Sea ice and snow – the key elements

Dr Marcel Nicolaus, sea ice physicist at the Alfred Wegener Institute and deputy head of the Sea Ice Physics Section.

Sea ice



Autonomous Systems


Sea ice and its snow cover are key elements of the global climate system. Their properties determine the majority of the exchange processes between the atmosphere and the ocean. The observed sea ice changes affect the climate and ecosystem, extending to our latitudes. The decline of Arctic sea ice in area and volume is one of the most striking indicators of climate change. In contrast, Antarctic sea ice has not yet experienced a similarly dramatic decline, even though it can be assumed that its condition is also changing.

We are studying these key regions in both hemispheres, applying a variety of methods on and under the sea ice, by remote sensing and by computer simulations. We find that snow on the sea ice is playing an increasingly important role. It has particularly characteristic physical properties, such as its high reflection of sunlight (albedo) and its high insulating effect. At the same time, the snow layer changes over time and is therefore difficult to describe in climate models.

For the better understanding of sea ice and snow properties, we are constantly developing research methods. By now, we obtain much of our data from a distance in real time using autonomous systems (buoys) installed on the sea ice and drifting with it. Likewise, we have further developed and established the non-destructive survey of sea ice from below using cabled diving robots (Remotely Operated Vehicle, ROV).

In future, we will focus more on studying the physical snow and ice properties in interaction with the ocean and atmosphere, but also study linkages to the ecosystem. We will continue expanding our measurements – from point measurements, to time series and regional data sets – in order to study seasonal and regional differences and changes. The contrast between the Arctic and Antarctic is a particular challenge. Here we need to better understand the dominant features and processes and then use models to better describe them. After all, such improved models are the only way to get more precise and reliable predictions.