Great minds don't always think alike...

For Monica Ionita and Frank Kauker, it all comes down to September – when they’ll finally find out which of their forecasts for the summertime sea-ice extent for September was more accurate: a battle of methods that spurs on research!

Dr Frank Kauker

Forecasting with the help of a ready-made model

Since 2008, sea-ice modellers at the AWI have taken part in the international competition for forecasting the Arctic sea-ice minimum. The contest always starts in June, when the ­Arctic summer settles in and the ice cover begins melting. We use the term sea-ice minimum to describe the smallest area the sea ice shrinks down to before the winter returns in late September, bringing fresh ice with it.

For the past ten years, I’ve been using a dynamic climate model of the Arctic for my forecasts. It’s a ready-made model and uses a number of mathematical equations to calculate the physical interactions between the sea-ice and the ocean in this region. Basically, I feed the latest sea-ice thickness data from the CryoSat-2 satellite into the model and then have it estimate the future changes.

The sea-ice modellers at the AWI are ­currently the only research group in the world using the CryoSat-2 data for this purpose – and we know that the data is highly reliable; after all, our colleagues verify it every spring in the course of our aerial and ship-based campaigns. When I made my first forecast, almost ten years ago, I actually felt it was impossible to make accurate prognoses for the September minimum in June or July, since the summer weather is so decisive for the sea-ice development. And, as everyone knows, we can only predict the weather for the next few days.

Today, I know that we can boost the accuracy of the model’s outcomes by determining the initial conditions as well as possible. There’s still a degree of uncertainty, of course – the summer weather can make a significant difference. That’s why Monica and I discuss our findings before either of us submits a forecast. Granted, her approach is fundamentally different from mine – the two have practi­cally nothing in common. But her method offers me a different perspective. Combining the lessons learned from the two approaches provides us with a more complete overall picture.

Dr Monica Ionita

Statistics based on numerous small climate parameters

Unlike the sea-ice modellers, I don’t need a climate model for my forecast. Instead, I use climate parameters like the air temperature, the temperature at the water’s surface, and the heat energy stored in the ocean. I use this data, which is gathered by satellites, in statistical calculations. This kind of approach has allowed me to more precisely forecast the water levels in the central European rivers, including droughts and floods. I’m eager to see whether the same approach can provide an accurate forecast for the sea-ice development in the Arctic and Antarctic.

Given the growing amount of ship traffic, we need dependable forecasting systems for both regions. For my sea-ice forecast of the Arctic September mini­mum, I don’t focus on the Arctic as a whole; I concentrate on those regions that are most important for sea-ice formation, which I use as the basis of my statistical calculations. In the past two years, we’ve managed to produce forecasts that quite accurately predicted the final sea-ice minimum – both with my statistical method and with a model-based approach. Thus, both approaches point in the same direction, which is a good indication for us scientists that our methods and findings are reliable.

My method can help identify those regions that are essential to the sea-ice development, so that modellers can focus on them more intensively during the modelling process. Frank and I come from different climate research disciplines. He’s a sea-ice expert, while I’m a climatologist. But we’re also both physicists, which means we speak the same language when it comes to our methods and findings. When two colleagues with different backgrounds work together, it’s always a win-win situation, and helps to move research forward.