Here is some alternative speculation as to how a concentration of water in excess of 40ppm (ie in excess of water concentration found in main tanks) not derived from the main tanks might end up in each fuel manifold and thereby be a contributory factor to the suspected restriction in each feed line at late stage in flight:
1. Common cause - water in the centre tank after back to back long cold soaked sectors in humid conditions.
2. Otherwise normal operation - on return sector, any concentration of water in centre tank fuel transferred in first 2.5 hours by OJ pumps to each fuel manifold would not restrict the fuel lines (and would be consumed in the engines before it could become a contributory factor) due to the relative warmness of the centre tank fuel as compared to, say, the much faster plummeting temperature of the main tank fuel (which remains in the main tanks until about the time the OJ pumps are turned off) and the Critical Icing Temperature.
3. Common event - EITHER a concentration of water in remaining centre tank fuel at time switched off OJ pumps left a concentration of water beyond the OJ/jettison check valves in part of each fuel manifold connected to, but not subject to the cold slow flow from, the main tank feed, OR an event (relative low pressure in that part of each fuel manifold as compared to centre tank sufficient to overcome OJ/jettison check valves for a few seconds or more) after the OJ pumps were originally switched off leads to a concentration of water left behind in the otherwise empty centre tank flowing through the OJ/jettison check valves into the much colder mass of main tank fuel in each fuel manifold.
It is probable that water and/or ice may accumulate under certain normal certified operating conditions in the centre tank. However, given the circumstances of flights similar to BA038, arguably the bare results of general sump testing of the centre tank may be unrepresentative of the probability of this common cause becoming one of the contributory factors to this accident. For example, to what extent has sump testing been done in warm hangar or only several hours after uplift of warm fuel, in each case before further flight after the return from long back to back cold-soaked sectors in humid conditions? How much free water might be found in an empty centre tank but not come out of centre tank sump due to where it is located behind a baffle or at an alternate low point in centre tank (or it being frozen at the time, it not having been in a warm hangar)?