The FAME flagship developing a toolkit for mathematical methods for the needs of society
Picture: ShutterstockThe mathematical methods used in different sectors are largely similar. Instead of developing computational methods in each field only for the needs of the sector, the flagship project is doing it together with researchers from different fields.
”The mathematical methods used in other fields can be useful for the Finnish Meteorological Institute, for example, when we develop weather forecasting and its methods. The methods we use can also be valuable for other sectors. For this reason, it is important to have discussions with experts from different fields," says Research Professor Anders Lindfors.
In addition to weather forecasts, the Finnish Meteorological Institute also applies computational methods in other areas, such as when measuring greenhouse gases using satellites. Satellites help to measure air pollution and greenhouse gas emissions from space, among other things. This can be used to monitor e.g. the progress of the objectives of the Paris Agreement on Climate Change. The European CO2M satellite, which will be launched in 2026, will enable more accurate measurement of this data.
"Mathematics is not separate from the rest of society, but is needed in all areas of life, especially in science and technology. The FAME flagship is an excellent example of cutting-edge research in mathematics, which has applications in different fields, from medicine to space," says Research Professor Marko Laine.
The aim is to study concrete everyday topics
FAME flagship project, which was launched this year, is a continuation of long-term research with Finnish inversion mathematicians. The difference with previously done work is that the FAME flagship carries out research on topics that concern everyday life.
One of the objectives of the flagship is to carry out research that will be of concrete benefit to both partners and society. In this context, it is essential to understand what kind of solutions companies need in relation to the flagship competence area.
One of the objectives of the flagship is to carry out research that will be of concrete benefit to both partners and society.
The research in the FAME Flagship project is based on inversion problems and modelling. Inversion modelling refers to the indirect identification of unknown variables in the model.
An example of an inversion problem is the satellite remote sensing of greenhouse gases.
"The satellite instrument detects the sun's light spectrum which has reflected from the ground and passed through the atmosphere. When we know how a certain amount of atmospheric gases, such as carbon dioxide, affect the course of sunlight, the observed spectrum can inversely determine the amount of different atmospheric constituents produced by this observation. This and many other typical solutions to inversion problems are ambiguous and very sensitive to minor changes in the state of the system", explains Johanna Tamminen, Head of Unit.
"It is rewarding to carry out inversion research, especially if a solution works for our problem and, in addition, someone else can utilise it in a completely different problem, possibly in another field", says Tamminen.
Studying the reliability of AI-generated weather forecasts
Research carried out in the flagship project not only seeks answers to individual problems, but also aims to broadly understand the uncertainties related to phenomena. Knowing the uncertainties helps to produce the most realistic understanding of the reliability of the models and the forecasts based on them. This is very important in many areas of the Finnish Meteorological Institute's operations, including weather forecasting.
Research carried out in the flagship project not only seeks answers to individual problems, but also aims to broadly understand the uncertainties related to phenomena.
Machine learning and artificial intelligence are new factors that affect many sectors. Traditional weather forecasts are based on modelled physical data on the atmosphere. In addition to the models, observations produced by e.g. satellites are needed. The latest trend is data-based neural network-based forecasting.
“We investigated whether we could use observations alone to predict the weather and we are examining how reliable data-based artificial intelligence is in weather forecasting. This also requires an understanding of mathematical methods," Laine explains.
More than 100 researchers are involved in the FAME flagship (Flagship of Advanced Mathematics for Sensing, Imaging and Modelling) project. The flagship is led by the University of Eastern Finland. In addition to the Finnish Meteorological Institute, participants include the University of Helsinki, LUT University, University of Jyväskylä, Aalto University, University of Oulu and University of Tampere.
Throughout the spring, we will publish articles on our website, focusing on the flagship projects in which the Finnish Meteorological Institute participates. Previously published articles:
Text Juliana Hulkkonen