In addition to the task of taking hourly surface observations, day and night, which have to be coded into 5 figure groups of numbers which have to be obtained from a comprehensive book, and passed by radio to the shore based Central Meteorological Station, the Ocean Weather ships based at Greenock, when on station, had to measure the upper atmosphere.
This was done 4 times a day by using a radio-sonde. The sonde is carried aloft by a balloon made of rubber and inflated by hydrogen. The sonde itself is a small cylinder about 12" high and 8" in diameter. Inside the sonde are the elements one needs to measure temperature, another to measure humidity and the third, an aneroid element to measure pressure.
Each element in turn is contacted by a rotor activated by a windmill on the outside of the sonde and a battery inside the sonde causes the element to emit a signal which is picked up by an officer aboard the ship.
In preparing for a launch, the sonde is calibrated in the office aboard the ship, meanwhile an assistant inflates the balloon in a large balloon shelter towards the stern. When the sonde is calibrated so that the elements are sending the same as the conditions on the sea surface, the officer carries it out to the balloon shelter and ties it to the end of 90ft. of string the other end of which is attached to a radar reflector which is hanging some 20ft. below the balloon.
The trick then is to get the whole apparatus into the air without hitting anything, either balloon shelter, bulwarks or ocean waves.
The operators then go below into their met office and plot on a graph the signals emitted by the sonde as the balloon rises.
At first the signals are quite frequent as the balloon rises at about a 1,000 ft/min.. As it ascends into rarer air the balloon expands and the rate of ascent slows. Eventually the balloon bursts and and the whole apparatus drops into the sea. The signals cease after about an hour and meanwhile the ship's radar has been plotting the range and bearing of the balloon and its radar reflector, and from these readings the wind strength and direction at heights obtained from the readings of the sonde can be calculated.