What is data assimilation?

Data assimilation is the combining of different sources of information to estimate at best the state of a system. These sources generally are observations and a numerical model.

Why not simply use observations?

The main reason is that observations are sparse and may not provide all the information needed. So we may need to interpolate the information from observations to unobserved regions or quantities - a numerical model naturally does that using our scientific understanding of the system, rather than purely by statistical interpolation.

A good example of data assimilation is the combining together of recent weather observations and a weather forecast to obtain as complete as possible a picture of the state of the atmosphere now (the so-called analysis). It is starting from this analysis that the next forecast can be made.

Our research

In DARC we work on data assimilation theory and applications in all areas of the geosciences, including meteorology, atmospheric chemistry, oceanography, land surface physics, coastal sediment transport, and space.

Research and development in DARC is spread across the Department of Meteorology and the Department of Mathematics in the School of Mathematical and Physical Sciences.


Background error covariance matrix for a tropospheric temperature profile

Page navigation