Land and Hold Short

While the readership for this blog is still…shall we say…extremely modest, the blog has somehow crawled its way up to the #1 result on Google for the search phrase “land and hold short,” ahead of more worthy targets such as the AOPA brief on land and hold short operations (perhaps my link will help push AOPA’s page back up). If there are any air accidents because pilots clicking on I’m feeling lucky got my page instead of the AOPA bulletin, well, I guess they weren’t so lucky after all.

It’s always fun to see a specialized use of a common word or phrase make it to the top of Mt. Google, as is the case with the Simple API for XML, beating out the common short form of “saxophone” for a Google search for “sax“, and Tim Bray’s Ongoing blog, topping out the Google results for “ongoing“. No points, of course, for coining something new, like “Slashdot” or “Kazaa”. Similarly, it’s to be expected that Cessna would have the top hit for the search “cessna“, but the fact that The New Piper, perpetually on the brink of bankruptcy, has the top Google result for “piper” is probably leaving a lot of bagpipers seething and stomping in their kilts.

As I mentioned in an earlier posting, Ottawa will be facing an American-style TFR (temporary flight restriction) during President Bush’s visit on Tuesday 30 November and Wednesday 1 December, covering nearly 10,000 square kilometers of Canadian airspace from the ground to 12,500 feet. I’m not too worried about violations in the inner core, which is controlled airspace (class D) anyway, but I am worried about innocent pilots getting in trouble inside the bigger circle flying from uncontrolled airports like Arnprior or Smith’s Falls. To try to help, I’ve made up a map of the TFR coverage (low-resolution and high-resolution), which is available in the full version of this posting on my Web site.

Please note that these charts are ENTIRELY UNOFFICIAL. Pilots are responsible for reading the actual TFR themselves and for confirming the boundaries — I may have any or all of them wrong. OK, here’s the low-resolution version, based on the U.S. Montreal Sectional Chart (since we’re not allowed to distributed copies of a Canadian VNC, just to add insult to injury):

(For a high-resolution version, click here.)

According to version of the NOTAM available at the time of this posting (it may change still), there are three areas of restricted airspace, all marked on the chart:

CYR537

This is the normal Parliament Hill class F on steriods: it has grown from a 0.5 nm radius and 1500 feet to a 12 nm radius and 12,500 feet, effectively closing CYOW, CYND, and CYRO to any non-airline/emergency/military traffic. This restriction is in force from 15:50z (10:50 EST) on Tuesday 30 November to 13:00z on Wednesday 1 December.

CYR539

This is an expanded version of the CYOW control zone, again, from the surface to 12,500 feet: basically, the airport will be closed while the President is arriving and while he is leaving. The restriction is in effect from 15:00z-15:50z (10:00-10:50 EST) on Tuesday for the arrival, and again from 12:30z-13:15z (7:30-8:15 EST) on Wednesday for the departure. Note that the airport is still covered by CYR537 in between these two closures.

CYR540

This is an expanded version of the Ottawa Terminal control space, again, from the surface to 12,500 feet. This is the one most likely to get pilots in trouble, since many pilots who fly planes and gliders from the smaller airports under it do not call ATC at all for normal flights. Note that there is no way to fly under this, unlike the outer ring of Ottawa Terminal airspace. To fly in this area, you have to be on a flight plan, talking to ATC, and squawking an assigned code. This restriction is in effect from 15:00z (10:00 EST) on Tuesday until 13:15z (8:15 EST) on Wednesday.

Good luck, everyone, and please be careful. Hopefully, COPA will let us know where we can vent our frustration when this is all over. I also hope that the government plans to compensate all of the small schools and other aviation businesses that cannot afford to lose a day’s revenue.

I am a big fan of NAFTA and Canada/US relations in general — free trade with the U.S. is the lifeblood of our economy — but sometimes the price is too high. President Bush is visiting Ottawa next week, and our government (which is usually quite calm and rational about security issues) is imposing American-style flight restrictions. Basically, if I don’t reposition my plane to an airport at least 60 km away from Ottawa, I will be grounded (along with hundreds of other non-military, non-commercial flights) for much of 30 November and 1 December. Here’s the NOTAM: [updated with COPA's reaction]

 040918 CZUL MONTREAL FIR PART 1 OF 2  CZUL DESIGNATED AIRSPACE HANDBOOK IS AMENDED AS FOLLOWS: 1) CYR537, PARLIAMENT HILL ON, REVISED TO READ: CLASS F RESTRICTED AIRSPACE IS ESTABLISHED WITHIN THE AREA BOUNDED BY A CIRCLE OF 12 NM RADIUS CENTRED ON 452529N 754159W, SFC TO 12,500 FT MSL.  NO PERSON SHALL OPR AN ACFT EXCEPT FOR STATE ACFT, MIL, POLICE OPS, REGULARLY SKED COMMERCIAL PASSENGER AND CARGO CARRIERS, HUMANITARIAN OR EMERGENCY FLTS AUTH BY ATC. 0411301550/0412011300 2) CYR539, OTTAWA ON, CLASS F RESTRICTED AIRSPACE IS ESTABLISHED WITHIN THE AREA BOUNDED BY A CIRCLE OF 10NM RADIUS CENTRED ON 451921N 754009W (OTTAWA/MACDONALD-CARTIER INTL).  SFC TO 12,500 FT MSL.  NO PERSON SHALL OPR AN ACFT WITHIN THE AREA DESCRIBED UNLESS THE FLT HAS BEEN AUTH BY THE RCMP AT (613) 993-1430.  0411301500/0411301550 AND 0412011230/0412011315 3) CYR540, OTTAWA ON, CLASS F RESTRICTED AIRSPACE IS ESTABLISHED WITHIN THE AREA BOUNDED BY A CIRCLE OF 30NM RADIUS CENTRED ON 451921N 754009W (OTTAWA/MACDONALD-CARTIER INTL) EXCLUDING CYR537 AND CYR539. SFC TO 12,500 FT MSL.  NO PERSON SHALL OPR AN ACFT WITHIN THE AREA DESCRIBED EXCEPT FOR STATE ACFT, MIL AND POLICE OPS, REGULARLY SKED COMMERCIAL PASSENGER AND CARGO CARRIERS, HUMANITARIAN OR EMERGENCY FLTS, ACFT ARR OR DEP LOCAL AERODROMES AND ACFT TRANSITING THROUGH THE AREA AUTH BY ATC: ALL ACFT ENTERING, EXITING OR OPR WITHIN CYR540 SHALL: -  BE ON AN ACTIVE IFR OR VFR FLT PLAN WITH A DISCRETE CODE ASSIGNED BY ATC 866-837-2633 AND SQUAWK THE DISCRETE CODE PRIOR TO DEP AND AT ALL TIMES WHILE IN CYR540. -  REMAIN IN TWO-WAY RADIO COMM WITH ATC.

040918 CZUL MONTREAL FIR PART 2 OF 2  0411301500/0412011315 ACFT PLANNING TO OPR WITHIN CYR537, CYR539 AND CYR540 CAN ANTICIPATE POSSIBLE DELAYS.  0411301500 TIL 0412011315 

This is … dare I say it … Uncanadian. There was one case of a similar overreaction when the G7 leaders met in Kananaskis, Alberta a few months after the September 11 attacks, but that case was a bit more understandable (if still wrong-headed), since everyone was pretty jumpy. Still, foreign leaders visit Ottawa all the time, and we have nothing like this kind of fuss. We usually have a tiny restriction over Parliament hill (0.5 nm circle up to 1,500 ft MSL) and a similar one over Rideau Hall, but that’s about it. I do believe that we should welcome President Bush to Canada, but I do not believe that our government should roll over and let the U.S. start imposing the kind of security hysteria they’ve foisted on their own citizens. If you’re not a pilot, imagine not being allowed to take your car out of the driveway for two days because a VIP is visiting your city, and you’ll understand. The worst thing is that other leaders might start asking for the same treatment now so that they do not seem less important than the U.S. president. I hope that COPA is ready to speak up, loudly — this is when we need you most, guys.

[Update: COPA has posted a reaction on their front page. Unfortunately, that's not a permalink, and the information will eventually disappear; so, in the spirit of persistence and hopefully under the umbrella of fair use, here's the most important part of the commentary:

When COPA first heard about the Bush visit, we contacted officials at Transport Canada, NAV CANADA and the Department of National Defence to determine if there would be any airspace restrictions and to press for an early disclosure of any restrictions so that the word could get out to everyone who may be affected. In past presidential visits, immediate airspace around the airport was closed for a short period before arrival and after departure. However, with the state of the world these days, there could be more extensive closures involving more airspace and time.

With just a few days to go until the visit, there was still nothing announced, except for vague rumours that something was coming. After all, COPA as well as any other organization, is not permitted to have input to these security operations. While we understand that there were several agencies involved and some confusion about the President's schedule and the security requirements, and while we appreciate and support the desire to provide protection, we expressed a strong concern that a last-minute NOTAM and sudden closure of airspace may go unnoticed by some pilots and there could be dire consequences as a result.

Following 9/11, COPA convinced the government that a better way to communicate and cooperate was needed in order to ensure that clear NOTAMs are produced and conveyed in a timely manner. To that end, COPA and CBAA have positions established near the Transport Canada Operations Centres. This resource was not employed for the development of these NOTAMs, although, in our opinion, it should have been. We will be pursuing this in hopes that a more coordinated effort can occur in the future.

In the meantime, check NOTAMs whenever you fly. Although these NOTAMs are not officially known as TFRs in Canada, what looks like a duck... .Let's hope that the "T" in TFR means "Temporary" in Canada rather than what it has come to be known in the US -- "Total".

]

To follow up on my last posting about ice, here’s an interesting piece by Philip Greenspun about dealing with light icing over the New York area in a Diamond Star. There’s nothing exciting here, no “ohmygodwe’regoingtodie” moments, just the practical, routine way that small plane pilots handle the light rime icing that can happen in any cloud. Philip made sure he had outs, he used one of them, it worked, and he completed his flight uneventfully: this is the kind of piece new IFR pilots have to read.

Ice in Ottawa

Philip did a good job using layperson’s language in his piece, so I’ll try to do the same here. In general (ignoring local effects like weather around the Great Lakes), there are four bands of weather that can hit Ottawa over the year, stacked from north to south like the layers of a cake. First, there’s the really cold, dry Arctic air furthest north. In the summer, that’s far out of reach; in the winter, it often pushes down into Ottawa giving us cold temperatures (too cold for icing) and good visibility. To the south of that, there’s usually a band of mucky, wet air that hovers just around freezing and causes most of the icing problems for airplanes. In the summer, that air is up around Hudson Bay; it passes through Ottawa in the late fall on its way down to the northeastern US, stays down there for much of the winter (occasionally poking up into Ottawa to annoy us), then passes back up through Ottawa in the early spring on its way back to the Arctic. South of that is the dry, moderate air that gives a nice spring, summer, or fall day. We get that for a lot of the summer in Ottawa, and (I’m guessing) it spends the winter down in Florida, giving nice weather for the snowbirds. Finally, the layer furthest south is the hot, wet air from the Caribbean and Gulf of Mexico. In summer, that air sometimes pokes up into Eastern Ontario bringing thunderstorms and other violent weather, but it seems to spend most of its summer southwest of us, tearing up the U.S. midwest.

So, the point of all this is that (if this winter goes like most) the wet, icy weather will move south soon. Philip may have to deal with it all winter, but we (I hope) will soon have more nice, safe, cold days for flying. Of course, with global warming, we could end up getting more of New England’s winter weather, and then winter flying in Ottawa will be much less pleasant. Keep it nice and cold, please.

Some time in the middle of October my instrument rating became a lot less useful, and it will stay that way until some time in the middle of next April. For half the year, the rating is almost a guarantee that I can fly. Sure, sometimes there’s a solid squall line that I cannot get around, and sometimes the weather is below my personal IFR minima, but last year, out of dozens of planned flights during the late-spring/summer/early-fall season, I had to postpone exactly one due to weather (a night flight with numerous thunderstorms enroute) — that’s better than I’ve done in the past travelling for business on the airlines. Come October, though, I might almost as well be VFR only.

Freezing levels

Once the freezing levels come down low enough, clouds can mean ice, and most single-engine planes (and many twins) are not equipped to deal with that. It doesn’t have to be winter weather for icing to be a problem: with the standard temperature lapse rate (i.e. no inversion), a nice warm fall/spring temperature of 8 degC on the ground means that I’ll be hitting icing conditions at about 4,000 feet, which gives me very little room for IFR flight (the lowest IFR obstacle-clearance altitudes are typically around 3,000 feet, and I like a 2,000 foot safety margin from the icing). If the clouds are high enough, I can go VFR underneath them; if the ceiling is low, I might just have to wait.

Actually, the story is more complicated than that. I will admit that last winter — my first with an instrument rating — I flew a lot in cloud at below-freezing temperatures. The only ice I ever saw was a tiny piece of clear ice starting to form on my outside air temperature probe, in drizzle around -1 degC. I immediately descended 1,000 feet, the ice melted off, and I finished my flight uneventfully aside from a slightly fast, no-flaps landing. What if it hadn’t melted? Well, I could have descended further below MEA to MOCA (the minimum safe IFR altitude); if that didn’t work, I had already listened to the weather from the nearest big airport, which was reporting surface temperatures of around 6-8 degC (memory is getting fuzzy), and I had the navaids for a nearby approach tuned in, so I could have shot the approach until I broke out from the clouds and then decided whether to land or continue VFR underneath. In other words, I had more than one very good out. Would I have tried the same thing with freezing temperatures down to near the ground or over hilly or mountainous terrain? Not a chance. There’s always got to be a way out, preferably several.

Too cold for icing

More typically, though, I can fly in IMC in the winter without worrying, because as the winter goes on it gets too cold for airframe icing. According to this bulletin to FAA controllers, at -16 degC, 95% of droplets in a cloud will have frozen into ice crystals (which won’t stick to the airframe); at -25 degC, 99.9% of the droplets will have crystalized. I live in Ottawa, where people refer to -16 degC on the ground in February as a mild day; I’ve even flown on -16 degC days without preheating (I promise never to do it again). And, when there’s no inversion, -16 degC on the ground might mean -30 degC at altitude, where there are definitely nothing but ice crystals in the clouds. So come January and February, whenever Ottawa is under a cold arctic airmass, I can happily go and fly in the clouds without worrying about icing at all. However, when that warm wet gulf air mass pushes up from the U.S. and the surface temperature rises to, say, -5 degC or even 0 degC, I’m back to VFR only. Once the surface temperature gets above 5 or 6 degC, I can fly again, because I know that I’ve got warmer air underneath to get rid of any ice I might pick up. That means that the worst time for icing is not winter but late October, November, March, and early April, where the surface temperature is just a bit above freezing.

I think it’s a big shame that training courses for instrument ratings do not spend much more time on icing. Simply saying to an IFR pilot “Don’t fly in clouds below freezing” is roughly equivalent to saying to a VFR pilot “Don’t fly when there are clouds in the sky” or saying to a teenager “Don’t have sex” — they’re probably going to do it anyway, so you might as well teach them how to do it safely. Sometimes flying in icing conditions is OK and sometimes it’s not, and we really need more resources to help new instrument-rated pilots make more informed decisions and develop strategies for flying safely during the icing months.

In general, pilots are a pretty smart bunch of people, so I’m always surprised reading aviation mailing lists and newsgroups to learn how many of them don’t seem to have the slightest understanding of how to control their planes’ airspeed and power. This ignorance will typically come out in a statement like “a Cessna 172p really only flies about 105 knots” (it can really fly around 120 knots true airspeed if the pilot knows how to operate it). I thought it would be interesting to look at how pilots actually set speed and power in planes with fixed-pitch propellers, and what they get for their trouble.

Marketing Writers

So the POH says that your plane will fly at 120 knots, but you never seem to get that — typically, pilots blame marketing writers for making up numbers so that their planes look better. In reality, the POH’s do usually try to put the plane’s speed in the best light, but the numbers are not made up. Normally, a plane with a normally-aspirated piston engine (like a 172 or Cherokee) will have its true airspeed will be calculated at between 7,000 and 8,000 feet density altitude, where the plane flies at its fastest: any lower, and the dense air slows you down; any higher, and there’s not enough oxygen for the engine to produce 75% power. If you normally fly below 7,000 feet or above 8,000 feet, you can expect to see a slower true airspeed. Some manufacturers will also test with the plane lightly loaded, providing a boost of a couple of extra knots (that’s the case for my Warrior) — if you put the whole family on board, you can expect to fly a bit slower.

One of the biggest problems, though, is the wind. It is a simple fact that you will spend longer (possibly much longer) flying with headwinds than tailwinds, because headwinds slow you down — for example, a trip might take three hours outbound against the wind and only two hours return with the wind, meaning that you spend 60% of your trip with headwinds. As a result, your average groundspeed will always be lower than your plane’s best true airspeed, possibly by 10% or more if you fly a lot with strong winds. That difference does not mean that the POH lied about the true airspeed, which should be as advertised (more or less), but just that wind is a big pain.

So how do pilots control true airspeed and fuel burn? It turns out that it’s easy to manage one or the other, but managing both can be a bit of a challenge.

The Constant-RPM Pilot

To start, consider the pilot who always flies with the same tachometer reading (say 2400 rpm), letting the indicated airspeed rise or fall as it will. Using a constant RPM will give a nearly constant true airspeed at any altitude, so this seems like a simple system: according to the POH, 2400 rpm will give a true airspeed of 109 knots at 2,000 feet density altitude, 107 knots at 6,000 feet density altitude, and 105 knots at 10,000 feet density altitude. That’s easily close enough, and makes flight planning simple: set the power to 2400 rpm and assume 105 knots true airspeed (to allow for old paint, chips in the propeller, draggy antennas, etc.), and everything will usually work out for any cruise altitude, plus or minus the wind.

Unfortunately, for this pilot fuel burn will vary. At 2,000 feet density altitude, the Skyhawk’s O-320 Lycoming engine spinning the propeller at 2400 rpm will be producing 69% power (110 hp) and burning 7.7 gallons of fuel per hour; at 12,000 feet density altitude, the engine will be producing 56% power (90 hp) and burning 6.3 gallons of fuel per hour — that’s an almost 20% difference in fuel consumption. Since fuel consumption seems unpredictable, the pilot has learned to fly short legs (so that there’s always lots of extra gas in the tank), making IFR flight difficult. The pilot also might decide to spend thousands on speed mods to get 3 or 4 extra knots, when simply using the right power setting at 7,000 or 8,000 feet density altitude would give an extra 12 knots without any modification to the plane. This is the pilot who goes on mailing lists and claims that his or her plane is much slower thant he POH says it should be.

The Constant-Indicated-Airspeed Pilot

Next, consider the pilot who always flies with the same indicated airspeed, varying the RPM as required. Let’s say that the pilot chooses 114 knots indicated (111 knots calibrated), which will give the maximum cruise performance for the Cessna 172p. At 2,000 feet density altitude, the pilot needs to set the engine to 2500 rpm to maintain this airspeed; at 8,000 feet, the pilot needs to set the engine to 2650 rpm (at 10,000 feet, the plane is no longer capable of this speed in level cruise). By using a fixed indicated airspeed, the pilot is actually using a fixed power setting, and that means a fixed fuel consumption: at either 2,000 feet or 8,000 feet density altitude, the fuel burn will be the same (about 8.5 gallons per hour according to the POH).

Unfortunately, for this pilot true airspeed will vary, making flight planning trickier. At 2,000 feet density altitude, the plane’s true airspeed will be 114 knots; at 8,000 feet density altitude, the plane’s true airspeed will be 121 knots. There’s also the problem that the plane might be draggier than the one used to calculate the POH numbers, so 114 knots might actually push the engine up to 80% power or higher, burning extra fuel and risking detonation.

Power Setting

So the constant-RPM pilots know how fast they will fly but not how much fuel they will burn, while the constant-indicated-airspeed pilots know how much fuel they will burn per hour, but not how fast they will fly.

Right about this point, a lot of people will argue that that’s not the case — after all, if a pilot always flies at about the same density altitude, he or she will find the fuel burn and true airspeed pretty predictable with either technique. The problem comes, however, when one of those pilots flies somewhere different than the normal summer cross-country at 3,000 feet (or whatever normal means for that pilot). For example, where I live, in Ottawa, it gets cold enough in the winter that the density altitude at 3,000 feet is sometimes still negative. A constant-RPM pilot who flies under these conditions will burn far more fuel than expected, and could end up landing with near-empty tanks when expecting a half-hour reserve; a constant-indicated-airspeed pilot who flies under these conditions will fly far slower than expected, and could end up landing with near-empty tanks (again) because of the extra travel time. I believe strongly that this is why some good, experienced pilots run out of fuel: something changes from their normal flying routine (colder weather, different cruise altitude, etc.), their normal technique produces abnormal results, and they do not understand how to compensate for it.

Knowing your power setting requires calculating your density altitude and then looking up your RPM in a table or graph, which is a big pain, but it does allow you to get maximum performance (true airspeed and range) out of your airplane safely. You do not have to do that for every flight, of course — once you know your indicated airspeed at any given power setting, and have confirmed your fuel burn, you can use a variation of the constant-indicated-airspeed approach, as long as you do the calculations to get your true airspeed. The alternative is believing that your plane is 10 knots slower than it really is, or never knowing quite how fast you’ll fly or how much fuel you’ll burn.

Mirabel airport is a beautiful facility. Located in the foothills of the Laurentians, it has two 12,000 foot runways (one with a CAT II approach to 100 ft) and a big, bright terminal building. Now, less than 30 years after it opened, the airport is mostly shut down, but there’s no outcry from the aviation community; in fact, no one except the employees losing their jobs seems to care much at all. I know I don’t.

The problem is that Mirabel reminds Canadians of some of the more unpleasant parts of recent Canadian history. Way back when Canada was celebrating its 100 birthday with Expo ‘67, Montreal was Canada’s unchallenged first city, a corporate, cultural, and industrial giant. The next year a Montrealer, Pierre Trudeau, became Prime Minister, and one of his priorities was to help Montreal keep growing into one of the world’s major cities. A new, larger airport built (too far) outside the city was one of the prerequisites, so Trudeau’s government grabbed an enormous area of land north of the city and evicted 10,000 people from their property (most of that was never used). By 1975, when the airport opened, it was not only massively over budget but already superfluous — Montreal’s downward spiral had begun with the October crisis of 1970 and Trudeau’s military intervention, which had alienated French Canadians and would soon lead to separatist governments, referendums, and the repressive language laws that would drive out much of Montreal’s business base to Toronto and Ottawa during the late 1970’s and early 1980’s. Furthermore, preparations for the (disasterous) 1976 Olympics were already pushing the city itself close to bankruptcy.

When my parents left me with my grandparents and drove to Montreal in the late 1960’s, they saw a bright, shining city of the future; when I went there on school trips around 1980, I saw a dirty city full of boarded-up buildings and a mind-bogglingly large collection of strip clubs advertising Girls! Young! Young! Young!. By then, Montreal had gone from Canada’s first city to Canada’s basket case: barely enough people were travelling to Montreal to keep the old Dorval airport busy, much less Mirabel. So the government tried different ways to force people to use Mirabel — for a long time, all international flights had to use the airport — damaging Montreal’s economy even more by inconveniencing travellers.

Eventually, long after Trudeau left office, the government accepted that Mirabel was never going to flourish and allowed international flights to use Dorval airport again. The city of Montreal is cleaner and more prosperous than it has been in a long time, but Dorval seems quite adequate for its needs: it is still a relatively quiet, general-aviation-friendly airport (there are no landing fees, and they’re always happy to let me do a touch-and-go: try that at Toronto/Pearson). For a while, charter flights kept using Mirabel, but on Monday 1 November 2004 the last of those departed and the passenger terminal shut down permanently.

That’s not to say that people haven’t found uses for Mirabel. It’s great for practice approaches — how often can you find an airport with multiple ILS approaches and 12,000 ft runways where the tower controllers are actually grateful for something to keep them busy? And flight instructors love setting up trick cross-country problems where Mirabel is the obvious fuel stop, to see if students read the Canada Flight Supplement closely enough to find out that Mirabel does not have 100LL fuel available. Mirabel should have been the obvious choice for charter flights to Mont Tremblant, a trendy ski resort a short drive to the north, but instead promoters fixed up an old military strip a bit closer and are amusingly calling it Mont Tremblant International Airport.

I haven’t heard the details, but I imagine that Mirabel’s tower will shut down now and the control zone will revert to class E. At least one of the runways will probably stay in operation — Mirabel is the home base of Bombardier, the world’s #3 airline manufacturer after Airbus and Boeing — but it might not be worth the money it will cost to maintain the ILS approaches (Ottawa is already a good, close alternate for Montreal) unless there’s a lot of freight traffic. Recently, Montreal’s Dorval airport, the one Trudeau tried to replace, was renamed Pierre Elliott Trudeau Airport in his honour, but most people think that in all justice the name should have gone to Mirabel.