Landslide inventories and their statistical properties

Landslides are generally associated with a trigger, such as an earthquake, a rapid snowmelt or a large storm. The landslide event can include a single landslide or many thousands. The frequency–area (or volume) distribution of a landslide event quanti?es the number of landslides that occur at different sizes. We examine three well‐documented landslide events, from Italy, Guatemala and the USA, each with a different triggering mechanism, and ?nd that the landslide areas for all three are well approximated by the same three‐parameter inverse‐gamma distribution. For small landslide areas this distribution has an exponential ‘roll‐over’ and for medium and large landslide areas decays as a power‐law with exponent ‐2·40. One implication of this landslide distribution is that the mean area of landslides in the distribution is independent of the size of the event. We also introduce a landslide‐event magnitude scale m_L = log(N_LT), with N_LT the total number of landslides associated with a trigger. If a landslide‐event inventory is incomplete (i.e. smaller landslides are not included), the partial inventory can be compared with our landslide probability distribution, and the corresponding landslide‐event magnitude inferred. This technique can be applied to inventories of historical landslides, inferring the total number of landslides that occurred over geologic time, and how many of these have been erased by erosion, vegetation, and human activity. We have also considered three rockfall‐dominated inventories, and ?nd that the frequency–size distributions differ substantially from those associated with other landslide types. We suggest that our proposed frequency–size distribution for landslides (excluding rockfalls) will be useful in quantifying the severity of landslide events and the contribution of landslides to erosion. Copyright © 2004 John Wiley & Sons, Ltd.     

Effects of forest harvesting on the occurrence of landslides and debris flows in steep terrain of central Japan

Landslides and debris flows associated with forest harvesting can cause much destruction and the influence of the timing of harvesting on these mass wasting processes therefore needs to be assessed in order to protect aquatic ecosystems and develop improved strategies for disaster prevention. We examined the effects of forest harvesting on the frequency of landslides and debris flows in the Sanko catchment (central Japan) using nine aerial photo periods covering 1964 to 2003. These photographs showed a mosaic of different forest ages attributable to the rotational management in this area since 1912. Geology and slope gradient are rather uniformly distributed in the Sanko catchment, facilitating assessment of forest harvesting effects on mass wasting without complication of other factors. Trends of new landslides and debris flows correspond to changes in slope stability explained by root strength decay and recovery; the direct impact of clearcutting on landslide occurrence was greatest in forest stands that were clearcut 1 to 10 yr earlier with progressively lesser impacts continuing up to 25 yr after harvesting. Sediment supply rate from landslides in forests clearcut 1 to 10 yr earlier was about 10‐fold higher than in control sites. Total landslide volume in forest stands clearcut 0 to 25 yr earlier was 5·8 × 10³ m³ km^?2 compared with 1·3 × 10³ m³ km^?2 in clearcuts >25 yr, indicating a fourfold increase compared with control sites during the period when harvesting affected slope stability. Because landslide scars continue to produce sediment after initial failure, sediment supply from landslides continues for 45 yr in the Sanko catchment. To estimate the effect of forest harvesting and subsequent regeneration on the occurrence of mass wasting in other regions, changes in root strength caused by decay and recovery of roots should be investigated for various species and environmental conditions. Copyright © 2007 John Wiley & Sons, Ltd.     

Landslide incidence zonation in the Rio Mendoza valley,Mendoza Province,Argentina

This paper presents a landslide incidence zonation map showing the percentage of underlying material involved in mass‐movement processes in the Rio Mendoza valley, Argentina. The landslide incidence zonation map was derived from an inventory map of landslides and reveals that many areas of the Rio Mendoza valley are implicated in this kind of process. A correlation has been found between the occurrence of landslides, earthquakes, and rainfall. The relation between lithology and landslides is clear: areas covered by friable sedimentary and volcanic rocks of the Choiyoi Group are prone to debris ?ows and complex landslides. The slope map has been ranked and a general relation between slope and type of event is shown. Falls commonly develop in high‐angle slopes. Copyright © 2004 John Wiley & Sons, Ltd.     

Slope control on the frequency distribution of shallow landslides and associated soil properties,North Island,New Zealand

The extrapolation of results from field trials to larger areas of land for purposes of regional impact assessment is an important issue in geomorphology, particularly for landform properties that show high stochastic variability in space and time, such as shallow landslide erosion. It is shown in this study, that by identifying the main driver for spatial variability in shallow landslide erosion at field scales, namely slope angle, it is possible to develop a set of generic functions for assessing the impact of landslides on selected soil properties at larger spatial scales and over longer time periods. Research was conducted within an area of pastoral soft‐rock Tertiary hill country in the North Island of New Zealand that is subject to infrequent high intensity rainfall events, producing numerous landslides, most of which are smaller than several hundred square metres in size and remove soil to shallow depths. All landslides were mapped within a 0·6 km² area and registered to a high resolution (2 m) slope map to show that few landslides occur on slopes < 20° and 95% were on slopes > 24°. The areal density of landslides from all historical events showed an approximately linear increase with slope above 24°. Integrating landslide densities with soil recovery data demonstrates that the average value of a soil property fluctuates in a ‘saw‐tooth’ fashion through time with the overall shape of the curve controlled by the frequency of landslide inducing storm events and recovery rate of the soil property between events. Despite such fluctuations, there are gradual declines of 7·5% in average total carbon content of topsoil and 9·5% in average soil depth to bedrock, since the time of forest clearance. Results have application to large‐scale sediment budget and water quality models and to the New Zealand Soil Carbon Monitoring System (CMS). Copyright © 2012 John Wiley & Sons, Ltd.     

Testing a model for predicting the timing and location of shallow landslide initiation in soil‐mantled landscapes

The growing availability of digital topographic data and the increased reliability of precipitation forecasts invite modelling efforts to predict the timing and location of shallow landslides in hilly and mountainous areas in order to reduce risk to an ever‐expanding human population. Here, we exploit a rare data set to develop and test such a model. In a 1·7 km² catchment a near‐annual aerial photographic coverage records just three single storm events over a 45 year period that produced multiple landslides. Such data enable us to test model performance by running the entire rainfall time series and determine whether just those three storms are correctly detected. To do this, we link a dynamic and spatially distributed shallow subsurface runoff model (similar to TOPMODEL) to an in?nite slope model to predict the spatial distribution of shallow landsliding. The spatial distribution of soil depth, a strong control on local landsliding, is predicted from a process‐based model. Because of its common availability, daily rainfall data were used to drive the model. Topographic data were derived from digitized 1 : 24 000 US Geological Survey contour maps. Analysis of the landslides shows that 97 occurred in 1955, 37 in 1982 and ?ve in 1998, although the heaviest rainfall was in 1982. Furthermore, intensity–duration analysis of available daily and hourly rainfall from the closest raingauges does not discriminate those three storms from others that did not generate failures. We explore the question of whether a mechanistic modelling approach is better able to identify landslide‐producing storms. Landslide and soil production parameters were ?xed from studies elsewhere. Four hydrologic parameters characterizing the saturated hydraulic conductivity of the soil and underlying bedrock and its decline with depth were ?rst calibrated on the 1955 landslide record. Success was characterized as the most number of actual landslides predicted with the least amount of total area predicted to be unstable. Because landslide area was consistently overpredicted, a threshold catchment area of predicted slope instability was used to de?ne whether a rainstorm was a signi?cant landslide producer. Many combinations of the four hydrological parameters performed equally well for the 1955 event, but only one combination successfully identi?ed the 1982 storm as the only landslide‐producing storm during the period 1980–86. Application of this parameter combination to the entire 45 year record successfully identi?ed the three events, but also predicted that two other landslide‐producing events should have occurred. This performance is signi?cantly better than the empirical intensity–duration threshold approach, but requires considerable calibration effort. Overprediction of instability, both for storms that produced landslides and for non‐producing storms, appears to arise from at least four causes: (1) coarse rainfall data time scale and inability to document short rainfall bursts and predict pressure wave response; (2) absence of local rainfall data; (3) legacy effect of previous landslides; and (4) inaccurate topographic and soil property data. Greater resolution of spatial and rainfall data, as well as topographic data, coupled with systematic documentation of landslides to create time series to test models, should lead to signi?cant improvements in shallow landslides forecasting. Copyright © 2003 John Wiley & Sons, Ltd.     

Terrain attributes of earthquake‐ and rainstorm‐induced landslides in orogenic mountain Belt,Taiwan

Landsliding induced by earthquakes and rainstorms in montane regions is not only a sculptor for shaping the landscape, but also a driver for delivering sediments and above‐ground biomass downstream. However, the terrain attributes of earthquake‐ and rainstorm‐induced landslides are less discussed comprehensively in Taiwan. As part of an island‐wide inventory, we here compare and contrast the landslide terrain attributes resulting from two catastrophic events: the Chi‐Chi earthquake (M _w = 7.6, September 1999) and typhoon Morakot (rainfall >2500 mm, August 2009). Results show that the earthquake‐induced landslides are relatively small, round‐shaped and prone to occur primarily in middle and toe of slopes. In contrast, the rainstorm‐induced landslides are larger, horseshoe‐shaped and preferentially occurring in slope toes. Also, earthquake‐induced landslides, particularly large landslides, are usually found at steeper gradients, whereas rainstorm‐induced landslides aggregate at gradients between 25° and 40°. Lithologic control plays a secondary role in landsliding. From an island‐wide perspective, high landslide density locates in the region of earthquake intensity ≥ VI or one‐day rainfall ≥600 mm day^?1. Through the landslide patterns and their terrain attributes, our retrospective approach sheds light on accessing the historical and remote events for close geophysical investigations. Finally, we should bear in mind that the landslide location, size, and terrain attributes varying with triggers may affect the landscape evaluation or biogeochemical processes in landslide‐dominated regions. Copyright © 2017 John Wiley & Sons, Ltd.     

Temporal and spatial variation of infilling processes in a landslide scar in a steep mountainous region,Japan

The duration of the soil‐depth recovery needed for reoccurrence of shallow colluvial landslides at a given site in humid regions is much longer than the return period of rainfall needed to generate sufficient pore water pressure to initiate a landslide. Knowledge of the rate of change in soil depth in landslide scars is therefore necessary to evaluate return intervals of landslides. Spatial variation in sediment transport at the Kumanodaira landslide scar in central Japan was investigated by field observations. Spatial distribution of the rate of change in soil depth was estimated using sediment transport data and geographic information system (GIS) analysis. Observations revealed that the timing of sediment transport differed for shallow and deep soil layers. Near‐surface sediment transport (mostly dry ravel and some shallow soil creep at depths ≤0·05 m) measured in sediment traps was active in winter and early spring and was affected by freezing–thawing; soil creep of subsoil (i.e. >0·05 m), monitored by strain probes, was active in summer and autumn when precipitation was abundant. Near‐surface sediment flux was estimated by a power law function of slope gradient. Deeper soil creep was more affected by relative location to the landslide scar, which influences soil depth, than by slope gradient. Our study indicated that the rate of soil‐depth recovery is high just below the head scarp of the landslide. Abrupt changes in the longitudinal slope topography immediately above, within and just below the head scarp became smoother with time due to degradation proximate to the landslide head scarp and flanks, as well as aggradation just below the head scarp. Copyright © 2014 John Wiley & Sons, Ltd.     

Landslide Amplification by Liquefaction of Runout‐Path Material after the 2008 Wenchuan (M 8·0) Earthquake,China

Here, we propose that an earthquake can trigger the failure of a landslide mass while simultaneously triggering liquefaction of runout‐path materials before the arrival of the landslide mass, thus greatly increasing the size and mobility of an overriding landslide. During the 2008 Wenchuan earthquake, about 60 000 landslides were triggered, directly resulting in about 20 000 casualties. While these landslides mainly originated from steep slopes, some landslides with high mobility formed in colluvial valley deposits. Among these, the most catastrophic was the Xiejiadian landslide in Pengzhou city, which traveled hundreds of meters before coming to rest. Through field investigation and laboratory testing, we conclude that this landslide primarily formed from colluvial deposits in the valley and secondarily from failure of slopes in granitic rock located uphill. Much of the granitic slope failure was deposited in the upper part of the travel path (near the slide head); the remainder was dispersed throughout the main landslide deposit. Superposition of deposits at the landslide toe indicates that landslide debris derived from colluvial soil was deposited first. The deposits at the landslide toe displayed flow characteristics, such as fine materials comprising basal layers and large boulders covering the deposit surface. We hypothesize that the main part of the landslide resulted from seismogenic liquefaction of valley colluvium, rather than from liquefaction potentially caused by undrained loading from the granitic slope failures impacting the colluvium. To examine the likelihood that seismogenic liquefaction occurred, we took samples from different areas of the landslide deposit and performed undrained cyclic shear tests on them in the laboratory. The results showed that the sandy soils that comprise most of the deposit are highly liquefiable under seismic loading. Therefore, we conclude that liquefaction of the colluvium in the valley during the earthquake was the main reason for this rapid (~46 m/s) long‐runout (1·7 km) landslide. Copyright © 2012 John Wiley & Sons, Ltd.     

Use of LIDAR‐derived images for mapping old landslides under forest

Large, deep‐seated landslides are common features in the Flemish Ardennes (Belgium). As most of these old (>100 years) landslides are located under forest in this hilly region, aerial photograph interpretation is not an appropriate landslide mapping method. This study tested the potential of LIDAR (Light Detection and Ranging) images for mapping old landslides under forest. Landslide inventory maps were created for a 125 km² area by applying the expert knowledge of seven geomorphologists to LIDAR‐derived hillshade, slope and contour line maps in a GIS environment. Each of the seven LIDAR‐based landslide inventories was compared (i) with the other six, (ii) with a detailed field survey‐based inventory, and (iii) with a comparable study in which topographic data were extracted from a topographical map. The combination of the percentage of field landslides indicated by an expert and the percentage of positional discrepancies (expressed in terms of positional mismatch) were used to evaluate the quality of the LIDAR‐based inventory maps. High‐quality LIDAR‐derived landslide inventory maps contain more than 70 per cent of the landslides mapped during the field survey, and have positional discrepancies smaller than 70 per cent when compared with the field survey‐based inventory map. Four experts and the combination map of all experts satisfied these criteria. Together the seven experts indicated all landslides mapped in the field. Importantly, LIDAR enabled the experts to find ten new landslides and to correct the boundaries of eleven (of the 77) landslides mapped during the field survey. Hence, this study showed that large‐scale LIDAR‐derived maps analysed by experienced geomorphologists can significantly improve field survey‐based inventories of landslides with a subdued morphology in hilly regions. Copyright © 2006 John Wiley & Sons, Ltd.     

Parameterizing a physically based shallow landslide model in a data poor region

Shallow landslides and consequent debris flows are an increasing concern in the Western Ghats of Kerala, India. Their increased frequency has been associated with deforestation and unfavourable land‐use practices in cultivated areas. In order to evaluate the influence of vegetation on shallow slope failures a physically based, dynamic and distributed hydrological model (STARWARS) coupled with a probabilistic slope stability model (PROBSTAB) was applied to the upper Tikovil River basin (55·6 km²). It was tuned with the limited evidence of groundwater conditions during the monsoon season of 2005 and validated against observed landslide activity in the hydrological year 2001–2002. Given the data poor conditions in the region some modifications to the original model were in order, including the estimation of parameters on the basis of generalized information from secondary sources, pedo‐transfer functions, empirical equations and satellite remote sensing data. Despite the poor input, the model captured the general temporal and spatial pattern of instability in the area. Sensitivity analysis proved root cohesion, soil depth and angle of internal friction as the most dominant parameters influencing slope stability. The results indicate the importance of root cohesion in maintaining stability and the critical role of the management of rubber plantations in this. Interception and evapotranspiration showed little influence on the development of failure conditions. The study also highlights the importance of high resolution digital terrain models for the accurate mechanistic prediction of shallow landslide initiation. Copyright © 2009 John Wiley & Sons, Ltd.     

Post-deforestation soil loss from steepland hillslopes in Taranaki,New Zealand

Soil erosion on steepland hillslopes in Taranaki, New Zealand, where landsliding is the dominant erosion form, was investigated by comparing mean regolith depths between first-order basins that have had their forest cover removed for different periods of time. Regolith depth and slope angle data were collected along 19 profile lines and 30 profile lines from steepland basins that had been deforested for 10 and 85 years, respectively. These profile lines were subdivided into a total of 236 profile segments of relatively linear slope angle and uniform regolith depth, that averaged 17·5 m in length. The depth of pre-existing regolith on post-deforestation landslide sites is estimated from a regression of regolith depth on slope angle for undisturbed (non-landslide) profile segments. Regolith depletion on landslide sites is in turn estimated by subtracting the depth of regolith on landslide sites from the estimate of pre-existing regolith depth. Regolith depletion by post-deforestation landslides, averaged over the entire length of profile lines, gives an estimate of average surface lowering. For the area deforested for 85 years, average surface lowering by post-deforestation landslides is 0·15 ± 0·04 m, and is the same as the difference in mean depth of 0·15 ± 0·11 m between this area and the area deforested for 10 years. Erosion of regolith from hillslopes by processes other than landsliding appears to be minimal. The 0·15 m average surface lowering represents a regolith depletion rate of 1·8 ± 0±5 mm yr^?1. For hillslopes steeper than 28°, where all post-deforestation landslides occur, average surface lowering is 0·20 ± 0·05 m, and the regolith depletion rate is 2±4 · 0±6 mm yr^?1. Average surface lowering is greatest at 0·23 ± 0·07 m on hillslopes steeper than 32° where most post-deforestation landslides occur. Here, the regolith depletion rate is 2·7 ± 0·8 mm yr^?1. A large-magnitude, low-frequency storm in March 1990, produced an average surface lowering of 0·041 m. There were proportionately more landslides in the area deforested for 10 years, illustrating the importance of previous erosion history of hillslopes on the spatial distribution of landslides. There were also comparatively few landslides on steeper hillslopes because previous lower magnitude storms had already removed much of the deeper regolith.     

Quantifying rainfall controls on catchment‐scale landslide erosion in Taiwan

Landslide erosion is a dominant hillslope process and the main source of stream sediment in tropical, tectonically active mountain belts. In this study, we quantified landslide erosion triggered by 24 rainfall events from 2001 to 2009 in three mountainous watersheds in Taiwan and investigated relationships between landslide erosion and rainfall variables. The results show positive power‐law relations between landslide erosion and rainfall intensity and cumulative rainfall, with scaling exponents ranging from 2·94 to 5·03. Additionally, landslide erosion caused by Typhoon Morakot is of comparable magnitude to landslide erosion caused by the Chi‐Chi Earthquake (M_W = 7·6) or 22–24 years of basin‐averaged erosion. Comparison of the three watersheds indicates that deeper landslides that mobilize soil and bedrock are triggered by long‐duration rainfall, whereas shallow landslides are triggered by short‐duration rainfall. These results suggest that rainfall intensity and watershed characteristics are important controls on rainfall‐triggered landslide erosion and that severe typhoons, like high‐magnitude earthquakes, can generate high rates of landslide erosion in Taiwan. Copyright © 2012 John Wiley & Sons, Ltd.     

How hydrological factors initiate instability in a model sandy slope

Knowledge of the mechanisms of rain‐induced shallow landslides can improve the prediction of their occurrence and mitigate subsequent sediment disasters. Here, we examine an artificial slope's subsurface hydrology and propose a new slope stability analysis that includes seepage force and the down‐slope transfer of excess shear forces. We measured pore water pressure and volumetric water content immediately prior to a shallow landslide on an artificial sandy slope of 32°: The direction of the subsurface flow shifted from downward to parallel to the slope in the deepest part of the landslide mass, and this shift coincided with the start of soil displacement. A slope stability analysis that was restricted to individual segments of the landslide mass could not explain the initiation of the landslide; however, inclusion of the transfer of excess shear forces from up‐slope to down‐slope segments improved drastically the predictability. The improved stability analysis revealed that an unstable zone expanded down‐slope with an increase in soil water content, showing that the down‐slope soil initially supported the unstable up‐slope soil; destabilization of this down‐slope soil was the eventual trigger of total slope collapse. Initially, the effect of apparent soil cohesion was the most important factor promoting slope stability, but seepage force became the most important factor promoting slope instability closer to the landslide occurrence. These findings indicate that seepage forces, controlled by changes in direction and magnitude of saturated and unsaturated subsurface flows, may be the main cause of shallow landslides in sandy slopes. Copyright © 2013 John Wiley & Sons, Ltd.     

Large scale landslides triggered by Quaternary tectonics in the Acambay graben,Mexico

Detailed analysis was conducted on large‐scale gravitational‐tectonic deformations and landslides in the Acambay graben, an intra‐arc basin in the trans‐Mexican volcanic belt (TMVB). Field mapping and remote sensing revealed the slope instability of the northern graben boundary induced by the Acambay‐Tixmadejé fault. Two major landslides of 0·1 km³ and 0·05 km³ in volume were identified and their characteristics were analyzed according to the role of tectonics, mechanism of slope failure, and possible triggering factors. Quaternary faulting played a major role in increasing the local relief, and the activity of the Acambay‐Tixmadejé fault represents the main geomorphic factor conditioning the gravitational movements. Moreover, displacements along this fault generated sliding surfaces and reduced the strength of the rock mass. The two landslides are classified as large‐scale rotational slides involving volcanic rocks of late Miocene‐Pleistocene age. Since the Acambay graben is a seismogenic area with a known maximum horizontal ground acceleration of 0·5 g, a strong earthquake could be ascribed as the possible triggering mechanism of the landslides. Our work represents the first analysis of large gravitational slope movements in tectonically active regions in Mexico, a process that can be common in the intra‐arc basins of the TMVB, where active tectonic, seismicity, weak altered volcanic rocks, and heavy rains affect the slope stability. Copyright © 2010 John Wiley & Sons, Ltd.     

Spatial Distribution of Seismic Landslides in the Areas of 1927 Gulang M8.0 Earthquake

The 1927 Gulang M8.0 earthquake has triggered a huge number of landslides, resulting in massive loss of people''s life and property. However, integrated investigations and results regarding the landslides triggered by this earthquake are rare; such situation hinders the deep understanding of these landslides such as scale, extent, and distribution. With the support of Google Earth software, this study intends to finish the seismic landslides interpretation work in the areas of Gulang earthquake (VIII-XI degree) using the artificial visual interpretation method, and further analyze the spatial distribution and impact factors of these landslides. The results show that the earthquake has triggered at least 936 landslides in the VIII-XI degree zone, with a total landslide area of 58.6 km². The dense area of seismic landslides is located in the middle and southern parts of the X intensity circle. Statistical analysis shows that seismic landslides is mainly controlled by factors such as elevation, slope gradient, slope direction, strata, seismic intensity, faults and rivers. The elevation of 2 000-2 800 m is the high-incidence interval of the landslide. The landslide density is larger with a higher slope gradient. East and west directions are the dominant sliding directions. The areas with Cretaceous and Quaternary strata are the main areas of the Gulang seismic landslides. The X intensity zone triggered the most landslides. In addition, landslides often occur in regions near rivers and faults. This paper provides a scientific reference for exploring the development regularities of landslides triggered by the 1927 Gulang earthquake and effectively mitigating the landslide disasters of the earthquake.     

Characteristics of landslides in forests and grasslands triggered by the 2016 Kumamoto earthquake

We examined the characteristics of landslides triggered by the 2016 Kumamoto earthquake (Mw = 7.0: focal depth=10.0 km) in forests and grasslands within two affected watersheds (Tokosegawa: 6.9 km² and Nigorigawa: 6.1 km²) in southwestern Japan. We identified 190 landslides using aerial photographs and analyzed their sizes by geographic information system (GIS). Field investigations were conducted to obtain landslide depth, volume and residual sediment for 38 selected landslides (21 in forests and 17 in grasslands). The minimum area of detected landslides in grasslands (400 m²) was smaller than in forests (1000 m²), probably because of reduced detectability of landslides under tree cover. The ratio of total area occupied by landslides for a given range of slope gradient in the watersheds increased from 3.2% on gentle grassland slopes (10–15°) to 15.5% on steep (>45°) slopes, whereas the maximum landslide-area ratio in forest sites (7.4%) occurred on relatively gentle slopes (25–30°). Estimated landslide volume ranged from 27 to 9622 m³, based on mean depth of each landslide measured around individual landslide scars. Moreover, the volumetric ratio of landslide deposit volume to total landslide volume exceeded 100% for 48% of the landslides within forests and 35% of the landslides within grasslands. Our findings show that land cover had extensive and recognizable effects on the characteristics of landslides and resulting in-channel sediment accumulations. Resetting sediment dynamics after earthquakes associated with different land cover distributions needs to be considered within watersheds. © 2019 John Wiley & Sons, Ltd.     

Assessment of shallow landslide susceptibility by means of multivariate statistical techniques

Several multivariate statistical analyses have been performed to identify the most influential geological and geomorphological parameters on shallow landsliding and to quantify their relative contribution. A data set was first prepared including more than 30 attributes of 230 failed and unfailed slopes. The performance of principal component analysis, t‐test and one‐way test, allowed a preliminary selection of the most significant variables, which were used as input variables for the discriminant analysis. The function obtained has classified successfully 88·5 per cent of the overall slope population and 95·6 per cent of the failed slopes. Slope gradient, watershed area and land‐use appeared as the most powerful discriminant factors. A landslide susceptibility map, based on the scores of the discriminant function, has been prepared for Ensija range in the Eastern Pyrenees. An index of relative landslide density shows that the results of the map are consistent. Copyright © 2001 John Wiley & Sons, Ltd.     

A spatiotemporal probabilistic modelling of storm‐induced shallow landsliding using aerial photographs and logistic regression

Landslides constitute one of the major natural hazards that could cause significant losses of life and property. Mapping or delineating areas prone to landsliding is therefore essential for land‐use activities and management decision making in hilly or mountainous regions. A landslide hazard map can be constructed by a qualitative combination of maps of site conditions, including geology, topography and geomorphology, by statistical methods through correlating landslide occurrence with geologic and geomorphic factors, or by using safety factors from stability analysis. A landslide hazard map should provide information on both the spatial and temporal probabilities of landsliding in a certain area. However, most previous studies have focused on susceptibility mapping, rather than on hazard mapping in a spatiotemporal context. This study aims at developing a predictive model, based on both quasi‐static and dynamic variables, to determine the probability of landsliding in terms of space and time. The study area selected is about 13 km² in North Lantau, Hong Kong. The source areas of the landslides caused by the rainstorms of 18 July 1992 and 4–5 November 1993 were interpreted from multi‐temporal aerial photographs. Landslide data, lithology, digital elevation model data, land cover, and rainfall data were digitized into a geographic information system database. A logistic regression model was developed using lithology, slope gradient, slope aspect, elevation, slope shape, land cover, and rolling 24 h rainfall as independent variables, since the dependent variable could be expressed in a dichotomous way. This model achieved an overall accuracy of 87·2%, with 89·5% of landslide grid cells correctly classified and found to be performing satisfactorily. The model was then applied to rainfalls of a variety of periods of return, to predict the probability of landsliding on natural slopes in space and time. It is observed that the modelling techniques described here are useful for predicting the spatiotemporal probability of landsliding and can be used by land‐use planners to develop effective management strategies. Copyright © 2003 John Wiley & Sons, Ltd.     

Geomorphic imprint of landslides on alpine river systems,southwest New Zealand

An inventory of 846 mass movements, mainly landslides, in two alpine regions of southwest New Zealand was created to explore the geomorphic impacts of slope‐failure processes on river channels and valley floors. In total, 213 (i.e. 27 per cent) of the slope failures descended to valley floors, affecting the geomorphology of trunk channels (catchment area A_C > 10 km²) and valley floors in recurring patterns. A nominal classification system is introduced for characterizing (a) the physical contact nature between landslides and river channels, and (b) the resulting geomorphic consequences for drainage. Although landslide area A is useful for estimating the length of channel directly impacted by debris, it does not necessarily predict the direction of fluvial response or type of impact. Dominant persistent geomorphic imprints of bedrock landslides include channel occlusions and landslide dams in South Westland and Fiordland, respectively. Differences in size distribution and geomorphic effects on river systems between the two study regions are attributed to bedrock geology, tectonics and sediment flux. Although South Westland rivers are more frequently affected by landslides, disrupting long‐term effects such as blockage are more persistent in Fiordland. Copyright © 2005 John Wiley & Sons, Ltd.