Continuous monitoring of landslides is of basic importance for understanding their behavior, defining their 3D geometry, and providing a basis for early warning purposes. While a number of instrumentations can be used for tracking surface displacement, only automatic or fixed multi-module inclinometers can be used for continuous monitoring of displacement at depth, providing valuable information for landslide geometry reconstruction. Since these instruments are very expensive, thus rarely used, a low-cost and multi-module fixed inclinometer for continuous landslide monitoring has been developed. In this paper, the electronics of the system, including sensor characteristics and optimization, controlling software, and structure are presented. For system development, a single module prototype was first developed and tested in the field to ensure sufficient measuring performance. Subsequently, the multi-module system was designed, assembled, and tested in controlled conditions. Test results indicate the good performance of the system with a displacement measuring accuracy of 0.37% of the length of the inclinometer chain. The linearity test indicates the high linearity of the measures, especially in the range ±20°, which is the typical operating range of such kinds of instrumentations. The thermal efficiency test indicates the high efficiency of the system in preventing measuring errors caused by thermal drifting.

Low-cost MEMS accelerometers have the potential to be used in a number of tilt-based monitoring applications but have the disadvantage of being very sensitive to temperature variation (thermal drift). In this paper, we analyze the thermal behavior of a low-cost sensor in the range −10 to +45 °C in order to provide a simple compensation strategy to mitigate this problem. For sensor analysis, we have developed a miniaturized thermal chamber, which was mounted on a tilting device to account for tilt angle variation. The obtained raw data were used to construct low degree polynomial equations that by relating the measurement error induced by thermal drift (i.e., acceleration residuals) to temperature and inclination (of each specific axis), can be used for thermal compensation. To validate our compensation strategy, we performed a field monitoring test and evaluated the compensation performance by calculating RMS errors before and after correction. After compensation, the RMS errors calculated for both the X and Y axes decreased by 96%, indicating the potential of using a simple set of equations to solve common drawbacks that currently make low-cost MEMS sensors unsuitable for tilt-based monitoring applications.

A key component of disaster management and infrastructure organization is predicting cumulative deformations caused by landslides. One of the critical points in predicting deformation is to consider the spatio-temporal relationships and interdependencies between the features, such as geological, geomorphological, and geospatial factors (predisposing factors). Using algorithms that create temporal and spatial connections is suggested in this study to address this important point. This study proposes a modified graph convolutional network (GCN) that incorporates a long and short-term memory (LSTM) network (GCN-LSTM) and applies it to the Moio della Civitella landslides (southern Italy) for predicting cumulative deformation. In our proposed deep learning algorithms (DLAs), two types of data are considered, the first is geological, geomorphological, and geospatial information, and the second is cumulative deformations obtained by permanent scatterer interferometry (PSI), with the first investigated as features and the second as labels and goals. This approach is divided into two processing strategies where: (a) Firstly, extracting the spatial interdependency between paired data points using the GCN regression model applied to velocity obtained by PSI and data depicting controlling predisposing factors; (b) secondly, the application of the GCN-LSTM model to predict cumulative landslide deformation (labels of DLAs) based on the correlation distance obtained through the first strategy and determination of spatio-temporal dependency. A comparative assessment of model performance illustrates that GCN-LSTM is superior and outperforms four different DLAs, including recurrent neural networks (RNNs), gated recurrent units (GRU), LSTM, and GCN-GRU. The absolute error between the real and predicted deformation is applied for validation, and in 92% of the data points, this error is lower than 4 mm.

Historical movement of the Montaguto earth flow in southern Italy has periodically destroyed residences and farmland, and damaged the Italian National Road SS90 and the Benevento-Foggia National Railway. This paper provides maps from an investigation into the evolution of the Montaguto earth flow from 1954 to 2010. We used aerial photos, topographic maps, LiDAR data, satellite images, and field observations to produce multi-temporal maps. The maps show the spatial and temporal distribution of back-tilted surfaces, flank ridges, and normal, thrust, and strike-slip faults. Springs, creeks, and ponds are also shown on the maps. The maps provide a basis for interpreting how basal and lateral boundary geometries influence earth-flow behavior and surface-water hydrology.