GOSI10: Review viscosity choices

[atb299]

Fix issue with eORCA1 where the 3D viscosity file uses a constant value of 20,000 m^2/s.

Revisit Alex and Dave's work which showed positive impact of increasing viscosity at eORCA025.

[atb299]

This file summarises my analysis of viscosity at eORCA1:
Lateral_viscosity_in_GOSI9p8_eORCA1.pdf

Suite u-cu895 used the eORCA1 'default' spatially varying viscosity that George had been using for the OSMOSIS development, but I merged it to create a 3D field that include the equatorial refinement. The validation note comparing to GOSI9p8.0 is here:
https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-cu895_nemo_vs_u-cn131_nemo/assess.html

Talking to George (early in 2023) he wasn't aware how close to the lower bounds for stability the default was. Without any concrete argument to justify it, he felt that something a bit higher would be a better choice, hence my proposal for something like the blue curve in Figure 5 of the document.