Deleting stanza life stage does not fix leading B, CB group

[jeroens]

When deleting a life stage, the leading B and CB retain the index of the original group. This may cause assertions when a model is reloaded. The user interface can take care of this, but perhaps it would be a good idea if the core or Ecopath become robust enough to deal with this eventuality in case groups are deleted from code (without a UI)