Ever since the first modern weather forecasting models were developed in the 1960s, the forecasting problem has been tackled using a numerical simulation. The numerical model solves physical and thermodynamic equations on a high resolution grid. This method has proven has been very succesful for predicting the weather up to two weeks ahead. The numerical weather prediction approach is being used for subseasonal-to-seasonal predictions. However, as the model predicts further ahead, the forecast skill is quickly lost due to the error that grows continuosly.
The growth in climate data and artificial intelligence has paved the way to solve this problem. Artificial intelligence allows us to systematically search for and discover connections in our climate system that can boost predictive power in S2S timescale. By learning from over more than 70 years of climate data and predicting from observations directly, we can circumvent the issues from numerical weather predictions.
We believe that a skillful forecast of the weather by itself is not enough. Trust in a forecast is as essential. Especially AI techniques can be 'right for the wrong reasons'. Over the last seven years of academic research, we have gained a lot of expertise. Our expert knowledge of the climate system, in combination with causal discovery and explainable AI techniques, allows us to interpret the sources of predictability. With S2S-AI forecasts, you will get forecasts that are reliable not only accurate, but also physically plausible and verified by experts.
Conventionally, weather forecasts have always been a one-way street. Due to the computational power needed to create numerical forecasts, forecasts have been created by the national meteorological institutes. As is their public task, they generate weather forecasts for society at large. Weather forecasts are therefore never created with a question from the end-user in mind.
With S2S-AI, we want to turn this into a two-way street. Why not create a forecast that starts with your interests? Because S2S-AI creates a completely independent weather forecast, we can flexibly tune our forecast towards your needs. We can focus on your area of interest, the variable that matters most to your business and the time of year that is most important to you.
In the end, we are not only interested in weather, but also its impact. With S2S-AI, we can take weather forecasting that takes that step. Due to our skillful predictions on the long range, we can directly forecast the phenomena that truly matters. In one of our latest researches, we directly forecast crop yields with astounding performance.
Paper waiting for publication.