Refine
Has Fulltext
- no (58) (remove)
Year of publication
Document Type
- Article (54)
- Other (2)
- Monograph/Edited Volume (1)
- Preprint (1)
Is part of the Bibliography
- yes (58)
Keywords
- landslide (5)
- Erosion (4)
- Himalayas (4)
- Earthquake (3)
- Himalaya (3)
- Landslide (3)
- hydropower (3)
- landslides (3)
- Alaska (2)
- Bayesian logistic regression (2)
Institute
- Institut für Geowissenschaften (58) (remove)
Deforestation is a prominent anthropogenic cause of erosive overland flow and slope instability, boosting rates of soil erosion and concomitant sediment flux. Conventional methods of gauging or estimating post-logging sediment flux often focus on annual timescales but overlook potentially important process response on shorter intervals immediately following timber harvest. We resolve such dynamics with non-parametric quantile regression forests (QRF) based on high-frequency (3 min) discharge measurements and sediment concentration data sampled every 30-60 min in similar-sized (similar to 0.1 km(2)) forested Chilean catchments that were logged during either the rainy or the dry season. The method of QRF builds on the random forest algorithm, and combines quantile regression with repeated random sub-sampling of both cases and predictors. The algorithm belongs to the family of decision-tree classifiers, which allow quantifying relevant predictors in high-dimensional parameter space. We find that, where no logging occurred, similar to 80% of the total sediment load was transported during extremely variable runoff events during only 5% of the monitoring period. In particular, dry-season logging dampened the relative role of these rare, extreme sediment-transport events by increasing load efficiency during more efficient moderate events. We show that QRFs outperform traditional sediment rating curves (SRCs) in terms of accurately simulating short-term dynamics of sediment flux, and conclude that QRF may reliably support forest management recommendations by providing robust simulations of post-logging response of water and sediment fluxes at high temporal resolution.
Lacustrine sediments have been widely used to investigate past climatic and environmental changes on millennial to seasonal time scales. Sedimentary archives of lakes in mountainous regions may also record non-climatic events such as earthquakes. We argue herein that a set of 64 annual laminae couplets reconciles a stratigraphically inconsistent accelerator mass spectrometry (AMS) C-14 chronology in a similar to 4-m-long sediment core from Lake Mengda, in the north-eastern Tibetan Plateau. The laminations suggest the lake was formed by a large landslide, triggered by the 1927 Gulang earthquake (M = 8.0). The lake sediment sequence can be separated into three units based on lithologic, sedimentary, and isotopic characteristics. Starting from the bottom of the sequence, these are: (1) unweathered, coarse, sandy valley-floor deposits or landslide debris that pre-date the lake, (2) landslide-induced, fine-grained soil or reworked landslide debris with a high organic content, and (3) lacustrine sediments with low organic content and laminations. These annual laminations provide a high-resolution record of anthropogenic and environmental changes during the twentieth century, recording enhanced sediment input associated with two phases of construction activities. The high mean sedimentation rates of up to 4.8 mm year(-1) underscore the potential for reconstructing such distinct sediment pulses in remote, forested, and seemingly undisturbed mountain catchments.
The Norwegian traffic network is impacted by about 2000 landslides, avalanches, and debris flows each year that incur high economic losses. Despite the urgent need to mitigate future losses, efforts to locate potential debris flow source areas have been rare at the regional scale. We tackle this research gap by exploring a minimal set of possible topographic predictors of debris flow initiation that we input to a Weights-of-Evidence (WofE) model for mapping the regional susceptibility to debris flows in western Norway. We use an inventory of 429 debris flows that were recorded between 1979 and 2008, and use the terrain variables of slope, total curvature, and contributing area (flow accumulation) to compute the posterior probabilities of local debris flow occurrence. The novelty of our approach is that we quantify the uncertainties in the WofE approach arising from different predictor classification schemes and data input, while estimating model accuracy and predictive performance from independent test data. Our results show that a percentile-based classification scheme excels over a manual classification of the predictor variables because differing abundances in manually defined bins reduce the reliability of the conditional independence tests, a key, and often neglected, prerequisite for the WofE method. The conditional dependence between total curvature and flow accumulation precludes their joint use in the model. Slope gradient has the highest true positive rate (88%), although the fraction of area classified as susceptible is very large (37%). The predictive performance, i.e. the reduction of false positives, is improved when combined with either total curvature or flow accumulation. Bootstrapping shows that the combination of slope and flow accumulation provides more reliable predictions than the combination of slope and total curvature, and helps refining the use of slope-area plots for identifying morphometric fingerprints of debris flow source areas, an approach used outside the field of landslide susceptibility assessments.
The growing body of research on large-scale mass wasting events so far has only scarcely investigated the sedimentology of chaotic deposits from non-volcanic terrestrial landslides such that any overarching and systematic terminological framework remains elusive. Yet recent work has emphasized the need for better understanding the internal structure and composition of rockslide deposits as a means to characterise the mechanics during the final stages of runout and emplacement. We offer a comprehensive overview on the occurrence of rock fragmentation and frictional melt both at different geographic locations, and different sections within large (>10(6) m(3)) rockslide masses. We argue that exposures of pervasively fragmented and interlocked jigsaw-cracked rock masses; basal melange containing rip-up clasts and phantom blocks; micro-breccia; and thin bands of basal frictionite are indispensable clues for identifying deposits from giant rockslides that may remain morphologically inconspicuous otherwise. These sedimentary assemblages are diagnostic tools for distinguishing large rockslide debris from macro and microscopically similar glacial deposits, tectonic fault-zone breccias, and impact breccias, and thus help avoid palaeoclimatic and tectonic misinterpretations, let alone misestimates of the hazard from giant rockslides. Moreover, experimental results from Mossbauer spectroscopy of frictionite samples support visual interpretations of thin sections, and demonstrate that short-lived (<10 s) friction-induced partial melting at temperatures >1500 degrees C in the absence of water occurred at the base of several giant moving rockslides. This finding supports previous theories of dry excess runout accompanied by comminution of rock masses down to gm-scale, and indicates that catastrophic motion of large fragmenting rock masses does not require water as a potential lubricant.
Late Quaternary valley infill and dissection in the Indus River, western Tibetan Plateau margin
(2014)
The Indus, one of Earth's major rivers, drains large parts of the NW Himalaya and the Transhimalayan ranges that form part of the western Tibetan Plateau margin. In the western Himalayan syntaxis, where local topographic relief exceeds 7 km, the Indus has incised a steep bedrock gorge at rates of several mm yr(-1). Upstream, however, the upper Indus and its tributaries alternate between bedrock gorges and broad alluvial flats flanked by the Ladakh and Zanskar ranges. We review the late Quaternary valley history in this region with a focus on the confluence of the Indus and Zanskar Rivers, where vast alluvial terrace staircases and lake sediments record major episodes of aggradation and incision. New absolute dating of high-level fluvial terrace remnants using cosmogenic Be-10, optically and infrared stimulated luminescence (OSL, IRSL) indicates at least two phases of late Quaternary valley infilling. These phases commenced before similar to 200 ka and similar to 50-20 ka, judging from terrace treads stranded >150 m and similar to 30-40 m above modern river levels, respectively. Numerous stacks of lacustrine sediments that straddle the Indus River >200 km between the city of Leh and the confluence with the Shyok River share a distinct horizontal alignment. Constraints from IRSL samples of lacustrine sequences from the Leh-Spituk area reveal a protracted lake phase from >177 ka to 72 ka, locally accumulating >50-m thick deposits. In the absence of tectonic faulting, major lithological differences, and stream capture, we attribute the formation of this and other large lakes in the region to natural damming by large landslides, glaciers, and alluvial fans. The overall patchy landform age constraints from earlier studies can be reconciled by postulating a major deglacial control on sediment flux, valley infilling, and subsequent incision that has been modulated locally by backwater effects of natural damming. While comparison with Pleistocene monsoon proxies reveals no obvious correlation, a lateor post-glacial sediment pulse seems a more likely source of this widespread sedimentation that has partly buried the dissected bedrock topography. Overall, the long residence times of fluvial, alluvial and lacustrine deposits in the region (>500 ka) support previous studies, but remain striking given the dominantly steep slopes and deeply carved valleys that characterise this high-altitude mountain desert. Recalculated late Quaternary rates of fluvial bedrock incision in the Indus and Zanskar of 1.5 +/- 0.2 mm yr(-1) are at odds with the longevity of juxtaposed valley-fill deposits, unless a lack of decisive lateral fluvial erosion helps to preserve these late Pleistocene sedimentary archives. We conclude that alternating, similar to 10(4)-yr long, phases of massive infilling and incision have dominated the late Quaternary history of the Indus valley below the western Tibetan Plateau margin. (C) 2014 Elsevier Ltd. All rights reserved.
The 2002 M-w 7.9 Denali Fault earthquake, Alaska, provides an unparalleled opportunity to investigate in quantitative detail the regional hillslope mass-wasting response to strong seismic shaking in glacierized terrain. We present the first detailed inventory of similar to 1580 coseismic slope failures, out of which some 20% occurred above large valley glaciers, based on mapping from multi-temporal remote sensing data. We find that the Denali earthquake produced at least one order of magnitude fewer landslides in a much narrower corridor along the fault ruptures than empirical predictions for an M 8 earthquake would suggest, despite the availability of sufficiently steep and dissected mountainous topography prone to frequent slope failure. In order to explore potential controls on the reduced extent of regional coseismic landsliding we compare our data with inventories that we compiled for two recent earthquakes in periglacial and formerly glaciated terrain, i.e. at Yushu, Tibet (M-w 6.9, 2010), and Aysen Fjord, Chile (2007 M-w 6.2). Fault movement during these events was, similarly to that of the Denali earthquake, dominated by strike-slip offsets along near-vertical faults. Our comparison returns very similar coseismic landslide patterns that are consistent with the idea that fault type, geometry, and dynamic rupture process rather than widespread glacier cover were among the first-order controls on regional hillslope erosional response in these earthquakes. We conclude that estimating the amount of coseismic hillslope sediment input to the sediment cascade from earthquake magnitude alone remains highly problematic, particularly if glacierized terrain is involved. (C) 2014 Elsevier Ltd. All rights reserved.
Quantifying volumes and rates of delivery of terrestrial sediment from island arcs to subduction zones is indispensable for refining estimates of the thickness of trench fills that may eventually control the location and timing of submarine landslides and tsunami-generating mega-earthquakes. Despite these motivating insights, knowledge about the rates of erosion and sediment export from the Japanese islands to their Pacific subduction zones remains patchy regardless of the increasing availability of highly resolved data on surface deformation, climate, geology, and topography. Traditionally, natural erosion rates across the island arc have been estimated from regression of topographic catchment metrics and reservoir sedimentation rates that were recorded over several years to decades. We review current research in this context, correct for a systematic bias in one of the most widely used predictions, and present new estimates of decadal to millennial-scale erosion rates of Japan's terrestrial inner forearc. We draw on several independent and unprecedented inventories of mass wasting, reservoir sedimentation, and concentrations of cosmogenic Be-10 in river sands. We find that natural Be-10-derived denudation rates of several mm yr(-1) in the Japanese Alps have been sustained over several centuries to millennia, and are, within error, roughly consistent with sediment yields inferred from artificial reservoir sedimentation. Local exceptions may likely result from release of sediment storage or regional landsliding episodes that trigger transient sediment pulses. Our synopsis further reveals that catchments draining Japan's eastern seaboard differ distinctly in their tectonic, lithological, topographic, and climatic characteristics between the Tohoku, Japanese Alps, and Nankai inner forearc segments, which is underscored by a marked asymmetric pattern of erosion rates along the island arc. Erosion rates are highest (up to at least 3 mm yr(-1)) in the Japanese Alps that mark the collision of two subduction zones, where high topographic relief, hillslope and bedrock-channel steepness foster rapid denudation by mass wasting. Comparable, if slightly lower, erosion rates characterise the Nankai inner forearc in southwest Japan, most likely due to higher typhoon-driven rainfall totals and variability rather than its high topographic relief. In contrast, our estimated erosion and flux rates are lowest in the Tohoku inner forearc catchments that feed sediment into the Japan Trench. We conclude that collisional mountain building of the Japanese Alps drives some of the highest erosion rates in the island arc despite similar uplift and precipitation controls in southwest Japan. We infer that, prior to extensive river damming, reservoir construction, and coastal works, the gross of Japan's total sediment export to the Pacific Ocean entered the accretionary margin of the Nankai Trough as opposed to the comparatively sediment-starved Japan Trench. Compared to documented contemporary rates of sediment flux from mountainous catchments elsewhere in the Pacific, the rivers draining Japan's inner forearc take an intermediate position despite high relief, steep slopes, very high seismicity, and frequent rainstorms. However, the average rates of millennial-scale denudation in the Japanese Alps particularly are amongst the highest reported worldwide.
Local mismatches between these late Holocene and modern rates emphasise the anthropogenic fingerprint on sediment retention that may have significantly reduced the island arc's mass flux to its subduction zones, as is the case elsewhere in east and southeast Asia. (C) 2014 Elsevier B.V. All rights reserved.
The Indus River, one of Asia's premier rivers, drains the western Tibetan Plateau and the Nanga Parbat syntaxis. These two areas juxtapose some of the lowest and highest topographic relief and commensurate denudation rates in the Himalaya-Tibet orogen, respectively, yet the spatial pattern of denudation rates upstream of the syntaxis remains largely unclear, as does the way in which major rivers drive headward incision into the Tibetan Plateau. We report a new inventory of Be-10-based basinwide denudation rates from 33 tributaries flanking the Indus River along a 320 km reach across the western Tibetan Plateau margin. We find that denudation rates of up to 110 mm k.y.(-1) in the Ladakh and Zanskar Ranges systematically decrease eastward to 10 mm k.y.(-1) toward the Tibetan Plateau. Independent results from bulk petrographic and heavy mineral analyses support this denudation gradient. Assuming that incision along the Indus exerts the base-level control on tributary denudation rates, our data show a systematic eastward decrease of landscape downwearing, reaching its minimum on the Tibetan Plateau. In contrast, denudation rates increase rapidly 150-200 km downstream of a distinct knick-point that marks the Tibetan Plateau margin in the Indus River longitudinal profile. We infer that any vigorous headward incision and any accompanying erosional waves into the interior of the plateau mostly concerned reaches well below this plateau margin. Moreover, reported long-term (>10(6) yr) exhumation rates from low-temperature chronometry of 0.1-0.75 mm yr(-1) consistently exceed our Be-10-derived denudation rates. With averaging time scales of 10(3)-10(4) yr for our denudation data, we report postglacial rates of downwearing in a tectonically idle landscape. To counterbalance this apparent mismatch, denudation rates must have been higher in the Quaternary during glacial-interglacial intervals.
Dealing with predicted increases in extreme weather conditions due to climate change requires robust knowledge about controls on rainfall-triggered landslides. We explore relationships between rainfall and landslide size throughout the Japanese archipelago. We test whether the total volume of landslides can be predicted directly from rainfall totals, intensity, and duration using a nationwide inventory of 4744 rainfall-triggered landslides recorded from A.D. 2001 to 2011. We find that larger landslides were more abundant at the expense of smaller ones when total, maximum, and mean rainfall intensity exceeded similar to 250 mm, similar to 35 mm/h, and similar to 4 mm/h, respectively. Frequency distributions of these rainfall parameters are peaked and heavily skewed. Yet neither the most frequent nor the most extreme values of these rainfall metrics coincide consistently with the maximum landslide volumes. A striking decrease of landslide volumes at both mean and maximum rainfall intensity, as well as duration, points to an exhaustion in hillslope geomorphic response regardless of sample size, landslide type, mobilized volume, dominant lithology, or reporting bias. Our results underscore substantial offsets between the peaks of rainfall metrics and maximum associated landslide volumes, thus complicating straightforward estimates of geomorphic work from metrics of rainstorm magnitude or frequency. Only the rainfall total appears to be a suitable monotonic predictor of landslide volumes mobilized during typhoons and frontal storms.
Projected scenarios of climate change involve general predictions about the likely changes to the magnitude and frequency of landslides, particularly as a consequence of altered precipitation and temperature regimes. Whether such landslide response to contemporary or past climate change may be captured in differing scaling statistics of landslide size distributions and the erosion rates derived thereof remains debated. We test this notion with simple Monte Carlo and bootstrap simulations of statistical models commonly used to characterize empirical landslide size distributions. Our results show that significant changes to total volumes contained in such inventories may be masked by statistically indistinguishable scaling parameters, critically depending on, among others, the size of the largest of landslides recorded. Conversely, comparable model parameter values may obscure significant, i.e. more than twofold, changes to landslide occurrence, and thus inferred rates of hillslope denudation and sediment delivery to drainage networks. A time series of some of Earth's largest mass movements reveals clustering near and partly before the last glacial-interglacial transition and a distinct step-over from white noise to temporal clustering around this period. However, elucidating whether this is a distinct signal of first-order climate-change impact on slope stability or simply coincides with a transition from short-term statistical noise to long-term steady-state conditions remains an important research challenge.