Implications of terrain resolution on modeling rainfall-triggered landslides using a TIN- based model
Introduction
Physically-based modeling is one of the approaches used to assess the vulnerability of natural basins to hillslope instability induced by extreme or prolonged precipitation. The increasing trend of weather-related disasters (Hoeppe, 2016) motivates the continuing interest in more reliable tools for prediction and analysis of precipitation-induced landslide events.
One of the issues extensively discussed in landscape modeling is the use of the appropriate grid Digital Elevation Model (DEM) resolution. Specifically, the question is whether adopting the finest available grid-DEM (hereinafter simply DEM) resolution is a justified choice, not only in terms of computational requirements, but also in terms of effective improvement of the model capability in predicting/determining the initiation of landslides (Cavazzi et al., 2013; Fuchs et al., 2014).
The DEM is used to extract morphological secondary attributes, such as slope, aspect, flow path, upstream contributing area, etc. Lack of accuracy in the primary attribute (i.e., elevation) would be propagated on the extracted morphological information (Wu et al., 2007; Vaze et al., 2010; Yang et al., 2014).
In landslide modeling, the local slope angle is the variable which most influences the calculation of the terrain stability, in both direct and indirect ways. Hydrological-stability approaches are based on the integration of distributed hydrological models with the simple infinite slope model (Montgomery and Dietrich, 1994; Iverson 2000; Claessens et al., 2005; Rosso et al., 2006; Arnone et al., 2011; Lepore et al., 2013). The landslide stability model computes the equilibrium of forces on a shallow soil prism. Gravity acts to initiate a slide as a function of the slope angle and the total wight of the soil, including water. Friction resists sliding and it is affected by soil moisture. The steeper the slope, the greater the component favoring slide initiation (direct effect). Catchment slope distribution also controls many of the hydrological terrain-based processes, such as the surface flow paths and the lateral redistribution of subsurface flows, which ultimately determine the local soil moisture, the duration of the transient regime after an event and thus the soil water pressures that impact the forces equilibrium.
Although high resolution digital terrain data allows a more realistic representation of topography and, consequently, a better analysis of hillslope and valley morphology, which are very important in the recognition of the topographic signature of valley incision by debris flows and landslides (Tarolli and Dalla Fontana, 2009), a high resolution DEM does not always imply a better performance in modelling the processes that lead to landslides. Several studies have explored how the grid-cell size of the input topography data may influence rainfall induced landslides. Some studies focus on landslide susceptibility (Chang et al., 1991; Lee et al., 2010; Grohmann et al., 2015; Arnone et al., 2016a; Cama et al., 2016) and others explore the impact of resolution on results from physically-based models (Zhang and Montgomery, 1994; Tarolli and Tarboton, 2006; Claessens et al., 2005; De Sy et al., 2013; Keijsers et al., 2011; Fuchs et al., 2014; Penna et al., 2014; Mahaigam and Olsen 2015; Viet et al., 2016). Most of the results of these studies agree that the coarser resolutions tend to smooth the terrain description, i.e., local slope angle decreases, thus reducing the number of unstable areas.
Specifically, Keijsers et al. (2011) used the LAPSUS-LS (Claessens et al., 2005) model and found that coarser resolutions reduced the ability to predict probability of failure at a particular location, yet stable areas were predicted correctly. However, many others concluded that the finest available resolution does not necessarily lead to better model performance (Arnone et al., 2016b; Fuchs et al., 2014), since modelling a physical process such as landslides, may depend on scales not detected with very high resolutions (Tarolli and Tarboton, 2006; Penna et al., 2014). At finer resolutions, the local surface topography is less representative of the process governing the landslide initiation and hence impacts the average size of the landslides (Freer et al., 2002; Tarolli and Tarboton, 2006). The availability of very-high resolutions DEMs (up to 1 m) (Yang et al., 2014; Noto et al., 2017; Francipane et al., 2020) resulting from the use of LIDAR begs the question of their value in landslide mapping (Wang et al., 2013; Fuchs et al., 2014; Ciampalini et al., 2016). Fuchs et al. (2014) found an improvement of 3% in determining slope instability by using < 10 m resolution, but they stated that such an improvement can have a small impact in applications where, for example, the soil terrain properties are poorly described and there is a lack of other data.
All studies mentioned so far make use of hydrological-landslide models that are grid-based, i.e., they require a grid-DEM to describe topography. Another class of hydrological and geomorphologic models uses Triangulated Irregular Networks (TINs) (e.g., CHILD by Tucker et al., 1999; tRIBS by Ivanov et al., 2004; tRIBS-Erosion by Francipane et al., 2012; CHM by Marsh et al., 2020), which make it possible to represent more efficiently the topography by increasing the number of nodes only where morphology is complex. TIN meshes can be built directly from measured elevation points but are more commonly derived from readily available grid-DEMs. Although the quality of simulations directly depends on the TIN mesh, the quality of the TIN discretization depends on the original DEM.
This study evaluates the influence of the DEM resolution on the slope stability analysis by using a distributed eco-hydrological-landslide model, which uses TINs derived from a DEM to describe the topography. Most hydrological-landslide models in the literature are grid-based and not much is written about the dependence of TIN- based models on terrain resolution. We use the tRIBS-VEGGIE-Landslide (Triangulated Irregular Network (TIN)-based Real-time Integrated Basin Simulator - VEGetation Generator for Interactive Evolution) (Lepore et al., 2013), which is capable of representing vegetation dynamics, and rainfall triggered landslides while simulating soil moisture evolution on the hillslope. The study addresses questions regarding the impact of the original DEM resolution on the landslide modeling, for given DEM-TIN conversion algorithm. Some of the questions are: How significant is the influence of the grid resolution on the estimation of slope distribution? How do the resolution impact terrain-driven hydrological processes, such as lateral redistribution, and then the landslide occurrence? How does the use of coarse resolutions modify the amount of the predicted total failure area?
The study area is the Mameyes basin, which is located in the Luquillo Experimental Forest (Puerto Rico), where numerous slope stability analyses have been carried out with the same model (Lepore et al., 2013; Dialynas et al., 2016; Arnone et al., 2016b). The impact of the original DEM resolution on tRIBS–VEGGIE landslide output is studied using different resampled DEMs at 20, 30, 50, and 70 m resolution (from the available 10 m DEM) to obtain the triangulated irregular network required by the model.
Section snippets
tRIBS-VEGGIE-landslide model
The tRIBS-VEGGIE-Landslide model (Lepore et al., 2013) couples the eco-hydrological model tRIBS-VEGGIE (Ivanov et al., 2008) and the infinite slope analysis in order to compute the factor of safety (FS) of a slope as a response to the soil moisture dynamics.
The hydrological component of the model reproduces essential hydrologic processes over the complex topography of a river basin (e.g., infiltration, evapotranspiration, interception, lateral redistribution and soil moisture dynamics). It
Basin description
The Mameyes basin is within the Luquillo Experimental Forest (LEF), in the northeast of the island of Puerto Rico, USA. It has an area of 16.7 km2, with an elevation ranging between 104.2 and 1046 m a.s.l. (Fig. 1a). About 30% of the basin has a slope greater than 25 deg (Fig. 1a). The basin is one of the wettest basins in Puerto Rico and is characterized by a high variability in rainfall and air temperature throughout the basin. The mean annual precipitation (MAP) ranges between 3000 and
Results
Slope stability in the model depends on terrain representation and simulated hydrological processes, both dependent on resolution. For given mechanical soil properties, three variables influence the local failure: depth of hypothetical plane of failure, slope, and soil moisture.
Summary and discussion
The effects of the original DEM size on the slope stability modeling have been explored by analyzing variables and processes that directly (i.e., slope) and indirectly (i.e., soil moisture dynamics) are involved in triggering failures. In contrast to other efforts, a distributed eco-hydrological-landslide model based on an irregular mesh, that is better suited to describe the topography, was used. A 10 m resolution DEM available for the study area was resampled to the resolutions of 20, 30, 50,
Conclusions
This study evaluated the hydro-geomorphological influences of DEM resolution on the slope stability analysis by using a distributed eco-hydrological-landslide model that uses a Triangulated Irregular Network (TIN) to describe the topography. The model has been applied to the Mameyes basin (Puerto Rico), where numerous landslide analyses have been carried out in the past (Lepore et al., 2013; Arnone et al., 2016b).
The results demonstrated that the use of a TIN-based hydrological-landslide model
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 and data availability
Map of soil is available from the “Soil Survey of Caribbean National Forest and Luquillo Experimental Forest, Commonwealth of Puerto Rico”, at the USDA Forest Service website www.nrcs.usda.gov. All data and calibrated parameters can be obtained from Dr. E. Arnone, [email protected]. The work of Dr. Dialynas while at the Georgia Institute of Technology and the collaboration with Drs. Arnone and Noto by Dr. Bras' research group has been supported by the National Science Foundation (Luquillo
References (52)
- et al.
Effect of raster resolution and polygon-conversion algorithm on landslide susceptibility mapping
Environ. Model. Software
(2016) - et al.
Physically-based and distributed approach to analyze rainfall-triggered landslides at watershed scale
Geomorphology
(2011) - et al.
Are fine resolution digital elevation models always the best choice in digital soil mapping?
Geoderma
(2013) - et al.
TRIBS-Erosion: a parsimonious physically-based model for studying catchment hydrogeomorphic response
Catena
(2012) - et al.
The effect of DEM resolution on the computation of the factor of safety using an infinite slope model
Geomorphology
(2014) Effects of spatial resolution on slope and aspect derivation for regional-scale analysis
Comput. Geosci.
(2015)Trends in weather related disasters – consequences for insurers and society
Weather Clim. Extremes
(2016)- et al.
Preserving, high-resolution surface and rainfall data in operationalscale basin hydrology: a fully-distributed physically- based approach
J. Hydrol.
(2004) - et al.
Calibration and resolution effects on model performance for predicting shallow landslide locations in Taiwan
Geomorphology
(2011) - et al.
Runoff cascades, channel network and computation hierarchy determination on a structured semi-irregular triangular grid
J. Hydrol.
(2001)
Integration of fuzzy logic and image analysis for the detection of gullies in the Calhoun Critical Zone Observatory using airborne LiDAR data
ISPRS J. Photogrammetry Remote Sens.
The role of soil processes in determining mechanisms of slope failure and hillslope development in a humid-tropical forest eastern Puerto Rico
Geomorphology
Impacts of DEM resolution, source, and resampling technique on SWAT-simulated streamflow
Appl. Geogr.
Hillslope to valley transition morphology: new opportunities from high resolution DTMs
Geophys. J. Roy. Astron. Soc.
Statistical analysis of drainage density from digital terrain data
Geomorphology
Impact of DEM accuracy and resolution on topographic indices
Environ. Model. Software
Systematic triangulations
Comput. Vis. Graph Image Process
Modeling the effects of elevation data resolution on the performance of topography-based watershed runoff simulation
Environ. Model. Software
A study on DEM-derived primary topographic attributes for hydrologic applications: sensitivity to elevation data resolution
Appl. Geogr.
What is the effect of LiDAR-derived DEM resolution on large-scale watershed model results?
Environ. Model. Software
Accounting for soils parameter uncertainty in a physically-based and distributed approach for rainfall-triggered landslides
Hydrol. Process.
Hydraulic properties of porous media
Exploring relationships between grid cell size and accuracy for debris-flow susceptibility models: a test in the Giampilieri catchment (Sicily, Italy)
Environ. Earth Sci.
The effect of dem resolution on slope and aspect mapping
Cartogr. Geogr. Inf. Syst.
The effectiveness of high-resolution LiDAR data combined with PSInSAR data in landslide study
Landslides
DEM resolution effects on shallow landslide hazard and soil redistribution modelling
Earth Surf. Process. Landforms
Cited by (5)
DEWS: A QGIS tool pack for the automatic selection of reference rain gauges for landslide-triggering rainfall thresholds
2023, Environmental Modelling and SoftwareAn open-source and QGIS-integrated physically based model for Spatial Prediction of Rainfall-Induced Shallow Landslides (SPRIn-SL)
2023, Environmental Modelling and SoftwareCitation Excerpt :Virtually, this should entail important advantages in SLs modelling, especially in understanding the role of instability factors controlled by topography. However, this is not as straightforward as would expected, since it was found that the performance of process-based models does not improve automatically by reducing the grid size (Claessens et al., 2005; Fuchs et al., 2014; Penna et al., 2014; Arnone et al., 2021). Moreover, in some cases, the use of very fine resolutions cannot be helpful if other factors, such as bedrock morphology or soil piping, drive the triggering mechanisms of SLs (Freer et al., 2002; Crosta et al., 2003; Tarolli and Tarboton, 2006; Penna et al., 2014).
On the Leonardo's rule for the assessment of root profile
2022, Ecological EngineeringCitation Excerpt :The former assumes a perfect binary graph where from every vertex only two branches can emerge; a more generalized form allows for more links emerging from the mother branch and a discrepancy from the cross-sectional area (van Noordwijk et al., 1994). The more generalized scheme has been adopted by Arnone et al. (2016) and Ielapi et al. (2019) to develop a Root Topological Model (RTM) that could serve as support to estimate the additional resistance exerted by roots in stabilizing hillslopes against shallow landslides (Lepore et al., 2013; Arnone et al., 2021; Pollen and Simon, 2005). The object of this communication is demonstrating the physical interpretation of the Lrule parameters by deriving analytical formulations of such parameters, valid in specific soil conditions.
Research in the Field of Natural Hazards Based on Bibliometric Analysis
2023, Natural Hazards Review