Archive and library special collections as proxy data: reconstructing the American chestnut blight through digitized collections

Abstract

Abstract This paper discusses the design and implications of a study that explored the potential for archives and library special collections to serve as historical environmental proxy data to support the reconstruction of the spatiotemporal spread of the American chestnut blight in Tennessee. By collecting, reconciling, and analyzing heterogeneous mundane primary source materials from 1904 to 1950, the major period of infection and tree loss, the case study reached beyond conventional evidence to ask new questions of nontraditional sources. QGIS and Python were used to reconcile and model nonstandardized and ambiguous natural-language keywords derived from these sources to identify trends and patterns that may not be evident from traditional document analysis. The paper argues that the contributions made by textual and visual information fragments found in these materials support an expansion of the term “proxy data” beyond what is currently understood as paleoclimate archives, i.e., physical, chemical, and biological materials preserved within the geologic record (USGS 2022). Such socially constructed records found in archives and library special collections offer additional qualitative and quantitative information about historical climate change to support modeling variable fluctuations over time. They can also provide a rich and dynamic context for the natural causes and human interventions that, in combination, act on the environment. However, the study also identifies significant limitations in the digital accessibility of relevant archival sources and a lack of specificity in their descriptions. These need to be addressed if integrating such source material into scientific studies is to become more widespread and scalable.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 1 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.