This Year is Difficult to Predict

by Kirstin Werner, Alfred Wegener Institute

The World Meteorological Organization is concerned about the decrease of airborne atmospheric observations due to the suspension of commercial air traffic since mid-March. The German Weather Service starts using additional data from radiosondes’ sinking to the ground. While many of the research campaigns planned for this and the coming season in the Arctic and Antarctic were cancelled and others are postponed, the one-year ice drift MOSAiC goes on. How does the COVID-19 pandemic affect the polar prediction community?

There are years when the weather can be forecast more easily than in other years. “The COVID-19 year is one with the summer being more difficult to predict than for example last year’s summer”, says Detlev Majewski, head of the German Meteorological Service’s department Meteorological Analysis and Numerical Modeling. In March, when the outbreak of COVID-19 in Europe forced people to move their entire lives to home, the weather was still dominated by longer-term stable high-pressure systems enabling nice and sunny conditions. However, “the current weather in Germany resembles more of a typical mid-European summer”, says Majewski, “with more instability due to rapidly moving short-term low-pressure areas over Europe, allowing at maximum some three to four days of sunny weather before chilly temperatures are back.” In Germany, this can be felt in particular at the coasts, for example close to the North Sea.

Missing an Historic Event while MOSAiC Continues
Close to the North Sea is home for Stefanie Arndt, sea-ice scientist at the German Alfred Wegener Institute (AWI). After 145 days being away from Bremerhaven, Arndt returned to Germany only four weeks ago. Together with more than one hundred German and international people, she spent the time when the pandemic peaked in Europe in the central Arctic aboard the German research vessel Polarstern. “While I definitely missed what can only be considered an historic event with entirely empty streets and people in Germany panic buying pasta and toilet paper, the pandemic had a huge impact to MOSAiC. Instead of three, I have now spent five months in total away from home, and for some time, there was a huge uncertainty amongst the team about how and when the next exchange of the expedition legs would happen. At the same time, at least I myself have never worried about anything but my family’s health.” Eventually, the MOSAiC project board and logistic team found a good solution for the teams’ exchange. After a number of international requests and negotiations, the two German research vessels RV Maria S. Merian and RV Sonne were finally able to assist in swapping the science teams, crews and captains. Participants of leg 4 had to stay quarantined in a hotel in Bremerhaven for two weeks in May before boarding the vessels at Bremerhaven port. In a Svalbard fjord, they met with RV Polarstern in early June for the transfer. MOSAiC leader and chief scientist of leg 4 Markus Rex is proud of the MOSAiC logistic team who assessed all possible options to make the transfer happen: “It had been a difficult time for MOSAiC, and sometimes it was not clear at all whether the expedition would go on, but in the end, while most of the international expeditions had to be cancelled due to the Corona virus, MOSAiC continues”. During the time when Polarstern left the central Arctic sea-ice observatory unwatched for some weeks, most of the research camp needed to be packed. However, some of the autonomous instruments left on the ice floe during the vessel’s absence continued measurements. “After our return to the floe, re-building of the research camp happened within just a week”, says Rex who came back to an Arctic Ocean bathed in 24-hour-daylight, much different from what he had experienced during the dark Arctic winter conditions during leg 1. This time it is a light version of the research camp as the ongoing instability of ice floe requires flexibility and the possibility to rapidly pack up instruments. “However, we might also be able to stay on this ice floe until the end of this phase of the expedition”, says Rex in the German podcast ‘Arctic Drift – Das Audiologbuch’.

Drastic Decrease in Airborne Atmospheric Observations
One huge impact of COVID-19 the World Meteorological Organization has been concerned about was the number of weather observations going down due to the drastic decrease in flight traffic. “Overall, the decrease in the number of commercial flights has resulted in a reduction of in certain regions up to ninety percent in observations of meteorological measurements from aircraft platforms”, states the WMO in their press release from 7 May 2020. Typically, commercial airliners of 43 airlines and several thousand aircraft contribute to the WMO Aircraft Meteorological Data Relay (AMDAR) programme with daily over 800,000 high-quality automatic observations of atmospheric temperature, wind speed and direction, and increasingly also adding humidity and turbulence measurements. In June 2019 about ten per cent of the data assimilated into the global DWD ICON model were from airborne observations, reducing to only about four per cent in May 2020. In a test run by the European Centre for Medium-Range Weather Forecasts (ECMWF) all airborne observations were removed from the forecast. “The sensitivity studies have shown that in particular the short-range wind and temperature forecasts at 11 to 12 kilometers height – which is a typical aircraft cruising altitude – would degrade by up to 15%, with significant degradations at all forecast ranges up to seven days”, states ECMWF on 24 March 2020 in a news article on their website.

Why not Measure while Radiosonde Sinks
In order to keep up the good quality of forecasts despite the decrease in airborne observations since mid-March, the German Weather Service (DWD) reacted quickly. Not only did they increase the number of weather balloon launches over Germany since early April, they used the humidity data from the Global Navigation Satellite System GNSS as well as radar volume data to measure radial wind and precipitation. The DWD also initiated the use of the radiosondes’ descents for obtaining additional atmospheric data. Previously, radiosondes had recorded meteorological data only during ascents. But why not measure atmospheric data also after the weather balloon has burst in the stratosphere, which usually happens around thirty kilometers elevation, and sondes sink to the ground? The radiosonde manufacturer Vaisala had already equipped their radiosonde model 41 with the ability to measure data while the sonde drops down. “Polarstern was the first location where we tested the data uptake during the radiosondes’ descent starting in September 2019”, says Alexander Cress, a senior research scientist in the data assimilation section of the German Weather Service‘s Meteorological Analysis and Numerical Modeling department. Cress was the one who initiated and supported from Germany this very first test run in the central Arctic. “It included a little bit of puzzling out how to install the new software to make sure the data is recorded in both the DWD and the Polarstern system and then being fed into the Global Telecommunication System (GTS) of the WMO where it is available for the national weather centers to use the data to initiate their forecasts”. Since May 2020, data from all radiosondes’ dives in Germany are recorded and used in the DWD’s global and regional forecast models. “MeteoSwiss, the UK Met Office and other European countries have also prompted to follow us and started to measure atmospheric data during their sondes’ descents”, says Cress.

Modest Impact at the Poles?
Recording the radiosondes’ drop down over the small research town Ny-Alesund on Svalbard would require both additional personnel to adjust the software and consultation with the Norwegian Meteorological Institute through which the atmospheric data is transmitted to the GTS. “But since there aren’t a lot of commercial overflights in the Arctic, we wouldn’t expect a lot of atmospheric data loss across the North Pole”, says Detlev Majewski. Indeed, weather observations have not been strongly impacted so far in the Arctic, confirms Marion Maturilli, head of the Meteorological Observatory of the AWIPEV Research Base in Ny-Alesund, Svalbard. “For research campaigns in February and March and during the YOPP Targeted Observing Period around and after Easter, we launched weather balloons every six hours. Otherwise, we did one radiosonde per day.“

However, the crisis already had impact on a research campaign using the German Polar 5 and 6 aircraft within the YOPP-endorsed (AC)³ project. “Our flight campaign to support MOSAiC was supposed to launch in March/April 2020, aligned to the YOPP Targeted Observations. But because during this time there was no way to travel to Svalbard from where we would have departed, the entire activity had to be cancelled”, says Manfred Wendisch who is a meteorology professor at the University of Leipzig and leads (AC)³. The second MOSAiC aircraft campaign is, however, scheduled to take place as planned in August/September 2020. Other atmospheric science field work in Greenland, the Canadian Arctic or in Alaska has been either cancelled or postponed. Remote stations such as Summit in Greenland might eventually run out of helium and therefore need to reduce their radiosounding frequency.

At the other side of the world, at the German Antarctic research station Neumayer, the impact of COVID-19 so far seems to be modest. Holger Schmithüsen who leads the meteorological observatory at Neumayer does not expect the pandemic to impact routine atmospheric measurements from the station. However, again, research activities for the next Antarctic summer season have already been cancelled or postponed. “The exchange of the overwintering team will take place during austral summer as scheduled. Routine meteorological measurements and maintenance of instruments will thus be secured also for the coming year at the Neumayer station”, says Schmithüsen.

“The COVID-19 pandemic has caused all National Antarctic Programmes to reconsider how they can support the coming field season, with a united focus on keeping Antarctica free from the virus”, is stated by Antarctica News Zealand which is New Zealand‘s government agency responsible for carrying out activities in Antarctica. Therefore, plans are to only support essential operational activities and planned maintenance in the austral summer season 2020/2021.

“Regarding ships‘ operations, to my knowledge all levels of weather forecast communication for merchant ships have been reported as ‘regular’”, says Thomas Viguier, a former Safety and Security Officer of Merchant Marine now working as a researcher at the Icelandic Arctic Cooperation Network in Akureyri, Iceland. “However, maritime traffic reduced drastically mainly due to the impossibility to call at ports, the economic situation and border control.” The cancellation of shipboard operations also affects this year’s research cruises that were already underway but are now cancelled. “Even for next year, all cruises in the Indian Ocean sector of Southern Ocean were cancelled already so opportunities to deploy buoys or moorings have been changing rapidly, and shipboard support measurements are drastically reduced”, reports Lynne Talley, professor for oceanography at the Scripps Institution of Oceanography, University of California, San Diego.

Impact to Research Operations Strongly Expected
Resulting from a recent, non-representative survey carried out by the YOPP International Coordination Office, there is concern among the polar prediction community regarding a potential gap in the collection of polar data and the subsequent deterioration of observational quality that might reduce the amount of data assimilated into numerical models which eventually could lead to a loss of forecast accuracy. “Several field campaigns have been cancelled or changed which will affect the availability for observational data to compare with numerical models”, says meteorology professor at Stockholm University Gunilla Svensson who leads the YOPPSiteMIP effort to evaluate model output with polar observations.

If travel restrictions continue and only limited staff is allowed to stay at the polar research stations, maintenance and service requirements for instruments will not be kept at a high level. “The installation of a new tall tower automatic weather station (AWS) at Byrd Station is now delayed, as is much of the field work to maintain the AWS network”, says Matthew Lazzara, research meteorologist and principal investigator of the United States’ Antarctic Automatic Weather Station Program that is maintained by the Antarctic Meteorological Research Center at the University of Wisconsin- Madison.

In the survey that was open from 6 to 20 July 2020, the majority of the 47 respondents indicated that the pandemic so far has impacted their daily operations, life and/or their research in the Arctic or Antarctic very much (37%) or at least to some extent (52%). So far, the working situation of more than 40% of the survey participants has not been impacted by COVID-19 while more than 30% (50% of female respondents) indicated that they would now have less time to carry out research due to additional duties (including childcare, household etc.). Only 13% of the survey participants expect the national lock downs to impact the quality of Arctic or Antarctic forecast while there exists a huge uncertainty (more than 45%). More than 30% do not consider the pandemic to have any impact to forecast quality. However, operations are expected to being impacted “very much” (more than 40%) or at least “a little bit” (more than 20%) during the next field seasons in the Arctic and Antarctic.

Much of the personal communication and exchange is now being compensated by video conferencing. “In some cases, this new way of networking might even be an advantage, with new collaborations between far-apart colleagues being established and online conferences being much more accessible to everyone”, says Helge Goessling, climate scientist at the Alfred Wegener Institute. “The reduced carbon footprint of our research community is also a positive aspect. Some aspects, however, will be difficult to compensate if strong constraints due to the pandemic remain. For example, the establishment of research networks by young scientists and intense workshops where people stick their heads together for a few days to advance the science. This can hardly be done online.”

Early-career scientists might indeed be seriously affected from the pandemic as field work for PhD projects is getting delayed and networking opportunities shrink to a minimum. Irlanda Mora Rosales did her Master in Antarctic Sciences at the Chilean University of Magallanes: “Doing science in South America is very difficult as education is not a priority. We already have social, economics, ethics, and now also ‘pandemic’ problems.”

Wind Speed Measured from Space
It can be considered pure luck that the Earth observation satellite ADM-Aeolus has been launched by the European Satellite Agency at the right time to compensate for the loss of wind data through the drop in airborne atmospheric observations during the COVID-19 pandemic. Every three seconds, a laser beam is sent from the satellite through the atmosphere where it is reflected by aerosols and other particles. The movements of the particles causes a frequency shift between the laser beam and its reflection, the so-called Doppler effect, which can be translated into a wind speed. Since May 2020, the German Weather Service has added this data to initiate their global forecast model ICON. Considering the positive experience which the national weather services are now gaining from using the laser measurements from space, it has a drop of bitterness when Alex Cress and Detlev Majewski mention that the laser onboard ADM-Aeolus is still a scientific mission: “After the laser will be down in probably one and a half years, it will take another four to five years to build another laser. But then it will be for operational use of wind data”.

The new data sources that have been added during spring and summer this year to initiate DWD’s global and regional forecast models will be pursued, even after airborne measurements might be back to a pre-pandemic state. “In a way, COVID-19 has forced us to speed up enhancements which were planned anyway to improve the quality of forecasts”, says Majewski. “With the additional data, we do not see any deterioration of the forecasts’ quality. However, as I mentioned – in some years, the weather forecast is more easy. This year is difficult anyway.”