The Future: Earth System Modelling

Abstract

Earth system modelling will be the kind of modelling that will be performed more and more. After some philosophical remarks, an overview will be given of some science questions that can only be properly treated by earth system models. The focus of these science questions will be on emissions. Concerning meteorology, the problems still existing with the atmospheric boundary layer will be presented. Data-assimilation will be addressed as an application issue, which can only be carried out more objectively with earth system models. The paper ends with ten rules for correct modelling.

Questions and Answers

Questioner: Valerie Garcia

Question: Regarding your remark always combining observations with modelling, are there not some applications where care should be taken? For example “correcting” the model in areas where you have data causing inconsistent bias, which affects the ability to discern a signal in health studies.

Answer: You are right, you should always be careful which observations you want to include, or not in your data-assimilation. My remarks was meant as: always use observations, unless..etc.

Questioner: Wouter Lefebvre

Question: How can we combine: 1) Including plants/biogenic reactions/feedbacks with 2) Do not make your model overcomplicated?

Answer: Good point. We should try to include plants and feedbacks as simply as possible, but we should include them. And always try to do that in a balanced way: not to include in one part of the model system a complicated process description, and at the other hand have rather uncertain input data to that process.

Questioner: Heinke Schluenzen

Question: If emissions are more relevant than meteorology, how are good emissions data that depend on meteorology ever achieved? Should not both be well described?

Answer: Of course, both should be well described. I meant to say that we know already a lot about meteorology, and many scientist are working in that area. We know still less about emissions, and we need more effort and people for the area of the science of emissions.

Questioner: Ivanka Stajner

Question: Emissions are essential for accurate AQ-modelling. Data assimilation for air quality is limited by availability of sufficient quality and type of observations. Another issue limiting progress in chemical data assimilation are model biases such as those due to the emissions.

Answer: I agree with you. However this should not lead to not to perform chemical data assimilation, we can still learn a lot by applying it, but with care.

Questioner: S.T.Rao

Question: Isn’t it obvious that emissions are more important than meteorology. You can control and abate emissions, but not meteorology

Answer: I cannot agree more