Small remaining issues with TWBR

[luyssaert]

See ticket #712 for the figure and more discussion on the topic. It looks like two small issues may remain for TWBR:

the "max" average twbr is -2E-9 mm/d (magenta point in the USA), and ca -2E-10 mm/d over Tibetan Plateau. This is acceptable within a CMIP6 context. Several of the deviating pixels are in the mountains or regions with a substantial snow season, there might be a (small) remaining snow problem.

For the times series of twbr (units have been converted to kg/m2/year). the numbers are small, there is no temporal trend but a clear bias (consistent loss of water). Note that FG1 cycles over the same climate drivers, hence, this is most likely the cause of the cyclic pattern.