We can only trust current Earth System Models with future
projections if they are able to reproduce past events. I
will use model realizations of glacial inception and Dangaard-
Oeschger events to illustrate what aspects of current ESMs are
trustworthy and which ones need more work. Two major areas
of future developments will emerge: marine biogeochemistry
and software engineering. The case will be made that we
need new approaches to numerical modelling to address the
challenges posed by the carbon cycle.
Markus Jochum received his PhD from MIT in 2002 for his work on
equatorial ocean dynamics. From 2005 to 2012 he joined NCAR to
contribute to the development of CESM, and for the past 6 years
he has been professor at the Niels Bohr Insitute, Copenhagen,
where he uses numerical models to test hypotheses related to
past climate events.