Revisiting “nutrient trapping” in global coupled biogeochemical ocean circulation models