What on earth is remote sensing?

Or: How do satellites contribute to climate forecasting?

Seven to eight hundred kilometers above our heads orbit more than 1,000 satellites whose purpose is to monitor conditions on Earth. The Earth observation satellites provide us with enormous amounts of data on the state of the atmosphere, the land, on the sea ice and glaciers and on the sea surface. Using advanced data processing methods, the satellite observations are converted into so-called values of temperature, salinity, sea level, chlorophyll, wind, waves, ocean currents, and sea-ice cover, thickness, movement, and extent.

When calculations are to be made of future ocean conditions, sea-ice conditions, or climate, we use numerical models. These are large computer programs based on mathematical descriptions of relationships in nature. Observational data from satellites can be fed into such models to improve accuracy and through this increase the quality of the model calculations. The models can calculate forward in time, and the results can be presented in the form of forecasts for a specific area.

The forecasts will never be more reliable or accurate than the quality of the tools, which observational data is available, and the methods used to produce them. Much of the research at the Nansen Center revolves around recalculation of satellite observations into geophysical values, that can be used for data assimilation and used to improve the calculation models, so that these can describe future conditions with a high degree of accuracy. Research is also being done on how we can extract new information from satellite measurements with the help of artificial intelligence, and find new application areas for data obtained via satellites. Examples of this are classification of sea-ice age and thickness, sea level changes in coastal areas, and calculation of surface currents in the ocean.

Model improvement

The method we use to take observational data into models is called “Ensemble Kalman Filter – EnKF”, which in short involves making a series of model calculations (creating an ensemble), and based on these, finding statistical relationships between different geographical areas and different variables in the model. Thus, observational data at a specific point can be used to improve the model over a larger area.

Connected research areas


Client: Mercator Ocean International
Project owner: Nansen Center
Project leader at the Nansen Center: Laurent Bertino
Client: Mercator Ocean International
Project owner: Nansen Center
Project leader at the Nansen Center: Anton Korosov