The data package holds the simulation, which includes a number of attributes which are really CMIP attributes (forcing index et al). It might be cleaner to put a CMIP subclass of simulation into this work package.
input_dataset is really just a "manually modified dataset", but the use as an input is a concept which belongs to whatever uses it as input, so the name is wrong.
downscaling seems to be an unnecessary specialisation, it could be hoovered up into a related_to/related by association class, unless we want to add some real downscaling attributes (method etc).
input_dataset
is really just a "manually modified dataset", but the use as an input is a concept which belongs to whatever uses it as input, so the name is wrong.downscaling
seems to be an unnecessary specialisation, it could be hoovered up into a related_to/related by association class, unless we want to add some real downscaling attributes (method etc).