"Dear Professor Jones"

Dear Professor Jones I have just read with great interest the reporting of your work on sea water temperatures. However, am a little confused by what has been reported. Firstly the canvas bucket v sea water intake differences. We were always taught by the met office that only surface temperatures should be taken using the bucket just dipped into the sea. The design of the bucket greatly reduced colling through evaporation as it was double skinned. The sea water intake is by definition below the waterline. This must remain below the waterline at all stages of loading. So if you take a ship with a lightship draught of 2m its sea water intake will be just below the surface. If you now load that ship, the intake is now some 10m below the surface. This must throw up an enormous variation in readings.

Secondly, the UK register after 1945 greatly increased to its heyday in the 1970s before declining again. During this time the number of reports from UK ships must have grown vastly, dwarfing the number of US sources. This must then through up some variation in temperature reading methods post say the mid 50s.

Thirdly, as a former navigator I remember how hit and miss the whole method was. Chucking this bucket over the side, pulling it back in (trying not to hit the side of the ship and wake the Captain), peering at the thermometer with a weak torch in the middle of a gale, spilling most of the water, making up readings, copying what others had written before (as it was dark, blowing a hooley and you couldn't be bothered to go onto the bridge wing). And I was a pretty conscientious officer.

The met reports from the observing ships is pretty dubious at times. I remember crossing the Atlantic one time and could not understand why there was this isolated fog bank following us across as indicated on the weather fax. Then checked the visibility codes for our reports and found that we had all just been copying the wrong number down. Would be interested to hear some thoughts about the above issues though.

(via)