AAOE remembered - weather to go
By Peter Salter
17th August 2017
The US strategic team charged with establishing the project in 1987 concluded that the UK Meteorological Office (UKMO) was best placed to provide the necessary meteorological support; especially given that it had developed earlier an operational global NWP model. The model was run at Bracknell twice dally together with a limited area model with a finer grid than the global model's grid length of c 300 km. In addition, in 1986 a major weather office was established on the Falkland Is to meet military needs; there Peter Salter was heavily involved. Punta Arena was an invaluable and nearest upstream station. With hindsight one wonders whether it was predestined that he would become the Mission Forecaster for the AAOE.
Early summer 1987 at Ames I met with the Head of ER-2 Operations (John Arvesen) and the ER-2 Chief Pilot (Ron Williams) so that issues regarding the aircraft safety flying into and out of Punta Arenas, and for the ER-2 overflying Antarctica, were discussed and better understood. A visit to see an ER-2 brought home the challenges ahead, if any were needed, and that the Mission Forecaster's advice, inter alia, on any potential flight from Punta Arenas would be pivotal in the pilot's decision to fly, or not. Following the Ames visit I joined up with Chief Scientist Dr Adrian Tuck, Jim Barrilleaux, Dr Estelle Condon and a few others to visit both Santiago and Punta Arenas where party members variously had meetings with third parties to review and thrash out logistical issues ahead of the project commencing in August. Adrian Tuck and myself flew back from Punta Arenas to visit the Chilean Met Service Forecasting Director so that they understood the rationale behind AAOE and what would be requested of the Chilean local forecasting office at Punta Arenas. In that fact-finding visit in the austral summer, I familiarised myself with the 'met' facilities available there: basically round-the-clock local weather observations, plotted synoptic PMSL charts, and TAFs (Terminal Aerodrome Forecasts) and radio-sonde programme. By August an enhanced satellite cloud imagery receiving facility had been obtained by the Chilean local forecast office. This permitted colour-slicing so that cloudtop height might be inferred. This became important when the incidence of cirrus cloud over Antarctica became evident from the very first ER-2 flight.
On that preliminary visit, I met a Jesuit priest from a local seminary whose weather records went back some years. He told me that the climate seemed to be changing as they had noticed that significant snow was now relative rare, and temperatures were not as low as in past years, but the wind could still blow hard; very interesting. The celebrated Chilean poet, Jose Grimaldi (whom also I met) wrote in a poem that the Punta Arenas wind could blow flags out straight.
AAOE required so much timely meteorological information from UKMO, and from ECMWF near Bracknell, that dedicated links from UKMO to Punta Arenas, and from ECMWF to Punta Arenas, were put in place by a US c9ntractor. There was also a dedicated link put in place between the two centres in the !JK. Data were sent in bespoke form to meet the forecasting requirement, all by docfax. The contractor set up their telecoms unit which was manned 24 hrs a day.
The UKMO NWP global coarse mesh model's output from its 00 GMT and 12 GMT operational runs, required for our chosen extensive geographical coverage was: relevant data for PMSL, winds and temperatures for FL0085, FL070, FL100, FL180, FL240, FL300, FL340, FL390, FL450, FL530, FL610, FL680 and FL780 for HH + 00, 06, 12, 18, 24, 30, 36, 48 and 72 hr ahead. The UKMO fine mesh model was run also twice daily for the same levels but only up to 36 hr ahead. For all the above Fls, data were plotted in conventional wind/ temp format familiar to aviators, on an appropriate lat-long grid. ECMWF data based on 00 GMT output covered the same fields in the format of the UKMO output, but extended to 240 hr ahead making it possible to identify potential windows of opportunity for the project. In the event of one of the links from the UK going down, the data could be passed across to the other centre in the UK and down the functioning leg. In the event of an untimely and total telecomms outage there was sufficient flexibility in data received from the previous run to provide information for a flight the following day. The AAOE products from the UKMO 00 GMT run were received usually by around 0200 LCT, and from the 12 GMT run by around 1400 LCT. In particular this permitted pilots to be briefed early in the morning pre-take off based on the very latest forecasts, and similarly for use at post-flight debriefings/ science meetings. Clearly the forecaster was working pretty hard as so many charts were analysed after receipt and then displayed. A complete data set was made for the project; several '00s of charts per 24 hr. The met cell looked like a wallpaper-hanging outlet! TOMS data indicating the location of the ozone hole as the days passed, were routed to the met cell as well as products from the Atmospheric Chemistry Group at UKMO; these were of key importance to Adrian Tuck.
So what were the meteorological issues for operating out of Punta Arenas downwind of high ground? The predominant wind regime at the airfield is from the west- southwest and thus most flights used RY25/07. The ER-2 weather requirements at Punta Arenas airfield were specifically: tenminute mean wind speeds to be ≤ 25 kn and the cross-runway component to be ≤ 12 kn, both for take-off and landing together with no expected wind rotoring. Cloud base and visibility were considerations too, with cloud base ≥ 200 ft and horizontal visibility ≥ 1500 m. The designated diversion airfield was Puerto Montt some 800 miles north in Chile, and where the weather issues were relatively sublime in comparison with Punta Arenas. The possibility of clear air turbulence at flight levels fly also needed consideration. Overall the ER-2 would be limited to operating no higher than around FL720, whilst upper temperatures of ≤ -85 °C or less could lead to problems for various lubricants used in the aircraft control mechanisms. In addition ER-2 flights were restricted to not flying south of around 72 °S in view of the distance back to the aircraft's diversion airfield. Ahead of a projected flight, the ER-2 pilot and flight planning team visited the met cell to be briefed in detail as to expected airfield weather and upper winds and temperatures together with any significant weather for the projected flight; formal flight documentation was issued to cover the flight. As with the meteorological briefing the Project Scientist briefed on the scientific rationale for the upcoming flight and thus to request that the pilot fly, as appropriate, particular zones over Antarctica in level, ascending or descending mode at specific altitudes and locations. Thus Adrian and myself worked,very closely in our office which conveniently was located close to the ATC tower. Our office had a good view of an airfield anemometer on an adjacent building; very useful at times!
For the DC-8 and its long flight duration, there were differences from ER-2 operations which had to be addressed as generally it could be flying over the ice no lower than FL240 and possibly up to FL450. Upper temperatures < -76 °C were to be avoided due to potential aircraft issues. There was set a limit of ten-minute wind speeds to be ≤ 34 kn for take-off or landing with a limiting crossrunway component of 23 kn. Cloud base should be ≥ 500 ft and visibility ≥ 800 m at the airfield. Rio Gallegos was the authorised diversion airfield. Prior to the flight the pilot and planning team visited the met cell for a full weather briefing. Flight documentation in a similar format to that for the ER-2 but for altitudes typically between FL240 and FL450, and of course to provide forecasts of significant weather/ cirrus cloud/max wind/CAT etc.
Both ER-2 and the DC-8 teams were asked to provide the Mission Forecaster with in-flight reports of winds and temperatures which he coded into suitable AIREPs and sent directly to the UKMO for inclusion in the next NWP run. A copy of each of these AIREPs was given to the airfield's Chilean forecast office to pass up into WMO's Global Telecom ms System.
Performance of forecasts for the ER-2? The ER-2 pilots had been alerted to expect occasions of winds of 100 kn, or more, and temperatures as low as -90 °C during AAOE. On the very first flight the wind reached that expected strength and the temperature dipped to a remarkable -87 °C aloft Adelaide Is before the aircraft returned as planned. UKMO NWP output was looking to be reliable. There were occasions of course when the surface winds produced surprises causing the projected flight to be cancelled. However for one returning flight, Punta Arena developed airfield mean wind speeds of a little more that 40 kn; the aircraft was landed safely, and recovered in the usual way by the chase crew. There was no doubt in anybody's mind that we were working in a challenging environment. A back-of-the-envelope calculation for ER-2 flying days at its operating altitiudes down south, the UKMO model's mean vector difference for forecast winds was ± 10 kn and ± 2 °C for temperatures. By the end of AAOE the ER-2 had made 12 successful flights, considerably more than thought possible.
Performance of forecasts for the DC-8? Such an aircraft flying at its possible range of altitiudes should expect strong winds at times and often very low temperatures, and this was the case. This aircraft with key instrumentation to make clear-view vertical measurements above it and clear-view runs horizontally, required navigation around regions of cirrus cloud. The enhanced cloud imagery system available in the Chileans' local forecast provided assistance to the Deputy Chief Scientist (Dr Brian Toon) and the DC-8 navigating team to steer their aircraft as far as possible away from significant cirrus.
As AAOE progressed It became evident from DC·8 wind reports, on occasions of very strong upper winds, that the actual forecast upper winds were stronger than UKMO model forecasts, especially where upper ridges occurred. These ridges often extended much further south than previously thought likely. On those occasions when it was practical to do so, flight documentation was amended by hand (red ink!); typically an NWP predicted wind of say 90 kn might have required an increase of 20 kn or more. When forecast winds for the DC-8 were so modified, they assisted usefully the overall flight, but note it was not always possible for the forecaster to modify the model winds as was evidenced on the DC-8's final departure from Punta Arenas. The DC-8 made a total of 13 flights including fts long run home via NZ. No overall comparison of upper winds and temperatures for DC-8flights was made in theatre.
Many memories come back whenever AAOE is mentioned to me. Just two will suffice here: that road to the airport, and the many happy faces when the aircraft returned with lots of data.