Search

Large anabranching sand-bed rivers are characterised by dynamic lateral channel migration, bar aggradation and floodplain accretion. On the Amazon River, variations in bank and floodplain sediments exert a primary control on channel migration rates. In a reach near Tefé, Brazil, the Solimões River shows different migration dynamics along its north and south banks, suggesting that bank strength plays a role in the large-scale and long-term channel and floodplain evolution. Here we present measurements of lower bank strength and sediment resuspension in 29 locations along and across the Solimões River to investigate their spatial variability in sediments of different age and origin by means of a cohesion strength meter and a Pilcon Shear Vane. Results show that the north bank consists mainly of late Holocene sandy and silty deposits, whereas the south bank is characterised by frequent Pleistocene outcrops of cohesive muds and diagenetic iron cements and concretions. The south-bank Pleistocene deposits have on average three times higher bank strength than the younger floodplain deposits along the north bank. When comparing the locations of the Pleistocene deposits with lateral migration rates along both banks for two 40 km reaches, we observe that these sediments occur mainly where the south bank has been eroded, suggesting that they are revealed when the river migrates. Our results suggest: (1) that lateral migration uncovers the less erodible layers that can then deflect the flow towards the north bank; (2) that outcrops of resistant Pleistocene deposits might be abundant underneath the northern alluvial floodplain; and (3) spatial variations in bank erodibility exert a first-order local control on the planform morphology and lateral dynamics of the river. We suggest that such variations in erodibility are equally important for the morphodynamics of other large sand bed rivers that show evidence for the presence of resistant Pleistocene sediments, such as the Nile, Mekong and Mississippi Rivers.

Dunes are ubiquitous features in most sand- and gravel-bed rivers worldwide and are key elements of sediment transport. Their variable height may also interfere with shipping routes and help dictate shipping loads. Knowledge of dune dynamics and spatio-temporal sediment transport is thus essential in understanding river dynamics and for the navigability, sustainable management and maintenance of rivers, especially in times of more extreme floods. To date, most morphodynamic studies of river-beds have been based on either bathymetric time series or sub-bottom profiling data, but not collected at the same time and the sub-bottom data not in time series. As such, these data do not allow for the identification of spatio-temporal variations of sediment storage in, and reactivation of, the shallow sub-surface as related to dune kinematics. Our field study, reported here, sought to address this gap in knowledge by investigating the stratification produced by dunes in the shallow subsurface through sub-bottom profiler time series in combination with bathymetric time series and vibracores. In three areas of varying grain size in the River Waal, Netherlands, we collected four 7-km long tracks of high-resolution sub-bottom profiler data (Parametric Echo Sounder, PES) and, simultaneously, multibeam echo sounder (MBES) data. In two repeat surveys in areas 2 and 3, and four repeat surveys in area 1, data were acquired to gain insight into the preservation and reactivation of dune deposits over short-term periods of 1 day to 3 weeks. Interpretation of the sub-bottom data is aided using 18 vibracores of 4 – 5 m depth. Initial analyses show the migration and morphological change of the large dunes, thereby obliterating dunes mapped during the first survey, and the presence of superimposed small dunes. The PES data of large dunes exhibit foresets, reactivation surfaces where superimposed dunes migrated down the lee slopes, and strong near-horizontal reflectors at the base of large dunes, interpreted as the lower bounding surface. The surveys also identified dune stratification preserved below the active dune scour depth, and several horizontal reflectors at depth. Coupling these sedimentary structures in the bed profile data to both the simultaneous MBES data and a unique longer-term MBES time series, comprising two-weekly surveys (2005-2021) and half-yearly surveys (from 1999), provides an unparalleled opportunity to date these sedimentary structures, (1) to investigate longer-term aggradation and dune preservation and (2) to link these to flood and depositional events over the past decades. Here, we present initial results. This field dataset and approach yield a unique, high-resolution, spatio-temporal reconstruction of sediment preservation that significantly contributes to the insight into sediment storage times and preservation of dune-scale sedimentary structures in river beds. These field data also help to improve data-driven modelling.

Decimeter-scale early-stage aeolian bedforms represent topographic features that differ notably from their mature dune counterparts, with nascent forms exhibiting more gently sloping lee sides and a reverse asymmetry in their flow-parallel bed profile compared to mature dunes. Flow associated with the development of these “protodunes,” wherein the crest gradually shifts downstream towards its mature state, was investigated by studying the perturbation of the turbulent boundary layer over a succession of representative bedforms. Rigid, three-dimensional models were studied in a refractive-index-matched experimental flume that enabled near-surface quantification of mean velocities and Reynolds stresses using particle-image velocimetry in wall-normal and wall-parallel measurement planes. Data indicate strong, topographically induced flow perturbations over the protodunes, to a similar relative degree to that found over mature dunes, despite their low-angled slopes. The shape of the crest is found to be an important factor in the development of flow perturbations, and only in the case with the flattest crest was maximal speed-up of flow, and reduction in turbulent stresses, found to occur upstream of the crest. Investigation of the log-linearity of the boundary layer profile over the stoss sides showed that, although the profile is strongly perturbed, a log-linear region exists, but is shifted vertically. A streamwise trend in friction velocity is thus present, showing a behavior similar to the trends in mean velocity. Analysis of the growth of the internal boundary layer on the dune stoss sides, beginning at the toe region, reveals a similar development for all dune shapes, despite clear differences in mean velocity and turbulent stress perturbations in their toe regions. The data presented herein provide the first documentation of flow over morphologies broadly characteristic of subtle, low-angle, aeolian protodunes, and indicate key areas where further study is required to yield a more complete quantitative understanding of flow–form–transport couplings that govern their morphodynamics.

Recently, simultaneous PET/MR imaging systems have shown promise to provide simultaneous molecular and morphological evaluation of a variety of diseases [1]. However, challenges in MR-based attenuation correction (MRAC) methods still pose a problem for accurate quantification of PET images. Due to the difficulty of imaging and segmenting bone with MR sequences, all commercially available MRAC methods for simultaneous PET/MR ignore bone in the attenuation maps for whole-body attenuation correction, with the exception of brain imaging. Since bone has the highest attenuation coefficient in the body, ignoring bone will lead to underestimation of PET standardized uptake values (SUV) in and immediately adjacent to bone structures [2].Traditional machine learning algorithms have been used to train atlas database to predict attenuation coefficients for MRAC [3]. However, deep learning techniques with convolutional autoencoder networks have never been studied for MRAC purposes. The autoencoder network [4] turns out to be an appropriate structure for this goal. In this work, we study the use of a modified autoencoder network for automatic generation of MR-based attenuation maps.

This article has been withdrawn as it was published elsewhere and accidentally duplicated. The original article can be seen here: 10.1108/09685229710167962. When citing the article, please cite: Jim Best, Thornton May, (1997), “Is your infrastructure ticking?”, Information Management & Computer Security, Vol. 5 Iss: 1, pp. 3 - 6.