James Brasington

Subsistence farming communities are dependent on the landscape to provide the resource base upon which their societies can be built. A key component of this is the role of climate and the feedback between rainfall, crop growth, land clearance and their coupling to the ...

ABSTRACT The estimation of fluvial sediment transport rate from measurements of morphological change has received growing recent interest in the last five years. The revival of the 'morphological method' reflects continuing concern over traditional methods of rate determination but also the availability of new survey methods capable of high-precision, high-resolution topographic monitoring. In particular, remote sensing of river channels through aerial digital photogrammetry is a potentially attractive alternative to labor intensive ground surveys. However, while photogrammetry presents the opportunity to acquire survey data over large areas, data precision and accuracy, particularly in the vertical dimension is lower than traditional ground survey methods. This paper presents results of recent research in which DEMs have been developed for a reach of a large braided gravel bed river in Scotland using both digital photogrammetry and high resolution RTK GPS ground surveys. For both approaches, a statistical level of detection of change is assessed by intercomparing surfaces with independent check points. The sensitivity of the annual channel sediment budget to this level of detection is presented. Preliminary results suggest that as much as 60% of channel deposition and 30% of erosion may be obscured by the lower level of precision associated with photogrammetric monitoring.

ABSTRACT Subsistence farming communities are dependent on the landscape to provide the resource base upon which their societies can be built. A key component of this is the role of climate, and the feedback between rainfall, crop growth and land clearance, and their coupling to the hydrological cycle. Temporal fluctuations in rainfall on timescales from annual through to decadal and longer, and the associated changes in in the spatial distribution of water availability mediated by the soil-type, slope and landcover determine the locations within the landscape that can support agriculture, and control sustainability of farming practices. We seek to make an integrated modelling system to represent land use change by coupling an agent based model of subsistence farming, and the associated exploitation of natural resources, to a realistic representation of the hydrology at the catchment scale, using TOPMODEL to map the spatial distribution of crop water stress for given time-series of rainfall. In this way we can, for example, investigate how demographic changes and associated removal of forest cover influence the possibilities for field locations within the catchment, through changes in ground water availability. The framework for this modelling exercise will be presented and preliminary results from this system will be discussed.

... Various possible experimental procedures may account for this apparent blunder. ... PHOTOGRAMMETRIC ANALYSIS OF DRAINAGE BASIN EVOLUTION 241 ... were initiated on surfaces created with a central valley to minimize the influ-ence of local surface roughness, the ...

The recognition that fluvial dynamics and process rates may be robustly inferred indirectly from the changing pattern of channel morphology now puts developments in topographic monitoring at the focal point of current research in geomorphology. A range of survey technologies can potentially contribute to this field of research, each applicable at different spatial scales. At the mesoscale, high-resolution, precision topographic

Significant recent advances in floodplain inundation modelling have been achieved by directly coupling 1d channel hydraulic models with a raster storage cell approximation for floodplain flows. The strengths of this reduced-complexity model structure derive from its explicit dependence on a digital elevation model (DEM) to parameterize flows through riparian areas, providing a computationally efficient algorithm to model heterogeneous floodplains. Previous

Significant advances in flood management have been achieved through the use of a new generation of 2d flood inundation numerical models. These offer the potential to predict the local pattern and timing of flood depth and velocity, enabling informed flood risk zoning and improved emergency planning. With the availability of high resolution DEMs derived from ALSM, these models could theoretically now be routinely parameterised to incorporate the topographic complexity of urban areas, offering the potential to account for flow patterns at the scale of individual buildings. Currently, however, computational constraints on conventional finite element codes typically require model discretization at scales well below those achievable with ALSM and are thus unable to make optimal use this new data stream. In this paper we review two strategies that attempt to address this mismatch between model and data resolution in an effort to improve urban flood forecasts. The first of these strives for a solution by simplifying the mathematical formulation of the numerical model by using a computationally efficient 2d raster storage cell approach coupled to a 1d channel model. By contrast, the second method employs a morphological analysis of ALSM data to determine a sub-grid scale model of topographic complexity. This is then applied in coarse resolution simulations to represent the first order effect of the loss of available storage due to sub-grid scale topographic features. These two strategies are evaluated through numerical experiments simulating the pattern of flooding in the city of Cambridge UK, following a major storm in October 2001. For this, ALSM data were acquired with an Optech ALTM 3033 with a point spacing of 0.25m and simulations are evaluated through comparison with an observed flood envelope digitised from photography flown at the flood peak. Results from both approaches are encouraging, with inundation forecasts using the reduced complexity model comparing well with both the observed flood patterns and a 2d CFD simulation yet offering significant gains in runtime efficiency. The sub-grid parameterization model is also shown to improve low-resolution simulations without contributing significant computational complexity, but cannot match the predictive performance obtained by explicitly incorporating urban features in the topographic boundary condition.

A terrestrial laser scanner (Leica ScanStation) has been used to capture high-resolution topography of the River Feshie, Scotland, a mobile gravel braided system. This technology represents a new opportunity in the reproduction of grain-scale geometry, which has important geomorphological and ecological potential. This work outlines a methodology for recording and processing TLS data streams and discusses approaches for managing scan artefacts and modelling surfaces. Scans were taken at two spatial scales. First, patches 1 m2 were scanned with a point spacing of approximately 0.002 m. Repeat scans were taken from multiple positions to ensure complete coverage and were registered together. At the second scale, lower resolution scans with point spacings of between 0.005- 0.015 m were acquired to cover a reach 800 m long. The data were processed to remove artefacts such as mixed pixels and those produced by range walk errors and multipathing. The strategy employed to remove these anomalous points was based on analysis of the variance of replicate scans, using a threshold to remove points with a high level of variance. Control experiments of simple geometric shapes were used to identify the level of variance associated with anomalous points. 2.5D surfaces were produced from the data. A cone-shaped filter was used to remove points on the underside of grains. Simple smoothing was then applied to the surfaces to remove the underlying random noise in the data. The expected magnitude of this noise was calculated from the control experiments. The surfaces were then used to measure various parameters of the gravel. These include surface roughness, porosity and properties relating to grain entrainment.

ABSTRACT In the last decade, advances in topographic survey and digital elevation modelling have enabled a revolution in the study of river morphology and fluvial processes. Prior to the late 1990s, the analysis of channel form and dynamics was typically restricted to inferences from cross-sections and planform mapping. Since then, advances in survey instrumentation (e.g., Brasington et al., 2000) and remote sensing (Westaway et al., 2000; Brasington et al., 2003; Charlton et al., 2003) have enabled the collection of dense topographic data and the opportunity to build high-resolution, precision DEMs, ideally suited to morphometric analysis, sediment budgeting and as boundary models for fluid simulations. Continuing developments in survey technology are now poised to reset the parameters of this field once again through the recent emergence of ruggedized, terrestrial laser scanners. Based on time-of-flight or phase-based laser ranging, these instruments are capable of acquiring unprecedented volumes of survey-grade observations at operating frequencies of between 5-500 kHz and over ranges 25-1000 m. This new technology offers the potential to acquire rapidly, reach-scale datasets which record topographic information at the resolution of bed grain-scale upwards. This hitherto unprecedented data-stream presents new opportunities for river science, but also creates significant challenges particularly associated with: data management; regularization of resolution; visualization; and data assimilation with parallel models and data- products. In this paper we present a new methodology designed to analyze large 3d point clouds generated by terrestrial laser scanning. Specifically, the approach generates multi-resolution gridded terrain products from scan data whilst retaining the sub-grid scale information as key statistical attributes. We apply the method to a 1 km reach of the River Feshie, Scotland which was scanned in 2007 and evaluate the results through a comparison with independently acquired, spatially dense, GPS surveys of the study reach. The results reveal significant differences in the topographic signatures recorded by the two methods and reveal the value of the enhanced spatial resolution for representing complex morphologies and highlight the potential to retrieve grain-scale sorting patterns from the statistical attributes of the TLS data. Keywords: Terrestrial Laser Scanning, DEM, multi-resolution models, grain-scale, reach-scale, River Feshie

ABSTRACT In the last decade, advances in topographic survey and digital elevation modelling have enabled a revolution in the study of fluvial morphodynamics. Despite this recent progress, our understanding of braided river dynamics remains limited by the time-space scale of studies. Hindered by high labour and flight costs, together with slow ground-based survey methods, studies to date have focused either on event-scale dynamics of morphological units (Ferguson and Ashworth, 1992; Lane et al., 1995; Milan et al., 2007) or seasonal-annual dynamics of larger system-scale reaches (sensu Lane, 2006; e.g., Brasington et al., 2003; Lane et al., 2003). Terrestrial Laser Scanning technology offers the potential to acquire rapidly, reach-scale datasets which record topographic information at the resolution of bed grain-scale upwards. However, as yet, no detailed 3d datasets exist that reveal the system-scale evolution of a braided river through a continuous sequence of floods. Such data are urgently required to address unresolved and fundamental questions concerning the controls and behaviour of braided rivers and are also needed to validate morphodynamic simulation models (Brasington and Richards, 2007). Our recent wok has demonstrated that TLS can be applied to recover centimetre-scale channel morphology, maps of particle size, sorting, packing and floodplain roughness (Brasington et al., 2007, 2008; Antonarakis, 2008a,b; Hodge et al., in review). This potential is illustrated by the results obtained in a field study conducted in January 2008. This used TLS to monitor the evolution of channel morphology and develop methods to derive models of bed roughness and facies in a small 500 x 300 m reach of the actively braided Rees River, New Zealand. Fieldwork comprised repeat surveys before and after 3 competent events, combining laser scans from eight positions with bathymetric data obtained by RTK GPS. The resulting point clouds incorporated between 48-110 million survey points, with mean densities 1000 pts m-2. Algorithms were developed to process these large datasets (5 Gb per survey) and derive statistics of bed elevations from which multi-resolution DEMs, together with estimates of grain-size and surface roughness were extracted. Differencing sequential DEMs was used to quantify the volumes of erosion and sedimentation revealing the sediment budget and connected sediment transfer pathways. Keywords: Terrestrial Laser Scanning, DEM, bed roughness, grain-size, multi-resolution models, River Rees

Dams cut the continuity of sediment and water transfer worldwide. Magnitude and frequency of competent events are reduced while important percentages of the sediment load of regulated rivers are trapped in reservoirs. These alterations cause morphological changes on downstream reaches and coastline ecosystems and may create important effects on river's habitat. The analyses of such alterations are relevant for water and sediment management purposes in regulated rivers and, in the case of the European Union, may inform actions to restore geomorphic integrity of fluvial systems under the Water Framework Directive. In this work we present the preliminary results of a research project with the aim to study long-term morphological alterations in four regulated rivers under different impounded configuration in Wales, United Kingdom. Impoundment configuration refers to the number and relative position of dams along a stream course of a channel network. The project involve (i) state-of-the-art high resolution topographic surveys upstream and downstream from dams acquired by means of RTK-GPS and Terrestrial Laser Scanning, (ii) determination of the morphological changes downstream from dams since they were commissioned using historical maps and aerial photographs, (iii) ground-based characterization of surface and subsurface bed material, (iv) hydrological modelling to asses the effects of dams on flow regimes and flood magnitude and frequency and (v) hydraulic modelling to study bed stability downstream from the dams.

ABSTRACT The estimation of fluvial sediment transport rate from measurements of morphological change has received growing recent interest in the last five years. The revival of the 'morphological method' reflects continuing concern over traditional methods of rate determination but also the availability of new survey methods capable of high-precision, high-resolution topographic monitoring. In particular, remote sensing of river channels through aerial digital photogrammetry is a potentially attractive alternative to labor intensive ground surveys. However, while photogrammetry presents the opportunity to acquire survey data over large areas, data precision and accuracy, particularly in the vertical dimension is lower than traditional ground survey methods. This paper presents results of recent research in which DEMs have been developed for a reach of a large braided gravel bed river in Scotland using both digital photogrammetry and high resolution RTK GPS ground surveys. For both approaches, a statistical level of detection of change is assessed by intercomparing surfaces with independent check points. The sensitivity of the annual channel sediment budget to this level of detection is presented. Preliminary results suggest that as much as 60% of channel deposition and 30% of erosion may be obscured by the lower level of precision associated with photogrammetric monitoring.

A digital particle tracking algorithm designed to monitor small (10−4–10−2 m) movements of gravel clasts from digital video imagery is presented. Using the image brightness values as a set of weights, the algorithm identifies the sub-pixel co-ordinates of black circular transfer markers applied to contrasting tracer clasts. Once identified, marker centroids are tracked from frame to frame at a sampling

ABSTRACT A video analysis method for monitoring sediment transport and sorting processes in a laboratory flume is presented. Video taken through the glass side-wall of a laboratory flume is captured using a digital CCD (charge-coupled device) camera and significant movements between individual frames are detected using image analysis. This method involves direct subtraction of the brightness numbers of pixels in sequential video frames, followed by thresholding to produce binarized images of significant change, above the inherent level of system noise. Experimental results showing dilation of a gravel framework and rapid infiltration of fines just prior to entrainment are discussed. Copyright © 2000 John Wiley & Sons, Ltd.

The availability of new geomatics technologies and methods has transformed the investigation of sediment transport rates using the morphological approach. Terrestrial laser scanning (TLS), in particular, has transformative potential for mapping river change in braided rivers since much of the bed is sub-aerially exposed at low flows. TLS offers the opportunity to collect high-precision and high-accuracy data over large spatial