Efficient use of compute resources.

[seabbs]

As we have access to large amounts of computing it may make sense to run each region sequentially but use all cores. At the moment all regions are run in parallel and this is only as effective when there are more regions with cores (there are rarely more than 48 regions -> only in the global nowcast). Should probably be added as an option(whether or not use furrr or purrr in regional_rt_pipeline would be an effective way to do this

[seabbs]

Done in master