Data journey

The Zooniverse platform

Digitised ship log books are loaded onto the Zooniverse online citizen science platform, ready for transcription by Old Weather citizen scientists.

The records are organised on the website in a small number of projects, for example Royal Navy records for 1914–22. In each project, the records are then organised by ship, and for each ship there might be a number of different voyages over a period of months or years.

Transcribing the log books

The Old Weather project’s citizen scientists select a ship and voyage to work on from those currently available, and then begin transcribing via the Zooniverse online platform. By the end of the first phase of transcribing Royal Navy log books in 2012, Old Weather volunteers had recovered around 1.6 million weather observations.

Watch the video below from the NOAA (National Oceanic and Atmospheric Administration, USA) to find out more.

The volunteers are free to choose which records to work on. Some people base their selection on having a personal interest in the ship or geographic region. Other reasons for selection include more practical concerns such as the ease of transcribing the sailor’s handwriting, or the need to complete a particular section of the project – for example, ensuring a series of records is fully transcribed:

“This one really I just kind of landed on accidentally, I think I quite liked the picture of it to be honest, which was not a scientific reason at all. And the writing didn’t look too bad, and so I kind of got on it, and now I feel sort of responsible really, particularly because of this very, very long one, and nobody much seems terribly interested. I mean it’s been done by one person all the way through. He’s done the full lot.

“I’m the second person going pretty much all the way through, and then we’re going to have to try and persuade somebody else to do the third, or various people to come on and do, you know, and do a bit. So yeah, so I feel kind of responsible in a sense.”

Once the volunteer has selected a record, they view the digitised logs on the website and transcribe key information into a simple online form. The type of information that they transcribe includes: date, time, location and the various weather observations such as temperature and humidity. Below is an example of a log book page:

Transcription challenges

Transcription can be a difficult task. The main challenge is deciphering the handwriting in the log books. There is significant variation between sailors – some having very clear handwriting and others an illegible scrawl.

Other challenges are decoding some of the words used in the notes section, and identifying place names accurately. Volunteers often turn to the Old Weather forum for help with these issues, posting snippets of log books so other volunteers can take a look and offer assistance.

Editing historical information

Written notes made by sailors about the weather and aspects of life on-board the ship are also optionally transcribed by some volunteers. Information captured in this way might include notes about on-board cargo, supplies, deaths of crew members, and even personal notes and doodles.

Once a log book is completed, these extra notes are edited by volunteers to remove duplication and improve consistency. These transcribed notes are useful for climate scientists to understand more about the context of the observations they are using, but they are also used by other researchers. Much of the historical data is uploaded to naval-history.net so it can be accessed by professional and amateur historians.

Whilst the editing work is often more interesting, there is a certain level of responsibility and importance attached to transcribing. Transcribing activity was viewed as more important than editing in terms of volunteers supporting the needs of climate science.

“I suppose at the moment I probably find the editing more satisfying. But I feel the transcribing is more the point of the whole thing. As I said, when I wasn’t transcribing I did feel a bit guilty. So at the moment I’m doing both, and I tend to alternate, you know I’ll do a day or two of transcribing and then a day or two of editing, and so on. And that feels like a reasonable balance.”

Quality control

To avoid transcription errors and the potential for manipulation of the data, each entry is separately transcribed by three volunteers. After transcription has finished the data is checked for accuracy and consistency, before being sent to the Met Office.

Sharing the data

The transcribed data is then standardised into a format called IMMA (International Maritime Meteorological Archive) and integrated into the International Comprehensive Ocean Data Set (ICOADS). From ICOADS the Old Weather data can be accessed by climate scientists around the world.

As climate scientist Philip Brohan explains, these Old Weather data are used in at least five major climate reconstructions including 20CR Version 3 (NOAA/University of Colorado), Era-Clim (ECMWF), SODA (Texas A&M), HURDAT (NOAA) and HadISST2 (Met Office).