91TV

Today, we’re making our available online via our platform. This is the result of a long project started in 2020, which included re-cataloguing, conserving and imaging all our meteorological data. The resource includes over 1,600 individual observation sets, dated between 1706 and 1915, and originating from across the globe, from the northern shores of Greenland to the southern ice of Antarctica.

At first glance you may think this collection is just a bunch of boring numbers and graphs. However, I hope that by the end of this article you’ll be convinced that – in addition to the stories my colleague Keith discovered in his scoping piece – the collection’s origin is fascinating in its own right, as it carries both an unfulfilled scientific purpose and exciting possibilities for the future.

On 21 November 1839, the Physical and Meteorological Committee of the Royal Society advised its Council – and its Librarian – that all meteorological observations in the archives ought to be extracted from their existing locations and consolidated in ‘a separate department’ to which ‘future observations & records’ should be added.

This vision of a department gathering together all meteorological observations precedes the creation of the first for Britain by 15 years. The Chair of the Royal Society Meteorological Committee at the time was Sir John Herschel FRS, the Victorian ideal of a ‘man of science’, who articulated a vision of data and archives which will resonate with many today:

‘[T]he record of one of the greatest, perhaps the very greatest scientific operation that has ever been undertaken, ought not … to be consigned to the casualties of time and accident in manuscript, nor indeed can their full import be extracted by one theorist, or in a single generation. Like astronomical observations, they ought to be at once secured from the possibility of destruction, and rendered accessible to computists of every age and nation by publication.’ (John Herschel to Edward Sabine, 27 January 1841, TNA, Correspondence of Sir Edward Sabine, /27/38).

Herschel was not only talking about rescuing meteorological data but also proposing that changing formats (from manuscript to print) was crucial when dealing with the great ‘big data’ crisis of the 1840s: the produced by a network of observatories and private scientists. Meteorological observations were by-products of this magnetic crusade. If it was obvious in Herschel’s pre-computer age that manuscript was insufficient for preservation, access and scientific computation, his point is only more pressing in today’s digital age, in which .

Herschel and the Committee were correct that large amounts of meteorological data could be found in the Royal Society’s collections. As far back as 1667, curator of experiments suggested that to ‘make a history of the weather’, scholars ought to keep observations. Then in 1723, Secretary of the Royal Society published in the Philosophical Transactions a call inviting observations from across the world to be sent in regularly to the Society. Although many of these early records were extracted to form the new meteorological archive ordered in 1839, information remains spread throughout other parts of the archive already digitised on Science in the Making:

But by far our largest collection of historical weather data is preserved in our (MA) series, which is the result of the 1839 Committee’s decision.

Looking at the date range in MA, 1706 to 1915, it is clear there are clusters of manuscripts around the times that calls for observations were put out – in 1723 by James Jurin, and most markedly in 1839 to coincide with the Meteorological Committee’s creation of the ‘department’ – and that most pre-date the 1854 establishment of the Meteorological Office, which would become the logical home for this type of record.

Some records are the results of one-off experiments, such as that around the summer and winter solstices of 1836 when John Herschel took observations at the Cape of Good Hope in South Africa for 24 hours. Long and consistent runs of information are best suited for analysis, though, so I would particularly note the observations of the Bishop of Durham, , which last from 1770 to 1823.

As an aggregate, the records sent by the Hudson’s Bay Company for various locations across Northern Canadian territories from the 1770s to the 1790s, and those from the which have partially been studied already, make for a unique source on North American eighteenth-century climate. Of course, the Royal Society’s own observations taken from Somerset House between 1812 and 1823, with instruments which served as standard references, provide particularly important information for London.

The geographical scope of the collection is also truly global. To make this abundantly clear, we’ve geotagged individual sets and created an interactive map on Science in the Making (above). For instance, there are over 300 runs of observations sent from across India, from Mahabaleshwar to Chennai, via Trivandrum and Mumbai, dating from 1729 to 1858. This treasure trove of meteorological data for India stems directly from the Royal Society’s involvement in colonial scientific networks.

Madras Observatory interior (unknown artist)

The creation of the archive in 1839 was accompanied by official requests to the Secretary of State for the Colonies, the and the Corporation of (a seafaring regulator) to help in collecting meteorological observations. Additionally, the Royal Society was directly involved in setting up observatories across the British Empire for the magnetic crusade, which means that we hold raw data from Montserrat, Mauritius, South Africa, India, Singapore and the Falkland Islands.

There are also ships’ logs that cover all the oceans, including a very early Arctic log from 1727 and the raw data from the of 1901-1904. To give more context to the magnetic and meteorological journals written on this expedition, we’ve also included in this release , plus the manuscripts which give insights into how the expedition was jointly run by the Royal Society and the Royal Geographical Society.

In addition to the wide range of time periods and locations, there is a real wealth of methods, instrumentation and scientific enquiries to be found in the Meteorological Archives: barometric readings, tidal variations, textual descriptions and much more. Even when clear guidance was issued on how to standardise the data, such as the 112-page report by the 1839 Committee, it often becomes obvious that the instructions were ignored. There are also varying levels of metadata regarding which instruments were used and where they were placed, as some observers sent detailed explanatory letters while others simply filled in pre-printed sheets with no additional information.

Whenever possible, as part of the re-cataloguing exercise which preceded our digitisation, the observers have been identified and named. The bulk of the daily work within observatory settings was done by assistants who are usually overshadowed by the more famous men who headed the observatories. It therefore feels meaningful to read the names of , who took the observations in Deesa, Gujarat and are now associated with their productions.

As part of this launch, we’ve registered the dataset as part of the supported by the World Meteorological Organization (WMO) and Copernicus Climate Change Service (C3S) Data Rescue Service, under the leadership of the (KNMI). While a few of the observations, for example a part of the dataset (), have previously been reconstructed into datasets, the majority have yet to be rescued – digitisation is only the first step in opening them up to scientific reconstruction.

My hope for today’s release of the Meteorological Archives online is that it moves us one step closer to fulfilling Sir John Herschel’s vision of delivering them to the ‘computists’ of our age, helping them refine our understanding of climate past, present and future.

Top image: Meteorological diagrams made at Karachi, Pakistan, by James Ekin and others, 1858 ()