Add exclusion for marine zones with high traffic

[irm-codebase]

Is your feature request related to a problem? Please describe.
Our potentials for offshore area are likely going to be a bit optimistic as we do not have a method to exclude ports with and areas with high ship traffic.

Describe the solution you'd like
This could be done using this dataset: https://datacatalog.worldbank.org/search/dataset/0037580/global-shipping-traffic-density

It details AIS positions between Jan 2015 and Feb 2021 (~6 years)
PyPSA-Eur implements this exclusion with the following heuristic:

$$Yearlysightings * DaysInYear * TotalYears = 400 * 8760 * 6 = 21,024,000$$
It's a bit of an ugly number, but it should be fully compatible with the module's methods.

An important detail is that this dataset should only be used if specified, as the unzipped file is 9 GB in total.

Example

I've been playing around with adding this, and it seems possible with relatively minor changes. One difference is that PyPSA uses nearest resampling, which is fast but not as good (I think...) as the average resampling method we are using in the module.

[irm-codebase]

Here are some results I got after playing around with a local implementation.
I'd say they are mostly equivalent, although nearest leads to overly pessimistic results (higher drops).
Signed difference looks like resampling noise, mostly.

Counts and fractions of valid maritime area:
  nearest excluded     24,836,038 cells  21.877278%
  average excluded     20,706,773 cells  17.964159%
  mask disagreement     4,318,521 cells   4.088983%
  nearest only          4,223,893 cells   4.001051%
  average only             94,628 cells   0.087932%
  both excluded        20,612,145 cells  17.876227%
  neither excluded     91,787,487 cells  78.034790%