**5. Discussion**

Landslides have been widely considered as principal mass-wasting agents in areas experiencing varied influence of several causative factors (i.e., lithology, geological setting, climate regime, etc.). However, patterns of landslides are rarely addressed as a surface manifestation of interrelationships between morphostructural setting, lithology, and climate. Here, we have attempted to understand such interrelationships in the context of landslide distribution patterns in the hilly piedmont area of Abruzzo Region. Historical landslides analysis allowed us to understand that the distribution, mechanisms, and types of mass movements in the study area strictly correlate with the different physiographic, lithological, and geological–structural settings. The work mostly focuses on three landslide case studies analyzed with the aim of highlighting the multitemporal evolution of the landslide phenomena, emphasizing the role of lithological and morphostructural features on landslide types and the interplay between such processes and the geomorphological evolution. Landslide density maps, directly combined with the inventories and databases from which they were obtained, allowed us to define and graphically show different sectors in the study area. In each sector, we have outlined the landslide types and the mechanisms that mostly determine the slope instability reflecting the geological–structural and geomorphological setting. Selected case studies are representative of the most characterizing and frequent slope instability processes over the hilly piedmont area, showing different influences on geomorphological dynamics according to the physiographic and litho-structural setting. In the northern and central sectors, landslide phenomena affect a gently hilly area made of clayey–sandy deposits (with sandstone-conglomerate sequence on top), gently dipping towards the northeast or horizontal. In the southernmost sector, landslide phenomena affect a landscape derived from exogenous processes (fluvial and slope processes) on mostly chaotic marly–clayey deposits or chaotic succession of calcareous-marly deposits.

Several previous studies in the Abruzzo Region [48,79,90,91] analyzed and described the widespread slope geomorphic processes, showing an organic correlation between the morphostructural/geological setting and landslide types as the result of the dynamic interaction between morphostructural factors, linked to the conflicting tectonic activity and regional uplift, and morphosculptural factors, linked to drainage network linear down– cutting and slope gravity processes. The slope evolution is mainly related to the interplay of different landslide types referable to lateral spreading, rockfall, earth flow, rotational and translational sliding, evolving into complex movements and systems.

In this framework, local features such as lithology and morphostructural framework should be noted to control the occurrence and distribution of landslides. Nonetheless, interrelationships of these factors have been rarely associated with spatiotemporally varying landslide distribution patterns. However, there are many limitations to infer temporally varying landslide distribution, such as delineation of individual failure events on the

reactivated landslide, loss of landslide scarp caused by the successive mass movement, etc. Different predisposing and triggering factors can influence the stability of slopes and can cause landslides, among which heavy rainfall events are intended to be a significant one. It is well known that extreme and localized heavy rainfalls constitute the main triggering causal factor of landslides. Rainfall pattern is strongly controlled and influenced by climate regime and its variations. Therefore, it is to be expected that climate changes could influence slope stability at different temporal and geographical scales. The frequency and the intensity of heavy rainfall events are also increasing, although both at local and regional scale the average annual rainfall is not showing significant changes. The assessment of the effects of climate change on the natural environment is an open issue for the scientific community trying to establish a relation between climate change and its potential effects on the occurrence, or lack of occurrence, of landslides. However, the effects of changes in climate regimes on landslides (as on other geo-hydrological hazards) remain difficult to quantify and predict.

This work represents a useful source for investigating landslide behaviors in terms of spatial and temporal distribution, as well as for analyzing and attempting correlation between climate regime, historical landslides, and present-day geomorphological activity.

In order to understand and quantify how climate regime and its variability could affect landslides, a climatic analysis was performed using a 65-year period rainfall gauges data. Figure 12a shows the spatial distribution of annual average rainfall in the study area, with minimum values (~700 mm/year) recorded along with the coastal areas and the southeastern sector of the Maiella Massif; these rainfall values are gradually increasing, moving towards the innermost areas, where the maximum values (about 1150 mm/year) are reached. Similarly, the analysis of the annual average rainfall diagram from 1950 to 2015 shows values ranging from ~530 to ~1130 mm/year, with a clear decreasing trend over the examined period (Figure 12b).

Taking into account the spatial distribution (landslide heatmaps), the location and abundance of landslides, and the geomorphological features of selected case studies, the landscape dynamics and activity of the hilly piedmont area have also been confirmed by the interferometric analysis. Considering that movements recorded by interferometric data can be due to different causes acting at different scales (i.e., uplift, subsidence, landslide, etc.), the PSInSAR technique was here used as a tool for systematic monitoring of ground deformation related to slope instability. The presence and temporal persistence of clusters of anomalies within the main landslide body act as the most important parameters that show present-day landscape changes linked to temporal landslide dynamic. Figure 12c shows the total number of persistent anomalies detected over the period 2002–2010, clipped by landslide bodies (green polygons) mapped by the IFFI project [59,60] over the hilly piedmont area of Abruzzo Region. Analyzed data show a spatial distribution of negative movements (lowering) and positive movements (raising), which reflect the extension of the investigated landslide phenomena with the highest values located in the central-southern sectors and locally in the northernmost coastal slopes.

Moreover, in order to attempt a general correlation between long-term rainfall trends and trends in landslide occurrence, a statistical analysis of the annual distribution of landslides was carried. This kind of analysis was completed collecting data from historical sources, technical reports, and updated catalogues [60,65,86] containing a variety of historical, geographical, geomorphological, and bibliographical information on landslides.

Reported diagram (Figure 12d) stores information regarding dates of occurrences of several landslides, starting from the year 1950 until the present, with non-homogeneous rates of recorded landslides per year. A detailed analysis shows that the frequency remains under the value of 10 landslides per year starting from 1950 to 1990, with unique exception years (i.e., 1954, 1956, 1986). Subsequently, growth rates, from 1991 onwards, clearly increase. Even if the variance of the number of reported landslides over time is also due to the different availability of sources of information and not necessarily linked to the real frequency of landslide occurrences [92], it is possible to consider this analysis

a reasonably true reflection of reality for the period 1950–2018. Despite the presence of a timespan with a lack of suitable and univocal data (i.e., year, day, hour, etc.) on landslides' activation–reactivation in the period 2002–2009, it is possible to note that the annual landslide distribution ranges from ~5 to ~75 individual events. Considering the complete distribution of the number of landslides during the years covered, the annual landslide distribution during this period shows different periods of landslide activity and abundance. It is possible to note a nearly stable trend in the first 20-year time record (1950–1970), followed by a general increasing trend in the 1970–2000-time record, also supported and corroborated by a weak increasing trend in the last decade (2010–2018). The identified trend should be considered in relation to both the incremental data availability and the rise in mass-wasting processes, as directly shown by historical information on past and current landslides. Moreover, regarding the study area, it is not correct to conclude that a lack of reported landslides in a given time interval would be due to a minor activity of gravitational mass wasting or to a gap in the documental source, as marked by the present-day geomorphological activity testified by the temporal persistence anomalies of movement related to slope instability (Figure 12c).

**Figure 12.** (**a**) Average annual rainfall map. Black dots represent rainfall gauges. (**b**) Average annual rainfall diagram from a 65-year time record (1950–2015). (**c**) InSAR observations for the selected area over the hilly–piedmont area. Mean line-of-sight (LOS) velocity for the period 2002–2010 from Envisat descending track. Only the Persistent Scatterers (PSs) that fall within the landslide areas (dark green polygons) have been selected and are represented as colored dots. Positive values represent the motion of the ground toward the satellite (raising), and negative values represent the motion away from the satellite (lowering). Green polygons represent landslide bodies detected by the IFFI Project [60]. (**d**) Distribution of annual landslide occurrences over the 1950–2015 period (derived from [60,65,86]).

The final combination and overlapping between the spatial and temporal landslide distribution pattern, the mismatch between landslide areas and sectors characterized by high rainfall density, the lack of correspondence between decreasing annual average rainfall trend and the increasing annual landslide distribution allowed us to highlight the interplay between the morphostructural/geological framework and landslide dynamics in the hilly piedmont area of Abruzzo Region. The present study allowed us to better characterize the present-day landscape setting of the study area, confirming that it is characterized by active geomorphological processes, mostly represented by slope instabilities (i.e., rotational and translational slides, complex landslides, earth flows, and rockfalls). This was obtained from historical information on past and current landslides. Currently, geomorphological activity and landslide dynamics are testified and supported by interferometric data (clusters of persistent anomalies, detected over the period 2002–2010, and clipped by landslide' polygons) with negative movements (values between −10 and −2) and positive ones (values between 2 and 10) heterogeneously distributed over the hilly piedmont area. Detailed multitemporal geomorphological analysis on selected case studies (San Martino sulla Marruccina, Roccamontepiano, and Montebello sul Sangro) show multiple activations of the main event since the 18th century onwards with large landslides (mostly dormant and/or abandoned) and small landslides (generally more recent and active) constituting the wide and complex landslide systems and reflecting the physiographic, geological– structural, and geomorphologic setting.

In conclusion, by summarizing data obtained from multitemporal and multidisciplinary, it is possible to sugges<sup>t</sup> that landslide occurrence and the dynamics of the hilly piedmont area of Abruzzo Region are not directly linked to climate regime variations, but the most influential factors are represented by the lithological and morphostructural setting. These predisposing factors are strictly related to a cuesta, mesa, and plateau landscape in which it is possible to outline the landslide types and the mechanisms that mostly determine the slope morphogenesis and are characterizing of the specific geological–structural setting. To these characterizing landslide types are obviously associated and sometimes super-imposed a set of landslides secondary or however controlled by local conditions, single factors (i.e., extreme heavy rainfall events), and not by the whole morphostructural setting. Moreover, considering the historical landslide events and the geomorphological activity of the area, most of the recorded landslides could be considered as reactivations of pre-existing ones (dormant slides and/or paleolandslides), which have occurred in periods of climatic and geomorphological conditions different from those of the present, evolving in complex movements and systems because of the absence of sustainable land planning and appropriate landslide hazard mitigation measures.
