Search

Satellite-based maps of glacial landforms reveal that the mountain landscapes of northeast Eurasia contain over one million km2 of glaciated terrain. Previous work has speculated on the existence of large ice masses during the Last Glacial Maximum (LGM) and the preceding cold phases, but the lack of age constraints means that little is known about the timing of past glaciations across this vast region.With an aim to gain a better understanding of the glacial history of this region, we collected samples for cosmogenic 10Be exposure dating of boulder erratics and moraines in the mountains of eastern Siberia. Here, we present the first results from two sites, both within the Chersky Range: (1) Malyk Sen, which contains a succession of three end moraines in a foreland setting; and (2) Ust-Nera, which features boulder erratics and glacial bedrock pavement exposed in a previously glaciated valley. At Malyk Sen, the relative positions and corresponding ages of the three moraines indicate progressive contraction of maximum glacier extent since termination of the Marine Isotope Stage (MIS) 6, with the innermost moraine dated to the LGM. Our preliminary results from Ust-Nera suggest exposure ages from glacially-transported boulders and bedrock pavement that are significantly older than the LGM. Both sites indicate limited extents of mountain glaciation during the LGM in eastern Siberia. And while the glacial chronology of our study does not extend beyond MIS 6, mapping of the surrounding areas indicates that even more expansive glaciers existed further back in time.Our findings confirm the trend of successively smaller glacial extent maxima’s in continental Eurasia towards the LGM, with at least one ice advance during MIS 5-3 larger than the LGM advance. This trend could to be linked to extreme continental settings such as in Eurasia and westernmost America, as it contrasts with larger parts of the Northern Hemisphere glaciations where Late Pleistocene maxima were reached during LGM.

By simulating erosion and deposition, landscape evolution models offer powerful insights to Earth surface processes and dynamics. These models are typically constructed from parameters describing drainage area (m), slope (n), substrate erodibility (K), hillslope diffusion (D), and a critical drainage area (Ac) that signifies the downslope transition from hillslope diffusion to advective fluvial processes. In spite of the widespread success of such models, the parameter values have high degrees of uncertainty mainly because the advection and diffusion equations amalgamate physical processes and material properties that span widely differing spatial and temporal scales. Here, we use a global catalogue of catchment-averaged cosmogenic 10Be-derived erosion (denudation) rates with the aim to optimise a set of landscape evolution models via a Monte Carlo based parameter search. We consider three model scenarios: advection-only, diffusion-only, and an advection-diffusion hybrid. In each case, we search for a parameter set that best approximates erosion rates at the global scale, and we directly compare erosion rates from the modelled scenarios with those derived from 10Be data. Optimised ranges can be defined for many LEM parameters at the global scale. In the absence of diffusion, n ~ 1.3, and with increasing diffusivity the optimal n increases linearly to a global maximum of n ~ 2. Meanwhile we find that the diffusion-only model somewhat outperforms the advection-only model and is optimised when concavity is raised to a power of 2. With these examples, we suggest that our approach provides baseline parameter estimates for large-scale studies spanning long timescales and diverse landscape properties. Moreover, our direct comparison of model-predicted versus observed erosion rates is preferable to methods that rely upon catchment-scale averaging or amalgamation of topographic metrics. We also seek to optimise K and D parameters in landscape evolution models with respect to precipitation and substrate lithology. These optimised models allow us to effectively control for topography and target specifically the relationship between erosion rate and precipitation. All models suggest a positive correlation between K or D and precipitation > 1500 mm yr–1, plus a local maximum at ~ 300 mm yr–1, which is compatible with the long-standing hypothesis that semi-arid environments are among the most erodible.

On 25 August 1933, a 7.5-magnitude earthquake struck the eastern margin of the Tibetan Plateau in Sichuan, China. The Diexi earthquake is among the largest known geohazard events worldwide and is frequently cited by those studying the effects of large earthquakes. And yet, the attention focused on this event has failed to deliver a clear picture of landslides and their geomorphic impacts—key attributes of the Diexi earthquake and its aftermath remain obscure and debated. By integrating present-day LiDAR topographic data with existing records (including studies published in Chinese), and a series of unique archival photographs (from 1910, 1920, and 1934), we present the first inventory of coseismic landslides from the epicentral region of this catastrophic 1933 event. We find that the earthquake-triggered landslides were mainly of shallow to deep rock/debris fall/avalanche type, containing mass detached from steep slopes at the top of the mountain ridge source. We reinterpreted three major landslide dams on the Minjiang River based on their geomorphology and sedimentology, and we reconstructed the processes of impoundment and the maximum area of the dammed lake that breached 45 days after the earthquake. Since 1933, we estimate that 43.15–47.68 million m3 of post-earthquake sediment has been delivered to valley floors from the erosion of the three landslide dams, with a sediment yield of up to 131.71 t km−2 year−1. Retrospective studies of the effects of historical earthquakes are challenging, but long-term observations like these are the cornerstone of emerging knowledge of earthquake-induced landsliding and related landscape response.

Former lakes and wetlands can provide valuable insights to the late Pleistocene environments encountered by the first humans to enter the Levant from Africa. Fluvial incision along Wadi Gharandal in hyperarid southern Jordan has exposed remnants of a small riverine wetland that accumulated as a sedimentary sequence up to ~20 m thick. We conducted a chronometric and sedimentological study of this wetland, including 10 optically stimulated luminescence dates. The wetland sequence accumulated during the period ~125 to 70 ka in response to a positive water balance coupled with a (possibly coseismic) landslide that dammed the outlet. The valley fill was dissected when the dam was incised shortly after ~36 ± 3 ka. Comparison of our ages with regional palaeoclimate indicates that the Gharandal oasis developed during the relatively humid Marine Isotope Stage 5. A minimum age of 74 ± 7 ka for two Levallois flakes collected from stratified sediments suggests that the oasis was visited by humans during the critical 130–90 ka time window of human migration out of Africa. Gharandal joins a growing network of freshwater sites that enabled humans to cross areas of the Levant and Arabia along corridors of human dispersal.

In this study, we apply six machine-learning algorithms (XGBoost, Cubist, BMARS, ANFIS, Cforest and Elasticnet) to investigate the susceptibility of the Jazmurian Basin in southeastern Iran to dust emissions. This research is the first attempt to apply several machine-learning techniques (e.g., BMARS, ANFIS, Cforest and Elasticnet) to mapping of dust emissions from land surfaces. Fourteen parameters associated with meteorology, lithology, soil, and human activity were considered as potentially effective dust emission factors implemented in our modelling. Collinearity among the parameters and their weighted importance were examined statistically. To evaluate the accuracy of our predictive models and their performance, we applied the Taylor diagram (involving RMSE and correlation coefficient), the Nash Sutcliffe coefficient (NSC), and mean absolute error (MAE). The prediction accuracy of the six algorithms for identifying susceptibility to dust emissions, as assessed by the Taylor diagram, was as follows: Cforest (NSC = 98% and MAE = 3.2%) > Cubist (NSC = 90% and MAE = 10.6%) > Elasticnet (NSC = 90% and MAE = 10.7) > ANFIS (NSC > 90% and MAE = 11%) > BMARS (NSC = 89% and MAE = 11.2%) > XGBoost (NSC = 89% and 11.3%). Based on the map produced by Cforest (i.e., the best-performing algorithm in our assessment), we identify four dust susceptibility classes, and their respective total areas ranging from low (32%), moderate (8.2%), high (10%), to very high (50%). We identify the dry lakebed of Hamun-e-Jaz Murian as the most productive area for dust emissions.

The intricate interplay between subglacial topography and ice-sheet dynamics is key to the evolution of large ice sheets, but in Greenland as elsewhere the effects of long-term glacial history on landscape evolution remain poorly constrained. Here we measure abundances of cosmogenic 10Be and 26Al in bedrock and transported boulders to unveil the glaciation and erosion history of Dove Bugt, northeast Greenland. In agreement with studies of west Greenland, we find that apparent exposure ages increase with elevation from 9 ka to 13 ka in low-lying valleys to 21 ka to 204 ka on high-elevation, blockfield-covered plateaus. We employ a Markov chain Monte Carlo inversion framework to constrain the probability of various erosion histories, and we quantify the residence time of samples within the upper 2 m of the bedrock subsurface—a measure defined as the cosmogenic nuclide memory. This cosmogenic nuclide memory exceeds 600 ka on the highest plateaus but is limited to less than 500 ka in most other high-elevation samples and to less than 100 ka at low-elevations. Our results define maximum limits for the fraction of ice cover during the past 1 Ma to ∼70% on the Store Koldewey peaks and ∼90% farther inland at Pusterdal, respectively. Minimum limits to ice cover, however, cannot be reliably constrained by the data. Finally, we propose that limited erosion on the highest plateaus of Store Koldewey since 0.6–1.0 Ma indicates a minimum age for fjord-plateau formation within this area of northeast Greenland.

The widespread existence of cosmogenic nuclides accumulated in bedrock prior to the last glaciation demonstrates the limited erosional efficacy of the most recent Fennoscandian and Laurentide ice sheets. Yet the deeper history of erosion in these landscapes repeatedly blanketed by ice remains essentially unknown. Here we present the first comprehensive ice sheet-wide analysis of cosmogenic 10 Be data (n = 953) from the Fennoscandian landscape. We find 64% of all sampled bedrock surfaces contain 10 Be inheritance, including >85% of blockfields and tors, and >50% of ice-carved terrain, in addition to 27% of ice-transported boulders. Recent ice sheets scoured landscapes well beyond glacial troughs and nuclide inventories reveal a patchy legacy of erosional effectiveness that diminishes at high elevations, such that 89% (n = 55) of bedrock samples retain inheritance above 1600 m. We exploit this widespread nuclide inheritance in a Markov chain Monte Carlo-based inversion model to estimate long-term erosion rates and surface exposure histories from 113 paired 10 Be– 26 Al bedrock samples. Nuclide inventories with or without inheritance convey equally important information about the erosional effectiveness of the last ice sheet. We define cosmogenic nuclide memory as the residence time of bedrock samples inside the nuclide-production window (≤2 m depth) where ∼ 80% of the total nuclide production occurs. The cosmogenic nuclide memory is set by mean erosion rate and varies from ∼10 ka for samples eroded >2 m during the last glaciation to > 1-Ma for the slowest erosion rates. We find that mean erosion rates are well constrained compared to the ratio of exposure to burial. The inclusion of bedrock erosion in our computations thwarts the capacity to constrain surface exposure history or identify former nunataks from paired 10 Be– 26 Al data. Ice-carved surfaces reflect diverse erosion histories that are not straightforward to interpret from surficial morphology alone. Relative to the ∼10 mm/kyr benchmark for polar ice masses, we report point-based mean erosion rates that vary by more than three orders of magnitude, with glacial troughs and areal-scour terrain eroding at ∼1 to >100 mm/kyr, blockfields at 0.8–16 mm/kyr, and tors at 0.8–7.7 mm/kyr (5 th –95th percentiles).

Uranium (U) isotopes are useful for constraining the timescales of weathering and erosion processes. The (234U/238U) activity ratio (parentheses denote activity ratio) of fine-grained detrital minerals is proposed to record the time elapsed since mineral grains were reduced to

Zhao et al. (2019) examine gigantic landslides in the Diexi area along the eastern margin of the Tibetan Plateau and propose their successive failure based on a knickpoint migration conceptual model. They postulate that a major river knickpoint (KpMJ) identified by them in the middle reaches of the Minjiang River (Sichuan, China) was initiated at the Longmenshan Fault and subsequently propagated ~ 85 km to its present position upstream of the Diexi lake. They then argue that this retreating knickpoint left in its wake an inner gorge that undercut and destabilized hillslopes, triggering a series of large landslides in the Diexi area. We question this interpretation based on our high-resolution landslide mapping, an analysis of knickpoints (i.e., profile convexities >30 m high) in the Minjiang channel network, and field observations of lacustrine sediments and epigenetic gorges associated with the Diexi landslides. We confront the model proposed by Zhao et al. (2019) with three key arguments: 1) Major profile convexities in the Diexi area, including KpMJ, are associated with landslide or debris flow deposits and there is no basis for connecting explicitly any of these to long-distance knickpoint retreat; 2) the giant Diexi paleolandslide predates the debris avalanches at KpMJ, therefore the latter cannot have been the trigger for landsliding in this area; and 3) the spatial distribution of 666 mapped knickpoints in the Minjiang River mainstem and tributaries is not consistent with simple long-distance propagation of an ‘incisional wave’ initiated at the Longmenshan Fault.

The origins and Pleistocene evolution of plateau landscapes along passive continental margins of the North Atlantic have been debated for more than a century. A key question in this debate concerns whether glacial and periglacial surface processes have substantially eroded plateau areas during late Cenozoic climatic cooling or whether the plateaus have mainly been protected from erosion by cold-based and largely nonerosive ice sheets. Here we investigate the Pleistocene evolution of a prominent plateau landscape in Reinheimen National Park, southern Norway. We estimate erosion rates across the plateau via inverse modeling of 141 new cosmogenic 10Be and 26Al measurements in regolith profiles and bedrock. We combine these results with sedimentological analyses of the regolith. In the vicinity of Reinheimen's regolith-covered summits, the combination of uniformly slow erosion (50 m/Myr), possibly due to episodic glacial erosion. Despite some indications of chemical alteration, such as grusic saprolite and small amounts of secondary minerals, the fine regolith comprises low clay/silt ratios and is dominated by primary minerals with no sign of dissolution. Together with our modeled erosion rates, this indicates that the regolith cover formed, and continues to develop, during the cold climate of the Late Pleistocene.

Determining sediment transfer times is key to understanding source‐to‐sink dynamics and the transmission of environmental signals through the fluvial system. Previous work on the Bolivian Altiplano applied the in situ cosmogenic 14C‐10Be‐chronometer to river sands and proposed sediment storage times of ~10–20 kyr in four catchments southeast of Lake Titicaca. However, the fidelity of those results hinges upon isotopic steady‐state within sediment supplied from the source area. With the aim of independently quantifying sediment storage times and testing the 14C‐10Be steady‐state assumption, we dated sediment storage units within one of the previously investigated catchments using radiocarbon dating, cosmogenic 10Be‐26Al isochron burial dating, and 10Be‐26Al depth‐profile dating. Palaeosurfaces appear to preserve remnants of a former fluvial system, which has undergone drainage reversal, reduction in catchment area, and local isostatic uplift since ~2.8 Ma. From alluvium mantling the palaeosurfaces we gained a deposition age of ~580 ka, while lower down fluvial terraces yielded ≤34 ka, and floodplains ~3–1 ka. Owing to restricted channel connectivity with the terraces and palaeosurfaces, the main source of channel sediment is via reworking of the late Holocene floodplain. Yet modelling a set of feasible scenarios reveals that floodplain storage and burial depth are incompatible with the 14C‐10Be disequilibrium measured in the channel. Instead we propose that the 14C‐10Be offset results from: (i) non‐uniform erosion whereby deep gullies supply hillslope‐derived debris; and/or (ii) holocene landscape transience associated with climate or human impact. The reliability of the 14C‐10Be chronometer vitally depends upon careful evaluation of sources of isotopic disequilibrium in a wide range of depositional and erosional landforms in the landscape.

Mountain rivers respond to strong earthquakes by rapidly aggrading to accommodate excess sediment delivered by co-seismic landslides. Detailed sediment budgets indicate that rivers need several years to decades to recover from seismic disturbances, depending on how recovery is defined. We examine three principal proxies of river recovery after earthquake-induced sediment pulses around Pokhara, Nepal's second largest city. Freshly exhumed cohorts of floodplain trees in growth position indicate rapid and pulsed sedimentation that formed a fan covering 150 km2 in a Lesser Himalayan basin with tens of metres of debris between the 11th and 15th centuries AD. Radiocarbon dates of buried trees are consistent with those of nearby valley deposits linked to major medieval earthquakes, such that we can estimate average rates of re-incision since. We combine high-resolution digital elevation data, geodetic field surveys, aerial photos, and dated tree trunks to reconstruct geomorphic marker surfaces. The volumes of sediment relative to these surfaces require average net sediment yields of up to 4200 t km–2 yr–1 for the 650 years since the last inferred earthquake-triggered sediment pulse. The lithological composition of channel bedload differs from that of local bedrock, confirming that rivers are still mostly evacuating medieval valley fills, locally incising at rates of up to 0.2 m yr–1. Pronounced knickpoints and epigenetic gorges at tributary junctions further illustrate the protracted fluvial response; only the distal portions of the earthquake-derived sediment wedges have been cut to near their base. Our results challenge the notion that mountain rivers recover speedily from earthquakes within years to decades. The valley fills around Pokhara show that even highly erosive Himalayan rivers may need more than several centuries to adjust to catastrophic perturbations. Our results motivate some rethinking of post-seismic hazard appraisals and infrastructural planning in active mountain regions.

For sediment-burial dating with a cosmogenic nuclide pair, the isochron burial method performs well provided that the sediment source has undergone (1) steady erosion and (2) continuous exposure to cosmic rays. These conditions exert important limitations on applications of the method. And yet, in mountainous fluvial and glacial landscapes, it is commonly found that the source area has experienced landsliding or glacial quarrying (i.e., non-steady erosion), and/or intermittent sediment storage or burial beneath glaciers (i.e., discontinuous exposure). As well as breaching the assumptions of the isochron method, such processes tend to yield low nuclide concentrations in the sample, which further limits its workability.Here we present a more flexible method that accommodates complex, non-steady pre-burial erosion and exposure histories: conditions that exclude the isochron burial method. P-PINI (Particle Pathway Inversion of Nuclide Inventories) is a Monte Carlo-based inversion model that employs a source-to-sink approach for estimating the depositional age of fluvial and glaciogenic sediments. This method has been successfully applied to the Deckenschotter in the northern Alpine foreland (see Knudsen et al. 2020, Earth & Planetary Science Letters 549, 116491). As with the isochron burial method, P-PINI exploits an ensemble of paired nuclide (e.g., 10Be-26Al) concentrations measured in different samples from the same depth in a sedimentary sequence. But unlike the isochron method, P-PINI applies a stochastic approach to simulate a wide range of possible pre-depositional exposure and erosion histories for each individual sample. These different pre-burial histories (unique to each sample) are then integrated with the constraint that all samples share a common burial history at the sink. Where cosmogenic nuclide data (or other chronometric data, e.g., OSL) are available for multiple sites, Bayesian inference modelling can impose a priori relative age constraints, or estimates on the maximum duration of sediment storage.In this presentation, we extend P-PINI to explore how sediment storage and reworking (i.e., a range of burial depths and durations) between source and sink affects burial age estimates. Significant intermediate storage is characteristic of large river systems, such as the Danube River. Using cosmogenic 10Be-26Al concentrations measured in fluvial gravels at Gänserndorf and Schlosshof, two terraces along the Danube River in the Vienna Basin (Braumann et al., 2019. Quat. Int. 509. 87-102), we examine how the burial ages at these two sites are a function of the pre-burial history experienced by the samples.

Catastrophic meltwater drainage from glacial Lake Agassiz has been hypothesised as a trigger for large-scale ocean circulation change initiating the Younger Dryas cold reversal. Here we quantify the flood discharge that formed the northwestern outlet of Lake Agassiz using a one-dimensional step-backwater model and a zero-dimension gradual-incision model. Applying these two independent models, we estimate a peak discharge range of 1.8-2.5 × 106 m3 s-1 and a flood volume of ∼21,000 km3. Such a discharge can only be derived from Lake Agassiz rather than one of the two smaller regional glacial lakes: Churchill or Meadow. When coupled with existing ice margin chronologies, these results demonstrate that the northwestern outlet of Lake Agassiz provides a viable link for catastrophic meltwater to drain to the Arctic Ocean over a 5-10 month period during the Younger Dryas, though it is unclear whether this was near its beginning., This research was funded by the Natural Sciences and Engineering Research Council (D.F.), the Canada Research Chairs Program (D.F.) and a Primus fellowship from Charles University (M.M.). We thank the editor, Mathieu Morlighem, and reviewers, Andy Breckenridge and Julian Murton. Data is available through Government of Alberta (2017). Government of Alberta (2017). LiDAR Data Archives. Data provided under license and with project specific data sharing agreement by the Archaeological Survey of Alberta, Culture, Multiculturalism, and the Status of Women. Edmonton, Alberta via geodiscoveralberta@gov.ab.ca through https://geodiscover.alberta.ca/geoportal/#searchPanel

Remote sensing of specific climatic and biogeographical factors is an effective means of evaluating the desertification status of dryland regions affected by negative human impacts. Here, we identify and analyse land desertification trends in Iran via a combination of three indices of vegetation (NPP—net primary production, NDVI—normalized difference vegetation index, and LAI—leaf area index); and two climate indices (LST—land surface temperature, and P—precipitation) during the period 2001–2015. The Mann-Kendall non-parametric test, the Theil–Sen estimator, and a simple linear regression method were then applied to identify trends and to map regions of Iran that are susceptible to desertification. Our results show that an area of 680,000 km2 (~ 56 %) of Iran is classified with a very high level, indicating that a large fraction of Iran is susceptible to land desertification. We suggest that spatial and temporal trends in the three vegetation indices (NPP, NDVI, and LAI) and the two climate indices (LST and P) are a cost-effective choice for the prediction and management of future environmental trends in the world’s developing regions, and are a step towards achieving land-use sustainability by helping to locate the most degraded areas.

The mantle convection accompanying plate motion causes vertical movements of up to a few hundred metres at Earth’s surface over wavelengths of 102–103 km. This dynamic topography appears to come and go at ~ 1–10 Myr timescales in areas that are often well away from plate margins, although its spatial and temporal characteristics are subject to ongoing debate. Since such motions are small and transient, discriminating convective signals from other drivers of relief generation and/or sediment dispersal remains tricky. An outstanding challenge is to detect these elusive, transient undulations from a tell-tale geomorphic imprint preserved in either drainage patterns or the stratigraphic record.In the intra-plate setting of central Australia, a 30 km long sinuous gorge is developed where the major regional drainage, Finke River, dissects a band of low hills. Remarkably, this gorge is intertwined with an abandoned and less deeply incised gorge that forms hanging junctions and shares similar width and sinuosity. This unusual overprinting of the two gorges remains unexplained.With an aim to investigate the history of the intertwined gorges, we measured cosmogenic 10Be and 26Al in fluvial gravels stored in the palaeovalley cutoffs. The gravels are remnants of major alluviation episodes that we surmise result from ongoing vertical motions associated with dynamic topography. We use a Markov chain Monte Carlo-based inversion model to test two hypotheses to explain the nuclide inventory contained within the stored fluvial gravels. In the first case, rapid alluviation and erosion since 1 Ma preserves the nuclide memory of the source area; in the second, the nuclide memory is erased during long-term fluvial storage (> 5 Myr) and is restored during exhumation of the palaeovalley gravel-pile. The two hypotheses are therefore limiting-case scenarios that constrain overall fast versus slow landscape evolution, respectively. Our model results suggest that long-term burial decouples the source-area signal from nuclide abundances measured in the palaeovalley gravels. This casts events into a Miocene timescale.

Cataclysmic outburst floods transformed landscapes and caused abrupt climate change during the last deglaciation. Whether such events have also characterized previous deglaciations is not known. Arctic marine cores hint at megafloods prior to Oxygen Isotope Stage (OIS) 2, but the overprint of successive glaciations means that geomorphological traces of ancient floods remain scarce in Eurasia and North America. Here we present the first well-constrained terrestrial megaflood record to be linked with Arctic archives. Based on cosmogenic-nuclide exposure dating and optically stimulated luminescence dating applied to glacial-lake sediments, a 300-m deep bedrock spillway, and giant eddy-bars > 200-m high, we reconstruct a history of cataclysmic outburst floods from glacial Lake Vitim, Siberia, to the Arctic Ocean over the past 60,000-years. Three megafloods have reflected the rhythm of Eurasian glaciations, leaving traces that stretch more than 3500 km to the Lena Delta. The first flood was coincident with deglaciation from OIS-4 and the largest meltwater spike in Arctic marine-cores within the past 100,000 years (isotope-event 3.31 at 55.5 ka). The second flood marked the lead up to the local Last Glacial Maximum, and the third flood occurred during the last deglaciation. This final 3000 km3 megaflood stands as one of the largest freshwater floods ever documented, with peak discharge of 4.0–6.5 million m3s−1, mean flow depths of 120–150 m, and average flow velocities up to 21 m s−1.

Low-relief plateaus separated by deeply incised fjords are hallmarks of glaciated, passive continental margins. Spectacular examples fringe the once ice-covered North Atlantic coasts of Greenland, Norway and Canada, but low-relief plateau landscapes also underlie present-day ice sheets in Antarctica and Greenland. Dissected plateaus have long been viewed as the outcome of selective linear erosion by ice sheets that focus incision in glacial troughs, leaving the intervening landscapes essentially unaffected. According to this hypothesis, the plateaus are remnants of preglacial low-relief topography. However, here we use computational experiments to show that, like fjords, plateaus are emergent properties of long-term ice-sheet erosion. Ice sheets can either increase or decrease subglacial relief depending on the wavelength of the underlying topography, and plateau topography arises dynamically from evolving feedbacks between topography, ice dynamics and erosion over million-year timescales. This new mechanistic explanation for plateau formation opens the possibility of plateaus contributing significantly to accelerated sediment flux at the onset of the late Cenozoic glaciations, before becoming stable later in the Quaternary.

To provide a new reconstruction of the deglaciation of the Fennoscandian Ice Sheet, in the form of calendar-year time-slices, which are particularly useful for ice sheet modelling, we have compiled and synthesized published geomorphological data for eskers, ice-marginal formations, lineations, marginal meltwater channels, striae, ice-dammed lakes, and geochronological data from radiocarbon, varve, optically-stimulated luminescence, and cosmogenic nuclide dating. This is summarized as a deglaciation map of the Fennoscandian Ice Sheet with isochrons marking every 1000 years between 22 and 13 cal kyr BP and every hundred years between 11.6 and final ice decay after 9.7 cal kyr BP. Deglaciation patterns vary across the Fennoscandian Ice Sheet domain, reflecting differences in climatic and geomorphic settings as well as ice sheet basal thermal conditions and terrestrial versus marine margins. For example, the ice sheet margin in the high-precipitation coastal setting of the western sector responded sensitively to climatic variations leaving a detailed record of prominent moraines and other ice-marginal deposits in many fjords and coastal valleys. Retreat rates across the southern sector differed between slow retreat of the terrestrial margin in western and southern Sweden and rapid retreat of the calving ice margin in the Baltic Basin. Our reconstruction is consistent with much of the published research. However, the synthesis of a large amount of existing and new data support refined reconstructions in some areas. For example, the LGM extent of the ice sheet in northwestern Russia was located far east and it occurred at a later time than the rest of the ice sheet, at around 17–15 cal kyr BP. We also propose a slightly different chronology of moraine formation over southern Sweden based on improved correlations of moraine segments using new LiDAR data and tying the timing of moraine formation to Greenland ice core cold stages. Retreat rates vary by as much as an order of magnitude in different sectors of the ice sheet, with the lowest rates on the high-elevation and maritime Norwegian margin. Retreat rates compared to the climatic information provided by the Greenland ice core record show a general correspondence between retreat rate and climatic forcing, although a close match between retreat rate and climate is unlikely because of other controls, such as topography and marine versus terrestrial margins. Overall, the time slice reconstructions of Fennoscandian Ice Sheet deglaciation from 22 to 9.7 cal kyr BP provide an important dataset for understanding the contexts that underpin spatial and temporal patterns in retreat of the Fennoscandian Ice Sheet, and are an important resource for testing and refining ice sheet models. Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)

Successively smaller glacial extents have been proposed for continental Eurasia during the stadials of the last glacial period leading up to the Last Glacial Maximum (LGM). At the same time the large mountainous region east of Lake Baikal, Transbaikalia, has remained unexplored in terms of glacial chronology despite clear geomorphological evidence of substantial past glaciations. We have applied cosmogenic Be-10 exposure dating and optically stimulated luminescence to establish the first quantitative glacial chronology for this region. Based on eighteen exposure ages from five moraine complexes, we propose that large mountain ice fields existed in the Kodar and Udokan mountains during Oxygen Isotope Stage 2, commensurate with the global LGM. These ice fields fed valley glaciers (>100 km in length) reaching down to the Chara Depression between the Kodar and Udokan mountains and to the valley of the Vitim River northwest of the Kodar Mountains. Two of the investigated moraines date to the Late Glacial, but indications of incomplete exposure among some of the sampled boulders obscure the specific details of the post-LGM glacial history. In addition to the LGM ice fields in the highest mountains of Transbaikalia, we report geomorphological evidence of a much more extensive, ice-cap type glaciation at a time that is yet to be firmly resolved. (C) 2015 Elsevier Ltd. All rights reserved.

Cosmogenic nuclides are typically used to either constrain an exposure age, a burial age, or an erosion rate. Constraining the landscape history and past erosion rates in previously glaciated terrains is, however, notoriously difficult because it involves a large number of unknowns. The potential use of cosmogenic nuclides in landscapes with a complex history of exposure and erosion is therefore often quite limited. Here, we present a novel multi-nuclide approach to study the landscape evolution and past erosion rates in terrains with a complex exposure history, particularly focusing on regions that were repeatedly covered by glaciers or ice sheets during the Quaternary. The approach, based on the Markov Chain Monte Carlo (MCMC) technique, focuses on mapping the range of landscape histories that are consistent with a given set of measured cosmogenic nuclide concentrations. A fundamental assumption of the model approach is that the exposure history at the site/location can be divided into two distinct regimes: i) interglacial periods characterized by zero shielding due to overlying ice and a uniform interglacial erosion rate, and ii) glacial periods characterized by 100% shielding and a uniform glacial erosion rate. We incorporate the exposure history in the model framework by applying a threshold value to the global marine benthic δ18O record and include the threshold value as a free model parameter, hereby taking into account global changes in climate. However, any available information on the glacial-interglacial history at the sampling location, in particular the timing of the last deglaciation event, is readily incorporated in the model to constrain the inverse problem. Based on the MCMC technique, the model delineates the most likely exposure history, including the glacial and interglacial erosion rates, which, in turn, makes it possible to reconstruct an exhumation history at the site. We apply the model to two landscape scenarios based on synthetic data and two landscape scenarios based on paired 10Be/26Al data from West Greenland, which makes it possible to quantify the denudation rate at these locations. The model framework, which currently incorporates any combination of the following nuclides 10Be, 26Al, 14C, and 21Ne, is highly flexible and can be adapted to many different landscape settings. The model framework may also be used in combination with physics-based landscape evolution models to predict nuclide concentrations at different locations in the landscape. This may help validate the landscape models via comparison to measured nuclide concentrations or to devise new effective sampling strategies.

The blocking of river courses by mass movements is very common in mountainous areas with deep and narrow valleys. Landslide dams may pose serious threats to people and their livelihoods downstream in the case of abrupt dam failure. Since the publication of benchmark reviews of Costa and Schuster (1988) and Korup (2002) , there is a growing number of studies focusing on the formation, stability, and short-term impacts of landslide dams. This review combines the insights of all these studies, builds on current concepts of landslide dams, and suggests ways to unify terminologies and classifications. We furthermore present a new worldwide database compiled from literature data. It contains 410 landslide dams >1 million m3 in volume that were formed since 1900 since these have the most complete data entries. These data show that dam longevity is, among other factors, correlated with the type of landslide forming the dam. Those formed by rock/debris avalanches and rockslides have longest lifespans. However, the influence of landslide type or material on dam longevity decreases with time after dam formation. To ensure consistency in the next database generation, we suggest guidelines for data collection to provide a solid basis for evaluating dam stability and governing factors. A preliminary classification matrix for landslide dam stability that combines topographic setting and the internal structure of the dam body is another outcome of our review. Furthermore, an evaluation of the various geomorphic stability indices proposed in the literature regarding their suitability and limitations in assessing dam formation and stability shows that they predict the probability of dam formation reasonably well, but that their application to longevity estimates requires further assessment. The geomorphic impacts of landslide dams in the short-, medium- and long-term are summarized and illustrated with key examples. Finally, for a better understanding of the factors controlling dam stability, we recommend to (1) include dam composition and sedimentary structures in future case studies, (2) maintain and update the worldwide database for sound statistical analyses, (3) refine landslide dam stability indices and test them for different landslide types, and (4) study hazard cascades related to multiple dams in one watershed. For long-term landscape evolution studies, we suggest to (5) quantify terrestrial sediment flux related to landslide dams, (6) detect ancient landslide dams in river profiles, and (7) further exploit the sediment archives in former impoundment areas.

Sediment-routing systems continuously transfer information and mass from eroding source areas to depositional sinks. Understanding how these systems alter environmental signals is critical when it comes to inferring source-area properties from the sedimentary record. We measure cosmogenic 10Be and 26Al along three large sediment-routing systems (~ 100,000 km2) in central Australia with the aim of tracking downstream variations in 26Al/10Be inventories and to identify the factors responsible. By comparing 56 new cosmogenic 10Be and 26Al measurements in stream sediments with matching data (n = 55) from source areas, we show that 26Al/10Be inventories in hillslope bedrock and soils set the benchmark for relative downstream modifications. Lithology is the primary determinant of erosion-rate variations in source areas and despite sediment mixing over hundreds of kilometres downstream a distinct lithological signal is retained. Postorogenic ranges yield catchment erosion rates of ~ 6–11 m/m.y. and silcrete-dominant areas erode as slow as ~ 0.2 m/m.y. 26Al/10Be inventories in stream-sediments reveal overall downstream-increasing minimum cumulative burial terms up to ~ 1.1 m.y. but more generally ~ 400–800 k.y. The magnitude of the burial signal correlates with increasing sediment cover downstream and reflects assimilation from storages with long exposure histories, such as alluvial fans, desert pavements, alluvial plains, and aeolian dunes. We propose that the tendency for large alluvial rivers to mask their 26Al/10Be source-area signal differs according to geomorphic setting. Signal preservation is favoured by i) high sediment supply rates, ii) high mean runoff, and iii) a thick sedimentary basin pile. Conversely, signal masking prevails in landscapes of i) low sediment supply, ii) discontinuous sediment flux, and iii) juxtaposition of sediment storages with notably different exposure histories.

Central to the debate over the extinction of many of Australia's last surviving megafauna is the question: Was climate changing significantly when humans arrived and megafauna went extinct? Here we present a new perspective on variations in climate and water resources over the last glacial cycle in arid Australia based on the study of the continent's largest lake basin and its tributaries. By dating paleoshorelines and river deposits in the Lake Eyre basin, we show that major hydrological change caused previously overflowing megalakes to enter a final and catastrophic drying phase at 48 ± 2 ka just as the giant bird, Genyornis newtoni, went extinct (50-45 ka). The disappearance of Genyornis and other megafauna has been previously attributed to "ecosystem collapse" coincident with the spread of fire-wielding humans. Our findings suggest a climate-driven hydrological transformation in the critical window of human arrival and megafaunal extinction, and the results call for a re-evaluation of a humanmediated cause for such extinctions in arid Australia.

The influence of major Quaternary climatic changes on growth and decay of the Greenland Ice Sheet, and associated erosional impact on the landscapes, is virtually unknown beyond the last deglaciation. Here we quantify exposure and denudation histories in west Greenland by applying a novel Markov-Chain Monte Carlo modelling approach to all available paired cosmogenic 10Be-26Al bedrock data from Greenland. We find that long-term denudation rates in west Greenland range from >50 m Myr−1 in low-lying areas to ∼2 m Myr−1 at high elevations, hereby quantifying systematic variations in denudation rate among different glacial landforms caused by variations in ice thickness across the landscape. We furthermore show that the present day ice-free areas only were ice covered ca. 45% of the past 1 million years, and even less at high-elevation sites, implying that the Greenland Ice Sheet for much of the time was of similar size or even smaller than today., Erosion rates and ice cover extent of present day fjords and summit plateau landscapes beyond the last deglaciation are virtually unknown. Here, the authors constrain the long-term denudation rates and glaciation history in west Greenland based on cosmogenic nuclides.

Three terrestrial climate proxies are used to investigate the evolution of Holocene palaeoenvironments in southern central Australia, all of which present a coherent record of palaeohydrology. Single-grain optically stimulated luminescence from sediments supplemented by 14C from charcoal and lacustrine shells was obtained to date shoreline deposits (Lake Callabonna) and the adjacent Mt Chambers Creek alluvial fan. Our findings are complemented by a U/Th-based record of speleothem growth in the Mt Chambers Creek catchment, which we interpret to reflect increased precipitation. Together, these archives shed light on the timing of, and possible sources of water for, Holocene pluvial intervals. We identified several phases of elevated lake levels dated at ~5.8–5.2, 4.5, 3.5–2.7 and 1 kyr, most of which correspond to fluvial activity resulting from increased precipitation in the adjacent ranges. The enhanced hydrology during phases of the late Holocene likely increased the reliability of resources for regional human populations during a time of reduced winter rainfall. When considered within the framework of the current understanding of Holocene palaeoclimate in central Australia, our data suggest that the pattern of landscape response was broadly synchronous with larger scale climatic variability and punctuated by pluvial periods greater than today.

In this paper we synthesise existing palaeoenvironmental data from the arid and semi-arid interior of the Australian continent for the period 40-0 ka. Moisture is the predominant variable controlli ...

The Cape Mountains of southern Africa exhibit an alpine-like topography in conjunction with some of the lowest denudation rates in the world. This presents an exception to the often-cited coupling of topography and denudation rates and suggests that steep slopes alone are not sufficient to incite the high denudation rates with which they are commonly associated. Within the Cape Mountains, slope angles are often in excess of 30 degrees and relief frequently exceeds 1 km, yet Be-10-based catchment-averaged denudation rates vary between 2.32 +/- 0.29 m/m.y. and 7.95 +/- 0.90 m/m.y. We attribute the maintenance of rugged topography and suppression of denudation rates primarily to the presence of physically robust and chemically inert quartzites that constitute the backbone of the mountains. Be-10-based bedrock denudation rates on the interfluves of the mountains vary between 1.98 +/- 0.23 m/m.y. and 4.61 +/- 0.53 m/m.y. The close agreement between the rates of catchment-averaged and interfluve denudation indicates topography in steady state. These low denudation rates, in conjunction with the suggestion of geomorphic stability, are in agreement with the low denudation rates (

A sudden drop in river base-level can trigger a knickpoint that propagates throughout the fluvial network causing a transient state in the landscape. Knickpoint retreat has been confirmed in large ...

Optically stimulated luminescence (OSL) ages from a relict shoreline on Lake Callabonna record a major pluvial episode in southern central Australia between 1050 70 and 1100 60 Common Era (CE), w ...

Optically stimulated and thermoluminescence ages from relict shorelines, along with accelerator mass spectrometer 14C ages from freshwater molluscs reveal a record of variable moisture sources supplied by northern and southern river systems to Lake Mega-Frome in southern central Australia during the late Quaternary. Additional lacustrine, palynological and terrestrial proxies are used to reconstruct a record that extends back to 105 ka, confirming that Lakes Mega-Frome and Mega-Eyre were joined to create the largest system of palaeolakes on the Australian continent as recently as 50–47 ka. The palaeohydrological record indicates a progressive shift to more arid conditions, with marked drying after 45 ka. Subsequently, Lake Mega-Frome has filled independently at 33–31 ka and at the termination of the Last Glacial Maximum to volumes some 40 times those of today. Further sequentially declining filling episodes (to volumes 25–10 those of today) occurred immediately prior to the Younger Dryas stadial, in the mid Holocene and during the medieval climatic anomaly. Southern hemisphere summer insolation maxima are a poor predictor of palaeolake-filling episodes. An examination of multiple active moisture sources suggests that palaeolake phases were driven independently of insolation and at times by some combination of enhanced Southern Ocean circulation and strengthened tropical moisture sources.

This study reports the nature and timing of Quaternary fluvial activity in the Fitzroy River basin, which drains a diverse 143,000 km(2) area in northeastern Queensland, before discharging into the ...

escarpment edge, deep gorges and distinct knickzones in river profiles characterize the landscape on the Western Escarpment of the Andes between ~5°S and ~18°S (northern Peru to northern Chile). Th ...

The nature of the Australian climate at about the time of rapid megafaunal extinctions and humans arriving in Australia is poorly understood and is an important element in the contentious debate as to whether humans or climate caused the extinctions. Here we present a new paleoshoreline chronology that extends over the past 100 k.y. for Lake Mega-Frome, the coalescence of Lakes Frome, Blanche, Callabonna and Gregory, in the southern latitudes of central Australia. We show that Lake Mega-Frome was connected for the last time to adjacent Lake Eyre at 50–47 ka, forming the largest remaining interconnected system of paleolakes on the Australian continent. The final disconnection and a progressive drop in the level of Lake Mega-Frome represents a major climate shift to aridification that coincided with the arrival of humans and the demise of the megafauna. The supply of moisture to the Australian continent at various times in the Quaternary has commonly been ascribed to an enhanced monsoon. This study, in combination with other paleoclimate data, provides reliable evidence for periods of enhanced tropical and enhanced Southern Ocean sources of water filling these lakes at different times during the last full glacial cycle.

A steep escarpment edge deep gorges and distinct knickzones in river profiles characterize the landscape on the Western Escarpment of the Andes between 5°S and 18°S (northern Peru to northern Chile). Strong north–south and east–west precipitation gradients are exploited in order to determine how climate affects denudation rates in three river basins spanning an otherwise relatively uniform geologic and geomorphologic setting. Late Miocene tectonics uplifted the Meseta/Altiplano plateau (3000m a.s.l.) which is underlain by a series of Tertiary volcanic-volcanoclastic rocks. Streams on this plateau remain graded to the Late Miocene base level. Below the rim of the Meseta streams have responded to this ramp uplift by incising deeply into fractured Mesozoic rocks via a series of steep headward retreating knickzones that grade to the present-day base level defined by the Pacific Ocean. It is found that the Tertiary units on the plateau function as cap-rocks which aid in the parallel retreat of the sharp escarpment edge and upper knickzone tips. 10Be-derived catchment denudation rates of the Rio Piura (5°S) Rio Pisco (13°S) and Rio Lluta (18°S) average 10mm ky-1 on the Meseta/Altiplano irrespective of precipitation rates; whereas downstream of the escarpment edge denudation rates range from 10mm ky-1 to 250mm ky-1 and correlate positively with precipitation rates but show no strong correlation with hillslope angles or channel steepness. These relationships are explained by the presence of a cap-rock and climate driven fluvial incision that steepens hillslopes to near-threshold conditions. Since escarpment retreat and the precipitation pattern were established at least in the Miocene it is speculated that the present-day distribution of morphology and denudation rates has probably remained largely unchanged during the past several millions of years as the knickzones have propagated headward into the plateau.

Many anabranching rivers are characterized by dynamic interactions between fluvial processes and riparian vegetation, but uncertainties surround the processes and time scales of anabranch development. We use geomorphological investigations and optically stimulated luminescence (OSL) dating to determine spatial and temporal trends in the development of anabranching along a 6.5-km-long reach of Magela Creek in the seasonal tropics of northern Australia. Many trees and shrubs that survive the wet-season floods establish on the sandy beds and lower banks, such that anabranches divide and rejoin around numerous ridges and islands that are formed mainly by accretion in the lee of in-channel vegetation and, less commonly, by excision from formerly continuous island or flood plain surfaces. Once ridges and islands form, colonizing vegetation maintains their stability by increasing sediment cohesion and decreasing flow erosivity. Over the Holocene, Magela Creek has vertically aggraded and extended in length by delta progradation into Madjinbardi Billabong, resulting in a time sequence of anabranches and associated ridges and islands from older (upstream) to younger (downstream). OSL ages for islands in the upstream and middle reaches are ca. 1.6 ka and older, and the narrow, deep anabranches (width/depth [w/d] typically ~10–30) have few in-channel obstructions. Farther downstream, island OSL ages are ca. 0.7 ka and younger, anabranches tend to be wider and shallower (w/d >30) with more obstructions, and splays and locally scoured island and floodplain surfaces are more common. Based on these findings, previous flow and sediment-transport measurements, and theoretical analyses, we posit that there is a decline in anabranch efficiency from an upstream equilibrium system in mass-flux balance to a downstream disequilibrium system characterized by bed aggradation and localized island and floodplain erosion. In the downstream reaches, inefficient (high w/d and obstructed) anabranches do not persist because they either aggrade and are abandoned, or they are subdivided into more efficient (lower w/d and less obstructed) anabranches as a result of the interactions between in-channel vegetation growth and ridge and island accretion or local excision. Consequently, a more efficient anabranching system gradually develops with characteristics similar to those in the upstream reaches. This enhances downstream sediment transfer, which enables ongoing delta progradation and provides fresh sediment surfaces for vegetation to colonize and initiate new anabranches. The OSL ages from Magela Creek demonstrate that a recognizable but relatively inefficient anabranching system can develop within a few centuries, while adjustment to a more efficient system occurs over a few millennia.

Mixed bedrock–alluvial rivers–bedrock channels lined with a discontinuous alluvial cover–are key agents in the shaping of mountain belt topography by bedrock fluvial incision. Whereas much research focuses upon the erosional dynamics of such rivers in the context of rapidly uplifting orogenic landscapes, the present study investigates river incision processes in a post-orogenic (cratonic) landscape undergoing extremely low rates of incision ( The mobility of coarse bed materials (up to 2 m diameter) during late Holocene palaeofloods of known magnitude and age is evaluated using step-backwater flow modelling in conjunction with two selective entrainment equations. A new approach for quantifying the formative flood magnitude in mixed bedrock–alluvial rivers is described here based on the mobility of a key coarse fraction of the bed materials; in this case the d84 size fraction. A 350 m3/s formative flood fully mobilises the coarse alluvial cover with τb∼200–300 N/m2 across the upper and lower gorge riffles, peaking over 500 N/m2 in the knickzone. Such floods have an annual exceedance probability much less than 10− 2 and possibly as low as 10− 3. The role of coarse alluvial cover in the gorge is discussed at two scales: (1) modulation of bedrock exposure at the reach-scale, coupled with adjustment to channel width and gradient, accommodates uniform incision across rocks of different erodibility in steady-state fashion; and (2) at the sub-reach scale where coarse boulder deposits (corresponding to τb minima) cap topographic convexities in the rock floor, thereby restricting bedrock incision to rare large floods. While recent studies postulate that decreasing uplift rates during post-orogenic topographic decay might drive a shift to transport-limited conditions in river networks, observations here and elsewhere in post-orogenic settings suggest, to the contrary, that extremely low erosion rates are maintained with substantial bedrock channel exposure. Although bed material mobility is known to be rate-limiting for bedrock river incision under low sediment flux conditions, exactly how a partial alluvial cover might be spatially distributed to either optimise or impede the rate of bedrock incision is open to speculation. Observations here suggest that the small volume of very stable bed materials lining Sandy Creek gorge is distributed so as to minimise the rate of bedrock fluvial incision over time.

Knickpoint behaviour is a key to understanding both the landscape responses to a base-level fall and the corresponding sediment fluxes from rejuvenated catchments, and must be accommodated in numerical models of large-scale landscape evolution. Knickpoint recession in streams draining to glacio-isostatically uplifted shorelines in eastern Scotland is used to assess whether knickpoint recession is a function of discharge (here represented by its surrogate, catchment area). Knickpoints are identified using DS plots (log slope versus log downstream distance). A statistically significant power relationship is found between distance of headward recession and catchment area. Such knickpoint recession data may be used to determine the values of m and n in the stream power law, E = KAmSn. The data have too many uncertainties, however, to judge definitively whether they are consistent with m = n = 1 (bedrock erosion is proportional to stream power and KPs should be maintained and propagate headwards) or m = 0·3, n = 0·7 (bedrock incision is proportional to shear stress and KPs do not propagate but degrade in place by rotation or replacement). Nonetheless, the E Scotland m and n values point to the dominance of catchment area (discharge) in determining knickpoint retreat rates and are therefore more consistent with the stream power law formulation in which bedrock erosion is proportional to stream power. Copyright © 2005 John Wiley & Sons, Ltd.

Channel-scale sedimentary units associated with bedrock-controlled riffle-pool morphology are examined in detail along Sandy Creek gorge, an ephemeral stream in arid south-eastern central Australia. Pool-fills comprise cut-and-fill assemblages of poorly sorted sediments ranging in texture from muds to boulders. Five unit types are defined based on particle size, sedimentary structures, geometry and bounding surface character: (1) coarse-grained bar platform; (2) fine-grained bar supraplatform; (3) fine-grained pool-fill; (4) fine-grained bench; and (5) modern pool-fill. The last coarse-grained unit currently lining the pools suggests an altered sedimentation style over the post-settlement period (post-ad 1860s). Situated at bedrock valley constrictions, pool-fills are compared with other sedimentary units associated with recirculating currents: eddy bars and slackwater deposits. But only the fine-grained bench units reflect eddy recirculation; the pool-fills are principally forced-bars associated with bedrock-controlled or 'forced' riffle-pool morphology. A late Holocene palaeoflood history is proposed based on radiocarbon ages from the pool-fills: multiple phases of cut-and-fill activity were preceded by a superflood 3400–1900 years ago that eroded the pool-fills to bedrock. The resilience of the pool-fills was illustrated by the passage of a 1-in-100-year flood in 1992, which caused only minor erosion. The presence of pool-fills may provide a window to past phases of river activity that cannot be extracted from either historical records/observations or palaeoflood slackwater sediment analyses. The formation and sedimentary preservation potential of these landforms reflect a combination of hydraulic and structural influences, but the occurrence of high-magnitude floods exerts the dominant control.

A total of 436 logs were used to create 20 engineered log jams (ELJs) in a 1.1 km reach of the Williams River, NSW, Australia, a gravel-bed river that has been desnagged and had most of its riparian vegetation removed over the last 200 years. The experiment was designed to test the effectiveness of reintroducing woody debris (WD) as a means of improving channel stability and recreating habitat diversity. The study assessed geomorphic and ecological responses to introducing woody habitat by comparing paired test and control reaches. Channel characteristics (e.g. bedforms, bars, texture) within test and control reaches were assessed before and after wood placement to quantify the morphological variability induced by the ELJs in the test reach. Since construction in September 2000, the ELJs have been subjected to five overtopping flows, three of which were larger than the mean annual flood. A high-resolution three-dimensional survey of both reaches was completed after major bed-mobilizing flows. Cumulative changes induced by consecutive floods were also assessed. After 12 months, the major geomorphologic changes in the test reach included an increase in pool and riffle area and pool depth; the addition of a pool-riffle sequence; an increase by 0.5-1 m in pool-riffle amplitude; a net gain of 40 m3 of sediment storage per 1000 m2 of channel area (while the control reach experienced a net loss of 15 m3/1000 m2 over the same period); and a substantial increase in the spatial complexity of bed-material distribution. Fish assemblages in the test reach showed an increase in species richness and abundance, and reduced temporal variability compared to the reference reach, suggesting that the changes in physical habitat were beneficial to fish at the reach scale.

Many of the water mills on Scotland's east coast streams, unlike those discussed recently by Downward and Skinner (2005 Area 37 138–47), are found in predominantly bedrock reaches immediately downstream of knickpoints (i.e. bedrock steps). Bedrock knickpoints in the lower reaches of Scottish rivers are a widespread fluvial response to the glacio-isostatic rebound of northern Britain. These steps in the river profile propagate headward over time, but for intervals of a few centuries or so they are sufficiently stable to be exploited for the elevational fall necessary to power the mill wheel. Many of these mills were apparently powered by ‘run-of-the-river’, as are some today that formerly had mill dams. The typical lack of sediment storage along the erosional lower reaches of many Scottish rivers means that failure of mill structures in Scotland will probably have less dramatic geomorphological and management implications than those suggested by Downward and Skinner for southern English rivers.

In four rivers in western Scotland for which there is a well constrained record of relative base-level fall, the rate of postglacial bedrock erosion is quantifi ed by measuring the concentration of in situ cosmogenic 10Be on strath terraces downstream of headward retreating knickpoints. Along-channel gradients in 10Be exposure age show two distinct trends: upstream younging and constant age, which we interpret as diagnostic of knickpoint retreat and diffusive transport-limited incision, respectively. We show that bedrock channel incision and regional formation of strath terraces began shortly after deglaciation (ca. 11.5 ka), and that knickpoint retreat rates peaked in the early to mid-Holocene. Erosion rates have since decreased by two orders of magnitude, converging in the late Holocene to low rates independent of stream power per unit channel area. We infer this regional slowing in postglacial knickpoint retreat to be the result of the depletion of paraglacial sediment supply over the Holocene, leading to a defi ciency in “tools” for bedrock erosion. Our results imply that episodes of major fl uvial erosion may be in tune with glacial cycles, and that sediment depletion following glacial-interglacial transitions may be an important cause of bedrock erosion rate variations in rivers draining glaciated landscapes.