Process Blog Home

Archiving Weather Data

It was mid-summer and excruciatingly hot in Washington, D.C. It wasn’t just me––July 2014 would go down as the fourth hottest July ever recorded in DC, as documented by NOAA’s National Center for Environmental Information. The record-making heat was ironic, considering my reason for being in Washington. I was there to look at a nineteenth-century archive of weather data––the Smithsonian Meteorological Project archive––that gave rise to the United States Weather Bureau in 1890 and, less than a century later, NOAA and its National Center for Environmental Information.

The Smithsonian Meteorological Project (1849–1870) was the first weather data collection effort in the United States that brought together a national coalition of volunteer observers based around the pursuit of citizen science.[1] As historian James Roger Fleming argues in Meteorology in America, the Smithsonian Meteorological Project was important because it “formed a ‘seedbed’ for the continued growth of theories rooted in data.”[2] In addition to this, the project also crystalized the very notion of data for volunteer observers as well as a larger American public eager for national weather and climate data. I was going to the Smithsonian because I wanted to know how this crystallization happened. What would an inquiry into the fits and starts of environmental data culture in the United States have to teach environmental humanists––from environmental historians to digital humanities folks––about data itself? For how long had data been part of the American landscape?

It would be easy enough to get access to the data, I thought. Record Unit 60 contained the Smithsonian Meteorological Project’s files, including a box titled “data from eighteen twenties to eighteen seventies.” But when I got to RU 60, I found everything but the data––I found letters, drawings of hail and eclipses. Where had fifty-five years of weather data gone?

The question of where weather and climate data goes is a question that has recently been asked with renewed urgency given national changes to climate change initiatives and data management under the Trump administration. Projects like Data Refuge have not only preserved massive collections of government-stored environmental data at risk of being deleted by the new administration, but have also illuminated questions about the vulnerability of data and its place in the archive more generally. By looking back to the mid-nineteenth-century Smithsonian Meteorological project and specifically the ways in which weather and climate data produced by this project was managed by U.S. archives in the nineteenth and twentieth centuries, we can gain perspective on how environmental data becomes archival matter but also on the longer history of environmental data management in the United States. These perspectives will be useful as we continue to ask what data needs to be safeguarded under the new administration and how, precisely, to safeguard it.

In the mid-nineteenth-century, volunteers submitted their weather data to the Smithsonian in the form of a table, or what was then called a meteorological blank form. When Smithsonian volunteers submitted their data tables, each table was accompanied by a letter from the observer: I enclose my Journals for the month of April or I am now using my last Blank. Will the Department please send me some more blanks for the records of Meteorological Observations.[3] Some of the letters were filled with explanations of weather events and materials that didn’t fit into the layout of the table itself––one volunteer embedded a map of a tornado in their letter; another slipped a live locust between the sheets, explaining that he suspected the locust had been brought to his grain fields because of the weather and therefore needed analysis by the meteorological project. The bulk of meteorological materials at the Smithsonian are linguistic or pictorial remnants––the letters attached to each data table, drawings attached to letters. A few hand-made data tables are sprinkled across the records, but there is no sign of the typeset and completed blank forms themselves.

Throughout the nineteenth and twentieth centuries, Smithsonian Meteorological Project data traveled between six repositories––the Smithsonian Institution, the Signal Service, the War Department, the Department of Agriculture, the Weather Bureau, and the National Archives, where it finally stayed. The data tables were detached from the letters, most likely when they were reduced by a team of male and female computers under the direction of James Coffin throughout the 1850s.[4] But they were never reunited with the letters and drawings. The letters were categorized and preserved, separated between two archives––the Smithsonian Institution Archives and the National Archives. The data tables had a different fate. They were combined with other nineteenth-century numerical weather records from the Medical Department of the Army, the War Department, the Land Office, the Department of Agriculture, the Signal Department, and the Weather Bureau. Miscellaneous Smithsonian data documents (pictured above) were stuffed into five boxes in RG 27, Records of the Weather Bureau, sometime between 1890 and 1936, where they would remain until their 2016 preservation.

Some of the Smithsonian’s volunteer data tables were microfilmed and those that were microfilmed are scattered among T907, a collection housed at the National Archives titled “Climatological Records of the Weather Bureau, 1819-1892.” All nineteenth-century weather data––without any distinction made between collection method or difference in graphical form––is grouped together in T907. The Smithsonian data, the small amount that can be found on T907, has been separated from the letters and drawings that the data traveled with in the nineteenth century. The data roamed, without permanent home, for over a century and this roaming reveals the unique difficulty of valuing and categorizing environmental data. It poses a larger question that historians must consider given today’s climate: should data’s context matter?

Yes, it should. Context helps us see that the history of data collection, acquisition, reduction, and display is not divorced from the human but rather uniquely tied to it. When we lose data’s context, we lose a sense of the communities of people who have made data possible, from white women computers in the 1850s to African American women computers at NASA in 1960. In the same way that data is made possible by the work of communities, its longevity can also be threatened by particular communities. We need to safeguard vulnerable data and the cultural histories of production that come along with this data. We are still struggling to figure out how to archive data and where data’s archives should be. The twisty travels of nineteenth-century weather data show this much. When the National Archives inherited the Weather Bureau material––which included the Smithsonian Meteorological Project documents––in 1936, the data might have seemed extraneous. But to those who study the history of data culture as well as climate change, it’s anything but that.

The vantage point from which we look at data has changed over the last twenty years, both because human-gathered climate data has become an essential feature of studying the Anthropocene and because digital humanists have shown that “[d]ata is everywhere, piling up in dizzy amounts.”[5] A look back to the nineteenth-century origins of data culture and twentieth-century data archiving practices shows that this isn’t a new phenomenon, though it has taken on a new digital urgency. Data is not new, nor is the concept of data overload. Both have firm roots in nineteenth-century America. Data has been piling up for a while and, if we learn from our past mistakes, we can arrive at better systems of data care and stewardship in and for the Anthropocene. The twisty travels of weather and climate data compel us to ask who the guardians and caretakers of data are, and, well, who they should be.

Sara J. Grossman is a Visiting Fellow at the Center for Humanities + Information at Pennsylvania State University. She is currently at work on her first book, Measuring the Face of the Sky, an environmental history of data in the United States. For an updated list of projects and publications, please visit

[1] Smithsonian Institution Archives, Record Unit 60, Smithsonian Institution, Meteorological Project, Records.

[2] James R. Fleming, Meteorology in America, 1800-1870, (Baltimore: Johns Hopkins University Press, 1990), 92.

[3] Letters from May 8 1852 and July 4 1859. RU 60.

[4] Smithsonian Institution Archives, Record Unit 7060, Coffin, James Henry, 1806-1873, James Henry Coffin Papers.

[5] Lisa Gitelman, “raw Data” Is an Oxymoron, (Cambridge: MIT Press, 2013), 1.