Lantini, Livia ORCID: https://orcid.org/0000-0002-0416-1077, Ardakanian, Atiyeh and Tosti, Fabio ORCID: https://orcid.org/0000-0003-0291-9937 (2026) Operationalising GPR Evidence for Urban Tree and Infrastructure Management. In: The 1st Faringdon Symposium - Sensing the Impact (FARSY 2026), 25 Jun 2026, London, UK. (In Press)

PDF Lantini et al. - FARSY 2026.pdf - Accepted Version Restricted to Repository staff only Download (43kB)

Abstract

Urban trees are increasingly recognised as critical components of climate adaptation strategies, while also representing a major source of interaction with buried infrastructure [1]. Ground Penetrating Radar (GPR) is widely used to investigate shallow subsurface conditions and detect root-related features in urban environments. Nevertheless, translating spatially complex spatial GPR observations into management-level outputs remains a major unresolved challenge [2]. In particular, limited attention has been devoted to how aggregation and classification processes influence the consistency and reliability of resulting operational interpretations.
This exploratory work builds on ongoing research into descriptor-based assessment frameworks for urban tree investigations using GPR [3]. GPR-derived descriptors are used to classify spatial zones around urban trees according to relative levels of anomalous subsurface response. These zone-level classifications are subsequently aggregated into broader management categories intended to support monitoring and intervention prioritisation.
The study investigates whether different spatial anomaly patterns may converge towards identical management outputs after aggregation. Particular attention is devoted to cases in which diffuse moderate anomalies and highly concentrated localised responses generate the same operational category despite representing substantially different subsurface conditions. The analysis further examines how threshold selection, aggregation logic, and spatial weighting influence classification stability, anomaly prominence, and sensitivity to isolated features.
Preliminary observations suggest that aggregation may mask relevant localised anomalies, particularly when strong responses are embedded within otherwise homogeneous spatial distributions. At the same time, simplification remains necessary for comparability and large scale asset management. The results therefore highlight the need for operational frameworks capable of synthesising spatial GPR evidence while preserving the subsurface characteristics most relevant to infrastructure behaviour and management priorities.
The study therefore supports the development of more transparent and traceable approaches for integrating GPR evidence into urban infrastructure management and climate adaptation planning by explicitly examining how spatial anomaly distributions influence management-level classifications.

References:

[1] S. Pauleit, T. Zölch, R. Hansen, T.B. Randrup and C. Konijnendijk van den Bosch, "Nature Based Solutions and Climate Change – Four Shades of Green," in Nature-Based Solutions to Climate Change Adaptation in Urban Areas: Linkages between Science, Policy and Practice, N. Kabish, H. Korn, J. Stadler and A. Bonn, Springer, 2017, pp. 29–50.
[2] F. Hou, X. Rui, X. Fan and H. Zhang, "Review of GPR Activities in Civil Infrastructures: Data Analysis and Applications," Remote Sensing (Basel, Switzerland), vol. 14, pp. 5972, Dec 1. 2022.
[3] L. Lantini, A. Ardakanian and F. Tosti, "From Detection to Interpretation: A Decision-Support Framework for GPR-Based Evidence in Urban Climate Adaptation," EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22982.

Item Type:	Conference or Workshop Item (Paper)
ISSN:	2673-4591
Subjects:	Construction and engineering > Digital signal processing
Related URLs:	https://sciforum.net/event/FARSY2026?sec...
Date Deposited:	01 Jul 2026
Dates:	Date Publication status 7 June 2026 Accepted 25 June 2026 Presented
School, department or research centre:	The Faringdon Research Centre for Non-Destructive Testing and Remote Sensing
URI:	https://repository.uwl.ac.uk/id/eprint/15224

Actions (admin access)