Suspended sediment concentration (SSC) is a crucial indicator for aquatic ecosystems and reservoir management but is challenging to predict at large scales. This study seeks to test the feasibility of deep-network-based models to predict SSC at basin outlets given basin-averaged forcings and basin-physiographic attributes as inputs and extract insights by interpreting the spatially-varying model performances. We trained long short-term memory (LSTM) deep networks either separately for each of the 371 sites across the conterminous United States (local models), or on all the sites collectively (Whole-CONUS). The local and Whole-CONUS models presented median Nash-Sutcliffe Efficiency (NSE) values of 0.72 and 0.57, respectively, which are state-of-the-art results. However, this comparison disagrees with our previous “data synergy” conclusion for LSTM models and suggests there are still important yet unavailable sediment-related attributes. Both local and Whole-CONUS models tended to be more successful where SSC-streamflow correlations (R_s-q) were high - typically in the humid Eastern US - and with lower SSC. Low R_s-q basins were often found in the arid Southwest with higher SSC. The highly-nonlinear SSC-streamflow relationship is arguably due to heterogeneity in land cover and rainfall or limitations in sediment supply, suggesting these basins need to be simulated at higher spatial resolution. The local models mostly outperformed the Whole-CONUS one due to the latter lacking critical attributes, but the latter can be competitive in high-SSC regions with enough flow events. Moreover, the Whole-CONUS model also performed well for basins not included in the training dataset (median NSE=0.55), supporting large-scale modeling.

How to cite: Song, Y., Chaemchuen, P., Rahmani, F., Zhi, W., Li, L., Liu, X., Boyer, E., Bindas, T., Lawson, K., and Shen, C.: Deep learning insights into suspended sediment concentrations across the conterminous United States: Strengths and limitations, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-16821, https://doi.org/10.5194/egusphere-egu23-16821, 2023.

Evaluating the influence of earthquakes on erosion, landscape evolution and sediment-related hazards requires quantifying the volume and velocity of post-seismic sediment cascades. However, accurate estimates of post-earthquake sediment transfers remain rare. Following the 2016 M_W7.8 Kaikōura earthquake in New Zealand, the volume of post-seismic erosion was quantified directly by measuring the ground surface change between 4 lidar surveys captured in 2016, 2017, 2019 and 2021 using the multiscale model-to-model cloud comparison (M3C2) algorithm. The lidar surveys covered the 62 km² Hapuku and 66 km² Kowhai river catchments within the Seaward Kaikōura Range, representing the two catchments with the highest density of co-seismic landsliding.

The total co-seismic landslide source volume for the Hapuku Catchment was 30 ± 6 M m³,the catchment being dominated by a 17 M m³ rock avalanche which dammed the Hapuku River. In the 5 years after the earthquake a total of 10.60 ± 0.22 M m³ of sediment was post-seismically eroded (equivalent to ~26% of the co-seismic landslide debris volume when considering bulking of the landslide deposit). A total of 9.71 ± 0.23 M m³ of sediment was delivered to the riverbed resulting in considerable riverbed aggradation and 3.58 ± 0.28 M m³ was inferred to have been transported beyond the rangefront of the Seaward Kaikōura Range (equivalent to ~9% of the co-seismic landslide debris). The total co-seismic landslide source volume for the Kowhai Catchment was only 13 +4/-3 M m³. Over the 5 years 2.02 ± 0.10 M m³ of sediment was post-seismically eroded, equal to ~13% of the co-seismic landslide debris volume within the catchment. The volume delivered to the riverbed, 1.29 ± 0.10 M m³ and 0.85 ± 0.13 M m³ is presumed to have been transported beyond the rangefront (equivalent to ~5% of the co-seismic landslide debris).

From these volumes, the rates at which the co-seismic landslide sediment was eroded from hillslopes, delivered off-slope to channels and exported from the range front were calculated. When projected, these rates of sediment conveyance suggest the volume of co-seismically generated sediment is likely to be evacuated from the rangefront within or close to the recurrence interval for ground motions equivalent to the Kaikōura earthquake. The Hapuku and Kowhai river catchments being examples of where co-seismic landsliding counterbalanced uplift.

How to cite: Jones, K., Howarth, J., Massey, C., Sirguey, P., Lague, D., and Bernard, T.: Accurate quantification of sediment conveyance following the 2016 Kaikōura earthquake, New Zealand, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-4588, https://doi.org/10.5194/egusphere-egu23-4588, 2023.

Storage of sediment on floodplains delays downstream sediment delivery, increasing the timescale of catchment responses to forcing by tectonics, climate changes, and watershed sediment management practices.  Including floodplain storage in catchment sediment routing models, however, is challenging because the long timescales involved exceed the duration of stream gaging station and other observational data sources.  As a result, floodplain storage is typically ignored in catchment sediment modeling.  To quantify timescales of sediment storage for mid-Atlantic U.S. floodplains since the early Holocene, floodplain sediment thickness distributions are defined for three time periods by analyzing stratigraphic data: presettlement (deposited before 1750), legacy (deposited 1750-1950), and modern (deposited after 1950).  These data are used to calibrate a model that predicts the thickness, age, and storage time distributions of floodplain deposits through time.  The model uses empirical equations to estimate changes in flood magnitude and duration caused by changes in forest cover and urban development.  Simple hydraulic models predict the occurrence of overbank flow based on channel geometry (which changes through time as floodplains accrete) and the potential for backwater induced by nearby milldams during the 19^th Century. Overbank deposition during overbank flows is predicted based on sediment concentration, sediment settling velocity, and overbank flow duration.  Sediment erosion is predicted based on the age distribution of stored sediment and a power law function that specifies the exposure of sediment to erosion by age category, an approach that is similar to the StorAge Selection Functions often used in catchment hydrologic modeling.   The calibrated model, “tuned” to reproduce observed stratigraphic data, predicts monotonically increasing fluvial sediment concentrations from presettlement to modern time periods, and sediment budget components (input and output fluxes and rates of sedimentation and erosion) that also increase through time.  Predicted sediment residence times (mean age of stored sediment) vary from ~450 years in 1750 to ~300 years in 2017, and the model accurately reproduces the full age distribution (0 to > 5000 yr) of stored sediment documented by contemporary stratigraphic data.  This calibrated model can accurately represent floodplain storage for improved watershed scale sediment routing computations in the mid-Atlantic region, improving our ability to manage Chesapeake Bay restoration and other important watershed sediment management issues.

How to cite: Pizzuto, J.: Stratigraphic Data Calibrates Predictive Modeling of Holocene-Present Floodplain Sediment Storage In the Mid-Atlantic U.S., EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-7591, https://doi.org/10.5194/egusphere-egu23-7591, 2023.

Identifying sediments sources is an important branch of catchment erosion modeling that uses multiple tracers in a robust set of statistical analysis techniques commonly known as the “fingerprinting approach”. The techniques employed in the fingerprinting approach follow two distinct stages of multivariate statistical analysis: discrimination and classification. The first one refers to determining the best set of tracers that have the potential to be selected as a tracer. The second stage consists of classifying the eroded sediment samples in the n-dimensional space defined by the tracer properties. In this step, the relative contribution of each source to the composition of the suspended sediment is calculated. One of the challenges for improving the “fingerprinting approach” is estimating the uncertainties of the results. In this sense, defining the number of samples used to characterize sources and eroded sediments is considered an important issue in terms of costs and source of uncertainties. Therefore, the main objective of this work is to present an alternative modeling with a focus on uncertainty analysis and sample number optimization based on the model developed by Clarke and Minella (2016). The advantages of the proposed model include 1) the calculus of the source apportionments, making it possible to evaluate the effects of reducing the sample number on the uncertainties; 2) takes account the collinearity between the tracers adding the variance-covariance matrix applied into the generalized least squares (GLS) method; and 3)  adds the calculus of uncertainty associated with the number of samples (sediment sources and the  sediments. To demonstrate the usefulness of the model, we used a dataset available from the Arvorezinha experimental catchment located in southern Brazil. The implementation of this model was carried out in the Phyton®, so that any user can evaluate the uncertainties in the reduction of the number of samples as well as the importance of collinearity in the set of available tracers. The results confirmed the assumption the increased uncertainty as the number of samples decreases in the sources or eroded sediment samples. Moreover, the addition of the variance-covariance matrix in the solution of the overdetermined system allows to take into account the deleterious effects of collinearity in the fingerprinting approach. With this tool, new perspectives are opened to systematically improve the definition of the number of samples needed based on the uncertainty analysis of the set of samples available, fundamental to the advancement of research in the area of environmental monitoring and modeling, as well as for the management of water resources and soil management in agricultural catchments.

How to cite: Buligon, L., Buriol, T., and Minella, J.: An alternative approach to sediment source identification: uncertainty analysis and sample number optimization, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-8866, https://doi.org/10.5194/egusphere-egu23-8866, 2023.

Fine sediment plays an important role in the healthy functioning of river ecosystems providing nutrients and contributing to habitat functioning. However, excessive sediment supply into rivers has several detrimental impacts on water quality and it causes sedimentation in river channels, reservoirs and estuaries. In addition, silts and clays are geochemically active and consequently are responsible for the transport of contaminants, including trace metals, phosphorus, pesticides and radionuclides among others which have high sorptive affinity for fine-grained particles. Hence, quantifying the timescales of sediment transfer throughout a river system is critical for understanding river basin sediment dynamics and the fate of their associated pollutants.

Fallout radionuclides (⁷Be, ²¹⁰Pb_ex and ¹³⁷Cs) have been used to assess sediment travel distances, sediment age and sediment residence times in a variety of landscapes. An advantage of using these radionuclides as sediment chronometers is their half-lives which can be used to model sediment residence time from days to decades in different catchment compartments.

The River Avon (Devon, UK) is a 40 km long gravel-bed river, draining rough moorland and with a catchment area of 110 km². The mean annual flow is 3.7 m³ s^-1 and is moderated by managed discharges from a reservoir upstream. Suspended and channel bed sediments were sampled in a 5 km section of the river during four seasonal surveys (January, March, July and November 2022) and suspended sediments during a stormflow event were also sampled.

Radionuclide activity concentrations of channel deposited sediments varied substantially within and between river bars and seasonally. Suspended sediment activity concentrations varied within the stormflow hydrograph and seasonally. Relationships between radionuclide activity concentrations and sediment storage, particle size, total organic carbon and C:N ratios were also evaluated. Channel sediment residence times obtained using ⁷Be/²¹⁰Pb_ex activity ratios ranged between 0 to 110 days, reproducing the high variability found in activity concentrations. Future research will assess the influence of sediment sources on ⁷Be/²¹⁰Pb_ex ratios and the relationship between sediment storage dynamics and sediment-bound contaminants. Sediment residence time modelling will allow an improved understanding of sediment dynamics in gravel-bed rivers which is essential to inform management decisions and prediction of the timescales of transfer and fate of associated contaminants.

How to cite: Munoz-Arcos, E., Millward, G., Clason, C., Bravo-Linares, C., and Blake, W.: Variability of Fallout Radionuclides in River Channels: Implications for Sediment Residence Time Estimations, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15828, https://doi.org/10.5194/egusphere-egu23-15828, 2023.

Sediment source fingerprinting is a technique for determining proportional contributions from different catchment sources to sediments in downstream receiving environments. The technique involves a) selecting tracers that discriminate sources based on their biogeochemical or isotopic properties and b) applying statistical mixing models to quantitatively determine source contributions. Tracer suitability varies depending on the characteristics of the study catchment and the source property or erosion processes being targeted and can include geochemical, fallout radionuclides (FRNs), or compound specific stable isotopes (CSSIs). For instance, the spatial variation in soil geochemical properties is largely determined by underlying geological and pedogenic processes, whereas CSSIs utilise δ¹³C isotopic properties of fatty acid biomarkers that bind to soils and vary based on plant communities associated with each land cover.

While the environmental basis for sediment fingerprinting is increasingly understood, methodological challenges continue to present limitations that may hinder wider catchment applications. Here, we draw from recent research in New Zealand to highlight some of the challenges to source apportionment accuracy using numerical mixture testing and catchment studies to represent a range of tracers and sources. Tracers include bulk geochemistry, fallout radionuclides (FRNs), and compound specific stable isotopes (CSSIs) and sources are defined by parent material, erosion processes, and land cover. We focus on the influence of source dominance and source discrimination by different tracer types on source apportionment accuracy, as well as uncertainties introduced from post-unmixing transformations associated with CSSIs.  

How to cite: Vale, S. and Smith, H.: Factors influencing source apportionment accuracy using sediment fingerprinting: observations from New Zealand, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-10427, https://doi.org/10.5194/egusphere-egu23-10427, 2023.

Hydropeaking in rivers changes the flow regime, increases river clogging, mobilizes fine sediment, and causes major stress to fish, macroinvertebrates, and aquatic plants that suffer from the rapid water level fluctuations. One in four medium- to large-sized rivers in Switzerland is affected by hydropeaking. In this study, we investigated the effect of hydropeaking on fine sediment transport during an experimental flood on the Spöl river, a tributary of the Inn river, in the canton of Graubünden, Switzerland. The study was a proof-of-concept for new smart turbidity sensors, which were developed in our laboratory, calibrated, and tested in mixing tank experiments in 2021 and again in 2022 with a range of different sediment types. These sensors were deployed at two locations on the Spöl during an experimental flood release by the upstream Ova Spinne hydropower dam. The collected data reveal sudden sediment concentration increases and decreases (pulsing) as the discharge increases steadily throughout the day. The highest concentration of sediment is much larger (4-5 g/L) than would be expected and appeared with the onset of the flood and again with the peak discharge. Our findings also reveal clockwise and counter-clockwise hysteresis loops in the stage-concentration relation, which point to a switch in the sediment supply between supply limited and unlimited conditions during the experimental flood. This study shows that high spatial- and temporal-resolution monitoring of suspended sediment is possible with a low-cost sensor network. The applications of such a network are plentiful: from identifying sediment source activation and transport in small streams, glacier networks and deltas, to environmental monitoring of maximum sediment concentration levels for the survival of fry fish, for prevention of river bed clogging, and for pollutant monitoring (binding to sediments).

How to cite: Droujko, J. and Molnar, P.: Sediment pulse propagation and identification using a low-cost sensor network: a hydropeaking study on the Spöl river, Switzerland, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-14351, https://doi.org/10.5194/egusphere-egu23-14351, 2023.

We report the use of Luminescence sensitivity as a proxy to understand sediment dynamics in the Ganga River and its major tributaries, viz., the Yamuna, the Chambal and the Ramganga rivers in India.  The Ganga River is one of the world’s largest dispersal systems that originates in the Himalaya and travels across central India to meet the Bay of Bengal. The basin size, catchment lithology, climatic conditions and geomorphic processes of these large rivers are diverse. The rivers are classified into reaches based on varied morphometric characteristics. Sampling strategies focussed on point bars and mid-channel bars mostly from the low-gradient reaches of the rivers, and intervals such that the influences of local dimensions such as hillslope processes and smaller tributary confluences get integrated. Luminescence sensitivity (photon counts/unit dose/unit mass) of quartz grains of 90-150 µm size are examined after check on their purity.

The results suggest the following:

• A gradual change in luminescence sensitivity in the downstream direction.
• Change is slower at the beginning, then it increases to nearly twice the initial rates after the confluence with R. Ramganga suggestive of change in sediment flux and sediment transportation rates. In the upstream reaches of the river, influences of a landslide zone and the dun (intermontane valley) rivers are discernible.
• Rates of sensitivity change is nearly four times higher in the case of Yamuna River suggestive of longer transport times.
• Samples after the confluence of the Ganga and the Yamuna suggest variable contribution from the two rivers through time.
• Sensitivity of quartz suggests influence of tributary confluences on the change in luminescence sensitivity along the trunk rivers and offer prospect of developing it as an additional parameter to quantify river processes through time.

This project is supported through DST SERB-YoSCP grant.

How to cite: Parida, S., Kaushal, R. K., Chauhan, N., and Singhvi, A. K.: Luminescence Sensitivity as Proxy for Sediment Source and Transport – A Case Study from the Ganga River, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-1016, https://doi.org/10.5194/egusphere-egu23-1016, 2023.

At steady-state, sediment fluxes out of a drainage basin equal its average erosion rate. Quantifying relative sediment fluxes is therefore key in estimating spatial erosional variability among sub-basins and the consequential landscape evolution. Traditional approaches to quantify such fluxes in drainage basins include using the mineral and elemental compositions of sediments as markers for the relative contribution from sub-basins. Such an approach often fails to distinguish among bedrock sources, and have been shown to suffer from transport-related biases.

Here, we aim to test and explore the combination of these traditional approaches together with oxygen, carbon and ‘clumped’ isotope analyses of detrital carbonate as a novel combined proxy for relative sediment fluxes in carbonate-dominated drainage basins. We test this approach at the Hatrurim Syncline in southern Israel, east of the Dead Sea margins. The area comprises of marine carbonate rocks of the Judea Gr., as well as Hatrurim Fm. rocks that have experienced different grades of combustion-metamorphism, and thereby registered a wide range of isotope values together with distinctive carbonate mineral assemblages – allowing for using both ‘traditional’ and isotope-informed approaches. We collected bedrock and sediment samples from the Morag Basin in the Hatrurim Syncline, and analyzed their mineral and isotope compositions in bulk and specific grain-size fractions.

Our results show that: (a) Hatrurim Formation’s bedrock samples have a wide range of mineral and isotope values consistent with two main assemblages – high temperature metamorphic carbonates, and low temperature re-crystallized carbonates; and (b) Mineral and isotope compositions of fine grain sediment fractions (<2mm) show binary mixing between un-metamorphosed Judea Group and Low-T Hatrurim end-member sources. Coarser sediment fraction show deviations from a binary mixing, which we associate with contribution from a High-T Hatrurim third source.

Based on these analyses, we compiled a mixing model for fine grained sediments, aiming to identify the mineral and isotope compositions of end-member sources and to predict the mixing-ratio for each sediment sample. Model-predicted mixing ratios of sediment samples agree with mixing ratios estimated based on the relative exposure areas of the Judea Gr. and the low-T Hatrurim Fm. within the drainage area of each sediment sample. This consistency suggests that the Morag Basin is evolving under spatially uniform erosion conditions, in which sediment is being contributed equally from each area unit in the basin, and the overall landscape morphology is preserved over time.

A long-profile analysis of the Morag Basin channel network revealed several slope-break knickpoints, separating continuous channel sections with variable steepness indices. Accounting for our finding of a spatially uniform erosion rate, we interpret the knickpoints as reflecting transitions between different lithology-dependent rock erodibility rather than transient signals driven by tectonic or climatic perturbations. The Morag Basin thus presents a unique case where the morphology of the fluvial network has adjusted to erode the surface uniformly despite the multitude of rock types exposed in the basin.

How to cite: Hagbi, R., Goren, L., Eiler, J. M., and Ryb, U.: Oxygen, carbon, and clumped isotope compositions of detrital carbonates: A new combined proxy for quantifying relative sediment fluxes in carbonate terrains, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-1084, https://doi.org/10.5194/egusphere-egu23-1084, 2023.