The impact of Covid-19 on weather observations used by weather models
Numerical weather prediction (NWP) systems depend critically upon observations of the meteorological state of the atmosphere. Weather observations come from commercial aircraft, satellites, ground observation networks, weather balloons, radar, ocean buoys and ships. But vertical profile observations in the troposphere by aircraft and weather balloon radiosonde ascents are considered particularly valuable.
Impact of reduction of commercial aircraft due to Covid-19 restrictions.
Commercial aircraft constitute an essential part of the global observation system. Due to the grounding of flights during the first wave of the Covid-19 pandemic in spring, at the end of April, planes were reporting 75% fewer observations than normal. Air traffic is currently reduced by closer to 50%, so weather model predictions could still be experiencing a reduction in accuracy, although probably not as extreme as this spring when lockdowns on movement were stricter.
Aircraft typically provide observations by carrying instruments that record information about air temperature, relative humidity, air pressure and wind along their flight path. Aircraft observation coverage is most dense over the contiguous United States and Europe, with secondary maxima in East Asia and Australia/New Zealand. Generally, several thousand aircraft record over 800,000 weather observations on a daily basis. Unfortunately, due to the current decrease in flights, only around 1,000 aircraft are collecting weather data daily. The reduced weather data decreases forecast accuracy.
Reduction in other observations as a result of the pandemic
The COVID-19 situation has also created reductions in other data sources used for weather forecasts. Although most ground-based weather stations or moored buoys are automatic, maintaining them requires people that can travel on-site regularly and working calibration labs. There are also drifting buoys which have a shorter lifespan than moored ones, with an average lifespan of 18 months, but these still require people and ships to deploy them regularly. Some older generation ship-borne measuring systems require shipmates to operate them. If less of these ships are plying, less data is produced in the Atlantic and the data sparse Arctic.
Shortfall in some weather observations by the pandemic made up by other sources
The shortfall of aircraft data has somewhat been made up somewhat by satellite data coverage, which accounts for up to 75% of all observations used by NWP, and an increase in radiosonde and ground observations. Satellite data provide a lot of information on temperature and humidity fields, but less on wind fields, with information on wind heavily reliant on aircraft observations. Though wind information has recently expanded with Aeolus satellite data which partly fills the gap due to fewer aircraft reports.
Satellite data cannot yet be regarded as a total solution for better weather forecasting. Whilst satellite observations tend to have a very good resolution in the horizontal plane, their vertical resolution is relatively coarse, thus rendering them less detailed relative to comparable in-situ measurements from radiosondes and aircraft. So, the reduced number aircraft observations around the world will still have some impact, despite the shortfall being covered by other observations, particularly around the polar jet stream level (10–12 km altitude).
In response to the reduction mostly in aircraft but also other data due to covid-19, to compliment satellite data there has been an enhanced programme of radiosonde launches across Europe and other parts of the world. An increase in the number of radio-soundings made each day will aid in the accuracy of global NWP centres, particularly as Radiosondes provide very high-quality observations of temperature, wind and humidity as they ascend from the ground to an altitude of up to 35km (115,000 ft). Although fewer in number compared to the huge quantities of data obtained from meteorological satellites, they are one of the most valuable sources of weather data.
It’s still not clear how much impact the loss of aircraft and other data due to the pandemic is having on weather model skill
However, at this time, there is no clear signal/trend the impact of lost aircraft and other data due to covid-19 restriction are having on the skill of the numerical guidance the models are producing. As weather model skill requires a rigorous and substantial scientific analysis. So, the full impact of the lack of aircraft data probably won’t be known until well after the pandemic ends.