First fruits: Weather records from HMS Acacia
The OldWeather project has only been live for a couple of weeks, but we’ve already finished digitising the logs of our first ship – congratulations to captain oneoftheguys and the crew of HMS Acacia. So I thought I’d demonstrate that the science team are also working hard on the project, and show some preliminary results. The figure shows the noon air temperatures experienced by HMS Acacia and extracted from her logs.
There’s only one correct temperature for each observation, of course; but, naval handwriting being what it is, the 5 people digitising each page often submit different readings. I’ve shown the different readings in the figure – each reading is a red circle, the more popular readings are bigger circles. The black line links the most popular reading for each observations.
What can we learn from this figure? Well, first of all, OldWeather works: We have produced a perfectly credible air temperature time-series. Those are sensible values, you can see the combined effect of it being colder in the winter (middle of the graph) and warmer in the summer (both ends), and of the ship’s trips to Svalbard (cold excursion in August 1915) and move to Gibraltar (warming in December 1915). We can also see that, almost all of the time, the most popular reading is is the right value, though there are a couple of interesting problems:
- There is tendency for people to read the numeral 7 as 4: The temperatures in spring and summer 1916, with the ship in the Mediterranean, are in the 70s, quite a few people are reading that as the 40s.
- Sometimes the log itself is wrong: on the 19th June, 1916, the Acacia’s log-keeping officer wrote ‘July’ in the month field, which we then went on to enter, correctly, as July. So the data for this entry is out by a month, producing the strange effect you can see in the figure.
There are a few obviously erroneous values in the series caused by misreading of the logs or misstatements by the writers of the logs, but we’ll be able to clean them up. These results are definitely good enough to be used in climate research.
So thanks to all involved, please keep up the good work … and watch those 4s – are you sure that one’s not a 7?