The flying laboratory
Stalking Antarctica's shrinking ice
23 November 2009
I spent yesterday afternoon working on this forecast; now comes the real-time validation
Editor's note: Last month, Seelye Martin served as chief scientist on a NASA DC-8 mission based out of Punta Arenas, Chile, studying sea ice and ice sheet properties in the Antarctic Peninsula and West Antarctica region. He sent us this report from up in the air.
Sunday, October 18, on the plane flying down to the Pine Island/Thwaites glaciers. As I write, I’m sitting in a first-class commercial aircraft seat in front of an instrument rack displaying navigational data, imagery from nadir and forward cameras, and a flight chart display. I’m wearing headphones for communications, and to protect my hearing. When I need to have a private conversation without headphones, we have to shout at each other; it’s like being in a 19th century factory.
Ahead of my rack is a pile of strapped-down N2 gas bottles used in the trace gas studies, and to my right is the cylindrical gravimeter, measuring about 1 meter in diameter and 0.6 meters high, covered in an attractive dark-blue form-fitting insulating blanket and looking like an expensive piece of furniture. Further behind me is a pallet of survival gear, rations for 44 people for ten days, and two satellite beacons. If we went down, our rescuers would fly from Rothera (home of the British Antarctic Survey) or McMurdo (its American counterpart), in either ski-equipped Twin Otters or C-130s. We are also fortunate to have two commercial aircraft toilets on the DC-8; on the NSF C-130s and the NASA P-3, we make do with a funnel for the men and a small chemical toilet for the women. We communicate with the ground staff at Punta Arenas via an Iridium XChat link.
We're up here to do the second in a series of glacier surveys. A year ago, I was the funding officer for the Cryosphere at NASA Headquarters, on a two-year leave from the University of Washington. During my last two months at Headquarters, management became concerned with the coming demise of the laser altimeter measurements of the ice sheets begun in 2003 by the ICESat satellite, which ceased operation on October 11 of this year. Because its replacement will not be launched until 2016, they asked me to lead a team to investigate the use of aircraft to fill the observational gap between satellites. After I returned to Seattle in late December, NASA approved the program, dubbed ICE BRIDGE, and asked if I would to serve as program chief scientist.
The DC-8 we're flying was built in the late 1960s for Alitalia. NASA bought it in the early 1980s, based it at the Dryden Flight Research Center in the Mojave Desert and spent two years converting it into a science platform. The crew consists of two pilots, a flight engineer, navigator, two safety techs, and the lead and assistant mission managers, who coordinate the scientists’ requests with the flight deck. Some of the windows have cloth covers to prevent glare; the safety techs have taken many of these and moved them to cover windows at the rear of the aircraft so that they can watch movies and sleep in darkness. The crew is a mix of laid-back Californians and more uptight ex-military. There is also a civilian expert on the navigation instruments, and two system administrators for the aircraft intranet. We even have an onboard copier/printer.
The science experiments consist of the following: four atmospheric trace gas experiments that suck air into cabin analyzers through window-mounted ports; two sets of profiling lasers for determining the elevations of the ice sheets and sea ice; a digital mapping camera; a gravimeter that measures changes in gravity to a tenth of a mille-gal; and two radars, one that measures snow thickness on sea ice, the other an ice penetrating radar that can see the bedrock through two kilometers of ice. There are fourteen scientists onboard from the US, Singapore, China, England, Canada and Germany: three women, eleven men, and a good mix of senior scientists, post-docs and graduate students. The gravity team is not onboard; they simply set up their instrument, then let it run while they stay grounded in Punta Arenas, since they have no way to do an onboard reboot. The bad news is that their experiment has to be heated at night, which means that two people must be on the aircraft to monitor the heater, one NASA Dryden person and one gravity person. This is a normal lab; we eat cup-of-noodle soups, oranges and apples, as well as the usual junk food. A NASA photographer who came on board for a couple flights said she had a hard time making videos and taking photographs because the scientists were always eating.
The question arises: why are we flying in a forty-year old plane, when we could fly in a modern Boeing or Airbus? Two reasons: first, modern aircraft are designed using numerical stress analysis, where for example, the designers assume that the maximum stress the wing will experience is two gravities, so that with a 150% fudge factor, they design the wing such that it will tolerate three gravities. These programs are good enough that if the wing experiences 3.1 gravities, it will fall off. Because the DC-8 was built before the application of rigorous stress analysis to aircraft design, its wings will tolerate at least five gravities. This is a valuable attribute for a plane used to probe hurricanes. Second, for modern commercial aircraft, the controls are designed to operate entirely by computer, such that a major electrical failure would turn the plane into a flightless bird. In contrast, the DC-8 has two separate control systems, computer controls and the original set of cables, which allow the flaps and ailerons to be controlled manually. If the DC-8 were to suffer a complete loss of electrical power, say from a lightning stroke in a hurricane, it could still be manually flown. In addition, because of the energy stored in the hydraulic accumulators, it could also land and deploy its thrust reversers, all without electrical power. There was an even older Douglas at the airport, a rebuilt DC-3, which – as one of the Russians scheduled to fly on it from Punta Arenas to King George Island indignantly told me – had a plaque inside the door stating that the plane had been built in 1944.
The reason I started this article was to comment on the problems of weather forecasting as described in the LabLit podcast featuring the novel Turbulence by Giles Foden. I haven’t read it, but from the podcast description, I’m concerned about the fictional portrait of the meteorologist I. F. Richardson. During WWII, he was a retired pacifist meteorologist, who in the 1920’s carried out the first attempt at a numerical weather forecast using a desk calculator. This effort failed, but he gets the credit for being first, and is honored by having the Richardson number – a measure of atmosphere and oceanic stability – named after him. Being down here working on mission planning, and trying to forecast good weather for the flights, makes me appreciate in a visceral sense Richardson’s work, and the difficulties of producing reliable forecasts for the Normandy invasion – especially at a time when meteorologists lacked numerical models, satellite observations, and of course weather stations within Europe.
For the DC-8 flights, we rely on numerical forecasts and satellite imagery to predict clear skies for our target areas. As I write, our aircraft has just finished a 2,800 km transit from Punta Arenas to the Pine Island/Thwaites region, and at noon local time, is about to begin our descent to 1,500 ft. Pine Island was the name of the US Navy ship that first explored the region in 1946. I spent yesterday afternoon working on this forecast; now comes the real-time validation. We are still at 35,000 ft above high uniform stratus and approaching Pine Island Bay from the east, but believe it or not, sea ice and open water are visible to the north of aircraft. We have just passed over Thurston Island at the edge of the bay, where an aircraft from the USN Pine Island crashed, and where the aircraft and several men now lie buried beneath 90 feet of ice. The proof of the forecast will be in less than one hour as we transit Pine Island Bay and begin our descent to the Thwaites glacier.
We are descending at a rate if 2,500 ft/min, and my ears hurt. Altitude now 14,000 ft. Open water, plus ice shelf visible to south, while high clouds persist. These clouds were the part of the forecast that I hoped would go away, but under the clouds, the laser altimeters should have no problem seeing the surface. Winds are predicted to be off the ice sheet, so the ice interior should be dry with no low clouds or fog. View to south continues to improve. Lots of sea ice; as we get lower, this ice should provide an opportunity for the Kansas group to see if their snow radar can actually measure snow depth on sea ice. Geek Heaven. 9,000 ft: sea ice and horizon all around, so we may have done it again. In eight minutes we will begin our first traverse.
Close to the start of line #1, the high clouds have pretty much dissipated. Earl Frederick, one of our forecasters, has a great feel for conditions and pushed us to go west, away from a nasty little low. Smart move. Goal here is to repeat the existing ICESat satellite laser altimeter lines, four of them, on the Thwaites, to get measure of mass loss by comparison of our elevation measurements with earlier ICESat lines. Historically, the Thwaites is diminishing in elevation by about 1 meter per year. At 2,000 ft, classic wet sea ice all round. The infrared surface temperature measured by the plane is -9°C, but ice looks warmer.
Waypoint 2, beginning of first line, we are about to transit on a 400-km line up the Thwaites to an elevation of 7,000 ft. Perfect flying conditions. Laser altimeters and the ice penetrating radar are all operating. The glacier ice near sea level is heavily crevassed, as we go up the Thwaites, the ice surface becomes smoother. Now at 5,000 ft over the glacier. We are flying into a thin stratus deck, but surface is still visible, lasers are working. Whoops, laser camera just died, will have to be replaced in Punta Arenas; laser measurements not affected. 7,000 ft: we are in the haze but still have horizon, surface visibility. 7,700 ft, above haze, nice visibility, surface still visible. Turning now and heading south on line 2; this is a delightfully boring mission. Lines 2-4 also go well, we finish the survey at 1600 Chile time, then climb for 35,000 ft and home.
At this point in the mission, I collapse, watch a video from my daughter (Flight of the Conchords), and think about tomorrow’s mission. Once back over Punta Arenas, we do a high and low altitude laser runs over the runway parking ramp for calibration, where we have previously surveyed the ramp with a truck-mounted GPS. We finally land at 2230 local, for a total of 10.5 hrs in the air. For our navigation system calibration, we keep a fixed-point GPS station running at 5-times normal in an airport building, and Rothera, McMurdo and South Pole also run their fixed point GPSes at a high repetition rate for additional calibration.
In West Antarctica, we concentrated our measurements on the Pine Island/Thwaites glaciers and their associated ice tongues, as these are the fastest changing glaciers in Antarctica. Our most unusual experiment in this region was a two-day series of flying a low-level 5 by 10 km grid over the Pine Island Glacier (PIG) ice tongue. This ice tongue measures approximately 60 km in length and 40 km in width, and serves to buttress the glacier. The shape of the cavity beneath the tongue controls the access of deep warm seawater to the base of the glacier. Last summer, a British underwater autonomous vehicle called Autosub explored the cavity; the sub did several long traverses under the tongue, measuring its size and shape. On the west side of the cavity however, a 300-m high ridge blocked the Autosub, so that the topography beyond the ridge remained unexplored. In our flights, because of the difference in gravitational attraction associated with the presence of water instead of rock beneath the ice tongue, our airborne gravimeter was sufficiently sensitive to duplicate the Autosub measurements and to determine the shape of the cavity beyond the ridge. Thus our airborne measurements repeated and extended those made by a submarine. Since the topography of the cavity is critical to numerical models of the stability of the ice front, the results of this survey may be our most important work.
Because we burn 120,000 lbs of fuel per mission, we are conservative in our choice of targets with respect to the weather, but to avoid a crew mutiny, we need to finish and return to Palmdale before Thanksgiving. In our forecasts, we rely on the web-based AMPS (Antarctic Meteorological Prediction System), which can forecast the movement of major systems out to 108 hours. From bitter experience, we know it cannot forecast the presence or movement of low clouds. We also use a variety of near real-time infrared and visible satellite images, the Chilean airport weather services, and additional input from Rothera and South Pole. However, and to paraphrase Earl, we are bound to screw up at some point, so I shouldn’t take it personally if the reality doesn’t agree with the forecast. So thank you, Professor Richardson, for laying the foundation of numerical weather prediction that makes our research possible.
Quite coincidentally, I’m informed by my spouse back home that the reading from the Episcopal lectionary this Sunday is from Job 38:1-7, where the Lord answers Job out of the whirlwind. This is the only part in the Bible where geophysics gets its proper due, with its mention of the long time and space scales of the earth and universe. It seems appropriate to the mission:
Where were you when I laid the foundation of the earth? Tell me, if you have understanding. Who determined its measurements – surely you know! Or who stretched the line upon it? On what were its bases sunk, or who laid its cornerstone when the morning stars sang together and all the heavenly beings shouted for joy?