Journal of Geophysical Research: Earth Surface最新文献

Berms composed of surface gravel and underlying sand and gravel mix (gravel berms) naturally form on beaches and can help mitigate coastal erosion and flooding. Previous studies suggest that subsurface sediments influence gravel berm behavior, although detailed investigations remain limited. Here, we present a novel integrated field approach to quantitatively monitor gravel berm subsurface sediments. From March to October 2023, surface and subsurface sediments were characterized using combined methods including mechanical excavation, image-based grain size analysis, GNSS, and ground penetrating radar (GPR) surveys, and compared with gravel berm topography at two beaches (South Carlsbad and Torrey Pines) in southern California with contrasting sediment characteristics. South Carlsbad consistently exhibited backshore gravel exposure and limited seasonal sand volume changes, whereas Torrey Pines exhibited varying backshore gravel exposure and more pronounced seasonal sand volume fluctuations. At both beaches, the gravel berms consistently had a surface pure-gravel (PG) layer overlying a mixed sand-gravel layer, with the PG layer thickness decreasing seaward. At Torrey Pines, the upper gravel berm profile changed seasonally from spring to late summer as sand gradually accumulated within the berm (resulting in a PG layer thickness decrease) and varied subsurface sediment composition. In contrast, in South Carlsbad, both upper gravel berms and subsurface sediment structures exhibited seasonal consistency, with little change in PG layer thickness. The combined approach enables detailed and repeatable assessment of gravel berm subsurface sediments, offering further insights into the links between internal sediment structure and surface morphology. The present results inform a new conceptual model of seasonal gravel berm evolution.

Landsliding in river valleys poses unique risks for cascading hazards and can damage infrastructure and cause fatalities. In postglacial valleys, many landslides are posited to occur in relation to lateral river erosion, but the dynamics of fluvial-hillslope interactions are not well understood. Here, we investigate a section of the Nooksack River in western Washington State where the channel is flanked by landslide-prone glacial terraces similar to those that failed in the 2014 State Route 530 “Oso” landslide. We map 216 landslides through time across 17 aerial imagery data sets (1933–2022) and analyze them in relation to river meandering and curvature. We observe dynamic feedbacks between lateral river meandering and valley-adjacent landsliding. Terrace lateral retreat rates of up to 25 m/year owing to combined fluvial erosion and slope failure occur on pinned, outer meander bends immediately downstream from peaks in river curvature (>0.0075 1/m); these locations are predisposed to both shallow and deep-seated landslides. Deep-seated landslides extending 17%–32% of the active valley width into the floodplain can displace the river away from the floodplain margin and change the channel planform. River-displacing landslides relocate meanders up- or downstream, thereby conditioning the location of subsequent landslides. This conceptual model of coupled landslide-driven meander displacement and valley-adjacent landsliding is exemplified across western Washington river systems. The distance between up- and downstream valley-adjacent landsliding scales with valley width, meander wavelength, and terrace height. Our results can advance our understanding of the river-hillslope interface in landscape evolution and can be used to inform hazard management in river corridors.

The rise in global temperatures is amplified in high-latitude regions, where snow and ice play a vital role in the hydrological cycle. Understanding the impacts of climate change on ecosystems and communities in Northern regions requires accurate hydrological data. Within Northern Canada, in situ data sparsity (in both spatial and temporal resolution) poses a challenge to robust characterization of hydrological trends. The increasing availability of satellite-derived data can provide an independent measure of terrestrial water storage. This study compares terrestrial water storage anomalies (TWSA) from Gravity Recovery and Climate Experiment (GRACE) and GRACE-FO to in situ and satellite-derived precipitation and evaporation products within the Mackenzie River Basin (MRB), Canada, a high-latitude basin characterized by low population density and significant contribution of freshwater to the Arctic Ocean. Declining trends in TWSA from GRACE/GRACE-FO in the MRB are not fully explained by corresponding trends in hydrological parameters. Water budget analysis reveals inconsistencies between GRACE/GRACE-FO derived TWSA and TWSA derived using precipitation, evaporation, and runoff data, which may be attributed to physical processes represented in the GRACE/GRACE-FO observations. Three models of glacial isostatic adjustment (GIA), namely the ICE6G_D (VM5a), Caron-18, and LM-17.3 models, were compared to examine the sensitivity of the GRACE/GRACE-FO-derived TWSA to the GIA model (correction) employed, revealing approximately ±1 cm of equivalent water height per year variability in the TWSA linear trend. The results suggest that robust characterization of regional mass processes (e.g., subsidence, residual GIA) within the MRB is necessary to isolate hydrological mass changes.

To investigate the long-term evolution of granitic regolith under cold temperate climate, we examined a 300 cm-thick regolith profile in the Oroqen Autonomous Banner, northeast China. We analyzed the mineralogy and U-series isotopic compositions of bulk regolith samples. Measurements of (²³⁴U/²³⁸U), (²³⁰Th/²³⁸U), and (²³⁰Th/²³⁴U) isotopic activity ratios indicate U-series disequilibrium, with complex variations in depth, ranging from 0.949 to 0.989, 0.906 to 1.036, and 0.926 to 1.059, respectively. The conventional “gain and loss” model could not be applied across the entire profile in a single simulation. By subdividing the profile into three subzones based on elemental and mineralogical depth variations, the “gain and loss” model was applicable to two subzones, excluding the middle portion. U-series disequilibrium-derived regolith production rates were 1.42 ± 0.03 m/Ma and 5.97 ± 3.98 m/Ma for these subzones. When compared to denudation rates (∼34 m/Ma) determined from in situ cosmogenic nuclides (¹⁰Be and ²⁶Al), the regolith production rates were substantially lower, suggesting that the profile is in a non-steady state. Our findings highlight the necessity of subdividing regolith profiles when applying the “gain and loss” model, and demonstrate the value of integrating U-series disequilibrium with in situ cosmogenic nuclides for assessing regolith evolution over long timescales. The evolution of regolith thickness, as a controlling factor of production rate, also has a significant impact on whether there is a coupling between the regolith production rate and the denudation rate.

At 23:59 (UTC + 8) on 18 December 2023, an earthquake of Ms 6.2 struck Jishishan County in Gansu Province, China, and triggered a large-scale, flow-like loess landslide in Zhongchuan Town, resulting in some 20 deaths. Originated from relatively gentle terrain, the loess flow displayed high mobility with a run-out distance of 3,200 m, suggesting that pore-water may play a critical role in the mobility of Zhongchuan flowslide. Following onsite investigations and soil sampling, we replicated the initiation process of the flowslide through dynamic back pressure direct shear tests under a constant shear stress condition. Two types of tests were conducted on saturated loess samples: elevated back pressure tests to simulate instability induced by high pore-water pressure, and dynamic loading tests to examine the evolution of pore-water pressure under seismic loading conditions. The experimental results, supported by microscopic analysis, indicate that elevated pore-water pressure is the key factor driving the progressive transformation of shear displacement from accelerated motion to instantaneous runaway. Meanwhile, dynamic loading substantially amplifies the generation of excess pore-water pressure. Moreover, the initial pore-water pressure was found to be a critical factor in both the initiation and high mobility of the Zhongchuan flowslide. These experiments quantitatively capture the in situ evolution of pore-water pressure throughout the liquefaction process, providing a physically based framework for understanding the mechanisms of loess landslides.

Mud dominates the particulate load of sediment and organic carbon from continents to oceans, but mud concentration and transport rate remain notoriously difficult to predict. In rivers, mud is thought to be transported as washload—particles so small that they are absent from the riverbed, washed through the river like passive tracers, and controlled by external inputs rather than local sediment entrainment from the bed. However, freshwater flocculation in rivers can aggregate mud grains into larger particles that behave hydrodynamically more like sand. If correct, this finding opens the door to describe mud transport as bed-material load—particles in dynamic interchange between the bed and water column—for which robust theory exists. Here we present evidence that mud behaves as flocculated bed-material load rather than washload in the freshwater Wax Lake Delta (WLD), a major distributary of the Mississippi River Delta. Grain size-specific concentration-depth profiles indicate that mud is flocculated in WLD. In situ turbidity sensors, airborne hyperspectral imaging (AVIRIS-NG), and concentration-depth profiles show that mud concentration varies temporally and spatially in response to shear stress variations, consistent with bed-material load dynamics. Furthermore, mud exists in the channel bed (median 14% mud by volume) and dominates the bed on deltaic islands (median 90%). Bed-material entrainment theory explains observed near-bed mud concentrations using a formulation that accounts for floc growth and densification near the bed. Together, these findings support a unified treatment of sand and flocculated mud as bed-material load in lowland rivers and deltas.

Human activities have a strong impact on global climate and natural ecosystems, yet the extent of their influence on long-term natural erosional processes remains poorly determined. A quantitative analysis is needed. The Ayeyarwady River, renowned for its large sediment flux ranking second in Asia, provides a compelling case study. We here show that extensive anthropogenic activities in the Ayeyarwady catchment have strongly accelerated erosion rates compared to natural benchmark levels, thereby contributing to its high sediment discharge. To highlight this point, we compared present-day erosion rates calculated from sediment fluxes with long-term natural erosion rates derived from detrital-apatite fission track (AFT) and cosmogenic ¹⁰Be data. Our findings reveal a stark contrast. Long-term natural erosion rates were notably higher in the Upper Ayeyarwady (0.06–0.34 mm/a) than in the Upper Chindwin (0.02 ± 0.005 mm/a), whereas present-day erosion rates are three times higher in the Upper Chindwin (0.63 ± 0.05 mm/a) than in the Upper Ayeyarwady (0.19 ± 0.02 mm/a). Particularly, noteworthy are the Upper Chindwin and Mu drainages, where erosion rates are calculated to have increased by more than an-order-of-magnitude relative to long-term natural background rates. Such a striking increase in erosion rate correlates positively with the spatial distribution of alluvial mining, especially for the Upper Chindwin catchment. The observed increases in sediment fluxes from long-term to present-day timescales may also be attributed to land-use expansion related deforestation, and intensified precipitation. These results underscore how human activities can drastically accelerate erosional processes, thus exerting a dramatic impact on natural systems.

Salt marshes are dynamic coastal ecosystems characterized by unstable edges. Root mats reinforce upper banks while exposing the underlying soil to erosion, leading to cantilever failure. This study presents a process-based model of marsh edge retreat that integrates collapse and root mat effects, validated with data from the Shangsha Wetland in the Yangtze Estuary. The model evaluates the impacts of root mat thickness, porewater seepage, and hydrodynamic erosion on marsh retreat dynamics. Results reveal that marsh collapse overhang angles increase with root mat thickness, leading to higher tensile failure along the failure plane and decreased shear failure. This results in root mat thickness having a nonlinear effect on the collapse width and frequency. There is an optimal root mat thickness that balances the benefits of edge reinforcement with the risk of forming overly large cantilevers, thereby minimizing collapse frequency and retreat distance. As root mat thickness increases, the contribution of edge collapse to retreat rises from 65% to 88% due to reduced bare soil exposure and limited hydraulic erosion. Low hydraulic conductivity improves edge stability by dampening stress–strain oscillations and reducing collapse frequency, contribution, and retreat distance. However, when the root mat thickness exceeds the optimal value, it increases the collapse width. Hydrodynamic forces, including alongshore currents, high tidal amplitudes, and wave action, elevate collapse frequency but underscore the significance of optimal root mat thickness in mitigating retreat. These findings suggest that targeted interventions, such as vegetation management to optimize root mat thickness and implementation of wave or drainage barriers, can reduce the rate of marsh retreat.

Quantifying sequences of events and materials involved in the growth and dispersal of post-wildfire debris flows across entire mountain catchments and piedmont fans is rarely possible. However, understanding these processes facilitates assessing future flow magnitudes and recurrence risk. This study analyzed the evolution of debris flows generated by a rainstorm following near-complete burning of vegetation in six mountain watersheds. The flows transported large volumes of boulders to downstream fans, devastating Montecito, California. With rainfall-runoff modeling, lidar, photogrammetry, and field surveys, we quantified the hydrological and sedimentological components of the debris flows as they evolved from hillslope runoff to boulder-rich fan deposits and ocean discharge. Runoff from burned soils drove larger amounts of rill erosion and slurry generation on shale hillslopes than on sandstones. Hillslope slurry mobilized ravel deposits and fine sediments stored in the channel network, mainly on shales. Channels draining sandstones mainly supplied the flows' boulder loads. One-quarter of the mountain-shed sediment escaped to the ocean, while all boulders settled on the fans. Flows confined to primary channels remained erosive across the fans except where channel gradients and dimensions decreased in response to fault-related topography and where bridges trapped boulders, intensifying in-channel and overbank deposition. Although the results derive from a single event, they illustrate how a sequence of processes and landscape conditions determine debris-flow evolution across catchments and fans. Identifying debris-flow components highlights useful measurement and modeling methods to improve prediction while highlighting current limits on understanding critical processes and transient antecedent conditions.

Glacial lake outburst floods (GLOFs) caused by mass movement into lakes are common disaster chains in High Mountain Asia (HMA). However, the volumes of potential avalanche sources and the associated overtopping flood processes remain inadequately understood, hindering GLOF hazard assessments. We developed a comprehensive framework to quantify mass movement volumes and simulate GLOF process chains by integrating remote sensing data with hydrological models. We applied our methodology to Jiongpu Co, the largest glacial lake in southeastern Tibet. First, analysis of optical images revealed lake expansion from 2000 to 2024. Second, we assessed the volume of potential glacier avalanche using three-dimensional glacier velocities from multi-track Synthetic Aperture Radar (SAR) images. The estimated volume is 1.8 ± 0.06 × 10⁸ m³. Third, deformation on the surrounding slopes was investigated based on the time-series InSAR method, revealing a potential landslide volume of 3.5 ± 0.2 × 10⁸ m³. Next, we retrieved overtopping volumes from the potential glacier avalanche and landslide, which are 1.94 ± 0.1 × 10⁷ m³ and 9.89 ± 0.6 × 10⁷ m³, respectively. Finally, we evaluated the GLOF process chain under these two scenarios using the HEC-RAS model. Our integrated approach enhances GLOF monitoring and modeling, offering applicability to other glacial lakes for risk assessment and mitigation.