What Will the Weather Be Like in Ten Years?
Meteorologists from Freie Universität participated in developing a prototype for mid-range climate forecasts.
Jun 08, 2018
These days, weather reports already provide adequate answers to the questions of whether people should trade in their winter coats for a transitional jacket the next day, whether an umbrella is a good idea, or whether commuting by bike or transit would be better. They are also highly accurate, and even more so the shorter the time in the future.
But these are not the only questions that matter. A lot of research is also being done on the aspect of how the climate will change overall over the next few centuries, where it will grow warmer, how much sea levels will rise, and in which regions the number of extreme weather events could increase.
Between short-range and long-range forecasts, however, there is currently a gaping hole: Medium-range forecasts – on the order of ten years – have been hardly possible at all so far. But these are exactly the kinds of predictions that are highly important to the business sector, for medicine, and to policymakers. “Just think about the energy sector: Will we need more wind and solar power facilities, or less? If we need more, where? How much energy will be generated where, and when?” says Ulrich Cubasch, a professor of meteorology at Freie Universität Berlin.
Insurers, Farmers, and Shipping Companies Benefit from Forecasts
Cubasch is currently working with a large group of colleagues to conclude the second phase of the “Mittelfristige Klimaprognosen” (MiKlip) project, which is being financed by the German Federal Ministry of Education and Research (BMBF) with almost 40 million euros. The teams of researchers are studying whether reliable predictions for the next ten years are possible.
In the first phase of the project, which concluded in 2015, they already showed that this is definitely possible with the data available today. The goal of the second phase, which is now drawing to a close, was to create a software prototype for these kinds of predictions and provide it to the German National Meteorological Service (DWD).
Two of the project’s five research areas are being coordinated at Freie Universität, where more than a dozen researchers are working on MiKlip. “Of course, we will never be able to say how hot it will be on a certain day in 2025, or whether it will be sunny or rainy that day,” Cubasch says. “But we can say how many degrees warmer it will be on average, how the areas of rainfall will shift, and we can predict the changes in the Gulf Stream.”
There is great demand for these kinds of forecasts, says Cubasch, who is also a member of the United Nations Intergovernmental Panel on Climate Change. Insurers, for example, adjust their premiums to reflect the likelihood of large storms or heavy rainfall. Farmers want to know whether they would be better off growing crops for drought resistance or hardiness to winter conditions in the future. Shipping companies would like to know when and where the Arctic will be navigable, and hospitals might need to prepare for hot periods and the circulatory problems they bring with them.
Diving Buoys Collect Data in the Oceans
To be able to come up with a reliable ten-year forecast, also known as a decadal forecast, the researchers developed a mathematical model that draws on a wealth of data. The data come from weather balloons, satellites, and countless sensors, especially those located in the ocean, where many data buoys now float.
These include subsurface buoys, which descend to measure underwater currents and temperatures, then rise to the surface at regular intervals to transmit their results to satellites. Many merchant ships also take measurements, which add to the vast pool of data available for climate researchers’ use. Now, at the end of the project, the model has been created, and so has the software prototype.
“We ran a series of calculations on past years as a test, for example the period from 1961 to 2013 – and the results were a very good match for reality,” Cubasch says. “We learned a lot in the process, especially how important the oceans are when it comes to accurate forecasting.” Only in about the past ten years have sufficient data and processing capacity become available in order to take the oceans into account. Since then, the climate expert says, the development of forecast models has leapt forward.
German National Meteorological Service Plans to Use the Software for Its Work
Plans now call for the German National Meteorological Service (DWD) to use the prototype newly developed in the MiKlip project for medium-range forecasts. The research findings will be published, but because of the immense computing power involved, the DWD is likely to be the only entity that uses the programs, Cubasch explains. “An insurer or a farm won’t be willing or able to afford to do this; they’ll prefer to use the forecasts produced by the weather service.”
MiKlip is thus in the home stretch. An initial test forecast for the period from 2018 to 2027 is already offered at www.fona-miklip.de. The final conference will be held at the Harnack House of the Max Planck Society in Berlin’s Dahlem district in late May. By then, the DWD may already be using the software in a pilot phase, Cubasch says.
The crucial questions now are whether there is truly interest in the decadal forecast in the business sector, and whether the interest is sufficient for the DWD to earn money with these forecasts. Cubasch thinks it is unlikely that a single company could install the necessary computing power and the software itself; this is also why the DWD was involved in the research project from the start, since it is the only place that also has climate specialists who can engage in this kind of modeling reliably in the first place. “It would really be regrettable if the product of so much work were to be shelved somewhere,” says Cubasch. Still, the scientific gain is already immense and indispensable.
Ulrich Cubasch, Professor of Meteorology, Freie Universität Berlin, Department of Earth Sciences, Institute of Meteorology, Email: firstname.lastname@example.org