University of Toronto Vegetation Survey

Data Rescue Intern: Claire Voss

In the summer of 2024, I participated in the Living Data Project as a data rescue intern for the University of Toronto Vegetation Survey. The goal of this 6-week internship was to digitize and encode historical forest data from the second half of the 20th century, provided from the vegetation surveys of the lab of the late professor Paul Maycock. Currently housed on paper, the data required encoding to allow for their future use by the scientific community. To accomplish this, I created a relational database in Microsoft Access focusing on representing the whole of the data collected, streamlined data entry, and data quality assurance and control. Additionally, working with Peter Rodriguez, a PhD student in the Fortin Lab at the University of Toronto, we created an app in RShiny to visualize site distribution, displaying site name, coordinates, and date surveyed.

Figure 1: Example of historical data sheets, in this case from 1970. Note the early signs of physical degradation of the sheets (yellowing, etc.). Photo credit: Claire Voss.

From 1956-1994, the members of the Maycock lab conducted vegetation surveys across much of the province of Ontario and Southern Quebec. Vegetation data, as well as tree summary statistics, were previously encoded, but a plethora of seedling/tree frequency and presence data had yet to be preserved, leaving my job as somewhat of an ‘archivist’ very straightforward. The frequency/presence data were collected via ground surveys and included various ecosystems such as prairies, savannas, alvars, fens, bogs, swamps, marshes, and other wetlands, emergent aquatics, subarctic woodlands, and some coastal tundra. The data provide a rich historical record of vegetation cover for the regions included in the surveys and will provide an excellent baseline for comparative studies in relation to the current anthropogenic transitions present in the landscape of Ontario and Quebec. In order to make the data available to a wide range of researchers, the most straightforward approach was to utilize Microsoft Access to encode the data and archive it in the Borealis online data repository.

Figure 2: Early version of the RShiny app showing locations of all the survey sites, colour coded by region: green is Southern Deciduous Forest, red is Northern Coniferous Forest, and blue is Boreal. Note that given the early version, there are likely errors in some location data. Photo credit: Claire Voss.

Utilizing Microsoft Access, I created a simple relational database to enter and house all of the data. Mainly, the data requiring encoding were the site information and field notes (including disturbance records), as well as the corresponding species presence and frequency data per site surveyed. The surveys were split into three distinct regions:

1.  Boreal: Boreal forests spanning Ontario and Quebec, spanning further north than the Northern Coniferous Forests.
2. Southern Deciduous Forest: Deciduous forests in Southern Ontario.
3. Northern Coniferous Forest: Coniferous forests in Ontario and Quebec.

During my internship, I was able to create the relational database, as well as implement the protocol on the Boreal dataset. This initial version of the data, as well as the previously encoded data and the metadata in the form of .csv and .txt files, will be uploaded to the online data repository Borealis, and will be updated as entry of each region is completed. This data will then be publicly available for all those interested in utilizing this historical Canadian data in their research.

I would like to extend a huge thanks to Marie-Josée Fortin, David Hunt, Bruno Soares, and Peter Rodriguez, who guided me throughout the entire process and provided me with invaluable insight during my time as a data rescue intern. I learned so much during this internship, and I am so proud of the work we did to preserve this important data!