Spatial dendrogeomorphic sampling based on the specific tree growth responses induced by the landslide mechanism
Introduction
Landslides are among the most dangerous geomorphic processes worldwide (Guzzetti et al., 1999; Geertsema et al., 2006; Pánek et al., 2008; Grahne and Jaldell, 2017). As important natural hazard, they annually cause severe damage to human infrastructure or even fatalities (Petley, 2012; Froude and Petley, 2018). The knowledge of past landslide activity is one of the most important aspects of landslide research (Pánek, 2015). Data about temporal landslide behaviour are crucial for assessing their triggers (Lopez-Saez et al., 2012a) or magnitude-frequency ratio (Corominas and Moya, 2010). The crucial aspect regarding current and future climate changes is the detailed knowledge of the relationship between landslide activations and causing meteorological triggers. This relationship can be effectively reconstructed using data about historical landslide activity (Lopez Saez et al., 2012a). Combination of climatic forecast models with knowledge of the past landslide-climate relationships is a key for prediction future landslide activity (Stoffel and Huggel, 2012). Data about the spatial distribution of landslide movements are crucial from a forest management or urban planning point of view (Martire et al., 2012). The reactivation of partial, spatially limited zones is typical for the behaviour of large complex landslides (Cockburn et al., 2016). Thus, the general view of landslide activity can be considerably influenced by the various behaviours of (in)active zones. From a risk assessment point of view, knowledge about the most active zones is crucial, particularly if accompanied by temporal characteristics of the movements (Raška et al., 2015). Data about the temporal aspects of past landslide activity are generally rare (Lopez Saez et al., 2012a), and detailed knowledge about past spatial landslide behaviour is even more rare and more difficult to obtain than knowledge about the simple chronology of past landslide events. The creation of a detailed spatial reconstruction of landslide activity is very limited if historical monitoring is missing. Historical orthophotos can serve as a possible data source, but their availability, chronological range or quality is usually limited (Walstra et al., 2007). Archives and newspaper reports are usually very incomplete and limited to only temporal information (Raška et al., 2015).
Dendrogeomorphic methods are good alternatives for the long-term monitoring of landslide activity (Alestalo, 1971). Past landslide behaviour can be reconstructed with good precision (Lopez Saez et al., 2012a; Šilhán, 2019) based on data from the tree ring series of trees occupying landslide surfaces. Moreover, the length of reconstruction can reach several centuries up to a millennium (Šilhán et al., 2015; Zhang et al., 2019). Trees can provide high-resolution temporal and spatial information via their positions on the landslide surface (Corona et al., 2014). Several methods of using tree position data for spatial landslide reconstruction have been employed in the past. The simplest and most frequently used method is the realization of “event-response maps”. The positions of all trees containing landslide signals in a particular year are highlighted on the map (e.g., Stefanini, 2004; Wiktorowski et al., 2017). This simple concept was extended by Lopez Saez et al. (2012b), who tested the spatial position of affected trees in the case of low indicators of landslide activity using the Moran index (Moran, 1950). The mentioned approach assumes that the spatial clustering of disturbed trees is affected by landslide reactivation (i.e., secondary movement of the spatially limited area sometime after the initial landslide). As noted, this approach was used only for event years with a limited number of trees with tree ring signals. Thus, the validity of the theory of disturbed tree clustering during landslide movements should be tested for all landslide events (i.e., even for events detected in the tree rings of a high number of trees). The next possibility of past spatial landslide reconstruction is the interpolation of recurrence intervals detected between two tree ring signals in each tree (Šilhán, 2016). This approach can be used to obtain information about landslide activity even from areas not covered by trees. The most developed approach for the tree ring-based reconstruction of landslide spatial activity is the creation of probability maps, as introduced by Lopez Saez et al. (2012b), who combined the spatial positions of trees with the calculated frequency of landslide events processed with a Poisson probability model.
The spatial reconstruction of landslide movements has been performed for various types (rotational, translational, shallow, etc.) of landslides in the past. However, landslide areas often express complex geomorphology containing various types of morphological zones (Hungr et al., 2013). Each landslide type or morphological zone with a specific movement mechanism is able to induce different growth disturbances with different intensities in the tree ring series of affected trees (Bollati et al., 2012). For example, block-type movements with rotational mechanisms usually generate reaction wood in coniferous trees (Šilhán et al., 2013), whereas shallow translational movements are typical of inducing abrupt growth suppression in tree ring series (Lopez Saez et al., 2012b). Spatially limited area or specific morphological zone within the entire complex landslide area is usually activated during the landslide events (Guida et al., 2008). As different movement mechanisms in various zones will most likely impact individual growth behaviour in different ways, one can suppose that trees from different zones will variously participate in the resulting total landslide chronology. Thus, trees from some types of landslide morphological zones are able to provide more valuable data about past landslide activity than trees from other zones. The spatial reconstruction of past landslide activity should reflect the specific morphology of the landslide area. Simple spatial interpolation of point values for the entire landslide area seems to be unsuitable, particularly for complex landslides. Some studies suggest this effect and recommend the analysis of various landslide morphological zones separately (Van Den Eeckhaut et al., 2009; Šilhán et al., 2012, 2014; Lopez-Saez et al., 2017), but none have exactly verified, quantified and studied it in detail.
Regarding the abovementioned limitations and gaps in the dendrogeomorphic reconstruction of past landslide behaviour, the aims of this study are i) to test the theory of disturbed tree clustering when affected by landslide movements, ii) to distinguish types of tree growth disturbances induced by various movement mechanisms in different morphological zones of landslides and iii) to detect trees from which morphological zones are able to provide the most valuable data for the final spatiotemporal reconstruction of landslide movements. Toincrease the robustness of the obtained results, the mentioned study aims were adressed for a group of ten individual landslides in the Outer Western Carpathians.
Section snippets
Study area
All of the studied landslides are located in the Hostýnsko-vsetínská hornatina Mts. (centred at 49.4° N and 18.0° E; Fig. 1), a region with a known high frequency of slope deformations, particularly landslides (Kirchner and Krejčí, 1998). The region with a prevalence of studied landslides covers approximately 300 km2 (approximately 30 × 10 km). Its relief has typical highland to mountain character, with the highest peaks reaching 1000 m a.s.l., steep slopes reaching up to 20° and mean
Geomorphic mapping and defining the types of morphological zones
Each studied landslide was geomorphically mapped (1:500) using LiDAR (light detection and ranging)-based digital elevation model (DEM) and orthophotos as the first step. The positions of all trees expressing signs of past landslide activity (tilted or bent stems, stretched or damaged tree roots) were recorded using a GNSS (Global Navigation Satellite System) device and specified using orthophotos. Field geomorphic observation and analysis of the resulting geomorphic map were used to determine
Morphology of the studied landslides
All landslides expressed similar morphological structures and spatial distributions of individual morphological zones (Fig. 4). The zone of landslide blocks was always situated just below the source zone. The only exception was the Vlčice landslide, whose entire area was dominated by shallow movements. The zone of shallow movements followed the zone of landslide blocks in the lower parts of the landslide and was absent in the case of the Soláň landslide. The zone of frontal lobes was mapped in
Discussion
Tree ring-based data about past landslide behaviour were collected from ten individual landslides in the Outer Western Carpathians. This dataset was used for detailed analysis of the spatial distribution of movement signals in trees regarding their positions in six different morphological zones of landslides. The analysis of 560 signals of past landslide activity (tree growth disturbances) coming from the 1030 tree ring series of 515 disturbed trees led to the detection of 110 historical
Conclusions
This study analysed various growth responses and the spatial aspects of sampled trees within the different morphological zones of landslides. Data from 1030 increment cores from ten different complex landslides were studied in detail to increase the robustness and universality of the obtained results.
The testing of the spatial patterns of 515 studied trees on all landslides provided evidence of a limited tendency for the clustering of disturbed trees during individual reactivation events.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The Czech Science Foundation project 19-01866S and the Specific Research Project of the University of Ostrava Nr. SGS02/PřF/2019–2020 supported this study. The language was reviewed by American Journal Experts.
References (56)
- et al.
The origin of shallow landslides in Moravia (Czech Republic) in the spring of 2006
Geomorphology
(2008) - et al.
Dendrochronological and geomorphological investigations to assess water erosion and mass wasting processes in the Apennines of Southern Tuscany (Italy)
Catena
(2012) - et al.
Contribution of dendrochronology to the determination of magnitude–frequency relationships for landslides
Geomorphology
(2010) - et al.
Defining optimal sample size, sampling design and thresholds for dendrogeomorphic landslide reconstructions
Quat. Geochronol.
(2014) - et al.
An overview of recent large catastrophic landslides in northern British Columbia, Canada
Eng. Geol.
(2006) - et al.
Landslide hazard evaluation: a review of current techniques and their application in a multi-scale study, Central Italy
Geomorphology
(1999) - et al.
The mapping of soil taxonomic units via fuzzy clustering – a case study from the Outer Carpathians, Czechia
Geoderma
(2018) - et al.
Probability maps of landslide reactivation derived from tree-ring records: Pra Bellon landslide, southern French Alps
Geomorphology
(2012) - et al.
Mechanism of the slow-moving landslide in Jurassic red-strata in the three Gorges Reservoir, China
Eng. Geol.
(2014) - et al.
Gigantic low-gradient landslides in the northern periphery of the Crimean Mountains (Ukraine)
Geomorphology
(2008)