Thermobaric Warheads on the Rise: North Korea and fuel-air explosives

Just what were the North Koreans firing at Yeonpyeong Island, and from what artillery? It's likely that some or all of the fire was from their variant of Russia's BM-21 Grad MLR, short for "multiple rocket launcher" system. These are angle-adjustable banks of launching tubes on a rotating turntable; all this bolted to the back of a flatbed truck. The Grad and its knockoffs are highly mobile and quick to set up for firing; while limited in accuracy, they can launch a fearsome barrage of 40 warheads in just a few minutes, striking targets more than 20 miles away, depending on the model. Then the crew either slaps on a reload pack with 40 more rockets, or if counter-battery fire is on the way, hits the road and hunts shelter.


Allegedly the North Koreans rehearsed the attack to time their fire (which originated from multiple locations) in such a way that the projectiles struck the island at about the same time. In military terms time-on-target coordination is an effective anti-personnel tactic since troops (and civilians, here) don't have time to seek shelter.


Each Grad rocket is about ten feet long including the propellant section, and 122 mm in diameter, or a little more than five inches across. That makes for a big warhead, holding about fifty pounds of payload.


According to an article in the Daily Yomiuri, dud rounds indicate that the warheads held a powerfully engineered explosive called a thermobaric weapon; aka fuel air explosive. ("Thermo" stands for heat, "baric" for pressure.) While engineering a fuel-air explosion from a fast-moving warhead is not easy, thermobaric rounds can be very effective in military terms, and now come in sizes large and small. A unit of Marines reported in 2003 that a relatively small 40mm thermobaric grenade blew apart a single-story building. Users of the rocket propelled grenade launcher called the RPG-7 can choose from a thermobaric round called the TBG-7V.


The Germans may have experimented with thermobaric explosives in World War II and the U.S. used them against them in Vietnam, such as to clear helicopter landing zones. While still not widely used in wartime (so far) there's nothing secret about the physics of thermobaric explosions, which lie behind every house blown apart in a natural gas explosion, and every factory or grain mill wiped out by a dust explosion. Following somebody's unwise decision in February 2010 to purge pipes at the brand-new Kleen Energy Energy Systems combined-cycle power plant with natural gas, a fuel-air explosion demolished that 620-MWe plant before it even started up.


Some experts regard thermobaric explosives as several times more powerful (pound for pound) than conventional high explosives like Comp B; perhaps up to six times as powerful. One reason: because thermobarics draw their oxygen from the air; they don't have to carry an oxidizer in the warhead, so all the filler can be given over to energetics like volatile liquid fuels and aluminum particles. So thermobarics aren't going to work underwater, in space, or at very high altitudes. To see how conventional explosives are burdened with oxygen, take a look at the RDX molecule diagrammed here.


Characteristic of the thermobaric warhead in action, when viewed in slow motion, is the double explosion. The first charge disperses a superheated cloud of fuel so that it can mix with air; the second and much larger explosion that follows is the fuel-air blast. Here's a film of a big thermobaric bomb being tested.


If Country A wants to attack Country B's vast tunnel complex, count on liberal use of thermobaric warheads after armor-piercing rounds open up the portal doors. The effect of a fuel-air explosion, which is plenty devastating in the open, is greatly magnified in confined spaces if the proper fuel-air mix can be achieved.

Helicopters and High-Rise Fires: McDonnell Douglas's "rescue angel"

Here's the second part of my post on helicopters at high-rise fires. This focuses on McDonnell-Douglas's flying rescue machine. The material is adapted from my book The God Machine on the social history of helicopters.

By 1976 it was clear that helicopters responding to high-rise fires like Andraus and Joelma were having problems with unruly crowds and smoke obscuring the roof. 

One solution, it seemed to fire chiefs at the time, was to keep the rescue helicopter high above, away from fires where pilots might lose their orientation, or machines might sustain damage from heat or suffer loss of power. In 1976 the chiefs of two dozen American fire departments wrote to McDonnell Douglas Corporation and described the problem. 

The aerospace contractor came back with plans for the Suspended Maneuvering System and went on to build a prototype. The SMS was a supercharged, sixteen-passenger rescue basket designed to hang from a large helicopter at the end of a 500-foot-long steel cable. 

The SMS had a gasoline engine and adjustable air-thruster nozzles so the operator could move the basket from side to side at the end of that tether, allowing it some horizontal freedom of movement when alongside a tall structure or above a flooding river. It had a little catwalk that could be lowered to bridge the SMS and a window.

The SMS was working in prototype by 1978. Here's a link to a Pop Mechanics profile about how MD's "angel" did in the early tests. Despite their early enthusiasm, though, fire departments dropped all plans to buy them. The chief obstacle was not the rescue device itself, which worked well in tests, but the logistical demands it placed on a fire department compared to the rarity of its use. 

Any department that would be depending on the SMS would have to arrange for fast response by a big twin-engine helicopter comparable to the Sikorsky H-60 Black Hawk. That's because of the weight to be carried and the need for extremely high reliability whenever people are being hauled around at the end of a cable hooked to the belly of a helicopter. (A hair-raising subject, and one for a later post.)

Also weighing against deployment of the SMS was a scarcity of high-rise fires after the early 1980s. The frequency plummeted after big cities began requiring sprinkler systems in skyscrapers, a move spurred by the disastrous conflagrations of the Seventies. 

Sprinklers and other fire-minded measures worked so well in stopping fires before they caught that if a fire chief had authorized an SMS unit in 1980 he could have worked another three decades and retired without seeing the SMS squad go into action, even once, at a high rise fire in his city. 

Only one department, the Los Angeles Fire Department, actually readied the SMS for action. That deployment lasted less than a year. I talked to a firefighter who participated in the effort, and he said it performed much as advertised. Here's a link to an image of the SMS as displayed in the LAFD's fire museum.

Would an SMS have been of any use at the World Trade Center that morning of September 11? I believe so ... After all, the SMS was designed specifically for getting people off of high-rises on fire -- including through the windows where the roof was untenable or unreachable. Still and all, the very short time span available would have been a problem. Everything about the mobilization would have had to operate without delay or error, with all forces from the Air National Guard to the NYPD's Aviation Unit to the FDNY acting as one.

Emerging reports about reckless practices behind the recent Shanghai high-rise apartment fire (death toll now up to 58) indicate that other high-rises there are at risk until if and when safety practices are cleaned up. So that city's FD might want to look into setting up an SMS unit; for one thing, helicopters are more reliable now and have better auto-hovering controls. It could be a lot better than nothing!
  

The Ring: Some Background

As I mentioned in an earlier post on cloudspotting, I've been snapping cloud pictures with digital cameras since mid 2008. It's a hobby, like bird watching. The catch adds up to many thousands of pictures in all kinds of weather, mostly in the US. With so much time spent looking out and up, sometimes I come across striking visual effects. They don't come along often, but add zest to the hunt.

This post is to give some context to one of those visual outliers, since I'll be talking briefly about Photo No. 147 tonight with host Ian Punnett at the top of Coast to Coast AM.

Some background: this spring I was getting ready for the summer cloud season, going back through the previous year's crop of cloud pictures to organize them and dispose of unneeded ones; it's like cleaning the electronic attic. That's when an oddity in one image caught my eye; the photo was one out of a series of snaps taken on a weekend afternoon in mid-May 2009. I didn't notice anything special when taking the pictures that day. It's one of those cameras with no viewfinder, just a screen on the back, so in daylight it's very hard to see what the camera is capturing. I was just snapping away because I liked the look of an airliner contrail alongside a set of cool clouds.

Here's the camera I used, of garden-variety:

Here's a portion of the photo in question (No. 147), which I've watermarked for copyright purposes, along with the enlargement below. The anomaly in question is a faint blue oblong in the lower left of the photo.

After enlargement and some added contrast, it looks like this:

It doesn't look like a flying object to me, somewhat more like a ring of plasma, but that doesn't help much either.


Whatever it was, it was fleeting. I had taken a string of pictures in the same general span of sky over a fifteen-minute interval with this camera, and the Ring doesn't show up in any of the others.

In particular, the immediately preceding photo, No. 146, was taken less than a minute before, probably less than thirty seconds before. No. 146 has the same clouds and a slightly shorter jet contrail, but no Ring.

I don't have any great insight about this; I just happened to be outside with a camera; but three things about Photo No. 147 pique my interest. First (and supporting the idea that this was a ring viewed at an oblique angle) is the fact that the "ends" of the ring on the left and right are brighter than other parts of the ring. Look in particular at the tip on the right, how bright it is. This would be the case if looking obliquely at a ring of plasma; the ends would be somewhat overlapped and therefore brighter. Second, there is some kind of central core inside the ring. Third (and this is barely visible until the contrast is raised), there appears to be a faint, bluish arc on the lower left. What the heck is that?

I checked with a few contacts in the world of physics, and they offered no opinion other than the ring might be a camera glitch or a lens flare. I'm no JPEG expert so for all I know, digital cameras produce such glitches randomly. This camera hasn't done so before or since, but maybe it hiccuped that time. 

But I'm pretty sure the ring is not a lens flare given the direction the camera was pointing: north or northeast. It takes a brilliant light to make a lens flare and in Minnesota in May, the sun is nowhere near the field of view of a camera that's pointed north at mid-day. Also, lens flares I've seen don't look like this. 

Date: Allowing for a mistake I made when setting up the camera clock upon purchase (for example, not changing AM to PM), I'm reasonably sure that I took Photo No. 147 on the early afternoon of Saturday, May 16, 2009. Location was the Twin Cities of Minnesota, looking north or northeast.

I'm hoping a high-energy physicist might have a look. In the meantime the only half-baked hypothesis I can offer is that it's a picture of something very distant exploding in the very high atmosphere, and the shock waves glowed so brightly they overcame the brightness of the sky at mid-day. But when comets come in and blow up there's a debris trail, and none is visible here, so my proposition has its own shortcomings.

If some kind of distant large explosion in the high atmosphere did occur and if it was within view of sensors aboard Defense Support Program reconnaissance satellites, it might have been recorded. In addition to looking down for incandescent ICBM booster plumes, DSP satellites allegedly have, over the years, detected rather large breakups of high-velocity objects in the high atmosphere. I have read that at least one was in the multi-kiloton yield. But as of now all such sensor information is classified so I can't check my hypothesis.

That's it for now; just wanted to add a little extra context for Coast to Coast AM listeners. Thanks for tuning in!

High-Rise Fires and Helicopter Rescue

A fire that engulfed a 28-story apartment in the Jing'an District of Shanghai early today appears under control, but 42 people died and over one hundred were injured. 

Initial reports indicate that some residents made it onto the roof and three helicopter crews were orbiting in hopes they could pull people off, but few if any people made it off the roof in helos due to smoke that blocked the way.

After 9/11, the mere mention of helicopter roof-rescue brought down criticism from the Fire Department of New York (FDNY), which under operating rules put into effect after the 1993 WTC bombing (and controversial use of helicopters there) would have to approve and supervise such any such effort by the NYPD's aviation unit.

Does it ever work? Yes, sometimes. Rescue helicopters have participated at more than two dozen high-rise rescues worldwide since 1963. Each of these rescues was unique, and most proceeded under difficult circumstances. Not all went as intended, but helicopters did pluck 1,200 people from roofs, ledges, and balconies.

Two of the earliest and most spectacular fires with elements of helo rescue happened in downtown Sao Paulo, Brazil, in 1972 and 1974. The first to burn was the 330-foot-high Andraus Building, a concrete-framed structure that held offices and a department store. The fire began in a pile of combustibles stored near the fourth floor and went up via the exterior walls, touching off furnishings with heat radiated through the windows as it climbed. Crowds gathered on the street, blocking fire apparatus, to watch. At its peak the wall of flame was 350 feet high, burning with the fury of an oil well fire. 

Three hundred occupants climbed the stairs to the rooftop heliport (coincidentally, the first of many heliports built in Sao Paulo, today the world's most helicopter-friendly city). But after they reached comparative safety, someone in that group went back and blocked the steel door from opening, apparently fearing that those already on the roof would run out of foot room.

Problem: there were two hundred more occupants still in the stairwell, and they jammed against the obstruction. Whether or not those below could be called in a state of “panic” (a subject for another post) they were very upset and pushed hard. They couldn't imagine why a flow up the stairs had stopped all of a sudden.

Help was on the way. A helicopter with firefighters approached, then banked away when the crowd charged it. But the helicopter came back and hovered out of reach to drop off firefighters, who quickly took control of the crowd. This allowed helicopters to land safely. 

Here's my point: The helicopters’ biggest contribution toward survival at the world-class disaster at Andraus that day was not getting people off the roof, but getting firefighters to the roof so they could break open the stairwell door. This action certainly prevented many dozens of people on the stairs from being crushed or asphyxiated by the upward pressure. As it was, some in the stairwell had broken bones or lay unconscious from smoke inhalation when help arrived. Survival was a very close thing. 

In the end sixteen people died at Andraus, but the toll would have been much higher had the rooftop doorway remained blocked.

Sao Paulo suffered another skyscraper fire two years later, at the Joelma Building, and helicopters also fluttered to that scene. But the rooftop area available to helicopters was so small, and the fire and smoke so intense, that 90 people died before the aircraft could move in. Some survivors told investigators afterward that they had remembered how helicopters had lifted people off the Andraus roof two years earlier and said that this memory persuaded them to go up the stairs instead of down, because they assumed a similar rescue would be possible. Such false hopes are one reason why the FDNY has been, and remains, very skeptical about relying on helicopters to pull people off high rises.

High-rise emergencies continued worldwide after the twin fires at Sao Paulo, each offering unique lessons in high-angle rescue. A 1980 fire at the 26-story MGM Grand Hotel in Las Vegas prompted the biggest gathering of helicopters at any high-rise emergency, before or since. Here's the NFPA write-up. Police helicopter pilots were the first to see the smoke and sent out a radio alert. Thirty public and private helicopters mobilized to move 300 people from the roof. 

Coincidentally and remarkably, Air Force rescue helicopters and crews were close at hand, having staged at nearby Nellis Air Force Base for an exercise. Fearing that the roof deck was about to catch fire, the pilots organized their ships into a racetrack pattern: approaching from the east side, picking up hotel guests on the roof, departing on the west, then dropping people off in a parking lot. Some who waved for help were on balconies well below the roof level, which ordinarily seem to be out of reach for helicopters; but rescues proceeded even at that difficult location after a flight engineer lowered himself on a “jungle penetrator” seat at the end of a cable, and tossed a rope to those on the balconies. The hotel guests then pulled the engineer to their balcony and, one by one, joined the engineer on the rescue seat. One volunteer pilot was Mel Larson, then a vice president at Circus Circus Casino and owner of a helicopter charter service. I interviewed him later.

“It was not orderly,” Larson told me about his first view of the roof, where people were trying to grab hold of the landing skids as helicopters departed. “It was panicky. We had to have police to get control. ” At one point a police officer drew his gun to make his point. Eighty-four people died at the MGM Grand, but none of the fatalities happened on the roof. (For years afterward, and at his suggestion, the management of Circus Circus cached evacuation equipment on the roof in case of such an emergency.)

One solution to unruly crowds and smoke obscuring the roof, it seemed to fire chiefs at the time, was to keep the rescue helicopter high above fire and smoke where pilots might lose their orientation, sustain aircraft damage from the heat plumes, or suffer loss of power. And rescue pilots have assured me that power loss -- even flame-outs -- due to oxygen starvation has happened. 

McDonnell Douglas took the fire chiefs' suggestions and created a sort of steerable, aerial lifeboat to be hung from a line under a helicopter. It was called the Suspended Maneuvering System. The SMS was a device of amazing abilities but was never put into regular use. A subject for a later post.

Qantas Flight 32: A reminder that turbofan engines still explode

Offering a few elaborations on the catastrophic engine failure on board Qantas Flight 32 last week, over Batam Island, Indonesia. The aircraft landed at Singapore's Changi Airport without injury to crew or PAX, but it could have been disastrous.

The engine in question (one of four mounted on the aircraft) is the Trent 972 model built by Rolls-Royce. It's a model specific to Qantas A380. Here's a link to a Rolls-Royce brochure on the Trent 900 family, with a helpful cutaway diagram.

This was the first such failure in the Trent 900 family. (An engine in the Trent 1000 family broke up in August while being tested, but I don't count that event, because testing is supposed to be rigorous and probe for problems).

The Australian Transportation Safety Board has focused on a breakup of a geared disk. The ATSB asking Batam Islanders to contact authorities if they come across a disk fragment as shown in this article at FlightGlobal.com.

The most alarming run of uncontained engine failures occurred in the late 1990s, mostly among one variant of General Electric's widely-used CF6 engine, which was exploding at an average of four times a year. Failures in later models of CF6 models remain a concern. But to be fair, the CF6 is an extremely popular engine so there are a lot of them.

A common pattern is for an uncontained breakup to happen at full power, during the takeoff run. Full power is when the disks experience the highest stress, rotating thousands of times per minute.

But there have been dozens of engine breakups in flight, as on QF32. The most famous was United Flight 232, which led to a crash-landing at Sioux City, Iowa, in 1989. The tail engine on this DC-10 exploded at altitude and its fragments cut all hydraulic lines to the control surfaces.The crew gained enough marginal control over the aircraft, using asymmetric engine thrust, to bring it down and save most of the people on board.

This reminded disaster historians of how Capt. Bryce McCormick innovated that technique 17 years earlier to bring his American Airlines DC-10 down safely to Detroit's airport, using just the engine thrust, after a cargo-door blowout over Windsor, Ontario, destroyed his ability to use the control surfaces. McCormick's plane didn't crash, it didn't end with any fatalities, and it took to the air again after repairs. See Chapter 5 of Inviting Disaster for how McCormick pulled it off.

McCormick's feat still stands as a landmark in crew emergency preparation and is one of only two airliner cases I know in which such a catastrophic loss of flight controls ended in a controlled landing with no fatalities. (The other was a DHL cargo flight in 2003 that landed safely at Baghdad after being hit by a missile that caused horrendous damage to the A300's wing and chopped the hydraulic lines). McCormick had the keen foresight, two months before the near-crash in June 1972, to train himself in a simulator on how to fly the DC-10 using only the engines. It had been his initiative alone; there was no formal curriculum on such a thing.

So that's one risk to safety, an uncontained failure severing essential controls. Another theoretical risk, which is extremely small given protective measures in modern aircraft, is that shrapnel will injure passengers in seats that happen to be in line with the shrapnel zone.

I believe the more significant risk is that an engine explosion early in the takeoff run will puncture a fuel tank and cause a fire around that wing, after the crew aborts and comes to a hard stop. This happened on a Boeing 737 after an aborted takeoff at Manchester, England. The cause of the fuel leak from a wing tank was catastrophic failure of the combustion-can casing.

But most uncontained engine failures don't cause any injuries at all and they're so exceedingly rare I wouldn't let it affect my travel plans. Major-carrier scheduled flights are much safer than driving, mile for mile. And how else is one to vault the ocean?
 
But this advice to passengers:

• If there is any sign of a sustained fire when the aircraft is on the ground report it to a flight attendant. The flight crew's view to the rear is very poor, and engine instruments sometimes lie or are misunderstood. There was the case with a British Midlands 737 in flight where the left-hand engine started tearing itself to pieces; while passengers on the left saw it happen the information didn't get to the front of the plane. The flight crew in a series of misunderstandings and missed signals concluded the problem was the engine on the right side. So they shut down the single remaining good engine instead of the faulty one on the left. When the left engine finally went to pieces on the descent there was no time for a restart and the plane crashed short of the runway at Kegworth, England. Point being: accurate, timely communication of critical information from clear-headed witnesses to decision makers can be a real life saver!
• Make sure you are familiar with the exit doors nearest to you and how to open them, say, in case the exit-row passengers haven't been paying attention. For example, sometimes removable over-wing doors become a problem if passengers behind you are pressing forward and denying you the space you need to get the door out of the way.
• Put your shoes on but leave the hand luggage behind, including laptops, briefcases, and wheelie bags. If the cabin starts filling with smoke (it will rise to the ceiling and be very visible), you are in an extremely marginal situation so saving the luggage should not be on your list of concerns. Passengers and crew will have to work closely and cooperatively to get everyone out.

Patterns in AQ-LeT Gun-team Attacks

Last month's attack by a three-man gun team on the Chechen Parliament in Grozny, and the attack earlier this week on Our Lady of Salvation Church in Baghdad, have many features in common with the Parliament House attack in India in December 2001. The principal movers at this time are Al Qaeda affiliates and Lashkar-e-Taiba (LeT).

It's important to maintain basic citizen awareness of this very dangerous tactic, which poses a real threat even though the coordinated gun-team attack has not materialized recently in the West. While a gun-team attack in a major Western city would be up against a fast and capable security response, attackers tend to hold a tactical advantage in the early moments. The best response is to stop these before the shooting starts, and a vigilant citizenry may well help alert the police to precursor events, such as hostile reconnaissance or cache preparation. I understand that now the press is zeroed in on the attempted package-bomb attack last week on cargo airliners, but I recommend that the gun-team attack move up in the public's awareness.

I checked Google News and press coverage of the package bombs is a great deal more intense than coverage of the gun-team attack on Sunday that killed 52 at Our Lady of Salvation in Baghdad. And there was even less press coverage of the October gun-team attack in Grozny.

These common features of a gun-team attack are a squad or squads who move quickly from target to target; who are trying to get past guards and into a closed area packed with people; and who act in such a way to cause maximum havoc and maximum publicity.

If the attack aims for a hostage standoff -- and most gun-team attacks do, as a means to stretch out the timeline -- there will certainly be attempts to use media  for publicity. Note that during the church attack, the gunmen tried to use Al-Baghdadiya television as an outlet for demands, until the government cut the transmitter. Extended attacks also see the gun teams carrying GPS units, satellite phones, and backpacks with ammunition, explosives, and concentrated food. There is some evidence of pre-positioned caches with additional munitions.

Gun teams typically rely on standard assault weapons (the AK-47 and a Chinese copy called the Type 56 have been common), semi-automatic handguns, grenades, and some form of suicide bomb. Lately the equipment has included cellphones linked to a VoIP address to make the controllers hard to trace.

I use the term “gun-team attack” because of the defining characteristic: one or more fast-moving assault teams relying mostly on firearms and other light weapons to attack multiple targets in a city with economic and cultural significance. This is to distinguish it from three other main types of historical terror attacks (mass hostage-taking attacks that are mostly static; assassinations; and explosive attacks on crowded buildings and mass-transit facilities).

Usually a gun team is two to four in size. Some attacks involved multiple gun teams, as in Mumbai 2008. The attack at Mumbai started out as five teams, two of which joined up on the sixth floor of the Taj Hotel (which recently reopened after two years of repairs and security upgrades). The reason for the rendezvous on the sixth floor likely was a hotel room that had been rented and stocked with a cache of weapons and supplies. (A cache was also part of preparations for the mass-hostage event at a school in Beslan, North Ossetia.)

Gun teams are not new. They were among the first in the modern terror tactics – see the 1972 Lod Airport attack by Japanese Red Army members who sympathized with the PFLP.

Three well-documented cases revealing gun-team tactics against soft targets are:

• Lashkar-e-Taiba (LeT): “Parliament House” attack, December 2001
• Al Qaeda (AQ): Attack on the Oasis executive compound, Al Khobar, Saudi Arabia, May 2004
• LeT: Attack on multiple targets in downtown Mumbai, November 2008

In general, a major gun-team attack like Mumbai 2008 develops in ten stages:

1. Hostile reconnaissance of the target, possibly a year or more ahead: these operatives walk around with cameras and GPS devices to note hallways and doors. A list of ~200 principal targets around the world is already known from interrogations, ELINT, and captured computer files and we can hope the security people are watching the cameras for such behavior. 
2. Gun-team training in a remote location, currently Pakistan but Yemen and Somalia are also likely. Gun teams receive intensive training based on such reconnaissance. They spend much time on tactical shooting and physical endurance. These men are young and highly motivated until captured, at which point their resistance seems to melt.
3. Acquisition of special gear, like satellite phones, IEDs, SIM cards for cellphones, and inflatable boats. The attackers at Grozny were in touch with a handler by cellphone, as were the attackers in Mumbai. Given this pattern it's likely that authorities in major cities are now ready to shut down local cellphone networks.
4. Sometimes, staging caches of supplies inside the primary targets.
5. Final selection of gun teams at the training camp. The organizers try to screen out those who will balk at indiscriminate killing.
6. Arrival in target city. Teams split up and try to reach the primary targets without detection. The idea is to penetrate deeply into the target without using most of the ammunition, leaving most of it for use in a confined space crowded with targets. This didn't work at Parliament House but only because the Indian Vice President's motorcade happened to be blocking the narrow gate the attackers were trying to use.
7. First stage of attack: “Large-space attacks”. Timed for high traffic hours in a crowded public place like a mall, tourist destination, or train station. Teams throw grenades and fire automatic weapons at anyone in sight, killing as many as they can and then running off. If police return fire the gun teams will retreat and move to a softer target. If the space is large enough and has enough exits this stage of shooting might go on for an hour, and so the death toll is high. The gun teams are hard to stop unless police can arrive very quickly.
8. Second stage of attack, “Small-space attacks”: A gun team leaves the first location and moves to someplace more confined. Most likely is a restaurant or club near a five-star hotel. They continue shooting there. They begin setting fires and taking hostages before they move to the final destination. If they don't break off the attack and try to escape, this leads to:
9. The third stage of attack, “hostage-holding”: Hostages are brought to a defensible location, likely to be upper floors of a hotel or apartment building. More fires are set, and boobytraps may be laid. At this point the teams may join up with each other and try to establish regular communication with leaders in a remote location. The organizers want to prolong the event as long as possible, adding fires and explosions. Often this involves checking identification and making a big show of releasing Muslims.
10. Fourth stage of attack, “martyrdom”: Die as martyrs in a firefight with police or in a bomb explosion.

The main goals of a gun team attack like Mumbai are to promote the organization as an effective fighting force, attract recruits, and polarize Muslim and non-Muslim societies. Since all these goals depend on receiving 24/7 worldwide publicity, organizers try for attacks in internationally-known locations, that are sustained over many hours, and that have visual elements suited for TV. Mumbai 2008 achieved all of these. Indications are that AQ affiliates and LeT have decided that classic suicide bombings are not able to achieve the high publicity profile that only a sustained drama can provide, though IEDs are likely to be in the list of weapons.

Why don't these attacks happen more often? Obviously there is a lot of preparation and planning but I think the main reason is that the organizers go to a great deal of extra effort to avoid failed missions. Mission failure and particularly capture of their fighters alive strike at the heart of the belief system. If the recruit base starts to think that such attacks aren't working, it's harder to get would-be martyrs to come forward. 

A large-scale collapse of martyr-willingness happened halfway through the Iran-Iraq War of 1980-1988. There is good evidence that the human wave attacks of Basij early in the war couldn't be sustained politically among the Iranian population because human waves just weren't working against massed Iraqi artillery, tanks, and Mi-24 gunships.

So it's important to capture enemy fighters alive during these events, then gather information through patient, humane interrogation. No Abu-Ghraib heavy-handed tactics are necessary. Most young fighters agree to talk, once separated from their handlers. Additional evidence that their resolve is weaker than commonly thought comes from cellphone calls recorded between the attackers in Mumbai and their controllers. Despite LeT's screening efforts, some of the attackers once on the ground repeatedly balked at controllers' orders to carry through with the carnage expected of them, during the hostage-taking stage.