From Models to Predictions
Models are indispensable in environmental research. They reproduce coupled systems and processes on various temporal and spatial scales as accurately as possible. They are fed with data from various environmental observations, measurements and analyses, based on which simulations and predictions can be made. But just how accurate are these predictions? Are the data sufficient? Is there not too much heterogeneity in environmental systems? What about scale gaps and uncertainty?
On these grounds we are working on a systematic modeling protocol, starting with a systematic conceptualisation of models including analyses on multi-scale models and scaling, analyses on the uncertainty, a consistent measuring design of model variables and finishing with the parameterisation of models. In this way we want to make predictions and statements of models comparable and more accurate. The models must be conceptualised and able to calculate on the regional scale, because this is the scale at which these systems are managed. This proves to be a challenge, because the regional scale is considerably larger than the scales at which process studies, which are used to formulate and parameterise models, are usually carried out.