Benchmarking Large Language Models for Geolocating Colonial Virginia Land Grants

TL;DR

This paper evaluates large language models' ability to convert historical land grant descriptions into accurate geographic coordinates, demonstrating their potential for scalable and cost-effective historical georeferencing.

Contribution

It introduces a new benchmark dataset and systematically compares LLMs against traditional geocoding methods for historical land records.

Findings

Top LLM achieved 23 km mean error, outperforming baselines.

Ensemble methods further reduced error to 19.2 km.

Cost-effective models maintained high accuracy at low cost.

Abstract

Virginia's seventeenth- and eighteenth-century land patents survive primarily as narrative metes-and-bounds descriptions, limiting spatial analysis. This study systematically evaluates current-generation large language models (LLMs) in converting these prose abstracts into geographically accurate latitude/longitude coordinates within a focused evaluation context. A digitized corpus of 5,471 Virginia patent abstracts (1695-1732) is released, with 43 rigorously verified test cases serving as an initial, geographically focused benchmark. Six OpenAI models across three architectures-o-series, GPT-4-class, and GPT-3.5-were tested under two paradigms: direct-to-coordinate and tool-augmented chain-of-thought invoking external geocoding APIs. Results were compared against a GIS analyst baseline, Stanford NER geoparser, Mordecai-3 neural geoparser, and a county-centroid heuristic. The top single-call model, o3-2025-04-16, achieved a mean error of 23 km (median 14 km), outperforming the median LLM (37.4 km) by 37.5%, the weakest LLM (50.3 km) by 53.5%, and external baselines by 67% (GIS analyst) and 70% (Stanford NER). A five-call ensemble further reduced errors to 19.2 km (median 12.2 km) at minimal additional cost (~USD 0.20 per grant), outperforming the median LLM by 48.7%. A patentee-name redaction ablation slightly increased error (~7%), showing reliance on textual landmark and adjacency descriptions rather than memorization. The cost-effective gpt-4o-2024-08-06 model maintained a 28 km mean error at USD 1.09 per 1,000 grants, establishing a strong cost-accuracy benchmark. External geocoding tools offer no measurable benefit in this evaluation. These findings demonstrate LLMs' potential for scalable, accurate, cost-effective historical georeferencing.