|
Data Rescue Intern: Yiduo “Harry” Zhang The 1957-1958 New Brunswick topographic map series carries valuable data of historical land cover, land use, and landscape feature locations throughout the province that are only available on paper. Many of these paper maps currently in possession of the Canadian Forest Service also contain handwritten annotations about historical field sample sites cumulated through decades of work done by various researchers from the agency. With its provincial scope and a resolution of 1:50,000 scale, this historical dataset could prove invaluable for studies of species distribution and habitat change. However, in order to effectively utilize, store, and distribute this spatial dataset, it first needs to be digitized.
The process of digitizing the maps includes scanning the paper maps, georeferencing the scanned images, and classifying land cover and extracting features. The paper maps were scanned into high-resolution full color images in PDF format in 2022 by Access Corp. Under the supervision of Forest Service researchers Robert Buchkowski and Shane Heartz, the current data rescue project primarily focused on developing workflows to georeference and classify raster maps, which can later be applied to the over 200 scanned documents of this map series. Over the course of 6 weeks, a georeferencing workflow specific to these scanned maps was developed for ArcMap 10.6 (ESRI, 2018). A raster classification/feature extraction workflow was also developed for R (R Core Team, 2023) based on HistMapR package (Auffret et al, 2017). A set of five maps were fully processed using these methods to produce historical land cover classification and road feature datasets for the Fredericton area. These methods will be further applied to the rest of the map series in future sessions. The completed data rescue products will be made available to the public via the provincial open data platform.
Through these data rescue efforts, the unique knowledge held in these historical maps will be securely stored in digital formats, and continue to impact future researchers across the province. References Auffret, A.G., Kimberley, A., Plue, J. et al. (2017) HistMapR: Rapid digitization of historical land-use maps in R. Methods in Ecology and Evolution. 8-11. https://doi.org/10.17045/sthlmuni.4649854 ESRI 2018. ArcGIS Desktop: Release 10.6. Redlands, CA: Environmental Systems Research Institute. R Core Team, 2023. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R-project.org Comments are closed.
|
RSS Feed