Fixing warm start and updating reference results for MCT

[ecCinziaMazzetti]

Hi,
in this PR you have:

• fixes to the warm start for kinematic, split and mct routing
• updated reference results for mct, both quick and slow tests
• fix to mct routing that prevents channels volume becoming negative when openwater and/or wateruse is in use

The following 'slow' tests were not run because reference results are no longer valid:

• test_latlon
• test_results
• test_inflow
• test_caching (I still have the old version with the bug)

For further information, see https://jira.smhi.tds.tieto.com/browse/EFCC-3975