Switch from precipitation to ocean data

[DamienIrving]

A number of improvements could be made by switching from a data analysis problem focused on precipitation data to once focused on an ocean variable:

1. It would allow the topic of regridding (i.e. from curvilinear to rectilinear) to be explored (see Add content on regridding? #10)
2. It would allow the topic of large arrays to be covered (see Add content on dealing with large arrays? #8)
3. It would be a more appropriate use of the cmocean library

(The vectorisation lesson could be a basin file rather than an sftlf file.)