From a climate forecast to a potential user product – an interactive process

Decadal climate forecasts have only been a subject of research for a bit over a decade, thus the concrete use of these forecasts for real applications is an entirely new field. Research shows that forecast skill for some variables and regions could be good enough for some practical applications and the question arises: How do we bring the possibilities of the forecast system and the interests of potential users on one page? The answer to this question is to a large extent the use of an interactive process, meaning a continuous exchange between researchers and the potential users of decadal climate forecasts.

The forecast of North Atlantic surface temperatures is an example of a climate forecasts that shows potential from a scientific perspective. The MiKlip system and also other decadal prediction systems have repeatedly shown that there is a relatively high forecast skill for this region. Additionally, the processes behind this higher skill is relatively well understood. It is however fair to assume that there are few users directly interested in the surface temperatures of a small region of the North Atlantic. A statement on how the predictability of the North Atlantic influences the temperatures of the coming years in Europe, is likely to have more applicability.

In order to reconcile the expectations of the users and the feasibility given by the science, an interactive process, bringing users and scientists together is essential.

In how far a forecast is interesting or useful depends on the user. Potential users, e.g., from the insurance and wind sector, civil protection, from agriculture or the transport sector could have very heterogeneous demands. To establish whether a forecast is useful a dialogue with different users must take place. User interests can diverge in many ways, e.g., with respect to

  • the variable, the time-scale or the region
  • the demands on the forecast skill: a value that is sufficient for one user, giving a “good” potential for the forecast and thus potentially resulting in a decision to act, may be too low for another user.

These demands can be used by the scientist, to analyse the decadal climate forecasts in terms of their usability, of course within the frames of scientific acceptability, i.e., the choice of variables, to adjust the requirements on the skill scores. Generally, compromises have to be made, since not all specific user interest can be met.


The development of the decadal climate prediction system is a process of several interactive steps.

At the beginning of 2017 MiKlip published its first decadal climate forecasts for the public. These can be found here.

Furthermore, climate forecasts from different institutions are exchanged within  the framework of the “Multi-model decadal forecast exchange”. These forecasts should only be used for furthering scientific understanding and should not be seen as fully-fledged forecasts or to be used as a basis for decision making.

The data of the MiKlip climate prediction system are currently not available to the public. For people who are interested there is a user newsletter, user workshops are organised at a yearly basis und you can also contact the user projects if you are interested in a cooperation towards the development of user-oriented climate forecasts. If you are interested in either of these, please contact the MiKlip Office.