Add headland inversion example

[cpjordan]

Add an inversion example where the velocity is used in the loss formulation, with different methods for constraining the Manning's field e.g. constant, region-based, independent points scheme

[cpjordan]

@stephankramer @thangel

There are three aspects of this pull request which is intended to provide another example for model calibration using the adjoint-based approach:

1. Calibrating with velocity magnitude instead of elevation
2. Showing how to calibrate with a constant value across the domain, regions, independent point scheme in addition to the default of allowing the calibration parameter (e.g. Manning) to vary at every node
3. Fixing the remaining hanging problems when running in parallel and adding a callback instead of using update_forcings or export_func to embed the loss function

Feedback would be appreciated as to whether this is valuable and anything to change. My thoughts:

• It would be nice if we didn't need a separate script for the inversion tools for velocity, but I'm wary of trying to cover all scenarios and overcomplicating things in one script
• The exporting is not very nice when we want to export for the region-based or independent points scheme. Exporting could perhaps be separate from the Inversion Manager class?

Please tag anyone else who this might concern/interest!