What can be learnt from the catastrophic failure of a check dam system? A forensic analysis of a cascading natural-anthropogenic hazard

What can be learnt from the catastrophic failure of a check dam system? A forensic analysis of a cascading natural-anthropogenic hazard | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article What can be learnt from the catastrophic failure of a check dam system? A forensic analysis of a cascading natural-anthropogenic hazard Eleonora Dallan, Lorenzo Marchi, Giorgio Rosatti, Daniel Zugliani, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5478044/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 13 May, 2025 Read the published version in Natural Hazards → Version 1 posted 5 You are reading this latest preprint version Abstract Check dams may be effective structures for reducing debris flow hazard, but their failure often results in serious consequences for people and infrastructures. The examination of these failures embracing a forensic engineering approach, still rather poorly represented in the scientific literature, would lead to important improvements in how residual risk is planned and managed. In this study, we developed a framework for the forensic analysis of check dam systems failures in terms of cascading natural-anthropogenic hazards, and we applied such framework to the catastrophic event occurred in October 2018 in the Rotian creek catchment (Eastern Italian Alps). The post-event survey and analysis gathered observations about rainfall, peak discharges, morphological impacts, and damaged check dams. Based on these data, we applied a newly developed coupled hydrologic-hydraulic debris flow model and we assessed the failure mode of the check dam system. Our results highlight important practical implications for improving residual risk management, namely: i) development of debris flow models capable of simulating the role of check dams and their failure in the debris flow dynamics, ii) the call for extensive datasets of check dam system failures, and iii) the necessity to develop methodologies for the prioritisation of field inspection and maintenance of existing check dam systems. debris flow check dam failure Forensic analysis post-event survey cascading hazards Figures Figure 1 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Highlights A forensic analysis of debris-flow related check dam system failures is proposed A catastrophic check dam system failure event is used to illustrate the framework Check dam failure greatly amplified the debris-flow total volume Coupled hydrologic-hydraulic model helps ascertain the event evolution and volumes We suggest advances in debris flow modelling, data collection, structure inspection 1. Introduction Thousands of mountain channels in the Alps are equipped with check dam systems which require costly maintenance to avoid structural deterioration (Mazzorana et al., 2014). While check dams are generally effective in stabilizing mountain channels and retaining sediment transported by floods and debris flows, thus reducing hazard for settlements and transport routes, the failure or collapse of these mitigation structures may result in serious consequences for people and infrastructures (Piton et al., 2016). Failure not only removes the protection offered by the dam, but it can also lead to an increased supply of sediment downstream (Baggio and D’Agostino, 2022; Benito et al., 1998; Sodnik et al., 2015; Wang, 2013). As a result, check dam systems failure is a rare but highly dangerous event that may result in much higher damages compared to a situation without structures, possibly also impacting areas that would otherwise be untouched by the event. The failure of check dams has a higher probability to occur and becomes potentially highly hazardous during large magnitude floods or debris flows: on the one hand, severe flow conditions induce stress on the structure that can lead to their collapse; on the other hand, their abrupt failure frees large sediment volumes which can be easily transported downstream under, typically reaching the debris flow fan. A close examination of check dam system failures can help to better quantify and manage the residual risk associated with structural measures for the mitigation of flood and debris flow risk (Jakob, 2019). The examination of flood protection system failures is part of forensic engineering (Huebl et al., 2024), and it is based on accurate post-event analyses of both the relevant hydro-geomorphic processes and the structures. Whereas forensic methodologies for water floods and debris flows are well developed (Borga et al., 2014; 2019), literature is scarce for the case of the failures of check dam systems as a cascading natural-anthropogenic hazard (Huebl et al., 2024 and references therein; Marchi and Cavalli 2007; Bossi et al., 2024). This is understandable, as failures are tantamount to an admission of guilt at worst, or poor design work at best. This research gap is important given the ubiquity of ageing check dams in regulated steep streams, their difficult maintenance and the increasing intensity of flood events in small catchments related to climate change (Bertola et al., 2020; Dallan et al., 2024a,b). Within this background, this paper proposes a forensic model for debris-flow related check dam system failures. The model is tested on the catastrophic event in the Rotian catchment (2.5 km 2 in area, Eastern Italian Alps) occurred during the so-called “Vaia” storm, on 27- 29 October 2018. . The storm was characterized by exceptionally high cumulative rainfall depths and record wind velocities, and triggered in the Rotian Creek a debris flow. whose initial volume was greatly amplified by the collapse of 16 check dams built in the 1970s. A forensic model is illustrated based on data from the post-event survey carried out a few months after the event. The application aims to show the added value of exploiting the observations to qualify the results from a newly developed coupled hydrologic-hydraulic debris flow mathematical-numerical model and to ascertain the failure mode of the check dam system collapse. We also identify aspects where forensic analysis of check dam failures will have important practical consequences on residual risk management. 2. Forensic analysis of coupled debris flow and check dam system failure Forensic hydrology provides information and knowledge for extreme events (mostly floods and flash floods) where no direct systematic hydrometeorological data and observations are available (Bronstert et al., 2018; Borga et al., 2019). Hence, a central place in forensic hydrology is placed by the generation of data and observations from post-event surveys. This analysis borrows the term “forensics” from the field of criminal investigation, because it denotes a consistent approach to develop a comprehensive analysis of an event and its causes (Keating et al., 2016). While enabling a systematic approach, forensic hydrology analysis provides a consistent approach that facilitates cross-event learning, with a focus on quantifying flood peaks, timing, and volumes. Debris-flow magnitude relates to the volume of sediment mobilized, peak discharge, and inundation area (Ballesteros-Cánovas et al., 2024). When compared to water floods, forensic activities for debris flows put, in general, less emphasis on the reconstruction of the peak discharge (which, anyway, remains an important parameter in debris-flow assessment) and more attention to the assessment of the sediment volume involved. Debris-flow volume has direct relevance for assessing the severity of a debris flow and the design of control measures, has been proposed as a one of the key variables for classifying debris flows (Jakob, 2005), and is related to other important debris-flow variables (e.g. Rickenmann, 2005). Other focal points in the post-event recognition of debris flows are the identification of the initiation point(s) and the triggering mechanism, and the recognition of the characteristics of the solid material involved, which is crucial for identifying the flow type (i.e., stony/granular vs muddy debris flow). The assessment of erosion and deposition volumes associated to debris flow events has undergone stark improvements in the last 25 years thanks to the increasing availability of high-resolution topographic data based on LiDAR and Structure from Motion (SfM), which makes it possible differencing pre- and post-event Digital Terrain Models (DTMs) (Scheidl et al., 2008; Bull et al., 2010; Cavalli et al., 2017; Cucchiaro et al., 2019). The catastrophic interaction of debris flows and protection works, leading to their collapse, requires a new methodological framework (Fig. 1), which integrates post-event data gathering from both the hydro-geomorphic processes and the structures involved. The post-event analysis of the hydro-geomorphic processes aims to characterize the triggering event and the main features of the debris flow (see flow-chart on the left in Figure 1). The post-event analysis of the collapsed structures aims to provide information about the failure mode, and the amount of sediment stored and mobilized. The comprehensive data-gathering approach permits to assess the contribution of the check dam failure to the overall debris flow characteristics (volumes, peak discharge and inundation pattern) and thus to the actual hazard posed to settlements and infrastructures, as well as it provides data for calibration/evaluation of improved modelling of the debris flow mitigation measures (Larese et al., 2023). Figure 1: Flowchart for integrated forensic analysis of the coupled debris flow – check dam collapse. 3. The Rotian creek river catchment The Rotian creek catchment (Fig. 2 b) is a forested watershed located in the Eastern Italian Alps, within the Autonomous Province of Trento (also called “Trentino”). The catchment has an area of 2.54 km 2 , with a range in altitude between 840 m a.s.l. at the fan apex to 2050 m a.s.l.; the mean angle of slope is 26.4°. The channel is 4.8 km long and has a mean angle of slope of 11.9°. The Rotian creek has built a large debris flow fan (0.43 km 2 , mean angle of slope 7.41°), which has forced the receiving stream (Noce River) to the opposite side of the valley. The Melton ruggedness number is 0.76, which is consistent with debris flows as the dominant sedimentary process on the fan. Mean annual precipitation is about 1100 mm, with major rainfall events typically occurring in the autumn season (Formetta et al., 2022). The mean value of the annual maxima precipitation ranges from 15 mm in 1 hour to 70 mm over 24 hours (Dallan et al., 2024). Recent research in Trentino has shown an upward trend in precipitation for both sub-daily (Libertino et al., 2019) and sub-hourly durations (Dallan et al., 2022) over the past few decades. The shape of the catchment is elongated along the main channel in a north – south direction. The axis of the basin corresponds to a transpressional fault that separates massive igneous rocks (tonalite) that build the left slope of the catchment from limestone that outcrops on the right bank of the Rotian channel (Dal Piaz et al., 2007). Tonalite also outcrops in the lower sector of the basin, where the Rotian flows through a narrow rocky gorge. Most of the basin (68%) is covered by Quaternary deposits, mostly würmian moraines, with less widespread presence of alluvial and colluvial deposits. In the moraines, boulders – up to 2-4 m – are embedded in a sandy-silty matrix. Consistently with the structural settings of the basin and the widespread presence of easily erodible moraines, the basin results in a canyon-like incision of around 10–50 m of elevation difference compared with the top of the banks. The steep valley sides (range 30–45°) bordering the Rotian Creek act as sediment source to the stream during intense runoff events. The channel reach from 1233 to 1030 m a.s.l. has been consolidated with a series of 16 check dams built from 1977 to 1986, with the seven located in the upper reach made of reinforced concrete. The check dam mean height is 5.3 m (range 2.9–7.8 m) for a cumulative sum of 78.9 m and a mean channel angle of slope of 13°. Immediately downstream of the last check dam, the channel bed has been widened in order to create a depositional area formed by a series of 5 retention basins (total length 250 m, mean width 25 m). In 2014, a 7.4 m high open check dam was constructed at 989 m a.s.l., immediately downstream of the retention basins (Fig. 2 b). In the past, the Rotian catchment produced two major recorded debris-flow events, in 1776 and 1882 (Baggio and D’Agostino, 2022). In the Eastern Italian Alps, where the Rotian is located, the failure of check dam systems is a rare but not unprecedented event. A major debris-flow disaster, enhanced by the failure of an array of stone masonry check dams, occurred in November 1966 in another catchment in Trentino, causing the loss of three lives and substantial economic damage (Marchi and Cavalli, 2007). Less catastrophic in its consequences, but still relevant, was the destruction of check dams built in the 1920s-30s by a debris flow that originated from the mobilization of a large rotational landslide in the Venetian pre-Alps in 1985 (Bossi et al., 2024), as well as the masonry check-dams destroyed by 2009 and 2010 events in South Tyrol (Dell’Agnese et al., 2013). Figure 2. Study area with (a) its location in Italy, (b) the rain gauges included in the 10-km search radius centered on the study area, and (c) the Rotian catchment-debris flow fan system with the indication of the different sub-catchments commented in the text alongside check dams and lower retention area with the final slit dam. 4. The forensic analysis for the 2018 event in the Rotian catchment 4.1 The precipitation event: amount, timing and severity The Vaia storm (October 2018) caused one of the most impacting floods in northeastern Italy in the last century together with the flooding event of 1966 (Giovannini et al., 2021; Sioni et al., 2023). The storm hit almost the entire Alpine region, as well as Liguria and central Italy for three days, from 27 to 30 October. In several Alpine areas the rainfall amounts in the three days reached 600-900 mm, representing the strongest event in the last 150 years (Giovannini et al., 2021). In particular, on 29 October 2018, the event was characterized by the passage of a cold front with a deep convective band, stretching from South-East to North-West, which followed a previous phase (from 22:40 CET of 26 October to 18:00 CET of 28 October), characterized by stratiform orographic type precipitation (Borga and Zaramella, 2020). In the evening of October 29, the intense precipitation in the upper Rotian catchment triggered a catastrophic cascading natural-anthropogenic hazard (Rosatti et al., 2023). The basin is ungauged; Fig 2b shows the four rain-gauge stations considered in this study, all located within 10 km from the Upper Rotian river catchment. Precipitation data at 5 minutes temporal resolution, collected and managed by Meteotrentino (Autonomous Province of Trento), are available, with a total coverage of at least 25 years starting from the early 90s (Fig. 2c). Precipitation amount over the basin during the event was estimated by integrating data from the rain gauge stations and radar-based rainfall estimates from three nearby radar systems: Monte Grande (Veneto Region), Monte Macaion (Autonomous Province of Trento) and the Weissfluh at Davos (Switzerland). Quantitative rainfall estimation at the ground based on radar reflectivity observations presents the usual challenges which characterize radar rainfall estimation in a mountainous context (Marra et al., 2014). Radar-based estimates of precipitation at the ground were obtained by applying a complex error correction chain (Marra et al., 2014) accounting for wavelength attenuation, beam blocking, vertical profile of radar reflectivity, wind effects and finally integrating the radar-based estimates with data from the rain gauge stations. Final estimates are characterised by 1 km grid spacing and 15 min temporal aggregation. Precipitation maxima over the upper Rotian catchment, where the debris flow initiated, range from 24.8 mm at 1 h duration to 104.2 mm at 12 h and 359 mm over the three event days. Based on a probabilistic modelling approach (Pesce et al., 2024, submitted) in the Mezzana station (the closer station to the Rotian catchment, with the longest recording period), the severity of the precipitation maxima recorded during the storm was also evaluated. Based on a time-stationary approach, the precipitation event return period ranges from 50-100 year for 1 h duration, to 200-300 year for 12 h and > 300 year for 3 days, thus classifying the event as exceptional across all durations. 4.2 Time evolution of the debris flow The time evolution of the debris flow was investigated using different approaches, including interviews with eyewitnesses, analysis of video recorded on the debris flow fan, and post-event geomorphological and sedimentological observations. A questionnaire on the time occurrence and other features of the debris flow was distributed to 14 people, ranging in age from 25 to 74 years, who were located in different sectors of the debris flow fan during the event. Their answers indicated the occurrence of three debris-flow surges, respectively at 19:00-19:02 (average 19:01), 19:47-20:00 (average 19:51), and 23:30-23:40 (average 23:37) (Borga & Zaramella, 2020). These accounts confirm the reports of people involved in the emergency interventions and rescue activities on the occurrence of three surges on the evening of 29 October. The recordings of seven security video cameras installed in the campsite in the sector of the debris flow fan most severely hit by the debris flow were also analyzed. Although the quality of the videos is poor and recordings are limited to the first surge (i.e. before the video cameras went out of order), they provide some elements helpful to characterize - even if only qualitatively - some features of the debris flow: - the time on the video frames of one of the cameras shows the arrival of the front at 19:59:47 CET, which is consistent with the accounts of the witnesses (19:01 CET); - the debris flow front outside of the channel occurred suddenly, without or with a very limited fluid precursory surge; - the velocity of the front was high, although a quantitative assessment is not possible. - large boulders were observed in the debris-flow front, at least one of them can be ascribed to a large fragment of a check dam. The presence in the middle sector of the channel of clast-supported lateral levees consisting of angular boulders at a level lower than the maximum flow depth attained during the event (Fig. 3) confirms that multiple surges occurred, the last one being smaller than the main one. In addition to cobbles and boulders, the deposits consist of a sandy matrix with a limited amount of silt and negligible clay: these features led us to classify the October 2018 event in the Rotian creek as a granular debris flow (Coussot and Meunier, 1996). Fig. 3. Lateral debris flow deposits in the middle sector of the Rotian creek. This levee was likely formed by the last surge, after the collapse of the check dams and the resulting channel incision. Note a boulder deposited on the top of the bank in the first part of the event. 4.3 Geomorphic impacts of the event and the amplification of sediment volume due to check dams’ failure The debris flow of 29 October 2018 caused substantial erosion along the channel, especially in the sector affected by the failure of the check dams. Erosion processes involved channel bed incision as well as erosion and destabilization of the sideslopes consisting of moraines and other Quaternary deposits. Field surveys have shown that shallow landslides occurred only along the channel banks and on valley slopes directly connected to the Rotian creek, whereas most of the catchment area remained stable. The retention basins built downstream of the check dams array favoured the deposition of part of the sediment transported by the flow, but they were insufficient to store the entire amount. Most of the material eroded from the upper and middle reaches of the Rotian creek was transferred through the rocky gorge to the debris flow fan. To assess the changes in the elevation of the surface caused by the debris flow of 29 October 2018 - and hence the volume of debris eroded and deposited - a Digital Terrain model (DTM) obtained from surveys done after the event was compared with pre-event DTMs. The Autonomous Province of Trento carried out an UAV-based photogrammetric survey of the debris flow fan two days after the event. The point cloud (estimated error of 0.05 m with reference to 15 control points) was created using the Agisoft Metashape © software, and a raster with cell size of 0.1 m was created. A few days later (7 November), the debris basins were surveyed using the same technique and resolution. The post-event survey was completed on 14-15 June 2019 with the aerial LiDAR survey of the entire channel. Pre-event topography of the Rotian creek consists of two LiDAR-derived DTMs from surveys that cover the entire territory of the Province of Trento. The LiDAR surveys had been done in September-October 2014 (DEM resolution of 0.5 m) and 2008 (DTM resolution of 1 m). Although they derive from surveys done several years before the debris flow of October 2018, both DTMs have been considered representative of pre-event conditions because no flow events or landslides that could have changed the topography of the basin-fan system have occurred. It should be remembered, however, that the 2014 LiDAR DTM (still not officially validated) shows some quality issues, especially in vegetated areas. We preferred to use the 2008 DTM as representative of the pre-event conditions, and we used the 2014 DTM only for the sector of the debris basins, which had not yet been built in 2008. The comparison of the pre- and post-event DEMs required remapping them to the lowest resolution (1 m), which is anyway fully satisfactory for the objectives of the analysis: this was done by computing the mean elevation of the pixels included in a 1-meter cell. The co-registration of the DEMs was necessary because of a systematic shift among them: it was implemented using the code GRD-CoReg (Cucchiaro et al., 2020). The extent of the areas where the debris flow produced topographic changes and the depth of such changes were assessed using the tool GCD (Geomorphic Changes Detection) 7.5, freely available at https://gcd.riverscapes.net/ . GCD includes several methods that enable DTM differencing taking into account the quality of the available DEMs (Wheaton et al., 2009). We used the fixed-threshold method based on uniformly propagated error, initially proposed by (Brasington et al., 2003). This method defines a minimum level of detection under which the elevation changes are deemed not significant. The minimum level of detection considers the errors in the pre-and post-event DEMs that propagate into the DoD. DTM differencing was computed separately for the debris flow fan and for the channel of the Rotian creek. The total volume of the deposits on the debris flow fan amounted to about 157000 m 3 ± 27000 m 3 (17% error). Erosion on the debris flow fan was much smaller but not negligible, and amounted to 5300 ± 2000m 3 : it is possible that erosion was due to just-post-event removal of debris that also affected the pre-event surface or localized incisions along the flow path. DTM differencing along the Rotian creek was computed on a mask that encompasses the channel bed and the side slopes that show evidence of erosion and/or shallow instability. Erosion amounts to 207000 ± 28000 m 3 (13% error), while the deposits, which occurred essentially in the debris basins, amounted to 39000 ± 7000 m 3 (17% error). Summing the volumes deposited on the debris flow fan and within the catchment we get a total volume of about 196000 m 3 , quite close to the total volume eroded from the catchment and within its uncertainty range. In addition to the approximations intrinsic to DoD computation, other sources of uncertainties affect the assessment of eroded and deposited volumes and their comparison: - the sediment that reached the Noce River was not computed in DTM differencing; - changes in porosity between the source material and the deposits are not accounted for; - large wood is computed in the deposits but not in the eroded volume because DoD is computed based on DTM derived from bare ground points, thus not considering the trees in the channel bed and banks that became the main source of large wood. The last of the above-mentioned issues are likely not relevant for the comparison of eroded and deposited volumes as the volume of large wood deposited on the debris flow fan amounts to 1500 steres (approx. 1000 m 3 ). While this volume indicates a substantial supply of large wood from forest-mantled channel bed and sideslopes, it is almost negligible when compared to the sediment volume mobilized by the 29 October debris flow in the Rotian creek. The comparison of the volume of the deposits of the 29 October 2018 debris flow in the Rotian creek with other debris flows in the Eastern Italian Alps may indicate the severity of this event. A dataset of 808 debris flows in northeastern Italy documented from historical archives, post-event surveys and - in a few cases - monitoring in instrumented catchments was used to explore, by means of quantile regression, the scaling relationships between catchment area and debris-flow volume (Marchi et al., 2019). Figure 4 compares the debris-flow deposits of the Rotian creek with debris-flow volumes collected in the same region and the scaling relationships between debris-flow volume and catchment area for selected quantiles. If the total volume of the deposits is considered (i.e. the sum of the deposits on the debris flow fan and in the debris basins within the catchment, about 196000 m 3 ), the debris flow of 29 October 2018 corresponds to the 99th percentile of debris-flow volumes in the Eastern Italian Alps. Considering that most historical debris flows in the sample utilized for developing the scaling relationships neglect within catchment deposits, the deposits on the debrtis flow fan (about 157000 m 3 ) are more consistent with the sample. In this case, the volume of the Rotian creek debris flow lies between the 98th and the 99th percentile. In either case, such results indicate the extreme severity of the debris flows under study. The minimum sediment volume contribution from the collapse of the check dams could be approximated as the volume stored (and released) by the check dams, about 115000 m³, and doesn’t consider its erosion/sedimentation effect. The total volume of the debris flow would have been about 81000 m³. A maximum value of the contribution from the system failure can be evaluated considering also its erosion/sedimentation effect, by the difference between the total volume of deposits and the volume mobilized upstream the first check dam (about 50000 m³ from the DoD analysis), thus resulting in about 146000 m³. Thus, the amplification factor, that is the ratio between the total volume of deposit (about 196000 m³) and the volume of the debris flow (50000-81000 m³), ranges between 2.4 and 3.9. Considering these volumes, in Fig. 4 the debris-flow event is located below the line of 98%, but still in the upper sector of the plot of debris-flow volumes versus basin area. Figure 4. Comparison of the volume of the Rotian creek debris flow with data on debris-flow volumes in the Eastern Italian Alps and scaling relationships with basin area (modified from Marchi et al., 2019), for 5 percentiles (1, 2, 50, 98, 99%). For the Rotian debris flow of October 2018, three volumes are indicated: total deposits, debris flow fan deposits, range of the estimated debris flow volume excluding the sediment released by check dam system failure. 4.4 Debris flow triggering and subsequent evolution According to post-event surveys, the debris flow was triggered by a shallow landslide mobilised on the left slope in the upper portion of the catchment, corresponding to a drainage area of 0.22 km 2 (Figure 2). The peak liquid discharge in the Rotian channel at that section, and its uncertainty, was quantified based on the indirect methodology developed by Amponsah et al. (2016), with a central value of 0.9 m 3 /s and uncertainty ranging from 0.55 to 1.25 m 3 /s. These values were used to evaluate the accuracy of the integrated hydrology-hydraulic debris-flow model developed by Rosatti and Zugliani (2024). Classical methodologies for debris-flow simulation in a given channel reach (Martinengo et al., 2021) usually start with the evaluation of the liquid discharge, then the solid-liquid mixture is considered (Takahashi, 1978; Rosatti et al., 2019), and finally the debris-flow dynamics is simulated (Rosatti et al., 2018). However, the Rotian case places specific challenges this three-steps methodology. Indeed, the debris flow should be simulated along the whole basin, where the liquid discharge continuously changes moving downstream, and a single input hydrograph is no more relevant in this case. Moreover, the morphological variations, especially the erosions, are significant and in some zones greater than 10 m, leading to an important modification in the flow dynamics, which makes the classical fixed-bed approach insufficient. To account for these challenges, a new integrated model, called TRENT2D MBRR (Mobile Bed Rainfall Runoff), was developed. The model combines the capabilities of the two-phase isokinetic TRENT2D model to simulate the dynamic of debris flow flowing over erodible and non-erodible zones (Amadii et al., 2022; Amaddii et al., 2023; Rosatti and Zugliani, 2015) with a description of the hydrological response along the river reach, by allowing the spatially distributed simulation of infiltration and subsurface-surface flow at the basin scale. The infiltration process is simulated by the well-known Green-Ampt equation, while the sub-surface flow is described by the using the Darcy law and the extended Dupuit-Forchheimer hypothesis. The model also includes the simulation of the debris-flow triggering mechanism (see Rosatti and Zugliani, 2024 for details). The integrated model does not account for the check dam system failure. Thus, the simulation of the debris flow is here limited to the basin closed at the first check dam of the protection system (dam number 16, see Figure 2, with a drainage area of 1.02 km 2 ). Check dam number 16 collapsed first and most upstream in the cascading failure, hence it represents the limit to the domain where the debris flow was not affected by the failure of the protection works. The model application to this part of the basin provides both mixture (solid plus liquid) hydrographs in different channel sections and maps of morphological variations. Figure 5a shows the liquid hydrograph simulated in the section just upstream the triggering zone, together with a comparison with the indirect peak estimate. Figure 5b reports the mixture and solid hydrographs obtained just upstream check dam 16. The liquid peak discharge at the triggering section amounts to 1 m 3 /s at 18:50 CET with a slightly lower second peak of 0.9 m 3 /s at 19.30 CET. The simulated solid discharge at check dam 16 starts around 17:30 CET during the most intense phase of the event and shows a peak slightly greater than 1 m 3 /s at 19:35 CET. The mixture discharge exhibits a well-defined peak of 3.3 m 3 /s at the same time of the solid discharge peak (19:35 CET). Figure 5. Simulated hydrographs: a) liquid hydrograph obtained in a section near the triggering zone, and comparison with the indirect peak estimate, b) mixture and solid hydrographs obtained near the check dam 16. Figure 6. Morphological variations in the Rotian creek upstream of check dam 16 (green line): a) by using the procedure described in Section 4.3, and b) as computed with the TRENT2D MBRR at the end of the simulation (3:00 CET of 30/10/2018). The yellow line indicates the location of a forest road crossing the creek. Figure 6 compares the morphological variations evaluated for the upper part of the Rotian creek following the procedure described in Section 2.3 (panel a) with those simulated by the numerical model at the end of the simulation (3:00 CET of 30/10/2018) (panel b). The comparison reported in Figure 6 shows that in the upper part of the creek, i.e. from the debris flow triggering zone until the crossing of a forest road (yellow line in Figure 6), the agreement is satisfactory, although with a moderate estimation in the modelled solid volume. The net solid volume mobilised in this area is 15400 m 3 , about 9% greater than the upper limit value estimated from the DoD. Furthermore, except for the largest shallow landslide present on the left, also the spatial distribution of the morphological variation is consistent with the DoD both in terms of eroded depth and involved area. Instead, in the lower part, i.e. from the forest road to the check dam 16, the results are less satisfactory since the simulated net solid volume mobilised in this zone (21700 m 3 ) is 22% less than the lower limit value of the DoD. 4.5 Check dam failure mode A close examination of the check dam system after the event revealed that the system failure was triggered by the collapse of the most upstream check dam 16. With a total width of 29 m and a height of 8.5 m, this check dam is the highest in the check dam system. The inspection revealed that the check dam was destroyed in the joint-to-joint section in the wing section on the left bank side of the dam, due to the combined action of slipping of the reinforcing bars and the overload caused by the debris flow, as simulated based on results from section 2.5. Approximately 10000 kN of heavy fragments of the check dam were transported by the debris flow and impacted the downstream check dams triggering their collapse due to internal stability failure. The peak discharge of the debris flow which resulted from the release of the sediment behind the check dam was much larger than the designed peak discharge calculated from the design rainfall intensity. Further, a field survey also indicated that the main destruction mode was overturning rather than sliding. Thus, the collapse of the check dam 16 led to the failure of the whole system of 16 check dams. Also, the open check dam downstream the sequence of 5 retention basins was damaged - and partly outflanked - by the debris flow. A view of the upper part of the collapsed system is reported in Figure 7. The figure also shows a check dam (no. 7, marked in Figure 2) that did not fail but was under-excavated, probably because of the failure of other dams. The deposits visible over the wing of the dam also suggest that it had been overtopped by the debris flow before its under-excavation. Figure 7. Check dams after the failure event: (a) view of the upper part of the collapsed system; (b) check dam no. 7: the deposits over the wing of the dam indicate that it had been overtopped by the debris flow before its under-excavation. 5. Discussion The collected post-event data and observations provide valuable insights into the performance of the integrated hydrologic-hydraulic-debris-flow model application. The simulated peak hydrograph timing in Figure 5 can be considered in good agreement with the debris-flow timing described in Section 2 (around 19:00 CET), whereas the simulated water discharge peak (1 m 3 /s) in well within the uncertainty band of the post-event based estimated peak (0.55-1.25 m3/s). Also, the model does not predict solid discharge along the event for the channel reaches located upstream of the debris flow triggering section, thus showing a good accuracy in the simulation of the relevant hydro-geomorphic processes in this area. An important result with regard to simulated hydrograph is the absence of the third wave, which means that the latter may be not directly related to the rainfall event, but rather to the partial collapse of one of the check dams or one of the small shallow-landslides that occurred just after the event. Comparing the liquid discharge predicted at the triggering section and check dam solid discharges in Figure 5, it appears that the modelled solid discharge starts one hour before the moment of the debris flow triggering. The explanation for this anticipated solid discharge might be related to some minor erosion processes occurring in the section upstream of check dam 16. On the other hand, the absence of significant solid transport in the phase preceding the debris flow was also noted by the operators performing the event monitoring. The solid discharge simulated near the check dam 16 (Figure 5b) shows a long decreasing phase that appears in contrast with the observations of impulsive debris flow waves reported for the debris flow fan (Section 2.3). However, we should note that (i) the magnitude of the solid discharge in the decreasing phase is limited (around 0.5 m3/s), (ii) the simulation was only carried out for the basin closed near check dam 16, thus far away from the debris flow fan since the collapse of the dams is not considered in the model yet, (iii) the check dam system collapse releases a massive amount of sediment in a short period, probably causing the high and impulsive sediment discharges recorded on the debris flow fan. These points suggest that the modelled decreasing phase of the solid discharge is reasonable and falls within the uncertainties of event reconstruction. The comparison of the morphological variations simulated by the hydraulic model for the Rotian creek upstream of check dam 16 (Figure 6) shows mixed results. In the upper part (until the crossing of a forest road (yellow line in Figure 6) the model matches the observations very well. This can be partially explained with the presence of a limited erodible depth. However, results reported for the lower reach (from the forest road to the check dam 16) are less satisfactory, with the spatial distribution of the morphological variation differing with respect to the results from the DoD in terms of eroded depth and involved area. The first zone where the discrepancy is relevant (i.e., just downstream the forest road) could be justified by the presence of a few large boulders under the road that concentrated the debris flow in a small area, a local effect not well captured by the model due to the spatial resolution of the used domain. Instead, the second discrepancy (i.e., the one close to the check dam 16) can be partly explained by the absence of the dam failure procedure in the numerical model, a phenomenon that had a significant effect on the extent of excavation in the areas close to these artefacts. However, the overall results of the model have to be considered satisfactory. The integrated analysis of both the debris flow and of the collapsed check dam system shows that a very likely cause of the critical failure of check dam 16 is the combination of a debris flow which is larger than the design load – the design only included liquid flood – and of inaccuracies in the building of the work. Results reported by Baggio and D’Agostino (2022) in their simulation of the Rotian creek debris flow largely agree with our findings. However, Baggio and D’Agostino (2022) interpreted the debris flow of Rotian creek in October 2018 as a muddy debris flow, a term that applies to debris flows rich in cohesive clayey material (Coussot and Meunier, 1996), whereas the morphological and sedimentological evidences of the deposits described in Section 2.3 (abundant presence of large cobbles and boulders in clast-supported lateral levees, and sandy matrix with very limited presence of fines) led us to classify the event under study as a granular (i.e., stony) debris flow, and to model it accordingly. 6. Conclusions This paper highlights the critical need to address the complexities surrounding check dam system failures and their implications for residual risk management. The proposed forensic model includes several analyses providing information of the evolution of the event, from the triggering rainfall to the reconstruction of the flood and hydrograph, and information on the check dam failure mode. Its application to the Rotian creek case study provides a framework for understanding how the collapse of check dams can significantly amplify debris flow magnitude, leading to severe consequences for downstream areas, and for developing similar future analysis. Also, it illustrated how this analysis could improve our understanding on the collapse mode, which could be used for advancing our understanding and preparedness to these events, as discussed in the following. We identify three aspects where the forensic model may have important practical consequences on residual risk management. i) Development of integrated debris flow model able to account for check dam failures The collapse of flood protection systems is generally ignored in current models of debris flow simulation. The processes by which check dam failures amplify the scale of a debris flow, altering the flow velocity, mixture discharge and volumetric solid concentration, are complex and may depend on the geometry and construction technique of the protective structure, as well as the topography and morphology of the channel. Integrated models able to incorporate the effect of check dam failures could shed light on the amplification effect, thus quantifying the changes in the debris flow magnitude and its destructive capacity. ii) Consolidating existing datasets of check dam system failures Failures of debris-flow mitigation works are rarely reported in the scientific literature, presumably due to embarrassment or the threat of potential legal action. This is lamentable, as the greatest advances in debris-flow mitigation may derive from a detailed examination of past failures. To benefit from these learning opportunities, there is a pressing need to consolidate physical processes knowledge, engineering analysis, and social datasets related to check dam system failures. Integrating these diverse datasets and developing post-event analysis can provide a more comprehensive understanding of the interaction between physical process and structures and the factors contributing to failures, including environmental conditions, design flaws, and community responses. By fostering collaboration among engineers, scientists, and local stakeholders, we can create a richer knowledge base that informs better design and management practices. iii) Methodologies for the prioritisation of field inspection and maintenance The post-event analysis conducted in this study underscores the critical role of check dam failure in exacerbating debris flow and its associated damages, thereby highlighting the urgent need for further site-specific analyses and/or maintenance works on these structures. Establishing priorities for field inspection and potential repairs of check dams necessitates an approach that is applicable at the regional scale. Methodologies for prioritizing inspection and maintenance are essential, as they enable the efficient allocation of resources, ultimately reducing risks to infrastructure and enhancing community resilience against natural disasters. To facilitate this, a straightforward index is under evaluation that incorporates both variables related to the check dam, such as age, height, and material, and site-specific factors, including the presence of landslides and geolithology. Forensic analysis of flow events that involved check dam failure, such as the Rotian creek 2018 debris flow, provides important information for the choice and weighting of the variables for the indices of check dam fragility. At the same time, the modelling techniques developed in post-event studies can be applied to evaluate management alternatives for check dam systems potentially susceptible to failure. Declarations Fundings This research was supported by the GPR Project (“Approfondimento delle strategie di Governo della Pericolosità alluvionale a seguito dell’evento del 29 ottobre 2018 sul rio Rotiano” – Accordo di Programma GPR) funded by the Autonomous Province of Trento and by Fondazione Cassa di Risparmio di Padova e Rovigo (Excellence Grant 2021 to the Resilience Project). E. Dallan was supported by the RETURN Extended Partnership and received funding from the European Union Next-GenerationEU (National Recovery and Resilience Plan – NRPP, Mission 4, Component 2, Investment 1.3 – D.D. 1243 2/8/2022, PE0000005). Competing Interests he authors have no relevant financial or non-financial interests to disclose. Data availability The tool GCD (Geomorphic Changes Detection) 7.5 is freely available at https://gcd.riverscapes.net/ (last access: November 2024). Authors’ contribution. Conceptualization: LM, GR, RV, MB; Data curation: all authors; Formal analysis: all authors; Funding acquisition: GR, LM, MB; Visualization: ED, LM, SC, DZ; Writing – original draft: ED, LM, DZ, MB; Writing – review & editing: all authors. References Amaddii, M., Rosatti, G., Zugliani, D., Marzini, L., Disperati, L., 2022. Back-Analysis of the Abbadia San Salvatore (Mt. Amiata, Italy) Debris Flow of 27–28 July 2019: An Integrated Multidisciplinary Approach to a Challenging Case Study. Geosciences, 385(12), 1-25. DOI: 10.3390/geosciences12100385 Amaddii, M., Rosatti, G., Zugliani, D., Marzini, L., Disperati, L., 2023. Modelling stony debris flows involving culverted streams: the Abbadia San Salvatore case (Mt. Amiata, Italy). Rendiconti Online della Società Geologica Italiana, 61, 108-115. DOI: 10.3301/ROL.2023.55 Amponsah, W., L. Marchi, D. Zoccatelli, G. Boni, M. Cavalli, F. Comiti, S. Crema, A. Lucía, F. Marra, M. Borga, 2016. Hydrometeorological characterisation of a flash flood associated to major geomorphic effects: Assessment of peak discharge uncertainties and analysis of the runoff response. J. Hydrometeorology, 17(12), 3063-3077 Baggio T., D’Agostino V.,2022: Simulating the effect of check dam collapse in a debris-flow channel. Sci. Total Environ., 816 (2022), Article 151660, 10.1016/j.scitotenv.2021.151660 Ballesteros-Cánovas, J.A., Stoffel, M., de Haas, T., Bodoque, J.M., 2024: Debris Flow Dating and Magnitude Reconstruction. In: Jakob, M., McDougall, S., Santi, P. (eds) Advances in Debris-flow Science and Practice. Geoenvironmental Disaster Reduction. Springer, Cham. https://doi.org/10.1007/978-3-031-48691-3_8 Benito G., Grodek T., Enzel Y., 1998:. The geomorphic and hydrologic impacts of the catastrophic failure of flood-control-dams during the 1996-Biescas flood (Central Pyrenees, Spain), Zeitschrift Fur Geomorphol., 42 (1998), pp. 417-437, 10.1127/zfg/42/1998/417 Bertola, M., Viglione, A., Lun, D., Hall, J., Blöschl, G., 2020: Flood trends in Europe: Are changes in small and big floods different? Hydrology and Earth System Sciences, 1805–1822. https://doi.org/10.5194/hess-24-1805-2020. Borga, M., M. Stoffel, L. Marchi, F. Marra, M. Jakob, 2014: Hydrogeomorphic response to extreme rainfall in headwater systems: flash floods and debris flows. Journal of Hydrology, 518, 194–205, http://dx.doi.org/10.1016/j.jhydrol.2014.05.022, ISSN: 0022-1694 Borga M, Comiti F, Ruin I, Marra F., 2019: Forensic analysis of flash flood response. WIREs Water. 6:e1338. https://doi.org/10.1002/wat2.1338 Borga, M., Zaramella, M., 2020. Evento di piena del 27-29 ottobre 2018 sul bacino del Rotian creek: stima della precipitazione e valutazione della sua severità, Project Report, https://www.tesaf.unipd.it/en/sites/tesaf.unipd.it.en/files/Progetto%20VAIA_R01.2%20F.pdf. Bossi G., Cavalli M., Mantovani M., Catelan F. T., Ballaera A., Ceccotto F., Marcato G., Pasuto A, 2024. Expecting the expected – learning from the past to provide forward scenarios through geomorphic change detection, monitoring and modeling. Geoenvironmental Disasters, 11, 35, https://doi.org/10.1186/s40677-024-00292-7. Brasington, J., Langham, J., & Rumsby, B., 2003: Methodological sensitivity of morphometric estimates of coarse fluvial sediment transport. Geomorphology, 53 (3-4), 299–316. https://doi.org/10.1016/s0169-555x(02)00320-3 Bronstert, A. Agarwal, B. Boessenkool, I. Crisologo, M. Fischer, M. Heistermann, L. Köhn-Reich, J.A. López-Tarazón, T. Moran, U. Ozturk, C. Reinhardt-Imjela, D. Wendi, 2018: Forensic hydro-meteorological analysis of an extreme flash flood: the 2016-05-29 event in Braunsbach, SW Germany, Sci. Total Environ., 630, pp. 977-991, 10.1016/j.scitotenv.2018.02.241 Bull J.M., Miller H., Gravley D.M., Costello D., Hikuroa D.C.H., Dix J.K., 2010: Assessing debris flows using LIDAR differencing: 18 May 2005 Matata event, New Zealand. Geomorphology, 124 (1-2), 75-84, DOI: 10.1016/j.geomorph.2010.08.011 Cavalli, M., Goldin, B., Comiti, F., Brardinoni, F., Marchi, L., 2017: Assessment of erosion and deposition in steep mountain basins by differencing sequential digital terrain models. Geomorphology, 291, 4-16, doi:10.1016/j.geomorph.2016.04.009 Coussot P., Meunier M., 1996: Recognition, classification and mechanical description of debris flows. Earth-Science Reviews, 40, 209-227. Cucchiaro, S., Cavalli, M., Vericat, D., Crema, S., Llena, M., Beinat, A., Marchi, L., & Cazorzi, F.,2019: Geomorphic effectiveness of check dams in a debris-flow catchment using multi-temporal topographic surveys. Catena, 174, 73–83. hiips://doi.org/10.1016/j.catena.2018.11.004 Cucchiaro, S., Maset, E., Cavalli, M., Crema, S., Marchi, L., Beinat, A., & Cazorzi, F., 2020: How does co-registration affect geomorphic change estimates in multitemporal surveys? GIScience & Remote Sensing, 57 (5), 611–632, https://doi.org/10.1080/15481603.2020.1763048 Dallan, E., Borga M., Zaramella M., Marra F., 2022: Enhanced summer convection explains observed trends in extreme subdaily precipitation in the Eastern Italian Alps, Geophysical Research Letters, 49, e2021GL096727, 2022. Dallan, E., Borga, M., Fosser, G., Canale, A., Roghani, B., Marani, M., Marra, F., 2024a: A method to assess and explain changes in sub‐daily precipitation return levels from convection‐permitting simulations, Water Resources Research, 60, e2023WR035969, 2024. https://doi.org/10.1029/2023WR035969 Dallan E., Marra F., Fosser G., Marani M., Borga M., 2024b: Dynamical factors heavily modulate the future increase of sub-daily extreme precipitation in the Alpine-Mediterranean region. Accepted in Earth’s Future. Dal Piaz, G.V., Castellarin, A., Martin, S., Selli, L., Carton, A., Pellegrini, G.B., Casolari, A., Daminato, F., Montresor, L., Picotti, V., Prosser, G., Santuliana, G., Cantelli, L., 2007: Note Illustrative della Carta Geologica d'Italia alla scala 1:50.000. Foglio 042 Malè. Provincia Autonoma di Trento, ISPRA. - System Cart Roma : Regione Lombardia, APAT, 2007 (in Italian). Dell’Agnese A., Mazzorana B., Comiti F., Von Maravic P., D’Agostino V., 2013:. Assessing the physical vulnerability of check dams through an empirical damage index. J Agric Eng [Internet]. 2013 Jun. 14;44(1):e2. Available from: https://www.agroengineering.org/jae/article/view/jae.2013.e2 Formetta G., F. Marra, E. Dallan, M. Zaramella, M. Borga, 2022: Differential orographic impact on sub-hourly, hourly, and daily extreme precipitation. Adv. Water Resour., 159 (2022), Article 104085, 10.1016/j.advwatres.2021.104085 Fuchs S., Heiss K., Hübl J., 2007: Towards an empirical vulnerability function for use in debris flow risk assessment. Nat. Hazards Earth Syst. Sci., 7, pp. 495-506, 10.5194/nhess-7-495-2007 Giovannini L., Davolio S., Zaramella M., Zardi D., Borga M., 2021: Multi-model convection-resolving simulations of the October 2018 Vaia storm over northeastern Italy, Atmospheric Res., 253,105455. Grodek, T. and Benito, G., 2024: Reevaluating Flood Protection: Disaster Risk Reduction for Urbanized Alluvial Fans, Nat. Hazards Earth Syst. Sci. Discuss. [preprint], https://doi.org/10.5194/nhess-2024-171, under review. Hübl, J., Suda, J., Uchida, T., & Nagl, G., 2024:. Check Dam Failures. In Advances in Debris-flow Science and Practice (pp. 565-588). Cham: Springer International Publishing. Jakob, M., 2005: A size classification for debris flows. Engineering geology, 79(3-4), 151-161. Jakob, M., 2019: Debris-flow hazard assessments – a practitioner’s view. 7th International Conference on Debris-Flow Hazards Mitigation, Golden, Colorado, USA. Keating, A., K. Venkateswaran, M. Szoenyi, K. MacClune, and R. Mechler. 2016. From event analysis to global lessons: Disaster forensics for building resilience. Natural Hazards and Earth System Sciences 16(7): 1603–1616 Larese, A. et al., 2023: Multiphysics simulation of the impact of natural hazards on structures and protection systems. 2 nd GACM-GIMC Joint Workshop, RWTH Aachen University, September 14-15, 2023 Libertino, A., Ganora, D., & Claps, P., 2019: Evidence for increasing rainfall extremes remains elusive at large spatial scales: The case of Italy. Geophysical Research Letters, 46(13), 7437-7446. Marchi, L., Brunetti, M.T., Cavalli, M., Crema, S., 2019. Debris-flow volumes in northeastern Italy: relationship with drainage area and size probability. Earth Surface Processes and Landforms, 44(4), 933-943, doi: 10.1002/esp.4546 Marchi, L., Cavalli, M., 2007: Procedures for the Documentation of Historical Debris Flows: Application to the Chieppena Torrent (Italian Alps). Environmental Management, 40, 493-503. Marra, F., E. I. Nikolopoulos, J. D. Creutin, M. Borga, 2014: Radar rainfall estimation for the identification of debris-flow occurrence thresholds. Journal of Hydrology, Volume 519, Part B, 1607-1619, http://dx.doi.org/10.1016/j.jhydrol.2014.09.039, ISSN: 0022-1694. Martinengo, M., Zugliani, D., Rosatti, G., 2021: Uncertainty analysis of a rainfall threshold estimate for stony debris flow based on the backward dynamical approach. Natural hazard and Earth System Sciences, 21(6), 1769-1784. DOI: 10.5194/nhess-21-1769-2021 Mazzorana, B., Trenkwalder-Platzer, H., Fuchs, S. et al., 2014: The susceptibility of consolidation check dams as a key factor for maintenance planning. Österr Wasser- und Abfallw 66, 214–216, https://doi.org/10.1007/s00506-014-0160-4 Pesce M., Dallan E., Marra F., Borga M., 2024: Increasing probability of record-breaking precipitation: a case-study in the Eastern Italian Alps. J. Hydrol., Regional studies, under review. Piton and co-authors, 2016: Why do we build check dams in Alpine streams? An historical perspective from the French experience. Earth Surf. Process. Landforms 42, 91–108. Ranzi, R., Barontini, S., Ferri, M, 2015: Structural Residual Risk Due to Levee Failures in Flood Mapping. In: Lollino, G., Arattano, M., Rinaldi, M., Giustolisi, O., Marechal, JC., Grant, G. (eds) Engineering Geology for Society and Territory - Volume 3. Springer, Cham. https://doi.org/10.1007/978-3-319-09054-2_92 Rickenmann, D., 2005: Runout prediction methods. In: Jakob, M., Hungr, O. (eds), Debris-flow hazards and related phenomena, Praxis Springer, Berlin Heidelberg, 305-324 Rosatti G., Zugliani D., Fraccarollo L., 2023: The debris flow event of 29 October 2018 in the Rio Rotiano (Italy) and its challenges for the mathematical and numerical modelling. E3S Web of Conf., 415 (2023) 05022. https://doi.org/10.1051/e3sconf/202341505022 Rosatti, G., Zorzi, N., Zugliani, D., Piffer, S., Rizzi, A., 2018. A Web Service ecosystem for high-quality, cost-effective debris-flow hazard assessment. Environmental Modelling & Software, 100, 33-47. DOI: 10.1016/j.envsoft.2017.11.017 Rosatti, G., Zugliani, D., 2015: Modelling the transition between fixed and mobile bed conditions in two-phase free-surface flows: The Composite Riemann Problem and its numerical solution. Journal of Computational Physics, 285, 226-250. DOI: 10.1016/j.jcp.2015.01.011 Rosatti, G., Zugliani, D., 2024: TRENT2DMBRR: an integrated mobile-bed rainfall-runoff model for the simulation of debris flow at the basin scale, from the triggering to the deposition phase. Proceedings of ICIRBM 2024 “Technologies for Integrated River Basin Management – Tecniche per la Difesa del Suolo e dall’Inquinamento” G. Frega and F. Macchione (eds.), ISBN: 978‐88‐97181‐90‐3, ISSN: 2282‐5517. Rosatti, G., Zugliani, D., Pirulli, M., Martinengo, M., 2019: A new method for evaluating stony debris flow rainfall thresholds: the Backward Dynamical Approach. Heliyon 5(6), e01994. DOI: 10.1016/j.heliyon.2019.e01994. Scheidl C., Rickenmann D., Chiari M., 2008: The use of airborne LiDAR data for the analysis of debris flow events in Switzerland. Natural Hazards and Earth System Science, 8 (5), 1113-1127, www.nat-hazards-earth-syst-sci.net/8/1113/2008/ Sioni F., Davolio S., Grazzini F., Giovannini L., 2023: Revisiting the atmospheric dynamics of the two century floods over north-eastern Italy. Atmos. Res., 286, 10.1016/j.atmosres.2023.106662 Sodnik, J., Martinčič, M., Mikoš, M., Kryžanowski, A., 2015: Are Torrent Check-Dams Potential Debris-Flow Sources?. In: Lollino, G., et al. Engineering Geology for Society and Territory - Volume 2. Springer, Cham. https://doi.org/10.1007/978-3-319-09057-3_79 Takahashi, T., 1978: Mechanical Characteristics of Debris Flow. Journal of the Hydraulics Division, 104 , 1153-1169. DOI:10.1061/jyceaj.0005046.664 Wang G.L., 2013: Lessons learned from protective measures associated with the 2010 Zhouqu debris flow disaster in China, Nat. Hazards, 69 (2013), pp. 1835-1847, 10.1007/s11069-013-0772-1 Wheaton, J. M., Brasington, J., Darby, S. E., & Sear, D. A., 2009: Accounting for uncertainty in DEMs from repeat topographic surveys: improved sediment budgets. Earth Surface Processes and Landforms, 136–156. hiips://doi.org/10.1002/esp.1886 Cite Share Download PDF Status: Published Journal Publication published 13 May, 2025 Read the published version in Natural Hazards → Version 1 posted Editorial decision: Major revisions 11 Feb, 2025 Reviewers agreed at journal 20 Dec, 2024 Reviewers invited by journal 20 Dec, 2024 Editor assigned by journal 20 Nov, 2024 First submitted to journal 18 Nov, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5478044","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":392911538,"identity":"d4f78cd8-075f-42a8-a7c6-3730487a9d6e","order_by":0,"name":"Eleonora Dallan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAUlEQVRIie3PsWrDMBCA4TMGeVHrVSHQvIKMoXQpeZUTHrQ0EOjioQSBQdnatYU+hPsGCoJOeQBDQ1EJePKQKWNpPDSlBcVrB/3TLR93BxAK/ceS44SRcQAXAJECCsxP4h8CBgHyb+I3f4lQ/UzBvyaN49V2Xr5PICmcwbuNrN8q7Tq4WvjIqCJF/ri+zRRtucHXdlZvVsvs+cRh3NLL8ZnGSDE8EGJndSP0+NQvU5vuezJVTO4MflrJhwiPKemJUOyGG6EtDhJmSZ7TNRaadnMj7m321IjDL5yNlIeky+pjS0u8fkjki9vt7eS8ka3rykXq23KM/J75IPDzUCgUCsEXvJtV9Yh+HG4AAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-2113-8861","institution":"University of Padova: Universita degli Studi di Padova","correspondingAuthor":true,"prefix":"","firstName":"Eleonora","middleName":"","lastName":"Dallan","suffix":""},{"id":392911539,"identity":"c390ad97-96a9-43d0-98c3-78e8c2d97935","order_by":1,"name":"Lorenzo Marchi","email":"","orcid":"","institution":"IRPI-CNR: Istituto di Ricerca per la Protezione Idrogeologica Consiglio Nazionale delle Ricerche","correspondingAuthor":false,"prefix":"","firstName":"Lorenzo","middleName":"","lastName":"Marchi","suffix":""},{"id":392911540,"identity":"34c9da6d-eb10-4db2-b1bb-9866f930ec31","order_by":2,"name":"Giorgio Rosatti","email":"","orcid":"","institution":"University of Trento: Universita degli Studi di Trento","correspondingAuthor":false,"prefix":"","firstName":"Giorgio","middleName":"","lastName":"Rosatti","suffix":""},{"id":392911541,"identity":"26c21635-cec1-4211-bc54-3ff8e9f9541e","order_by":3,"name":"Daniel Zugliani","email":"","orcid":"","institution":"University of Trento: Universita degli Studi di Trento","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Zugliani","suffix":""},{"id":392911542,"identity":"ac6b05e8-cd92-451b-83d9-369545ff9276","order_by":4,"name":"Marco Cavalli","email":"","orcid":"","institution":"IRPI-CNR: Istituto di Ricerca per la Protezione Idrogeologica Consiglio Nazionale delle Ricerche","correspondingAuthor":false,"prefix":"","firstName":"Marco","middleName":"","lastName":"Cavalli","suffix":""},{"id":392911543,"identity":"a10279e6-d5ac-4ff9-87b5-c189a46c4e42","order_by":5,"name":"Stefano Crema","email":"","orcid":"","institution":"IRPI-CNR: Istituto di Ricerca per la Protezione Idrogeologica Consiglio Nazionale delle Ricerche","correspondingAuthor":false,"prefix":"","firstName":"Stefano","middleName":"","lastName":"Crema","suffix":""},{"id":392911544,"identity":"095ebc0c-0fbe-4ae8-b629-de3576633af5","order_by":6,"name":"Ruggero Valentinotti","email":"","orcid":"","institution":"Provincia Autonoma di Trento","correspondingAuthor":false,"prefix":"","firstName":"Ruggero","middleName":"","lastName":"Valentinotti","suffix":""},{"id":392911545,"identity":"0636a323-20bc-46c1-84e0-71e5b2260fd7","order_by":7,"name":"Marco Borga","email":"","orcid":"","institution":"University of Padova: Universita degli Studi di Padova","correspondingAuthor":false,"prefix":"","firstName":"Marco","middleName":"","lastName":"Borga","suffix":""}],"badges":[],"createdAt":"2024-11-18 18:16:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5478044/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5478044/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11069-025-07301-4","type":"published","date":"2025-05-13T15:57:29+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":72141549,"identity":"a9597f85-4f42-4f09-8713-157be238eacc","added_by":"auto","created_at":"2024-12-23 06:43:41","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1250482,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart for integrated forensic analysis of the coupled debris flow – check dam collapse.\u003c/p\u003e","description":"","filename":"Fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/17e4d102f5ad621e2d00fe5e.jpg"},{"id":72140393,"identity":"406613b8-9a62-470c-9553-90546e10c92b","added_by":"auto","created_at":"2024-12-23 06:35:41","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":10476544,"visible":true,"origin":"","legend":"\u003cp\u003eLateral debris flow deposits in the middle sector of the Rotian creek. This levee was likely formed by the last surge, after the collapse of the check dams and the resulting channel incision. Note a boulder deposited on the top of the bank in the first part of the event.\u003c/p\u003e","description":"","filename":"Fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/cefc0c25019687279ef9ffc1.jpg"},{"id":72141740,"identity":"ad3ed9fe-38cb-4c9e-9d73-d1416c6304ca","added_by":"auto","created_at":"2024-12-23 06:51:41","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":178317,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of the volume of the Rotian creek debris flow with data on debris-flow volumes in the Eastern Italian Alps and scaling relationships with basin area (modified from Marchi et al., 2019), for 5 percentiles (1, 2, 50, 98, 99%). For the Rotian debris flow of October 2018, three volumes are indicated: total deposits, debris flow fan deposits, range of the estimated debris flow volume excluding the sediment released by check dam system failure.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/e4d6e84b8e4915cb919a946a.png"},{"id":72140394,"identity":"0697a1da-32c6-481b-8bf6-541a78a26d25","added_by":"auto","created_at":"2024-12-23 06:35:41","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":406192,"visible":true,"origin":"","legend":"\u003cp\u003eSimulated hydrographs: a) liquid hydrograph obtained in a section near the triggering zone, and comparison with the indirect peak estimate, b) mixture and solid hydrographs obtained near the check dam 16.\u003c/p\u003e","description":"","filename":"Fig5new.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/9ebba50bc0f156efcd0c5747.jpg"},{"id":72140397,"identity":"093ae701-f656-4de4-892d-02f457be7db2","added_by":"auto","created_at":"2024-12-23 06:35:42","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3116429,"visible":true,"origin":"","legend":"\u003cp\u003eMorphological variations in the Rotian creek upstream of check dam 16 (green line): a) by using the procedure described in Section 4.3, and b) as computed with the TRENT2D\u003csup\u003eMBRR\u003c/sup\u003e at the end of the simulation (3:00 CET of 30/10/2018). The yellow line indicates the location of a forest road crossing the creek.\u003c/p\u003e","description":"","filename":"fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/3a3517a7cb5082d80de47096.png"},{"id":72140392,"identity":"6b308c65-5cea-46f0-933c-7fe1d72a2c33","added_by":"auto","created_at":"2024-12-23 06:35:41","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":5919546,"visible":true,"origin":"","legend":"\u003cp\u003eCheck dams after the failure event: (a) view of the upper part of the collapsed system; (b) check dam no. 7: the deposits over the wing of the dam indicate that it had been overtopped by the debris flow before its under-excavation.\u003c/p\u003e","description":"","filename":"Fig7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/509e5634ccfb05d7c93c73ed.jpg"},{"id":83067940,"identity":"9826c886-f72c-4e23-82d5-4babd8b2697c","added_by":"auto","created_at":"2025-05-19 16:08:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":22169641,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5478044/v1/36c39343-4441-4a65-a969-77f21dc9bdf0.pdf"}],"financialInterests":"","formattedTitle":"What can be learnt from the catastrophic failure of a check dam system? A forensic analysis of a cascading natural-anthropogenic hazard","fulltext":[{"header":"Highlights","content":"\u003cul\u003e\n \u003cli\u003eA forensic analysis of debris-flow related check dam system failures is proposed\u003c/li\u003e\n \u003cli\u003eA catastrophic check dam system failure event is used to illustrate the framework\u003c/li\u003e\n \u003cli\u003eCheck dam failure greatly amplified the debris-flow total volume\u003c/li\u003e\n \u003cli\u003eCoupled hydrologic-hydraulic model helps ascertain the event evolution and volumes\u003c/li\u003e\n\u003cli\u003eWe suggest advances in debris flow modelling, data collection, structure inspection\u003c/li\u003e\u003c/ul\u003e"},{"header":"1.\tIntroduction","content":"\u003cp\u003eThousands of mountain channels in the Alps are equipped with check dam systems which require costly maintenance to avoid structural deterioration (Mazzorana et al., 2014). While check dams are generally effective in stabilizing mountain channels and retaining sediment transported by floods and debris flows, thus reducing hazard for settlements and transport routes, the failure or collapse of these mitigation structures may result in serious consequences for people and infrastructures (Piton et al., 2016). Failure not only removes the protection offered by the dam, but it can also lead to an increased supply of sediment downstream (Baggio and D\u0026rsquo;Agostino, 2022; Benito et al., 1998; Sodnik et al., 2015; Wang, 2013).\u0026nbsp;As a result, check dam systems failure is a rare but highly dangerous event that may result in much higher damages compared to a situation without structures, possibly also impacting areas that would otherwise be untouched by the event. The failure of check dams has a higher probability to occur and becomes potentially highly hazardous during large magnitude floods or debris flows: on the one hand, severe flow conditions induce stress on the structure that can lead to their collapse; on the other hand, their abrupt failure frees large sediment volumes which can be easily transported downstream under, typically reaching the debris flow fan.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA close examination of check dam system failures can help to better quantify and manage the residual risk associated with structural measures for the mitigation of flood and debris flow risk (Jakob, 2019). The examination of flood protection system failures is part of forensic engineering (Huebl et al., 2024), and it is based on accurate post-event analyses of both the relevant hydro-geomorphic processes and the structures. Whereas forensic methodologies for water floods and debris flows are well developed (Borga et al., 2014; 2019), literature is scarce for the case of the failures of check dam systems as a cascading natural-anthropogenic hazard (Huebl et al., 2024 and references therein; Marchi and Cavalli 2007; Bossi et al., 2024). This is understandable, as failures are tantamount to an admission of guilt at worst, or poor design work at best. This research gap is important given the ubiquity of ageing check dams in regulated steep streams, their difficult maintenance and the increasing intensity of flood events in small catchments related to climate change (Bertola et al., 2020; Dallan et al., 2024a,b).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWithin this background, this paper proposes a forensic model for debris-flow related check dam system failures. The model is tested on the catastrophic event in the Rotian catchment (2.5 km\u003csup\u003e2\u003c/sup\u003e in area, Eastern Italian Alps) occurred during the so-called \u0026ldquo;Vaia\u0026rdquo; storm, on 27- 29 October 2018. . The storm was characterized by exceptionally high cumulative rainfall depths and record wind velocities, and triggered in the Rotian Creek a debris flow. whose initial volume was greatly amplified by the collapse of 16 check dams built in the 1970s.\u003c/p\u003e\n\u003cp\u003eA forensic model is illustrated based on data from the post-event survey carried out a few months after the event. The application aims to show the added value of exploiting the observations to qualify the results from a newly developed coupled hydrologic-hydraulic debris flow mathematical-numerical model and to ascertain the failure mode of the check dam system collapse. We also identify aspects where forensic analysis of check dam failures will have important practical consequences on residual risk management.\u003c/p\u003e"},{"header":"2.\tForensic analysis of coupled debris flow and check dam system failure","content":"\u003cp\u003eForensic hydrology provides information and knowledge for extreme events (mostly floods and flash floods) where no direct systematic hydrometeorological data and observations are available (Bronstert et al., 2018; Borga et al., 2019). Hence, a central place in forensic hydrology is placed by the generation of data and observations from post-event surveys. This analysis borrows the term \u0026ldquo;forensics\u0026rdquo; from the field of criminal investigation, because it denotes a consistent approach to develop a comprehensive analysis of an event and its causes (Keating et al., 2016). While enabling a systematic approach, forensic hydrology analysis provides a consistent approach that facilitates cross-event learning, with a focus on quantifying flood peaks, timing, and volumes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDebris-flow magnitude relates to the volume of sediment mobilized, peak discharge, and inundation area (Ballesteros-C\u0026aacute;novas et al., 2024). When compared to water floods, forensic activities for debris flows put, in general, less emphasis on the reconstruction of the peak discharge (which, anyway, remains an important parameter in debris-flow assessment) and more attention to the assessment of the sediment volume involved. Debris-flow volume has direct relevance for assessing the severity of a debris flow and the design of control measures, has been proposed as a one of the key variables for classifying debris flows (Jakob, 2005), and is related to other important debris-flow variables (e.g. Rickenmann, 2005). Other focal points in the post-event recognition of debris flows are the identification of the initiation point(s) and the triggering mechanism, and the recognition of the characteristics of the solid material involved, which is crucial for identifying the flow type (i.e., stony/granular\u0026nbsp;vs muddy debris flow). The assessment of erosion and deposition volumes associated to debris flow events has undergone stark improvements in the last 25 years thanks to the increasing availability of high-resolution topographic data based on LiDAR and Structure from Motion (SfM), which makes it possible differencing pre- and post-event Digital Terrain Models (DTMs) (Scheidl et al., 2008; Bull et al., 2010; Cavalli et al., 2017; Cucchiaro et al., 2019).\u003c/p\u003e\n\u003cp\u003eThe catastrophic interaction of debris flows and protection works, leading to their collapse, requires a new methodological framework (Fig. 1), which integrates post-event data gathering from both the hydro-geomorphic processes and the structures involved. The post-event analysis of the hydro-geomorphic processes aims to characterize the triggering event and the main features of the debris flow (see flow-chart on the left in Figure 1). The post-event analysis of the collapsed structures aims to provide information about the failure mode, and the amount of sediment stored and mobilized. The comprehensive data-gathering approach permits to assess the contribution of the check dam failure to the overall debris flow characteristics (volumes, peak discharge and inundation pattern) and thus to the actual hazard posed to settlements and infrastructures, as well as it provides data for calibration/evaluation of improved modelling of the debris flow mitigation measures (Larese et al., 2023).\u003c/p\u003e\n\u003cp\u003eFigure 1: Flowchart for integrated forensic analysis of the coupled debris flow \u0026ndash; check dam collapse.\u003c/p\u003e"},{"header":"3.\tThe Rotian creek river catchment ","content":"\u003cp\u003eThe Rotian creek catchment (Fig. 2 b) is a forested watershed located in the Eastern Italian Alps, within the Autonomous Province of Trento (also called \u0026ldquo;Trentino\u0026rdquo;). The catchment has an area of 2.54 km\u003csup\u003e2\u003c/sup\u003e, with a range in altitude between 840 m a.s.l. at the fan apex to 2050 m a.s.l.; the mean angle of slope is 26.4\u0026deg;. The channel is 4.8 km long and has a mean angle of slope of 11.9\u0026deg;. The Rotian creek has built a large debris flow fan (0.43 km\u003csup\u003e2\u003c/sup\u003e, mean angle of slope 7.41\u0026deg;), which has forced the receiving stream (Noce River) to the opposite side of the valley. The Melton ruggedness number is 0.76, which is consistent with debris flows as the dominant sedimentary process on the fan. Mean annual precipitation is about 1100 mm, with major rainfall events typically occurring in the autumn season (Formetta et al., 2022). The mean value of the annual maxima precipitation ranges from 15 mm in 1 hour to 70 mm over 24 hours (Dallan et al., 2024). Recent research in Trentino has shown an upward trend in precipitation for both sub-daily (Libertino et al., 2019) and sub-hourly durations (Dallan et al., 2022) over the past few decades.\u003c/p\u003e\n\u003cp\u003eThe shape of the catchment is elongated along the main channel in a north \u0026ndash; south direction. The axis of the basin corresponds to a transpressional fault that separates massive igneous rocks (tonalite) that build the left slope of the catchment from limestone that outcrops on the right bank of the Rotian channel (Dal Piaz et al., 2007). Tonalite also outcrops in the lower sector of the basin, where the Rotian flows through a narrow rocky gorge. Most of the basin (68%) is covered by Quaternary deposits, mostly w\u0026uuml;rmian moraines, with less widespread presence of alluvial and colluvial deposits. In the moraines, boulders \u0026ndash; up to 2-4 m \u0026ndash; are embedded in a sandy-silty matrix. Consistently with the structural settings of the basin and the widespread presence of easily erodible moraines, the basin results in a canyon-like incision of around 10\u0026ndash;50 m of elevation difference compared with the top of the banks. The steep valley sides (range 30\u0026ndash;45\u0026deg;) bordering the Rotian Creek act as sediment source to the stream during intense runoff events.\u003c/p\u003e\n\u003cp\u003eThe channel reach from 1233 to 1030 m a.s.l. has been consolidated with a series of 16 check dams built from 1977 to 1986, with the seven located in the upper reach made of reinforced concrete. The check dam mean height is 5.3 m (range 2.9\u0026ndash;7.8 m) for a cumulative sum of 78.9 m and a mean channel angle of slope of 13\u0026deg;. Immediately downstream of the last check dam, the channel bed has been widened in order to create a depositional area formed by a series of 5 retention basins (total length 250 m, mean width 25 m). In 2014, a 7.4 m high open check dam was constructed at 989 m a.s.l., immediately downstream of the retention basins (Fig. 2 b).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the past, the Rotian catchment produced two major recorded debris-flow events, in 1776 and 1882 (Baggio and D\u0026rsquo;Agostino, 2022). In the Eastern Italian Alps, where the Rotian is located, the failure of check dam systems is a rare but not unprecedented event. A major debris-flow disaster, enhanced by the failure of an array of stone masonry check dams, occurred in November 1966 in another catchment in Trentino, causing the loss of three lives and substantial economic damage (Marchi and Cavalli, 2007). Less catastrophic in its consequences, but still relevant, was the destruction of check dams built in the 1920s-30s by a debris flow that originated from the mobilization of a large rotational landslide in the Venetian pre-Alps in 1985 (Bossi et al., 2024), as well as the masonry check-dams destroyed by 2009 and 2010 events in South Tyrol (Dell\u0026rsquo;Agnese et al., 2013).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigure 2. Study area with (a) its location in Italy, (b) the rain gauges included in the 10-km search radius centered on the study area, and (c) the Rotian catchment-debris flow fan system with the indication of the different sub-catchments commented in the text alongside check dams and lower retention area with the final slit dam.\u003c/p\u003e"},{"header":"4.\tThe forensic analysis for the 2018 event in the Rotian catchment","content":"\u003cp\u003e\u003cstrong\u003e4.1\u0026nbsp;The precipitation event: amount, timing and severity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Vaia storm (October 2018) caused one of the most impacting floods in northeastern Italy in the last century together with the flooding event of 1966 (Giovannini et al., 2021; Sioni et al., 2023). The storm hit almost the entire Alpine region, as well as Liguria and central Italy for three days, from 27 to 30 October. In several Alpine areas the rainfall amounts in the three days reached 600-900 mm, representing the strongest event in the last 150 years (Giovannini et al., 2021). In particular, on 29 October 2018, the event was characterized by the passage of a cold front with a deep convective band, stretching from South-East to North-West, which followed a previous phase (from 22:40 CET of 26 October to 18:00 CET of 28 October), characterized by stratiform orographic type precipitation (Borga and Zaramella, 2020).\u003c/p\u003e\n\u003cp\u003eIn the evening of October 29, the intense precipitation in the upper Rotian catchment triggered a catastrophic cascading natural-anthropogenic hazard (Rosatti et al., 2023). The basin is ungauged; Fig 2b shows the four rain-gauge stations considered in this study, all located within 10 km from the Upper Rotian river catchment. Precipitation data at 5 minutes temporal resolution, collected and managed by Meteotrentino (Autonomous Province of Trento), are available, with a total coverage of at least 25 years starting from the early 90s (Fig. 2c). Precipitation amount over the basin during the event was estimated by integrating data from the rain gauge stations and radar-based rainfall estimates from three nearby radar systems: Monte Grande (Veneto Region), Monte Macaion (Autonomous Province of Trento) and the Weissfluh at Davos (Switzerland). Quantitative rainfall estimation at the ground based on radar reflectivity observations presents the usual challenges which characterize radar rainfall estimation in a mountainous context (Marra et al., 2014). Radar-based estimates of precipitation at the ground were obtained by applying a complex error correction chain (Marra et al., 2014) accounting for wavelength attenuation, beam blocking, vertical profile of radar reflectivity, wind effects and finally integrating the radar-based estimates with data from the rain gauge stations. Final estimates are characterised by 1 km grid spacing and 15 min temporal aggregation. Precipitation maxima over the upper Rotian catchment, where the debris flow initiated, range from 24.8 mm at 1 h duration to 104.2 mm at 12 h and 359 mm over the three event days. Based on a probabilistic modelling approach (Pesce et al., 2024, submitted) in the Mezzana station (the closer station to the Rotian catchment, with the longest recording period), the severity of the precipitation maxima recorded during the storm was also evaluated. Based on a time-stationary approach, the precipitation event return period ranges from 50-100 year for 1 h duration, to 200-300 year for 12 h and \u0026gt; 300 year for 3 days, thus classifying the event as exceptional across all durations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2\u0026nbsp;Time evolution of the debris flow\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe time evolution of the debris flow was investigated using different approaches, including interviews with eyewitnesses, analysis of video recorded on the debris flow fan, and post-event geomorphological and sedimentological observations.\u003c/p\u003e\n\u003cp\u003eA questionnaire on the time occurrence and other features of the debris flow was distributed to 14 people, ranging in age from 25 to 74 years, who were located in different sectors of the debris flow fan during the event. Their answers indicated the occurrence of three debris-flow surges, respectively at 19:00-19:02 (average 19:01), 19:47-20:00 (average 19:51), and 23:30-23:40 (average 23:37) (Borga \u0026amp; Zaramella, 2020). These accounts confirm the reports of people involved in the emergency interventions and rescue activities on the occurrence of three surges on the evening of 29 October.\u003c/p\u003e\n\u003cp\u003eThe recordings of seven security video cameras installed in the campsite in the sector of the debris flow fan most severely hit by the debris flow were also analyzed. Although the quality of the videos is poor and recordings are limited to the first surge (i.e. before the video cameras went out of order), they provide some elements helpful to characterize - even if only qualitatively - some features of the debris flow:\u003c/p\u003e\n\u003cp\u003e- the time on the video frames of one of the cameras shows the arrival of the front at 19:59:47 CET, which is consistent with the accounts of the witnesses (19:01 CET);\u003c/p\u003e\n\u003cp\u003e- the debris flow front outside of the channel occurred suddenly, without or with a very limited fluid precursory surge;\u003c/p\u003e\n\u003cp\u003e- the velocity of the front\u0026nbsp;was high, although a quantitative assessment is not possible.\u003c/p\u003e\n\u003cp\u003e- large boulders were observed in the debris-flow front, at least one of them can be ascribed to a large fragment of a check dam.\u003c/p\u003e\n\u003cp\u003eThe presence in the middle sector of the channel of clast-supported lateral levees consisting of angular boulders at a level lower than the maximum flow depth attained during the event (Fig. 3) confirms that multiple surges occurred, the last one being smaller than the main one. In addition to cobbles and boulders, the deposits consist of a sandy matrix with a limited amount of silt and negligible clay: these features led us to classify the October 2018 event in the Rotian creek as a granular debris flow (Coussot and Meunier, 1996).\u003c/p\u003e\n\u003cp\u003eFig. 3. Lateral debris flow deposits in the middle sector of the Rotian creek. This levee was likely formed by the last surge, after the collapse of the check dams and the resulting channel incision. Note a boulder deposited on the top of the bank in the first part of the event.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.3 Geomorphic impacts of the event and the amplification of sediment volume due to check dams\u0026rsquo; failure\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe debris flow of 29 October 2018 caused substantial erosion along the channel, especially in the sector affected by the failure of the check dams. Erosion processes involved channel bed incision as well as erosion and destabilization of the sideslopes consisting of moraines and other Quaternary deposits. Field surveys have shown that shallow landslides occurred only along the channel banks and on valley slopes directly connected to the Rotian creek, whereas most of the catchment area remained stable. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe retention basins built downstream of the check dams array favoured the deposition of part of the sediment transported by the flow, but they were insufficient to store the entire amount. Most of the material eroded from the upper and middle reaches of the Rotian creek was transferred through the rocky gorge to the debris flow fan. To assess the changes in the elevation of the surface caused by the debris flow of 29 October 2018 - and hence the volume of debris eroded and deposited - a Digital Terrain model (DTM) obtained from surveys done after the event was compared with pre-event DTMs. The Autonomous Province of Trento carried out an UAV-based photogrammetric survey of the debris flow fan two days after the event. The point cloud (estimated error of 0.05 m with reference to 15 control points) was created using the Agisoft Metashape \u0026copy; software, and a raster with cell size of 0.1 m was created. A few days later (7 November), the debris basins were surveyed using the same technique and resolution. The post-event survey was completed on 14-15 June 2019 with the aerial LiDAR survey of the entire channel.\u003c/p\u003e\n\u003cp\u003ePre-event topography of the Rotian creek consists of two LiDAR-derived DTMs from surveys that cover the entire territory of the Province of Trento. The LiDAR surveys had been done in September-October 2014 (DEM resolution of 0.5 m) and 2008 (DTM resolution of 1 m). Although they derive from surveys done several years before the debris flow of October 2018, both DTMs have been considered representative of pre-event conditions because no flow events or landslides that could have changed the topography of the basin-fan system have occurred. It should be remembered, however, that the 2014 LiDAR DTM (still not officially validated) shows some quality issues, especially in vegetated areas. We preferred to use the 2008 DTM as representative of the pre-event conditions, and we used the 2014 DTM only for the sector of the debris basins, which had not yet been built in 2008.\u003c/p\u003e\n\u003cp\u003eThe comparison of the pre- and post-event DEMs required remapping them to the lowest resolution (1 m), which is anyway fully satisfactory for the objectives of the analysis: this was done by computing the mean elevation of the pixels included in a 1-meter cell. The co-registration of the DEMs was necessary because of a systematic shift among them: it was implemented using the code GRD-CoReg (Cucchiaro et al., 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe extent of the areas where the debris flow produced topographic changes and the depth of such changes were assessed using the tool GCD (Geomorphic Changes Detection) 7.5, freely available at \u003cu\u003ehttps://gcd.riverscapes.net/\u003c/u\u003e. GCD includes several methods that enable DTM differencing taking into account the quality of the available DEMs (Wheaton et al., 2009). We used the fixed-threshold method based on uniformly propagated error, initially proposed by (Brasington et al., 2003). This method defines a minimum level of detection under which the elevation changes are deemed not significant. The minimum level of detection considers the errors in the pre-and post-event DEMs that propagate into the DoD.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDTM differencing was computed separately for the debris flow fan and for the channel of the Rotian creek. The total volume of the deposits on the debris flow fan amounted to about 157000 m\u003csup\u003e3\u003c/sup\u003e \u0026plusmn; 27000 m\u003csup\u003e3\u0026nbsp;\u003c/sup\u003e(17% error). Erosion on the debris flow fan was much smaller but not negligible, and amounted to 5300 \u0026plusmn; 2000m\u003csup\u003e3\u003c/sup\u003e: it is possible that erosion was due to just-post-event removal of debris that also affected the pre-event surface or localized incisions along the flow path. DTM differencing along the Rotian creek was computed on a mask that encompasses the channel bed and the side slopes that show evidence of erosion and/or shallow instability. Erosion amounts to 207000 \u0026plusmn; 28000 m\u003csup\u003e3\u003c/sup\u003e (13% error), while the deposits, which occurred essentially in the debris basins, amounted to 39000 \u0026plusmn; 7000 m\u003csup\u003e3\u003c/sup\u003e (17% error). Summing the volumes deposited on the debris flow fan and within the catchment we get a total volume of about 196000 m\u003csup\u003e3\u003c/sup\u003e, quite close to the total volume eroded from the catchment and within its uncertainty range.\u003c/p\u003e\n\u003cp\u003eIn addition to the approximations intrinsic to DoD computation, other sources of uncertainties affect the assessment of eroded and deposited volumes and their comparison:\u003c/p\u003e\n\u003cp\u003e-\u0026nbsp; \u0026nbsp; \u0026nbsp;the sediment that reached the Noce River was not computed in DTM differencing;\u003c/p\u003e\n\u003cp\u003e- changes in\u0026nbsp;porosity between the source material and the deposits are not accounted for;\u003c/p\u003e\n\u003cp\u003e-\u0026nbsp; \u0026nbsp; \u0026nbsp;large wood is computed in the deposits but not in the eroded volume because DoD is computed based on DTM derived from bare ground points, thus not considering the trees in the channel bed and banks that became the main source of large wood.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe last of the above-mentioned issues\u0026nbsp;are\u0026nbsp;likely not relevant for the comparison of eroded and deposited volumes as the volume of large wood deposited on the debris flow fan amounts to 1500 steres (approx. 1000 m\u003csup\u003e3\u003c/sup\u003e). While this volume indicates a substantial supply of large wood from forest-mantled channel bed and sideslopes, it is almost negligible when compared to the sediment volume mobilized by the 29 October debris flow in the Rotian creek.\u003c/p\u003e\n\u003cp\u003eThe comparison of the volume of the deposits of the 29 October 2018 debris flow in the Rotian creek with other debris flows in the Eastern Italian Alps may indicate \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;the severity of this event. A dataset of 808 debris flows in northeastern Italy documented from historical archives, post-event surveys and - in a few cases - monitoring in instrumented catchments was used to explore, by means of quantile regression, the scaling relationships between catchment area and debris-flow volume (Marchi et al., 2019). Figure 4 compares the debris-flow deposits of the Rotian creek with debris-flow volumes collected in the same region and the scaling relationships between debris-flow volume and catchment area for selected quantiles. If the total volume of the deposits is considered (i.e. the sum of the deposits on the debris flow fan and in the debris basins within the catchment, about 196000 m\u003csup\u003e3\u003c/sup\u003e), the debris flow of 29 October 2018 corresponds to the 99th percentile of debris-flow volumes in the Eastern Italian Alps. Considering that most historical debris flows in the sample utilized for developing the scaling relationships neglect within catchment deposits, the deposits on the debrtis flow fan (about 157000 m\u003csup\u003e3\u003c/sup\u003e) are more consistent with the sample. In this case, the volume of the Rotian creek debris flow lies between the 98th and the 99th percentile. In either case, such results indicate the extreme severity of the debris flows under study.\u003c/p\u003e\n\u003cp\u003eThe minimum sediment volume contribution from the collapse of the check dams could be approximated as the volume stored (and released) by the check dams, about 115000 m\u0026sup3;, and doesn\u0026rsquo;t consider its erosion/sedimentation effect. The total volume of the debris flow would have been about 81000 m\u0026sup3;. A maximum value of the contribution from the system failure can be evaluated considering also its erosion/sedimentation effect, by the difference between the total volume of deposits and the volume mobilized upstream the first check dam (about 50000 m\u0026sup3; from the DoD analysis), thus resulting in about 146000 m\u0026sup3;. Thus, the amplification factor, that is the ratio between the total volume of deposit (about 196000 m\u0026sup3;) and the volume of the debris flow (50000-81000 m\u0026sup3;), ranges between 2.4 and 3.9. Considering these volumes, in Fig. 4 the debris-flow event is located below the line of 98%, but still in the upper sector of the plot of debris-flow volumes versus basin area.\u003c/p\u003e\n\u003cp\u003eFigure 4. Comparison of the volume of the Rotian creek debris flow with data on debris-flow volumes in the Eastern Italian Alps and scaling relationships with basin area (modified from Marchi et al., 2019), for 5 percentiles (1, 2, 50, 98, 99%). For the Rotian debris flow of October 2018, three volumes are indicated: total deposits, debris flow fan deposits, range of the estimated debris flow volume excluding the sediment released by check dam system failure.\u003c/p\u003e\n\u003cp\u003e4.4 \u003cstrong\u003e\u0026nbsp;Debris flow triggering and subsequent evolution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccording to post-event surveys, the debris flow was triggered by a shallow landslide mobilised on the left slope in the upper portion of the catchment, corresponding to a drainage area of 0.22 km\u003csup\u003e2\u003c/sup\u003e (Figure 2). The peak liquid discharge in the Rotian channel at that section, and its uncertainty, was quantified based on the indirect methodology developed by Amponsah et al. (2016), with a central value of 0.9 m\u003csup\u003e3\u003c/sup\u003e/s and uncertainty ranging from 0.55 to 1.25 m\u003csup\u003e3\u003c/sup\u003e/s. These values were used to evaluate the accuracy of the integrated hydrology-hydraulic debris-flow model developed by Rosatti and Zugliani (2024). \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eClassical methodologies for debris-flow simulation in a given channel reach (Martinengo et al., 2021) usually start with the evaluation of the liquid discharge, then the solid-liquid mixture is considered (Takahashi, 1978; Rosatti et al., 2019), and finally the debris-flow dynamics is simulated (Rosatti et al., 2018). However, the Rotian case places specific challenges this three-steps methodology. Indeed, the debris flow should be simulated along the whole basin, where the liquid discharge continuously changes moving downstream, and a single input hydrograph is no more relevant in this case. Moreover, the morphological variations, especially the erosions, are significant and in some zones greater than 10 m, leading to an important modification in the flow dynamics, which makes the classical fixed-bed approach insufficient.\u003c/p\u003e\n\u003cp\u003eTo account for these challenges, a new integrated model, called TRENT2D\u003csup\u003eMBRR\u003c/sup\u003e (Mobile Bed Rainfall Runoff), was developed. The model combines the capabilities of the two-phase isokinetic TRENT2D model to simulate the dynamic of debris flow flowing over erodible and non-erodible zones (Amadii et al., 2022; Amaddii et al., 2023; Rosatti and Zugliani, 2015) with a description of the hydrological response along the river reach, by allowing the spatially distributed simulation of infiltration and subsurface-surface flow at the basin scale. The infiltration process is simulated by the well-known Green-Ampt equation, while the sub-surface flow is described by the using the Darcy law and the extended Dupuit-Forchheimer hypothesis. The model also includes the simulation of the debris-flow triggering mechanism (see Rosatti and Zugliani, 2024 for details).\u003c/p\u003e\n\u003cp\u003eThe integrated model does not account for the check dam system failure. Thus, the simulation of the debris flow is here limited to the basin closed at the first check dam of the protection system (dam number 16, see Figure 2, with a drainage area of 1.02 km\u003csup\u003e2\u003c/sup\u003e). Check dam number 16 collapsed first and most upstream in the cascading failure, hence it represents the limit to the domain where the debris flow was not affected by the failure of the protection works. The model application to this part of the basin provides both mixture (solid plus liquid) hydrographs in different channel sections and maps of morphological variations. Figure 5a shows the liquid hydrograph simulated in the section just upstream the triggering zone, together with a comparison with the indirect peak estimate. Figure 5b reports the mixture and solid hydrographs obtained just upstream check dam 16. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe liquid peak discharge at the triggering section amounts to 1 m\u003csup\u003e3\u003c/sup\u003e/s at 18:50 CET with a slightly lower second peak of 0.9 m\u003csup\u003e3\u003c/sup\u003e/s at 19.30 CET.\u0026nbsp;The simulated solid discharge at check dam 16 starts around 17:30 CET during the most intense phase of the event and shows a peak slightly greater than 1 m\u003csup\u003e3\u003c/sup\u003e/s at 19:35 CET.\u003c/p\u003e\n\u003cp\u003eThe mixture discharge exhibits a well-defined peak of 3.3 m\u003csup\u003e3\u003c/sup\u003e/s at the same time of the solid discharge peak (19:35 CET).\u003c/p\u003e\n\u003cp\u003eFigure 5. Simulated hydrographs: a) liquid hydrograph obtained in a section near the triggering zone, and comparison with the indirect peak estimate, b) mixture and solid hydrographs obtained near the check dam 16.\u003c/p\u003e\n\u003cp\u003eFigure 6. Morphological variations in the Rotian creek upstream of check dam 16 (green line): a) by using the procedure described in Section 4.3, and b) as computed with the TRENT2D\u003csup\u003eMBRR\u003c/sup\u003e at the end of the simulation (3:00 CET of 30/10/2018). The yellow line indicates the location of a forest road crossing the creek.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigure 6 compares the morphological variations evaluated for the upper part of the Rotian creek following the procedure described in Section 2.3 (panel a) with those simulated by the numerical model at the end of the simulation (3:00 CET of 30/10/2018) (panel b). \u0026nbsp;The comparison reported in Figure 6 shows that in the upper part of the creek, i.e. from the debris flow triggering zone until the crossing of a forest road (yellow line in Figure 6), the agreement is satisfactory, although with a moderate estimation in the modelled solid volume. The net solid volume mobilised in this area is 15400 m\u003csup\u003e3\u003c/sup\u003e, about 9% greater than the upper limit value estimated from the DoD. Furthermore, except for the largest shallow landslide present on the left, also the spatial distribution of the morphological variation is consistent with the DoD both in terms of eroded depth and involved area. Instead, in the lower part, i.e. from the forest road to the check dam 16, the results are less satisfactory since the simulated net solid volume mobilised in this zone (21700 m\u003csup\u003e3\u003c/sup\u003e) is 22% less than the lower limit value of the DoD.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.5 Check dam failure mode\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA close examination of the check dam system after the event revealed that the system failure was triggered by the collapse of the most upstream check dam 16. With a total width of 29 m and a height of 8.5 m, this check dam is the highest in the check dam system. The inspection revealed that the check dam was destroyed in the joint-to-joint section in the wing section on the left bank side of the dam, due to the combined action of slipping of the reinforcing bars and the overload caused by the debris flow, as simulated based on results from section 2.5. Approximately 10000 kN of heavy fragments of the check dam were transported by the debris flow and impacted the downstream check dams triggering their collapse due to internal stability failure. The peak discharge of the debris flow which resulted from the release of the sediment behind the check dam was much larger than the designed peak discharge calculated from the design rainfall intensity. Further, a field survey also indicated that the main destruction mode was overturning rather than sliding. Thus, the collapse of the check dam 16 led to the failure of the whole system of 16 check dams.\u003c/p\u003e\n\u003cp\u003eAlso, the open check dam downstream the sequence of 5 retention basins was damaged - and partly outflanked - by the debris flow. A view of the upper part of the collapsed system is reported in Figure 7. The figure also shows a check dam (no. 7, marked in Figure 2) that did not fail but was under-excavated, probably because of the failure of other dams. The deposits visible over the wing of the dam also suggest that it had been overtopped by the debris flow before its under-excavation.\u003c/p\u003e\n\u003cp\u003eFigure 7. Check dams after the failure event: (a) view of the upper part of the collapsed system; (b) check dam no. 7: the deposits over the wing of the dam indicate that it had been overtopped by the debris flow before its under-excavation.\u003c/p\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eThe collected post-event data and observations provide valuable insights into the performance of the integrated hydrologic-hydraulic-debris-flow model application. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe simulated peak hydrograph timing in Figure 5 can be considered in good agreement with the debris-flow timing described in Section 2 (around 19:00 CET), whereas the simulated water discharge peak (1 m\u003csup\u003e3\u003c/sup\u003e/s) in well within the uncertainty band of the post-event based estimated peak (0.55-1.25 m3/s).\u003c/p\u003e\n\u003cp\u003eAlso, the model does not predict solid discharge along the event for the channel reaches located upstream of the debris flow triggering section, thus showing a good accuracy in the simulation of the relevant hydro-geomorphic processes in this area.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAn important result with regard to simulated hydrograph is the absence of the third wave, which means that the latter may be not directly related to the rainfall event, but rather to the partial collapse of one of the check dams or one of the small shallow-landslides that occurred just after the event.\u003c/p\u003e\n\u003cp\u003eComparing the liquid discharge predicted at the triggering section and check dam solid discharges in Figure 5, it appears that the modelled solid discharge starts one hour before the moment of the debris flow triggering. The explanation for this anticipated solid discharge might be related to some minor erosion processes occurring in the section upstream of check dam 16. \u0026nbsp;On the other hand, the absence of significant solid transport in the phase preceding the debris flow was also noted by the operators performing the event monitoring.\u003c/p\u003e\n\u003cp\u003eThe solid discharge simulated near the check dam 16 (Figure 5b) shows a long decreasing phase that appears in contrast with the observations of impulsive debris flow waves reported for the debris flow fan (Section 2.3). However, we should note that (i) the magnitude of the solid discharge in the decreasing phase is limited (around 0.5 m3/s), (ii) the simulation was only carried out for the basin closed near check dam 16, thus far away from the debris flow fan since the collapse of the dams is not considered in the model yet, (iii) the check dam system collapse releases a massive amount of sediment in a short period, probably causing the high and impulsive sediment discharges recorded on the debris flow fan. These points suggest that the modelled decreasing phase of the solid discharge is reasonable and falls within the uncertainties of event reconstruction.\u003c/p\u003e\n\u003cp\u003eThe comparison of the morphological variations simulated by the hydraulic model for the Rotian creek upstream of check dam 16 (Figure 6) shows mixed results. In the upper part (until the crossing of a forest road (yellow line in Figure 6) the model matches the observations very well. This can be partially explained with the presence of a limited erodible depth. However, results reported for the lower reach (from the forest road to the check dam 16) are less satisfactory, with the spatial distribution of the morphological variation differing with respect to the results from the DoD in terms of eroded depth and involved area.\u0026nbsp;The first zone where the discrepancy is relevant (i.e., just downstream the forest road) could be justified by the presence of a few large boulders under the road that concentrated the debris flow in a small area, a local effect not well captured by the model due to the spatial resolution of the used domain. Instead, the second discrepancy (i.e., the one close to the check dam 16) can be partly explained by the absence of the dam failure procedure in the numerical model, a phenomenon that had a significant effect on the extent of excavation in the areas close to these artefacts. However, the overall results of the model have to be considered satisfactory.\u003c/p\u003e\n\u003cp\u003eThe integrated analysis of both the debris flow and of the collapsed check dam system shows that a very likely cause of the critical failure of check dam 16 is the combination of a debris flow which is larger than the design load \u0026ndash; the design only included liquid flood \u0026ndash; and of inaccuracies in the building of the work. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eResults reported by Baggio and D\u0026rsquo;Agostino (2022) in their simulation of the Rotian creek debris flow largely agree with our findings. However, Baggio and D\u0026rsquo;Agostino (2022) interpreted the debris flow of Rotian creek in October 2018 as a muddy debris flow, a term that applies to debris flows rich in cohesive clayey material (Coussot and Meunier, 1996), whereas the morphological and sedimentological evidences of the deposits described in Section 2.3 (abundant presence of large cobbles and boulders in clast-supported lateral levees, and sandy matrix with very limited presence of fines) led us to classify the event under study as a granular (i.e., stony) debris flow, and to model it accordingly.\u003c/p\u003e"},{"header":"6. Conclusions","content":"\u003cp\u003eThis paper highlights the critical need to address the complexities surrounding check dam system failures and their implications for residual risk management. The proposed forensic model includes several analyses providing information of the evolution of the event, from the triggering rainfall to the reconstruction of the flood and hydrograph, and information on the check dam failure mode. Its application to the Rotian creek case study provides a framework for understanding how the collapse of check dams can significantly amplify debris flow magnitude, leading to severe consequences for downstream areas, and for developing similar future analysis. Also, it illustrated how this analysis could improve our understanding on the collapse mode, which could be used for advancing our understanding and preparedness to these events, as discussed in the following.\u003c/p\u003e\n\u003cp\u003eWe identify three aspects where the forensic model may have important practical consequences on residual risk management.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ei) Development of integrated debris flow model able to account for check dam failures\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe collapse of flood protection systems is generally ignored in current models of debris flow simulation. The processes by which check dam failures amplify the scale of a debris flow, altering the flow velocity, mixture discharge and volumetric solid concentration, are complex and may depend on the geometry and construction technique of the protective structure, as well as the topography and morphology of the channel. \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIntegrated models able to incorporate the effect of check dam failures could shed light on the amplification effect, thus quantifying the changes in the debris flow magnitude and its destructive capacity.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eii) Consolidating existing datasets of check dam system failures\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFailures of debris-flow mitigation works are rarely reported in the scientific literature, presumably due to embarrassment or the threat of potential legal action. This is lamentable, as the greatest advances in debris-flow mitigation may derive from a detailed examination of past failures.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo benefit from these learning opportunities, there is a pressing need to consolidate physical processes knowledge, engineering analysis, and social datasets related to check dam system failures. Integrating these diverse datasets and developing post-event analysis can provide a more comprehensive understanding of the interaction between physical process and structures and the factors contributing to failures, including environmental conditions, design flaws, and community responses. By fostering collaboration among engineers, scientists, and local stakeholders, we can create a richer knowledge base that informs better design and management practices.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eiii) Methodologies for the prioritisation of field inspection and maintenance\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe post-event analysis conducted in this study underscores the critical role of check dam failure in exacerbating debris flow and its associated damages, thereby highlighting the urgent need for further site-specific analyses and/or maintenance works on these structures. Establishing priorities for field inspection and potential repairs of check dams necessitates an approach that is applicable at the regional scale. Methodologies for prioritizing inspection and maintenance are essential, as they enable the efficient allocation of resources, ultimately reducing risks to infrastructure and enhancing community resilience against natural disasters.\u003c/p\u003e\n\u003cp\u003eTo facilitate this, a straightforward index is under evaluation that incorporates both variables related to the check dam, such as age, height, and material, and site-specific factors, including the presence of landslides and geolithology. Forensic analysis of flow events that involved check dam failure, such as the Rotian creek 2018 debris flow, provides important information for the choice and weighting of the variables for the indices of check dam fragility. At the same time, the modelling techniques developed in post-event studies can be applied to evaluate management alternatives for check dam systems potentially susceptible to failure.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFundings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by the GPR Project (\u0026ldquo;Approfondimento delle strategie di Governo della Pericolosit\u0026agrave; alluvionale a seguito dell\u0026rsquo;evento del 29 ottobre 2018 sul rio Rotiano\u0026rdquo; \u0026ndash; Accordo di Programma GPR) funded by the Autonomous Province of Trento and by Fondazione Cassa di Risparmio di Padova e Rovigo (Excellence Grant 2021 to the Resilience Project). E. Dallan was supported by the RETURN Extended Partnership and received funding from the European Union Next-GenerationEU (National Recovery and Resilience Plan \u0026ndash; NRPP, Mission 4, Component 2, Investment 1.3 \u0026ndash; D.D. 1243 2/8/2022, PE0000005).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ehe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe tool GCD (Geomorphic Changes Detection) 7.5 is freely available at https://gcd.riverscapes.net/ (last access: November 2024).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contribution.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: LM, GR, RV, MB; Data curation: all authors; Formal analysis: all authors; Funding acquisition: GR, LM, MB; Visualization: ED, LM, SC, DZ; Writing \u0026ndash; original draft: ED, LM, DZ, MB; Writing \u0026ndash; review \u0026amp; editing: all authors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAmaddii, M., Rosatti, G., Zugliani, D., Marzini, L., Disperati, L., 2022. Back-Analysis of the Abbadia San Salvatore (Mt. Amiata, Italy) Debris Flow of 27\u0026ndash;28 July 2019: An Integrated Multidisciplinary Approach to a Challenging Case Study. Geosciences, 385(12), 1-25. DOI: 10.3390/geosciences12100385\u003c/li\u003e\n \u003cli\u003eAmaddii, M., Rosatti, G., Zugliani, D., Marzini, L., Disperati, L., 2023. Modelling stony debris flows involving culverted streams: the Abbadia San Salvatore case (Mt. Amiata, Italy). Rendiconti Online della Societ\u0026agrave; Geologica Italiana, 61, 108-115. DOI: 10.3301/ROL.2023.55\u003c/li\u003e\n \u003cli\u003eAmponsah, W., L. Marchi, D. Zoccatelli, G. Boni, M. Cavalli, F. Comiti, S. Crema, A. Luc\u0026iacute;a, F. Marra, M. Borga, 2016. Hydrometeorological characterisation of a flash flood associated to major geomorphic effects: Assessment of peak discharge uncertainties and analysis of the runoff response. J. Hydrometeorology, 17(12), 3063-3077\u003c/li\u003e\n \u003cli\u003eBaggio T., D\u0026rsquo;Agostino V.,2022: Simulating the effect of check dam collapse in a debris-flow channel. Sci. Total Environ., 816 (2022), Article 151660, 10.1016/j.scitotenv.2021.151660\u003c/li\u003e\n \u003cli\u003eBallesteros-C\u0026aacute;novas, J.A., Stoffel, M., de Haas, T., Bodoque, J.M., 2024: Debris Flow Dating and Magnitude Reconstruction. In: Jakob, M., McDougall, S., Santi, P. (eds) Advances in Debris-flow Science and Practice. Geoenvironmental Disaster Reduction. Springer, Cham. https://doi.org/10.1007/978-3-031-48691-3_8\u003c/li\u003e\n \u003cli\u003eBenito G., Grodek T., Enzel Y., 1998:. The geomorphic and hydrologic impacts of the catastrophic failure of flood-control-dams during the 1996-Biescas flood (Central Pyrenees, Spain), Zeitschrift Fur Geomorphol., 42 (1998), pp. 417-437, 10.1127/zfg/42/1998/417\u003c/li\u003e\n \u003cli\u003eBertola, M., Viglione, A., Lun, D., Hall, J., Bl\u0026ouml;schl, G., 2020: Flood trends in Europe: Are changes in small and big floods different? Hydrology and Earth System Sciences, 1805\u0026ndash;1822. https://doi.org/10.5194/hess-24-1805-2020.\u003c/li\u003e\n \u003cli\u003eBorga, M., M. Stoffel, L. Marchi, F. Marra, M. Jakob, 2014: Hydrogeomorphic response to extreme rainfall in headwater systems: flash floods and debris flows. Journal of Hydrology, 518, 194\u0026ndash;205, http://dx.doi.org/10.1016/j.jhydrol.2014.05.022, ISSN: 0022-1694\u003c/li\u003e\n \u003cli\u003eBorga M, Comiti F, Ruin I, Marra F., 2019: Forensic analysis of flash flood response. WIREs Water. \u0026nbsp;6:e1338. https://doi.org/10.1002/wat2.1338\u003c/li\u003e\n \u003cli\u003eBorga, M., Zaramella, M., 2020. Evento di piena del 27-29 ottobre 2018 sul bacino del Rotian creek: stima della precipitazione e valutazione della sua severit\u0026agrave;, Project Report, https://www.tesaf.unipd.it/en/sites/tesaf.unipd.it.en/files/Progetto%20VAIA_R01.2%20F.pdf.\u003c/li\u003e\n \u003cli\u003eBossi G., Cavalli M., Mantovani M., Catelan F. T., Ballaera A., Ceccotto F., Marcato G., Pasuto A, 2024. Expecting the expected \u0026ndash; learning from the past to provide forward scenarios through geomorphic change detection, monitoring and modeling. Geoenvironmental Disasters, 11, 35, https://doi.org/10.1186/s40677-024-00292-7.\u003c/li\u003e\n \u003cli\u003eBrasington, J., Langham, J., \u0026amp; Rumsby, B., 2003: Methodological sensitivity of morphometric estimates of coarse fluvial sediment transport. Geomorphology, 53 (3-4), 299\u0026ndash;316. https://doi.org/10.1016/s0169-555x(02)00320-3\u003c/li\u003e\n \u003cli\u003eBronstert, A. Agarwal, B. Boessenkool, I. Crisologo, M. Fischer, M. Heistermann, L. K\u0026ouml;hn-Reich, J.A. L\u0026oacute;pez-Taraz\u0026oacute;n, T. Moran, U. Ozturk, C. Reinhardt-Imjela, D. Wendi, 2018: Forensic hydro-meteorological analysis of an extreme flash flood: the 2016-05-29 event in Braunsbach, SW Germany, Sci. Total Environ., 630, pp. 977-991, 10.1016/j.scitotenv.2018.02.241\u003c/li\u003e\n \u003cli\u003eBull J.M., Miller H., Gravley D.M., Costello D., Hikuroa D.C.H., Dix J.K., 2010: Assessing debris flows using LIDAR differencing: 18 May 2005 Matata event, New Zealand. Geomorphology, 124 (1-2), 75-84, DOI: 10.1016/j.geomorph.2010.08.011\u003c/li\u003e\n \u003cli\u003eCavalli, M., Goldin, B., Comiti, F., Brardinoni, F., Marchi, L., 2017: Assessment of erosion and deposition in steep mountain basins by differencing sequential digital terrain models. Geomorphology, 291, 4-16, doi:10.1016/j.geomorph.2016.04.009\u003c/li\u003e\n \u003cli\u003eCoussot P., Meunier M., 1996: Recognition, classification and mechanical description of debris flows. Earth-Science Reviews, 40, 209-227.\u003c/li\u003e\n \u003cli\u003eCucchiaro, S., Cavalli, M., Vericat, D., Crema, S., Llena, M., Beinat, A., Marchi, L., \u0026amp; Cazorzi, F.,2019: Geomorphic effectiveness of check dams in a debris-flow catchment using multi-temporal topographic surveys. Catena, 174, 73\u0026ndash;83. hiips://doi.org/10.1016/j.catena.2018.11.004\u003c/li\u003e\n \u003cli\u003eCucchiaro, S., Maset, E., Cavalli, M., Crema, S., Marchi, L., Beinat, A., \u0026amp; Cazorzi, F., 2020: How does co-registration affect geomorphic change estimates in multitemporal surveys? GIScience \u0026amp; Remote Sensing, 57 (5), 611\u0026ndash;632, https://doi.org/10.1080/15481603.2020.1763048\u003c/li\u003e\n \u003cli\u003eDallan, E., Borga M., Zaramella M., Marra F., 2022: Enhanced summer convection explains observed trends in extreme subdaily \u0026nbsp; \u0026nbsp;precipitation in the Eastern Italian Alps, Geophysical Research Letters, 49, e2021GL096727, 2022.\u003c/li\u003e\n \u003cli\u003eDallan, E., Borga, M., Fosser, G., Canale, A., Roghani, B., Marani, M., Marra, F., 2024a: A method to assess and explain changes in sub‐daily precipitation return levels from convection‐permitting simulations, Water Resources Research, 60, e2023WR035969, 2024. https://doi.org/10.1029/2023WR035969\u003c/li\u003e\n \u003cli\u003eDallan E., Marra F., Fosser G., Marani M., Borga M., 2024b: Dynamical factors heavily modulate the future increase of sub-daily extreme precipitation in the Alpine-Mediterranean region.\u0026nbsp;Accepted in Earth\u0026rsquo;s Future.\u003c/li\u003e\n \u003cli\u003eDal Piaz, G.V., Castellarin, A., Martin, S., Selli, L., Carton, A., Pellegrini, G.B., Casolari, A., Daminato, F., Montresor, L., Picotti, V., Prosser, G., Santuliana, G., Cantelli, L., 2007: Note Illustrative della Carta Geologica d\u0026apos;Italia alla scala 1:50.000. Foglio 042 Mal\u0026egrave;. Provincia Autonoma di Trento, ISPRA. - System Cart Roma : Regione Lombardia, APAT, 2007 (in Italian).\u003c/li\u003e\n \u003cli\u003eDell\u0026rsquo;Agnese A., Mazzorana B., Comiti F., Von Maravic P., D\u0026rsquo;Agostino V., 2013:. Assessing the physical vulnerability of check dams through an empirical damage index. J Agric Eng [Internet]. 2013 Jun. 14;44(1):e2. Available from: https://www.agroengineering.org/jae/article/view/jae.2013.e2\u003c/li\u003e\n \u003cli\u003eFormetta G., F. Marra, E. Dallan, M. Zaramella, M. Borga, 2022: Differential orographic impact on sub-hourly, hourly, and daily extreme precipitation. Adv. Water Resour., 159 (2022), Article 104085, 10.1016/j.advwatres.2021.104085\u003c/li\u003e\n \u003cli\u003eFuchs S., Heiss K., H\u0026uuml;bl J., 2007: Towards an empirical vulnerability function for use in debris flow risk assessment. Nat. Hazards Earth Syst. Sci., 7, pp. 495-506, 10.5194/nhess-7-495-2007\u003c/li\u003e\n \u003cli\u003eGiovannini L., Davolio S., Zaramella M., Zardi D., Borga M., 2021: Multi-model convection-resolving simulations of the October 2018 Vaia storm over northeastern Italy, Atmospheric Res., 253,105455.\u003c/li\u003e\n \u003cli\u003eGrodek, T. and Benito, G., 2024: Reevaluating Flood Protection: Disaster Risk Reduction for Urbanized Alluvial Fans, Nat. Hazards Earth Syst. Sci. Discuss. [preprint], https://doi.org/10.5194/nhess-2024-171, under review.\u003c/li\u003e\n \u003cli\u003eH\u0026uuml;bl, J., Suda, J., Uchida, T., \u0026amp; Nagl, G., 2024:. Check Dam Failures. In Advances in Debris-flow Science and Practice (pp. 565-588). Cham: Springer International Publishing.\u003c/li\u003e\n \u003cli\u003eJakob, M., 2005: A size classification for debris flows. Engineering geology, 79(3-4), 151-161.\u003c/li\u003e\n \u003cli\u003eJakob, M., 2019: Debris-flow hazard assessments \u0026ndash; a practitioner\u0026rsquo;s view. 7th International Conference on Debris-Flow Hazards Mitigation, Golden, Colorado, USA.\u003c/li\u003e\n \u003cli\u003eKeating, A., K. Venkateswaran, M. Szoenyi, K. MacClune, and R. Mechler. 2016. From event analysis to global lessons: Disaster forensics for building resilience. Natural Hazards and Earth System Sciences 16(7): 1603\u0026ndash;1616\u003c/li\u003e\n \u003cli\u003eLarese, A. et al., 2023: Multiphysics simulation of the impact of natural hazards on structures and protection systems. 2 nd GACM-GIMC Joint Workshop, RWTH Aachen University, September 14-15, 2023\u003c/li\u003e\n \u003cli\u003eLibertino, A., Ganora, D., \u0026amp; Claps, P., 2019: Evidence for increasing rainfall extremes remains elusive at large spatial scales: The case of Italy. Geophysical Research Letters, 46(13), 7437-7446.\u003c/li\u003e\n \u003cli\u003eMarchi, L., Brunetti, M.T., Cavalli, M., Crema, S., 2019. Debris-flow volumes in northeastern Italy: relationship with drainage area and size probability. Earth Surface Processes and Landforms, 44(4), 933-943, doi: 10.1002/esp.4546\u003c/li\u003e\n \u003cli\u003eMarchi, L., Cavalli, M., 2007: Procedures for the Documentation of Historical Debris Flows: Application to the Chieppena Torrent (Italian Alps). Environmental Management, 40, 493-503.\u003c/li\u003e\n \u003cli\u003eMarra, F., E. I. Nikolopoulos, J. D. Creutin, M. Borga, 2014: Radar rainfall estimation for the identification of debris-flow occurrence thresholds. \u0026nbsp;Journal of Hydrology, Volume 519, Part B, 1607-1619, http://dx.doi.org/10.1016/j.jhydrol.2014.09.039, ISSN: 0022-1694.\u003c/li\u003e\n \u003cli\u003eMartinengo, M., Zugliani, D., Rosatti, G., 2021: Uncertainty analysis of a rainfall threshold estimate for stony debris flow based on the backward dynamical approach. Natural hazard and Earth System Sciences, 21(6), 1769-1784. DOI: 10.5194/nhess-21-1769-2021\u003c/li\u003e\n \u003cli\u003eMazzorana, B., Trenkwalder-Platzer, H., Fuchs, S. et al., 2014: The susceptibility of consolidation check dams as a key factor for maintenance planning. \u0026Ouml;sterr Wasser- und Abfallw 66, 214\u0026ndash;216, https://doi.org/10.1007/s00506-014-0160-4\u003c/li\u003e\n \u003cli\u003ePesce M., Dallan E., Marra F., Borga M., 2024: Increasing probability of record-breaking precipitation: a case-study in the Eastern Italian Alps. J. Hydrol., Regional studies, under review. \u0026nbsp;\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePiton and co-authors, 2016: Why do we build check dams in Alpine streams? An historical perspective from the French experience. Earth Surf. Process. Landforms 42, 91\u0026ndash;108.\u003c/li\u003e\n \u003cli\u003eRanzi, R., Barontini, S., Ferri, M, 2015: Structural Residual Risk Due to Levee Failures in Flood Mapping. In: Lollino, G., Arattano, M., Rinaldi, M., Giustolisi, O., Marechal, JC., Grant, G. (eds) Engineering Geology for Society and Territory - Volume 3. Springer, Cham. https://doi.org/10.1007/978-3-319-09054-2_92\u003c/li\u003e\n \u003cli\u003eRickenmann, D., 2005: Runout prediction methods. In: Jakob, M., Hungr, O. (eds), Debris-flow hazards and related phenomena, Praxis Springer, Berlin Heidelberg, 305-324\u003c/li\u003e\n \u003cli\u003eRosatti G., Zugliani D., Fraccarollo L., 2023: The debris flow event of 29 October 2018 in the Rio Rotiano (Italy) and its challenges for the mathematical and numerical modelling. E3S Web of Conf., 415 (2023) 05022. https://doi.org/10.1051/e3sconf/202341505022\u003c/li\u003e\n \u003cli\u003eRosatti, G., Zorzi, N., Zugliani, D., Piffer, S., Rizzi, A., 2018. A Web Service ecosystem for high-quality, cost-effective debris-flow hazard assessment. Environmental Modelling \u0026amp; Software, 100, 33-47. DOI: 10.1016/j.envsoft.2017.11.017\u003c/li\u003e\n \u003cli\u003eRosatti, G., Zugliani, D., 2015: Modelling the transition between fixed and mobile bed conditions in two-phase free-surface flows: The Composite Riemann Problem and its numerical solution. Journal of Computational Physics, 285, 226-250. DOI: 10.1016/j.jcp.2015.01.011\u003c/li\u003e\n \u003cli\u003eRosatti, G., Zugliani, D., 2024: TRENT2DMBRR: an integrated mobile-bed rainfall-runoff model for the simulation of debris flow at the basin scale, from the triggering to the deposition phase. Proceedings of ICIRBM 2024 \u0026ldquo;Technologies for Integrated River Basin Management \u0026ndash; Tecniche per la Difesa del Suolo e dall\u0026rsquo;Inquinamento\u0026rdquo; G. Frega and F. Macchione (eds.), ISBN: 978‐88‐97181‐90‐3, ISSN: 2282‐5517.\u003c/li\u003e\n \u003cli\u003eRosatti, G., Zugliani, D., Pirulli, M., Martinengo, M., 2019: A new method for evaluating stony debris flow rainfall thresholds: the Backward Dynamical Approach. Heliyon 5(6), e01994. DOI: 10.1016/j.heliyon.2019.e01994.\u003c/li\u003e\n \u003cli\u003eScheidl C., Rickenmann D., Chiari M., 2008: The use of airborne LiDAR data for the analysis of debris flow events in Switzerland. Natural Hazards and Earth System Science, 8 (5), 1113-1127, www.nat-hazards-earth-syst-sci.net/8/1113/2008/\u003c/li\u003e\n \u003cli\u003eSioni F., Davolio S., Grazzini F., Giovannini L., 2023: Revisiting the atmospheric dynamics of the two century floods over north-eastern Italy. Atmos. Res., 286, 10.1016/j.atmosres.2023.106662\u003c/li\u003e\n \u003cli\u003eSodnik, J., Martinčič, M., Miko\u0026scaron;, M., Kryžanowski, A., 2015: Are Torrent Check-Dams Potential Debris-Flow Sources?. In: Lollino, G., et al. Engineering Geology for Society and Territory - Volume 2. Springer, Cham. https://doi.org/10.1007/978-3-319-09057-3_79\u003c/li\u003e\n \u003cli\u003eTakahashi, T., 1978: Mechanical Characteristics of Debris Flow. Journal of the Hydraulics Division, 104 , 1153-1169. DOI:10.1061/jyceaj.0005046.664\u003c/li\u003e\n \u003cli\u003eWang G.L., 2013: Lessons learned from protective measures associated with the 2010 Zhouqu debris flow disaster in China, Nat. Hazards, 69 (2013), pp. 1835-1847, 10.1007/s11069-013-0772-1\u003c/li\u003e\n \u003cli\u003eWheaton, J. M., Brasington, J., Darby, S. E., \u0026amp; Sear, D. A., 2009: Accounting for uncertainty in DEMs from repeat topographic surveys: improved sediment budgets. Earth Surface Processes and Landforms, 136\u0026ndash;156. hiips://doi.org/10.1002/esp.1886\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"debris flow, check dam failure, Forensic analysis, post-event survey, cascading hazards","lastPublishedDoi":"10.21203/rs.3.rs-5478044/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5478044/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCheck dams may be effective structures for reducing debris flow hazard, but their failure often results in serious consequences for people and infrastructures. The examination of these failures embracing a forensic engineering approach, still rather poorly represented in the scientific literature, would lead to important improvements in how residual risk is planned and managed.\u003c/p\u003e\n\u003cp\u003eIn this study, we developed a framework for the forensic analysis of check dam systems failures in terms of cascading natural-anthropogenic hazards, and we applied such framework to the catastrophic event occurred in October 2018 in the Rotian creek catchment (Eastern Italian Alps).\u003c/p\u003e\n\u003cp\u003eThe post-event survey and analysis gathered observations about rainfall, peak discharges, morphological impacts, and damaged check dams. \u0026nbsp;Based on these data, we applied a newly developed coupled hydrologic-hydraulic debris flow model and we assessed the failure mode of the check dam system. Our results highlight important practical implications for improving residual risk management, namely: i) development of debris flow models capable of simulating the role of check dams and their failure in the debris flow dynamics, ii) the call for extensive datasets of check dam system failures, and iii) the necessity to develop methodologies for the prioritisation of field inspection and maintenance of existing check dam systems.\u003c/p\u003e","manuscriptTitle":"What can be learnt from the catastrophic failure of a check dam system? A forensic analysis of a cascading natural-anthropogenic hazard","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-23 06:35:36","doi":"10.21203/rs.3.rs-5478044/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revisions","date":"2025-02-11T18:35:06+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2024-12-20T16:24:38+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-12-20T15:02:31+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-11-20T05:25:56+00:00","index":"","fulltext":""},{"type":"submitted","content":"Natural Hazards","date":"2024-11-19T03:21:51+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c674b658-4206-4e0f-bc20-61f5abc7f6e0","owner":[],"postedDate":"December 23rd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-05-19T16:03:34+00:00","versionOfRecord":{"articleIdentity":"rs-5478044","link":"https://doi.org/10.1007/s11069-025-07301-4","journal":{"identity":"natural-hazards","isVorOnly":false,"title":"Natural Hazards"},"publishedOn":"2025-05-13 15:57:29","publishedOnDateReadable":"May 13th, 2025"},"versionCreatedAt":"2024-12-23 06:35:36","video":"","vorDoi":"10.1007/s11069-025-07301-4","vorDoiUrl":"https://doi.org/10.1007/s11069-025-07301-4","workflowStages":[]},"version":"v1","identity":"rs-5478044","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5478044","identity":"rs-5478044","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}