Data Rescue Citizen Science

The Hunt for Historical Data

Since the 1960s, most weather and climate records have been routinely computerised. Before that, all records were written on paper. Scattered and often lost in old cupboards, stacked on shelves, all-but forgotten in archives, there is a rich source of documents containing weather observations going back more than 250 years. These include extensive ship-based measurements and observations taken at weather stations, observatories, on ships, in schools, at churches and farms and towns. This storehouse of old data can help us understand today’s weather and climate and predict how climate might change in future. Once digitised, the data can be fed into computerised weather reanalysis systems, such as those run by ACRE partners and facilitated by data collected and digitised by ACRE, which dynamically recreate in great detail 4D global weather going back 150 years or more. Locating, imaging and digitising old weather data is a time consuming activity. One of ACRE’s core activities is to support citizen science initiatives that carry out the vital work of building a global database of historical, dense and quality-checked climate data. Citizen science is scientific work conducted generally by amateur or nonprofessional scientists. Citizen scientists are typically volunteers who donate their time to projects that have short to medium timelines, some working from home and others working in an office. Citizen scientists are crucial to ACRE as few funding sources would cover the extensive costs of paid-staff required for the 10,000’s hours involved rescuing millions of historic weather readings. Without the help of citizen scientists, much of our climate heritage would be lost.

Examples of ACRE-allied citizen science initiatives are:

Old Weather

This web-based initiative uses a form of citizen science called ‘crowd sourcing’, a format that attracts citizen scientists to contribute their work over the Internet. It specialises in recovering data from ship logbooks, data which is needed to complement historical observations from land-based readings. The internet volunteers “join the crew” of a ship that once plied the oceans and work to digitise it’s logbooks. As images are presented on-screen, the “crew members” type the data they see. In this way they have recovered 1.6 million new weather observations from 28,000+ pages of Royal Navy logbooks covering 1914-1923. Their work has considerably added to the historical depth of the International Comprehensive Ocean-Atmosphere Data Set (ICOADS) marine data repository. In the second phase of this project, volunteers are transcribing the logbooks of US Government ships that travelled in the Arctic and other destinations in the mid-19th century. To date over 85,000 pages have been transcribed adding marine weather readings to our climate history. How crowd sourcing helps climate analysis (video) Oldweather digitises whaling ships Oldweather digitises US naval ships Weather Detective The Australian project was configured like Oldweather, but concentrated on the logbooks of commercial ships sailing to and from Australia in the 1890’s to early 1900’s. To date 400,000 lines of weather data have been digitised. First outputs (video)

International Environmental Data Rescue Organization (IEDRO)

IEDRO’s mission is to locate, rescue, image and digitize historic climate data, particularly in developing countries. Run by volunteers, the organization works on site with owners of climate data, providing “in-a-box” solutions to image and digitise their historical weather records. The data owners, typically national meteorological services, are provided with computers, camera equipment and training to image their records. Data on the images is then digitised by data entry personnel, or if the images are strip charts, the digitising is done automatically by software. The data are then made available worldwide through environmental World Data Centers.

Data Rescue: Archives and Weather (DRAW)

Volunteers participate in this Canadian project by transcribing the data contained in historical weather logs held at the McGill weather observatory. Going back to 1874, the data will be made available for scientific research. The project team includes atmospheric science experts, information scienctists experts, archivists, programmers, volunteer coders and students.

Australian Weather Folios Project

For the past 20 years, The Australian Meteorological Association has a self-directed team of citizen science volunteers hosted by the Adelaide office of the Australian Bureau of Meteorology. Working with the South Australian Weather Folios and records (1879-1957) and numerous Lighthouse journals, they rescue weather records that cover Australia and the southwestern Pacific. They have created 110,000 images of the records and recovered almost ½ million data points, mainly covering the latter half of the 19th Century. The team specialises in creating and curating commercial-grade images of weather journals, including badly degraded paper records. The original journals contain so much handwritten data (possibly a billion data points), contemporary digitising events will never be able to capture it all. Thus, the images must be curated in digital form for future reference and capture. The journal images are available at the team’s website , A smaller set of 19th century Canadian images are also hosted for the DRAW Project.

Operation Weather Rescue

Operation Weather Rescue This citizen science activity is aimed at rescuing land-based station observations using volunteers. Phase 1 recovered 1.5 million hourly weather observations from three nearby sites in Scotland between 1883-1904, with more than 3500 volunteers contributing. Phase 2 rescued 1.8 million sub- daily and daily observations from more than 70 stations across the UK, Ireland and western Europe that were recorded in the UK Met Office’s Daily Weather Reports during 1900-1910. Phase 3 completed the rescue of 0.5 million observations from the UK Daily Weather Reports for 1861-1875. The project will be relaunched in 2020 to recover more data from pre-1900, with a focus on pre-1850 temperature observations from both land and sea.