PSI - Issue 78

Francesco Testa et al. / Procedia Structural Integrity 78 (2026) 905–912

911

4. Potential biases in the regional models While the differences in the behaviour of towers across various locations may be attributed to traditional construction and mitigation practices, seismic history, and local seismic response, such aspects are not directly identifiable from the features recorded in the survey forms. Consequently, arriving at a conclusive explanation for the observed discrepancies requires ensuring that the interpretation is not biased by other parameters that are strongly correlated with the damage level. To support this, three key aspects that influence the susceptibility to seismic damage and which can be readily extracted from the survey forms were considered: the maintenance level and the presence of modern strengthening interventions, both prior to the seismic events, and the number of severe seismic shocks preceding the inspection. Focusing on the date of the inspection, compared to the date of the most severe shocks of the sequence, it was possible to identify that most of the towers were inspected between the first and second shocks, or after all shocks of the seismic sequence, this suggesting the possibility of presenting cumulative damage. In what concerns the state of maintenance, the inspection form provides a qualitative assessment of the general condition of the churches, using the following categories: Good, Acceptable, Poor, Very Poor, and Work in Progress. Regarding the presence of modern strengthening interventions, a binary classification (yes/no) is included in Da.D.O. to characterise the sample. Although both parameters in the form do not refer specifically to the bell towers but to the entire church, it is hereafter assumed that the evaluation may also be indicative of the condition of the towers. The distributions of these three factors in the datasets from the four geographical areas were investigated. The results are summarised in Table 3. It is worth noting that information on the relevant factors is not always available in the database for every tower, particularly with regard to the presence or absence of interventions. As previously noted, the Marche and Umbria subsets represent 61% and 26% of the total dataset, respectively. Despite the difference in sample size, the higher vulnerability observed in towers located in the Marche region compared to those in Umbria may be linked to the greater proportion of towers in Marche that were affected by cumulative damage, as 70% were inspected after the final seismic event, compared to 53% in Umbria. Similarly, fewer towers in Marche were inspected after the first event (22%) compared to Umbria (31%). Comparable trends were observed at the provincial level when analysing towers in the Macerata and Perugia provinces. In contrast, classifications based on the level of maintenance did not reveal such pronounced disparities. Finally, discrimination based on the structural interventions indicated remarkable discrepancies between the sample located across the affected regions and provinces. However, in this case, the number of entries with missing information is significant and may affect the reliability of the results. These findings suggest that further investigation into the cumulative effects of repeated seismic events and the presence of strengthening interventions could help clarify the observed variations in tower vulnerability. This represents a key direction for future research. Table 3. Statistics of the subsets considering seismic scenarios prior to inspection, maintenance state and evidence of strengthening interventions. Subsets Region Province Location Entire set Marche Umbria Macerata Perugia 486 203 295 183 Seismic scenarios First Event 106 (22%) 63 (31%) 52 (18%) 61 (33%) Last Event 338 (70%) 108 (53%) 224 (76%) 92 (50%)

Good 251 (52%) 114 (56%) 149 (51%) 105 (57%) Acceptable 152 (31%) 55 (27%) 98 (33%) 49 (27%) Bad 50 (10%) 19 (9%) 26 (9%) 16 (9%) 55 (11%) 24 (12%) 35 (12%) 20 (11%)

Maintenance state

No Yes

Strengthening

34 (7%) 100 (49%) 24 (8%)

90 (49%)

5. Conclusions This study aimed to investigate the seismic vulnerability of masonry bell towers, using a comprehensive dataset of churches affected by the 2016–2017 Central Italy seismic sequence, as recorded in the Italian Da.D.O. database.