by starting the red line not from the zero coordinate origin but from a point on the airspeed axis a distance to the right that is equivalent to the headwind speed. In theory the reverse situation applies for a tailwind. In practice there are more important tailwind aspects for concern, which will be covered further on.

Vmp — the speed for minimum rate of sink

The final approach speed that will allow minimum kinetic or impact energy conditions is Vmp. What is apparent from the curve is any glide speed between Vmp and Vbg will provide a lower forward speed than Vbg, together with a slight reduction in rate of sink. Of course the glide path will be steeper, thus distance achieved from any particular height will be less than that achievable at Vbg. For example with Vbg of 60 knots (30 m/s) and a sink rate of 3 m/s an aircraft at a height of 60 metres would remain airborne for 20 seconds and travel forward 600 metres in nil wind. At Vmp of 50 knots (25 m/s) and a sink rate of 2.75 m/s the same aircraft would remain airborne for 22 seconds and travel forward 550 metres. (To convert feet per minute to metres per second divide by 200.)

At speeds greater than Vmp there is a possibility of converting glide momentum into maintaining height for a short period — at dive speeds there may even be enough energy to loop the aircraft. However, at any speed less than Vmp there is no possibility of converting glide momentum into short-period maintenance of height; whatever control change you make will result in an increased sink rate.

In the airspeed/sink rate diagram stall point is shown at Vs1, Vmp is shown at around 1.2 times Vs1 and Vbg around 1.4 Vs1. The angle of attack at Vbg may be around 4–5° and perhaps 7–8° at Vmp. The increasing aoa at the sub-Vbg speeds reduces the safety margin between flight speed and stall speed, so airspeed should only be reduced to Vmp in a stabilised approach after all significant manoeuvring is complete. Descent at Vmp in poor visibility lessens impact a little if contact is inadvertently made before flaring. In turbulent conditions the pilot must balance the possible safety of a higher airspeed against the higher impact forces brought about by that extra speed. We discuss the effects of low-level turbulence and wake vortices in 'Wind shear and turbulence'. Also pilots, particularly of low-momentum recreational light aircraft, should be aware that if a wing tip is first to make contact at low forward speed there is the possibility of cartwheeling.

Kinetic energy can really hurt!

You may find pencil and paper a helpful back-up from here.

The kinetic energy [KE] of a body is due to its motion and equals ½ mass × speed squared [½mv²], thus as speed changes linearly KE changes exponentially. For example a 540 kg aircraft with a stall speed of 42 knots CAS might have a Vmp around 50 knots CAS, so in nil wind conditions the KE when touching down at 50 knots (about 25 m/s) is ½ × 540 × 25 × 25 = 169 000 newton-metres [N-m]. Vbg for the same aircraft might be 60 knots CAS (near enough to 30 m/s) and touchdown KE at that speed would be ½ × 540 × 30 × 30 = 243 000 N-m; a 44% increase in energy at touchdown because of a 20% increase in speed. The distance required to bring the aircraft to a safe stop is directly proportional to the touchdown energy, as is the impact energy, if the aircraft and occupants are stopped prematurely.

The pilot of a low-momentum recreational light aircraft is exposed to much less KE. For example consider a 270 kg aircraft with a stall speed of 28 knots, Vmp 34 knots (17 m/s) and Vbg 40 knots (20 m/s). At Vmp touchdown KE = ½ × 270 × 17 × 17 = 39 000 N-m, while at Vbg touchdown KE = ½ × 270 × 20 × 20 = 54 000 N-m. The aircraft weight is half that of the heavier aircraft but impact KE is one-fifth.

Kinetic energy is substantially affected by wind velocity

Usually when we discuss in-flight energy management the aircraft speed (in the equation KE=½mv²) is that which is relative to the air — the true airspeed. However, for the purpose of measuring impact energy or the work that has to be done — i.e. the energy expended — to bring the aircraft and occupants to a stop, the speed is not true airspeed but the velocity resultant of ground speed and rate of sink; thus touching down into wind with a low sink rate will make a very favourable difference to energy level.

If the 270 kg aircraft's Vmp ground speed is reduced by 6 knots (18% reduction) to 28 knots (14 m/s) by landing into a 6-knot wind then KE is reduced 33% to 26 500 N-m. In the same conditions if that aircraft was landed downwind then ground speed would be 40 knots or 20 m/s, and the KE to be subsequently dissipated would be 54 000 N-m — the possible impact would be twice as great as landing into wind — and 6 knots is just a pleasant light breeze. The figures for the 540 kg aircraft landing at Vmp with a 6-knot headwind and tailwind are 131 000 N-m and 212 000 N-m respectively; quite a difference from the nil wind impact of 169 000 N-m even with just that light breeze. (It also underlines the fairly obvious expectation that low-momentum recreational light aircraft landing into wind at minimum speeds don't have a lot of energy.) There is a very significant advantage in a low-speed into-wind forced landing, but other conditions — such as clear landing distance available — may modify this.

A worse case is when the aircraft touches down both fast and downwind. For example if our 540 kg aircraft touches down somewhat fast, say 65 knots CAS, with a 6-knot tailwind then the KE at touchdown is ½ × 540 × 35.5 ×35.5 = 340 000 N-m; twice the energy at 50 knots in nil wind conditions. It also reinforces the point that, with the ever-present possibility of engine stoppage or degraded performance, it's a silly decision to take off downwind, no matter how long the distance available — unless it's a one-way downhill strip.

But let's look at the case where only the pilot is on board, our aircraft weight is 16% below MTOW at 454 kg, and the pilot lands at a Vmp reduced by 8% to 46 knots into a 6-knot headwind. Then the KE at touchdown is ½ × 454 × 20 ×20 = 91 000 N-m, a very significant decrease from the previous 340 000 N-m.

Density altitude adds kinetic energy

Density altitude also affects KE because TAS is about 1.5% higher than CAS for each 1000 feet of density altitude. For example the density altitude at Armidale, New South Wales (elevation 3500 feet) with temperature around 30 °C might be around 6000 feet, which means that the TAS will be about 9% greater than CAS. So using the preceding example of 540 kg touching down at 65 knots CAS, the TAS at 6000 feet density altitude would be 71 knots. So adding the 6-knot tailwind for a touchdown groundspeed of 77 knots the KE is then ½ × 540 × 38.5 ×38.5 = 400 000 N-m. The density altitude adds 60 000 N-m; quite a lot of energy that you need to be aware of.

Doing the KE calculations for the aircraft you fly

Kinetic energy calculation is easy if you first halve airspeeds and windspeeds to convert from knots to metres per second and express the operating weight (mass really) in kilograms; i.e. KE= ½ operating weight × groundspeed squared — the result is in newton-metres or joules if you prefer.

3.2 Plan and control the potential crash!

There is some element of chance in every emergency landing (Murphy's Law suggests that what can go wrong will go wrong, and at the worst possible time), but being well prepared and keeping cool (but not so cool that you freeze up) are by far the most important factors in deciding the outcome. A safe outcome greatly depends on placing occupant safety before airframe loss, knowing your aircraft, and on fully controlling the approach and landing. The latter depends, firstly, on carefully flying an approach (having selected the best readily attainable landing site) that finally minimises the forward speed and the sink rate at a nose up/wings level touchdown; thus minimising impact angles and better distributing the initial impact forces. It is best to dissipate excess energy as drag while still in the approach by using full flaps or side-slipping; both together if acceptable, though some aircraft are downright dangerous if side-slipped with full flap. Secondly, plan the direction of the subsequent ground travel so the remaining energy is substantially dissipated before the occupant enclosure hits something large and unyielding, or a barbed wire cattle fence, or the aircraft overturns. In short; you must plan and totally control the flight all the way into the potential crash.

So what final approach speed should be chosen?

Comment from CAR 35 engineer Dafydd Llewellyn: "It is often difficult to hold a steady speed on short final, even if there is no wind shear; it is not commonly appreciated that most aeroplanes have considerably degraded longitudinal stability in the landing configuration. Firstly, they have a greatly increased pitching moment coefficient due to flaps etc and secondly, they are close to or below the minimum drag speed — and both of these effects tend to make them speed-unstable. This is reflected in FAR 23.175, which requires a positive stick-force gradient in the landing configuration, with power off or only with sufficient power to maintain a 3 degree angle of descent. Try it with more power that this, and you will often find — especially if the cg is somewhat aft — that the thing cannot be trimmed to any given speed; it's actually negatively stable (this is the reason such care is needed in a baulked landing). If disturbed in speed, by turbulence for example, it will continue to slow down or speed up. The stick force versus speed gradient may be positive below this power, but it's usually pretty small, so the thing is a lot less stable than in cruise. This is why it's so critical to keep an eye on the ASI, in those phases of flight, and almost certainly the root cause of the high proportion of accidents that occur in the landing and aborted landing phases. Most pilots are blissfully unaware of this."

After considering the trade-off between adequate controllability, margin from stall, and excess kinetic energy the optimum speed over the fence is probably 1.3 times Vso (the stalling speed in the landing configuration at the particular operating weight); it will be close to the corresponding Vmp. Add no more than 5 knots in gusty conditions and resist any compulsion to add any additional 'safety' margin. Make sure Vso and the approach speed are both determined in terms of CAS rather than IAS, as the ASI may have significant, but not comparable, position error corrections at Vso and 1.3 × Vso.

Flare with care!

In the final approach, the aircraft aoa may be somewhere between 5° and 8°, and the flare to arrest the rate of descent may raise aoa close to critical, and the increased drag will slow the aircraft quickly — thus a rapid increase in rate of sink will follow. It is usually essential that the aircraft is flared gently, smoothly and at the height appropriate for a consequent near-stalled or fully stalled touchdown — but see 'Alighting in tree tops'. In some aircraft the loss in slipstream (compared with that from an idling engine in a normal landing) may significantly reduce the elevator authority and thus the stick must be pulled back further to flare successfully.

(The term 'dead-stick landing' — an 'in-word' used to describe a forced landing following complete loss of power (or a training exercise where the engine is shutdown on approach) — originated during World War I. It is thought to describe the decreased elevator authority following loss of the slipstream in some of those early aircraft. Nowadays the use of the term is deprecated — no one describes a normal glider landing as 'dead-stick'.)

The possibilities of landing safely downslope may range from difficult to impossible. A strong headwind may make a downslope landing feasible though it is difficult to judge the degree of slope until you are close to the surface and thus committed.

Moderate upslope is good if the pre-touchdown flare is well judged. There is a much greater change in the flight path during the flare; for example if the upslope has a one-in-six gradient (about 10°) and the aircraft's glide slope is 6° then the flight path has to be altered by 16° so that the aircraft is flying parallel to the upslope surface before final impact. A higher approach speed is needed because the increased wing loading during the very pronounced roundout/flare (a turn in the vertical plane) increases stall speed. If the wind is upslope then a crosswind landing may be feasible.

3.3 Deceleration forces and energy absorption

Load factors

The KE of a 540 kg aircraft touching down at 45 knots groundspeed is 137 000 N-m. If, in a normal landing on a prepared airstrip, the aircraft is uniformly decelerated to stop 100 metres from touchdown then the deceleration force — the total forces applied to stop the aircraft — is KE/distance (N-m/m) = 137 000/100 = 1370 newtons.

The deceleration forces place a load on the aircraft and the airframe transfers a load to its occupants. It is usual to compare such load factors in terms of the non-dimensional 'g ratio' calculated, in this case, by dividing the uniform deceleration force by the aircraft's weight in newtons (its mass in kg multiplied by the acceleration of gravity — close to 10 metres per second per second) which, in the example, would be 540 × 10 = 5400 N. Thus the horizontal deceleration factor is 1370/5400 = 0.25g — just a slight load which probably wouldn't register with the occupants; it can also be seen that the aircraft is decelerating at the rate of 2.5 m/s² (i.e. 10 × 0.25). If the family sedan is brought to a controlled stop under sustained braking the occupants would be unlikely to experience more than a 1g deceleration.

If the aircraft, under uniform deceleration, came to rest in 10 metres then the deceleration force is 13 700 newtons and the load factor is 13 700/5400 = 2.5g. But if uniformly brought to a halt in 5 metres by landing in dense, light scrub then the deceleration force is 27 400 newtons and the forward deceleration load factor is 27 400/5400 = 5g. Of course the aircraft's velocity at impact includes a vertical component, but we will look at that later.

It is unlikely that in the early stages of ground travel, after a planned and controlled forced landing approach, a light aircraft would slam head-on into a large unyielding object, such as a large tree trunk or a very large boulder. On the other hand, it is also unlikely that an aircraft will be uniformly decelerated — the surface conditions may be such that varying impact loads (from contact with brush, saplings, stumps, roots, stones, holes, furrows) are intermittently applied to the airframe and occupants from near touch-down until coming to a stop, making it impossible to control direction, or even keep feet on the rudder controls. These multiple impacts result in a series of peak deceleration loads applied for very short periods, probably a few hundredths of a second, and felt as severe jolts; many of these will have a sideward load component.

A note of caution. A firm touchdown with no float in ground effect is the aim but if you are forced into dissipating excess airspeed by holding off half a metre above the surface and the undercarriage strikes a rock or stump then the consequences are likely to be more traumatic than if you had pegged it down earlier at the higher speed and then run on into the object. Ground-assisted deceleration is better than ground effect float. The consequences may also not be good if you are holding off and pull back on the stick to avoid tripping over an obstruction. So if the terrain is cluttered with unavoidable obstructions of that nature then it may be best to place the main wheels on the ground earlier even though the velocity, and thus kinetic energy, is higher. If the distance between relatively high obstructions is less than the wing span try to steer a course that will equally distribute the impact forces on each wing so that the cockpit enclosure is not spun around into something unyielding. Of course if landing on a clear surface the aircraft will slow faster with its wheels on the ground than if held in ground effect, but the faster the speed at touchdown the greater the possibility of bouncing.

Airmanship is about making and implementing the wisest choice in such difficult situations. For example, when faced with an obstacle such as a rural powerline many pilots might choose to pull up over it rather than taking the possibly safer path under it. That natural tendency, when faced with some unexpectedly hostile surface or the inability to clear a previously unseen obstacle, to 'stretch' the glide distance by raising the nose excessively, may lead to an uncontrolled impact in a most unfavourable attitude. A controlled collision is far preferable to control loss 50 feet above the surface — the latter generally results in severe injury or worse. It is probably better to put it between obstacles that are closer together than the wingspan, than to stretch the glide and then drop-in nose first. Protect the cockpit by sacrificing the wing structure. It is best to avoid higher-speed impact with a strong barbed wire fence by ground looping if possible.

The following is an extract from a detailed incident report by a Boorabee pilot who did everything right when the engine packed up:

"... the positioning and timing seemed to come together almost at a crawling pace, but it must have been just a few seconds. The turn onto final had to be made at low level so I made a definite intent to ensure good speed into the turn. Turned onto the final approach high enough to clear the barbed wire fence and fast enough to have full control and touched down beyond the fence parallel with the ploughed furrows ... recall pushing the nose down just enough to ensure longest distance possible for ground roll as the dirt paddock would retard the motion a lot faster than flaring and easing onto the ground halfway up the paddock ... noisy and bumpy ride with underside of pod sliding along top of furrow ... ground looped to halt the aircraft when getting close to the end fence and into cross ploughing ..."

Tailwheel aircraft have an advantage over nosewheel aircraft on rough ground. The tailwheel is likely to be pulled over obstacles but even if it is knocked off, the aircraft remains stable and is converted into a true 'taildragger' with its built-in arresting effect. On the other hand recreational light aircraft nosewheel structures are not very strong and if a nosewheel can't be held off then it tends to be pushed into holes and may not ride across or over obstacles — the consequences may be loss of the nosewheel strut and of aircraft ground stability. In the worst case the aircraft nose may dig in and the aircraft flip onto its back; in which case ensure you are in an aircraft where the design includes a structure that rests on itself rather than the occupants heads, and there is an escape route from the inverted cockpit. Some aesthetically pleasing bubble canopies with unobstructed views may be death traps; steel roll cages/bars or high-wing aircraft are safer. The accident/incident reports indicate a surprising number of aircraft end up inverted following a forced landing or other landing mishap, but certainly for the high-wing aircraft the damage to the airframe is generally not total and injuries are low. If your heart is set on a low-wing or mid-wing aircraft first figure how you and your passenger will escape when it's inverted.

Energy absorption

From the foregoing it is evident that very little distance is required to bring the aircraft to a safe halt IF the kinetic energy can be dissipated uniformly during ground travel. For example the occupants of the 540 kg aircraft touching down at 45 knots and uniformly brought to a stop over 20 metres would experience about 1.25g deceleration. (In the days of heavy piston-engined aircraft conducting carrier landings the arresting load was 2–3g, which was not particularly uncomfortable when well strapped in.)

So where there is no clear, open space to land the aircraft, more or less normally, then an option is to choose an area where the vegetation is of sufficient height and density to absorb much of the kinetic energy and retard the aircraft. If that vegetation is weaker than the aircraft structure so much the better, but the primary consideration is occupant safety so energy absorption by sacrifice of non-vital aircraft structure — i.e. all that outside the occupant zone — is warranted. The requirement of course is to set up the touchdown so the aircraft is moving in a direction where the vital structure is unlikely to slam into an unyielding obstruction at speed.

High and dense crops, sugar cane, brush and light scrub all provide good energy-absorbing properties and good cushioning is provided if the aircraft is put down in the proper nose-high attitude so the impact forces have more spread over the aircraft's under-surfaces, rather than just catching at the undercarriage and overturning the aircraft. But even an unfavourable impact angle may not be particularly dramatic; e.g. here is an extract from a forced landing incident report: "The Jabiru impacted the sugar cane in a 20 degree left wing low attitude and came to rest upright after sliding 20 metres."

Again — the emphasis is on controlling the crash

" ... I decided to land in a cane field ... the Jabiru was held off until it stalled when full left rudder was applied to slew the aircraft sideways to prevent it from going over on its back ... the event was successful and neither of us were injured.

Impact forces are less if you touchdown at Vmp or a little higher and then run on into obstructions at the far side of a clearing rather than stall/spin at the near end. The aircraft structure will withstand longitudinal impact forces much better than lateral impact forces, so generally avoid touchdown with substantial drift towards a lowered wing.

'Alighting' in tree tops is certainly extremely hazardous and always results in total aircraft write-off. But if the aircraft is flown into a selected, dense crown in a reasonably nose-high attitude (and into wind) — so that some of the initial impact is absorbed by the under-surfaces of the fuselage, tailplane and wings — then the hazard to occupants may be reduced. It is important that the aircraft is not stalled above the tree crowns, because of the possibility of the nose and/or wing dropping into the crown before impact; rather, it should be flown into the canopy at the minimum sink glide speed. The greatest hazard may come from the subsequent fall from the trees.

Easier said than done, but certainly the aircraft must be flown all the way into the crash. The following is a summary of an accident report; the aircraft was a Skyfox CA22, the pilot had 16 000 hours experience and rescue was fast:

'While on cruise at 1100 feet agl the engine failed completely. The pilot set the aircraft for a forced landing into heavily timbered terrain and transmitted two mayday calls. The second call was answered and he gave details of his situation and position. He then maintained control of the aircraft until it touched the top of the tree canopy where he flared steeply, as the aircraft entered the trees, to present the underside of the aircraft for speed reduction and impact damage minimisation.

The pilot suffered bruising to one knee and was transported to hospital by an RAAF helicopter. The aircraft was severely damaged.'

The following is a report from a Jabiru passenger in another treetop alighting. The pilot, using a runway downslope advantage, took off toward the north with a five to eight knot south-east wind. The pilot had just turned crosswind at about 350 feet agl to avoid a noise-sensitive residential area when the engine died. There was no clear area within gliding distance — only a full expanse of mature eucalyptus trees some 20 to 40 feet high. The pilot lowered the nose to maintain 65 knot best glide speed and turned the aircraft slowly into wind for the landing. The passenger (an RA-Aus CFI along for a ride home) takes up the story.

"The trees were rapidly getting closer. They changed from a mass blur to individually defined trees and I prepared myself for the worst. The pilot tried the engine again unsuccessfully, then turned fuel and magnetos off. He started making a mayday broadcast, but before the transmission was complete, there were sounds of trees and aircraft breaking bits off each other as first contact was made.

Full marks to my pilot as he unerringly and unhesitatingly flew us right into the tree canopy, our only survival option. This went by the book. The pilot flew us onto a selected tree top, raised the nose a little to slow down our forward airspeed and then expertly used the top of the tree to slow us down still further from an estimated 60 to 40 knots, yet left us with sufficient forward speed to prevent the deadly vertical fall to the awaiting ground immediately below.

Seemingly incredible deceleration forces tried to rip me out of my seat — thank goodness for seat belts. We hit a sapling trunk with the inboard section of the starboard wing and decelerated rapidly as we were flung violently to the right. The engine cowling disappeared, exposing the engine which ripped outwards right in front of my field of view, dragging the firewall and windscreen area with it.

Next I was aware that we were bouncing backwards, then falling vertically some 10 to 15 feet. We hit the ground with an almighty thud. I became aware of smoke and electrical zapping noises coming from the distorted centre console/instrument panel, which lay on its back with the gauges pointed to the sky. "Get out! Get out!" I yelled to the pilot, who was still sitting in the left hand seat looking a little stunned. I couldn't exit my right-hand side door as something was jamming it. I was keen to get out as quickly as possible as sparks, smoke and arcing were coming from the damaged central instrument console and we had at least 30 litres of fuel sitting in the cockpit right behind us.

I followed the pilot out his side door and touched the ground with relief. However, I discovered I was in immense pain, which worsened when I tried to stand up. Due to my fear about being in an 'aircraft inferno' I left the immediate scene of the accident and headed towards the nearest road. I yelled to the pilot to follow me and hobbled off."

Deceleration effect of sink rate

In a forced landing there is normally no power available to vary the rate of descent or arrest sink, and often high sink rates are not recognised early enough, but it is extremely important that the downward component of the aircraft's velocity at touchdown be minimised. So in the last stages of the approach, after all manoeuvring is completed, the airspeed should be close to Vmp. A high sink rate at touch-down can result in an uncontrolled crash rather than a controlled landing. If the descent rate of our 540 kg aircraft was 300 fpm (1.5 m/s) and this was not arrested in the flare before touch-down, then the kinetic energy of the vertical component of the aircraft's velocity would be 7% (1.5/22.5) of 137 000 = 10 000 N-m. If the undercarriage (which held the aircraft 0.5 m above the ground) collapsed and the downward movement was arrested in 0.5 m then the downward deceleration force is 20 000 newtons and the downward load factor is 20 000/5400 = 3.7 g.

Here is a witness comment in an RA-Aus double fatality report:

"The take-off on the 1200 metre runway was sluggish and the engine was mis-firing. (Fuel was later found to contain a substantial amount of diesel!) The Murphy Rebel continued climbing and turned toward the east then right again onto a low downwind leg. In the mid-downwind position a loud bang was heard and the aircraft then descended flatly ... it appeared that the aircraft had a slow forward velocity but a high rate of descent when it struck ... came to rest 16 metres from the initial impact ... cockpit area and engine bay badly damaged by fire ... the reason for the inadequate airspeed in the forced landing may have been false horizon effect as the aircraft was approaching rising ground."

What is most concerning, from a perusal of the accident/incident reports, are the occasions where a forced landing precursor has been the engine displaying ample warning of a problem before take-off or while in the circuit area (as described above) but the pilot seems to have been hoping that it would fix itself and opted to press on — why? What could possibly be gained?

Aircraft design regulations

The Federal Aviation Regulations Part 23 lays down some crashworthiness requirements for normal category light aeroplanes to give each occupant every reasonable chance of escaping serious injury when the occupant experiences forward loads up to 9g and sideward loads up to 1.5g. So, in theory, if a certificated aircraft touches down under control and decelerates at a constant 9g forward, the occupants should escape serious injury — provided the lapbelts and shoulder harnesses are properly used and cockpit intrusions are avoided. FAR Part 23 also has a 6g downward load requirement.

What this means is that it is possible that a normal category aircraft, touching down at 45 knots, running into something sufficiently yielding (for example scrub and small saplings) and decelerating at 9g will come to a halt over a distance of just 3 metres (during a time of one second) with the occupants only suffering body bruising from properly fitting harnesses and perhaps some minor injuries to the legs and arms; provided the occupant zone remains reasonably intact and nothing intrudes into it. Of course, the rest of the aircraft itself will not come out of it so well. Many of the top-end recreational light aircraft fit into that FAR Part 23 'normal' category. By the way, rocket-deployed aircraft emergency parachute recovery systems generally aim for a maximum descent rate of six metres per second (1200 fpm).

Minimum ultralights and powered parachutes

At the very light-weight, low-speed (55 knots maximum level flight speed) end of the ultralight aircraft spectrum are the homebuilt single-seat minimum aircraft, the airframes of which are often constructed from aluminium tubing and sailcloth. The design, the structural integrity and the impact resistance of such aircraft will certainly not provide the 9g occupant protection required of the type-certificated aircraft but their kinetic energy, when touching down into wind, is very low — in the range 10 000 to 18 000 N-m. In a controlled landing, if such an aircraft was uniformly decelerated to a stop over 5 metres the force would be less than 1.5g. Powered parachutes are in the low end of this minimum category.

3.4 Your final defence perimeter — a safe occupant zone

The airframe density per cubic metre of finished structure is generally homogenous but the engine, fuel and occupant bodies have higher densities — thus a higher momentum (momentum = m × v) than the rest of the aircraft — and should all be properly restrained. The engine by very strong mountings, particularly if mounted behind the occupants; and the occupants by an adequate seat/restraint system so that the core fuselage structure, engine and occupants all decelerate at the same rate even though a considerable part of the aircraft may be sacrificed along the way. Re-read the preceding description of alighting into tree tops. There should be no loose objects in the aircraft — they will become a harmful missile.

If the adult body is properly restrained, human organs and their attached blood vessels, will cope with transverse deceleration loads very much greater than 20g — applied for short periods. However the spinal column has a much lower tolerance to downward deceleration loads; i.e. loads applied parallel to the spinal column. In this aspect the skeletal structure is much weaker than the aircraft under-structure and downward deceleration loads may result in serious injury — thus the importance of minimising the vertical velocity at impact.

Perusal of the accident/incident reports shows that few pilots or passengers were wearing head protection at the time and no doubt some are now wondering why they chose comfort before safety. Using the correct type of helmet and inner energy-absorbing pads will provide considerable protection from serious head injury if the occupant zone be deformed, intrusions occur or the restraint system fails. Children must always wear an appropriate helmet and liner. Why do parents who won't allow their child to ride a bicycle without a helmet take them flying with no such protection? Helmets also reduce the chances of being knocked unconscious in a wreck that subsequently catches fire.

Occupant restraint system

The pelvis-hip girdle is the strongest part of the body structure and the body's centre of gravity lies between the hip bones. The better occupant restraint systems usually consist of a seat and seat-back, a firm seat cushion, a lapbelt angled to hold the hips into the internal corner of the seat and a shoulder harness system to prevent forward/sideways movement of the upper torso. Shoulder harness systems are usually an adjustable webbing strap over each shoulder that clip to the lapbelt; the shoulder harness may incorporate an inertia reel system. A lap and sash belt, as used in the family sedan, is effective and might be used in a minimum aircraft. The following should be noted:

The occupant restraint system must be designed for the occupant(s) and the conditions likely to be experienced in that aircraft.
The airframe or seat attachment positions and angles for the harness system must be such that the adjusted lapbelt will remain across the occupant's hips and the shoulder straps on the shoulders during impact. The diagram is an example of the harness geometry guidance material from BCAR S.
If the seat should collapse or the seat back fails during impact, there is a possibility that the occupant's body may then slip forward beneath the lapbelt. The same problem — submarining — can occur if the lapbelt hasn't been tightened sufficiently or if a badly designed seat slopes downward from back to front. If submarining occurs and the forward slide is sufficient that the lapbelt is repositioned above the pelvic girdle, then consequent impact loads (and the rotation of the body about the lapbelt) can cause abdominal and spinal injuries. Extended submarining has resulted in strangulation by the harness. These problems are readily overcome by incorporating a fifth strap (the crotch strap) into the harness, locking the lapbelt in position against the pull of the shoulder straps and preventing bodily slip.
When fastening the harness, first position the body correctly in the seat then follow the correct fastening and tightening sequences — lapbelt across the hips not the abdomen, pulled very tight (then the crotch strap if fitted), and then the shoulder harness reasonably tight but without displacing the lapbelt position.
Don't totally release the shoulder harness straps in flight to facilitate easier movement; you may neglect to re-attach them properly in an emergency.
Perhaps the best value available in safety harness is a Confederation of Austalian Motor Sport approved 5-strap automotive racing harness, which costs less than an "approved" aircraft harness, and probably offers better protection, but any harness is only as good as its anchor points and their position relative to the occupant and their surrounding airframe structure.
In a sudden deceleration, momentum carries the upper body forward, stretching the shoulder harness, then stops and the upper body and head whip back. A wide, deep headrest will provide some whiplash injury protection, but a large number of aircraft seats lack headrest protection.
During pre-flight inspection check the webbing, inertia reel and fastener condition and integrity and the seat mounting integrity. If the seat is the moveable type check the rail holes or slots; if they are deformed the seat may slide back on take-off or may twist and detach under impact forces.
When initially settling in to the seat make sure that you can comfortably — i.e. without straightening your leg — apply full left and right rudder. If you cannot adjust the seat or rudder bar to achieve this, do not fly that aeroplane because you will not have the full rudder authority judged to be needed and provided by the designer. Also there is a high probability that, with the knee joint locked while applying full rudder — to steer the vehicle on the ground or initiate a ground loop — any impact forces transmitted via the rudder bar may severely damage the hip socket. You must be able to apply full rudder with the knee still bent.
If the aircraft is fitted with adjustable seats make doubly sure that the seats are locked and in a comfortable position. Do not attempt to relocate the seat position during an emergency landing.

Great care must be taken with child restraint systems

Infants cannot be carried in ultralights and it is most unwise to carry small children (say under 10 kg) as a passenger; there is no satisfactory restraint system. They cannot extricate themselves and they cannot go for help. Children weighing between 10 and 20 kg should use a government-approved child restraint system [CRS]. The US approved types have 'This restraint is certified for use in motor vehicles and aircraft' printed on them. The CASA advisory publication CAAP 235-2(1) contains more information on CRS standards.

Children over 20 kg and 145 cm tall might be restrained safely within a normal fully adjustable four-strap (preferably five-strap) seat harness. A booster cushion might be used if they don't quite make the height. Safest approach is to never carry children who are not old and strong enough to extricate themselves safely (and unaided) from the harness and a wrecked cockpit.

Crush zones

In theory the distance over which the aircraft is brought to a stop is the distance over which the aircraft's centre of gravity travels. Thus provision of an energy absorbing crush zone or deformable structure in the front of the occupant zone, even if it just adds half a metre to the distance, adds to the occupant stopping distance and reduces the deceleration forces on the occupants.

Generally the structure of the under fuselage does not incorporate a crush zone to provide some occupant protection from spinal injury, so it is advisable that aircraft with a retractable undercarriage should be landed with the gear down. That will absorb quite a lot of vertical load before collapsing. Some aircraft seats may be designed as a deforming, load-absorbing system in which case it is important that nothing is stowed beneath the seat. In a low-wing aircraft the pilot/passenger seats are probably directly over the main spar — which is obviously built not to collapse — so if there is no crushable structure between the seats and the spar the occupants' spinal columns will be directly exposed to the full vertical deceleration. In that case ensure the seat includes something similar to a body-conforming, energy-absorbing, 3-inch thick seat cushion laminated from three layers of Confor urethane foams.

Nearly all light aircraft have a fixed undercarriage; there may be a problem with some low-wing aircraft fitted with in-wing fuel tanks if the collapse of the undercarriage causes penetration of those tanks.

In a high-wing aircraft the overhead structure provides a crush zone sufficient to allow exit room from the cockpit if the aircraft pitch-poles onto its back. Also the cockpit is fitted with doors that can generally be forced open when the fuselage is inverted. In many low-wing or mid-wing aircraft fitted with a bubble-type cockpit canopy there is no structure, except possibly the vertical stabiliser, to prevent crushing of the cockpit canopy in a capsize. If this occurs, exit of the pilot/passenger may be impossible until the aircraft is lifted. If the aircraft is fitted with a canopy that can be jettisoned in flight, make sure the front canopy bow cannot drop down during the jettison process and scalp the occupants.

The next article in this series discusses the worst situation for engine failure — that occurring soon after take-off.

'Decreasing your exposure to risk' modules

| Introduction | Don't fly real fast | Don't stall and spin in from a turn |

| Don't land too fast in an emergency | Engine failure after take-off |

| The turn back, possible or impossible — or just unwise? | Wind shear and turbulence |