PSI - Issue 84

Lorenzo Brezzi et al. / Procedia Structural Integrity 84 (2026) 473–480

474

Keywords: Landslide Risk Assessment; Indirect Interaction; Smart Inspection; Geomorphological Indicators; Rockfall; Debris Flow.

1. Introduction Natural hazards such as earthquakes, floods, and landslides pose a constant threat to bridges and viaducts, particularly those located in geomorphologically complex areas or built to cross natural obstacles such as rivers and valleys (Scala et al., 2025; Gabrieli et al., 2025). In Italy, where a large portion of the infrastructure network consists of aging structures embedded in mountainous and hilly environments, landslide-related processes represent one of the most critical sources of risk for existing bridges (Mitoulis et al., 2022; Yang et al., 2022; Proske, 2019). Within the framework introduced by the Italian Guidelines for the classification and management of bridge risk (MIMS, 2022), landslide hazard is explicitly considered among the relevant risk sources to be assessed during the census and inspection phases. In particular, Level 2 assessments play a key role in refining the preliminary classification of attention classes through on-site inspections and expert judgment. However, landslide risk remains one of the most challenging components to evaluate, due to the wide variety of possible interaction mechanisms and the intrinsic difficulty of identifying slope instabilities in the field. During the census phase (Level 0), available cartographic information and documentation provided by asset managers are collected to obtain an initial understanding of the territorial context and to identify mapped landslide phenomena potentially affecting the structure. This information is then verified or revised during on-site inspections (Level 1), which may lead to the recognition of previously unmapped instabilities or to a more detailed interpretation of known phenomena, supporting the assignment of a landslide-related attention class. Despite this structured framework, experience shows that Level 2 assessments may be affected by significant uncertainty, which is not necessarily related to a lack of information, but rather to the interpretation of the evidence collected during inspections. In particular, some landslide processes interact with bridges through what can be defined as indirect mechanisms, which differ substantially from the more familiar direct interaction typically associated with slow-moving, large-volume slope instabilities. In the latter case, progressive ground deformations involve a significant portion of the foundation system, often producing cumulative and persistent structural distress that becomes detectable over time during routine inspections. By contrast, rapid rainfall-triggered phenomena such as debris flows, shallow slides, rockfalls, or topples, which are particularly frequent in alpine and mountainous environments (Hungr, 2007), may develop in slopes that appear stable under ordinary conditions. In these situations, the bridge and its foundations may show no visible signs of distress prior to the triggering event, because no load has yet been transferred to the structure. The interaction becomes effective only if and when the rapid process is activated, potentially travelling long distances and impacting the structure suddenly. At that stage, the effects may be severe and include significant structural damage or even collapse. Therefore, when referring to indirect interaction mechanisms, the main challenge lies in the intrinsic difficulty of recognizing the hazard in advance based solely on structural damage or visible environmental evidence, especially when potential source areas are partially hidden by vegetation. In many cases, unless previous events have already occurred and left visible traces such as debris deposits or impact marks, the structure may appear entirely undisturbed up to the moment when a rapid landslide is triggered. Nevertheless, even when involving relatively small volumes, the high energy and impact forces of these processes may cause severe damage to the infrastructure and, in extreme cases, lead to structural failure (Lu et al., 2012). The presence of mitigation works, such as rockfall nets, debris-flow barriers, or artificial galleries, may be interpreted as evidence of an already controlled hazard. However, field experience suggests that these elements may instead represent valuable indicators of past or recurrent instability, whose current effectiveness and relevance should be critically reassessed during Level 1 inspections. The reliability of Level 2 landslide risk assessment therefore strongly depends on the ability to correctly interpret geomorphological indicators and contextual evidence, rather than on the sole identification of mapped phenomena or the observation of structural damage. Based on a series of field inspections carried out in the province of Belluno, northeastern Italy, this paper aims to identify the critical aspects that should be explicitly checked during Level 2 assessments in order to reduce the risk of overlooking indirect landslide threats. The contribution focuses on lessons learned from real inspection experiences, highlighting how specific elements observed in the field can support a more robust interpretation of landslide