A distributed multi‐architecture design for massively parallel hyperbolic solvers is herein introduced and benchmarked. A unified object‐oriented central processing unit (CPU) + graphics processing unit (GPU) approach is complemented with an inter‐device communication layer, enabling both coarse and fine‐grain parallelism on hyperbolic solvers. The approach involves the combination of three different programming platforms, namely OpenMP, CUDA and MPI. The efficiency of this distributed‐heterogeneous approach is quantified under static and dynamic loads on consumer and professional grade CPUs and GPUs. An asynchronous communications scheme is implemented and described, showing very reduced overheads and a nearly linear scalability for multiple device combinations. For simulations (or systems) with non‐homogeneous workloads (or devices) the domain decomposition algorithm incorporates a low‐frequency load‐to‐device fitting function to ensure computational balance. A real‐world application to high‐resolution hydrodynamic modelling is presented: the propagation of a tsunami in the estuary a large river and its run‐up in an urban mesh. The proposed implementation shows speedups of up to two orders of magnitude, opening new perspectives for solvers with high‐demand requirements but relatively simple hardware in multi‐architecture machines.

Lake ecology can be affected by exchange flows driven by horizontal temperature gradients in lake–wetland interfaces. In this work, we investigate the hypothesis that thermally driven flows modulate the horizontal migration patterns of freshwater zooplankters. A 48-h field campaign in a shallow lake (Lake Vela, Quiaios, Portugal) was carried out to test this hypothesis. Thermal differences between the littoral and limnetic areas were measured along two transects featuring a Schoenoplectus lacustris and a Myriophyllum aquaticum stand in the littoral. In parallel, the physiochemistry and chlorophyll a, as a proxy for food availability differences between the littoral and the limnetic zones, were monitored. Zooplankton samples were collected for assessing overall and group-specific number-density differences. The diel period (day or night) and the site (littoral or limnetic zone) did not interact significantly to modulate the variation patterns for the studied physiochemical variables, indicating that these parameters should not explain horizontal zooplankton distribution patterns. The expected patterns for zooplankton diel horizontal migration as driven by the presence of visual predators were occasionally confirmed by our limnetic versus littoral abundance records through time, depending on the transect. Group-specific abundance patterns indicate particular features: copepods always preferred the littoral over the limnetic zone regardless of the diel period; chydorids always preferred the littoral zone regardless of the macrophyte stand involved; bosminids tended to preferentially concentrate in the limnetic zone. No consistent relationship was identified between the expected flow direction due to temperature differences and zooplankton abundance changes, although it occasionally occurred through the dataset.

We present novel velocimetry algorithms based on the hybridization of correlation-based Particle Image Velocimetry (PIV) and a combination of Lucas–Kanade and Liu–Shen optical flow (OpF) methods. An efficient Aparapi/OpenCL implementation of those methods is also provided in the accompanying open-source QuickLabPIV-ng tool enabled with a Graphical User Interface (GUI). Two different options of hybridization were developed and tested: OpF as a last step, after correlation-based PIV, and OpF as a substitute for sub-pixel interpolation. Hybridization increases the spatial resolution of PIV, enabling the characterization of small turbulent scales and the computation of key turbulence parameters such as the rate of dissipation of turbulent kinetic energy. The method was evaluated using both synthetic and real databases, representing flows that exhibit a variety of locally isotropic homogeneous turbulent scales. The proposed hybrid PIV-OpF results in a 3-fold increase in the PIV density for synthetic images. The analysis of power spectral density functions and auto-correlation demonstrated the impact of PIV image quality on the accuracy of the method and its ability to extend the turbulence range. We discuss the challenges posed by optical noise and tracer density in the quality of the vector map density.

The aim of the present work is to investigate, numerically and experimentally, the interaction of a density current with an array of in-line rigid and emergent cylinders. The laboratory tests were carried out employing the lock-exchange technique. The dense fluid was obtained by a mixture of fresh water, salt and rhodamine while the ambient fluid was obtained by a solution of clear denatured ethanol and freshwater. The numerical modelling was performed using a LES OpenFOAM (Open Field Operation and Manipulation). The results show that the current restructures macroscopically, gaining potential energy, when arriving at the vicinity of the array. The entrainment within the array was increasing at a smaller rate compared to the case of an unobstructed current, indicating that the drag promoted by the array does not contribute to the increase in the entrainment.

We investigate the interaction between lock release gravity currents and a vertical emergent cylinder. Two-dimensional instantaneous velocity fields were measured in vertical planes with Particle Image Velocimetry and mean velocity and Reynolds stresses are analyzed. The presence of the obstacle produces a deceleration in the longitudinal direction associated to the adverse pressure gradient. It was found that the cylinder does not change appreciably the structure of the mean flow. Turbulence is more expressively affected. Reynolds stresses are suppressed in the inner part of the head as the current approaches the cylinder. Third order moments are affected in a way that suggests that turbulence produced at the shear layer is not effectively fed back into the current. This is compatible with pressure diffusion counter-acting the effect of turbulent diffusion. This effect of the adverse pressure gradient on the current head has not been reported before.

Flood events are becoming more severe, causing significant problems to human communities, including physical, psychological, and material damage. For both flood forecasting in emergency response situations and flood mapping, georeferencing and data curation are paramount in the context of prevention or preparedness. Hence, data display, data management, and articulation with numerical simulation results must occur on GIS platforms. Our research is motivated by recent advances in Web and GIS technologies, social sensing and high-performance computing, and an envisaged wider availability of remote sensing data. This paper presents and discusses an innovative Web GIS platform named “RiverCure Portal” or “RCP” for short. This platform combines observations and hydrodynamic modelling tools to support various stages of the flood risk management cycle, including operational response, emergency preparedness, and risk assessment. RCP is a multi-organization, multi-context digital platform with flexible configuration features to define and support multiple sensor types and modelling options, satisfying the various needs of different organizations and stakeholders. In addition, this paper discusses the RiverCure Approach, which encompasses the following tasks directly supported by the RCP platform: defining the context and involved geometries, associating sensors to the context, pre-processing and generating the context mesh, defining the simulation event, running the simulation event, and analyzing the results from the simulation event. Thus, the RCP streamlines and simplifies data analysis and simulation procedures to meet decision-makers’ needs. The novelties discussed in this paper include the design and discussion of a Web GIS platform that allows (i) to manage flood data and results of simulations at several contextual levels by different stakeholders such as domain experts, decision-makers, researchers, or students; (ii) to process and curate sensed data obtained from physical and social sensors; and (iii) update the state and values of the parameters of simulation tools through continuous data assimilation techniques for forecasting purposes. Finally, this paper supports the explanation and discussion with a running example, “Águeda 2016 flood” event, which dataset is publicly available for further study and experimentation.

<p><strong></strong>: Gravity currents are flows generated by density differences within two contacting fluids. In this work the interaction between lock-release gravity currents propagating over a horizontal rectangular channel and an emergent cylinder is analyzed through velocity measurements obtained through PIV. Two-dimensional instantaneous velocity fields are measured in a plane perpendicular to the bottom along the center axis of the channel upstream of the obstacle. The experiments were also conducted without the cylinder for comparison purposes and ten repetitions were carried out for each configuration. The analyses focus on the effects that the presence of an adverse pressure gradient has on both the mean velocity field and the turbulence of the leading part of the current, the head, before the impact. The mean velocity field is not affected by the presence of the obstacle and since no differences were found in the spatial distribution of the mean velocity components, the necessary cylinder-induced deceleration occurs uniformly. Turbulence is studied through the components of the Reynolds stress tensor and their fluxes within the head. In the configuration with the cylinder, there are no fluxes of Reynolds stresses in the inner part of the section. Consequently, the Reynolds stress intensity decreases inside the head compared to the configuration without the obstacle. In conclusion, the presence of an adverse pressure gradient stops the mechanism of Reynolds stress distribution from the main source of production, i.e. the front region, to the inner region of the flow. This leads to a decrease in Reynolds stresses in the inner part of the head and an increase in the frontal region.</p> <p><strong>Acknowledgements</strong>: This work was partially supported by Foundation for Science and Technology's through funding UIDB/04625/2020 (CERIS research unit).</p> <p><strong>Keywords: </strong>Gravity currents, lock release, Particle Image Velocimetry, adverse pressure gradient, Reynolds stress.</p>

We address the capability of large eddy simulation (LES) to predict the physics of density currents interacting with bluff obstacles. Most density currents of interest in engineering and geophysical applications interact with obstacles or topographic features. Validating LES solutions in these contexts is crucial to establish it as a trusted tool. We thus propose a validation effort based on simple geometries that nonetheless pose challenges common to more complex systems, including boundary layer separation and convective instabilities. We focus on lock-exchange gravity currents in the slumping phase interacting with an emergent vertical circular cylinder. Our main investment was in ensuring that the comparison of experimental data and numerical results include, at least, the velocity and the density fields , and derived quantities (e.g., second order moments). Measurements of both density and velocity fields were performed in the side and plan views for cylinder Reynolds numbers, $$Re_d$$ , in the range 1300 to 3475. It was found that the LES accurately predicts the temporal evolution of the current front position. The computed front velocity exhibits a maximum relative error less than 8%. A good agreement between the LES and the experimental size and shape of the current head, and billows was found. The overall features upstream the cylinder, including a reflected wave, adverse pressure gradient and backflow, and downstream the cylinder, including the backflow, wake and the formation of a new head are well reproduced by LES. The agreement between the LES and the experimental time-space evolution of current spanwise- and depth-averaged density contours and the instantaneous velocity fields are not affected by $$Re_d$$ .

Optical Flow (OpF) approaches are quite diverse. In Fluid Mechanics, OpF methods are frequently applied to images that have been acquired for Particle Image Velocimetry, with laser or led illumination and seeded flows. In these cases, it is important to know which OpF methods are more adequate for which combinations of flow types, imaging conditions and image characteristics. This paper addresses this practical need by assessing the performance of three OpF methods, Lucas and Kanade (1981), Horn and Schunck (1981) and Farneback (2003), combined with the Liu and Shen (2008) algorithm. We evaluate the accuracy of different OpF methods as the difference between their values and of ground-truth. Hydrodynamic conditions include deformation dominated, rotation-dominated and uniform flows. For each flow type, relative and absolute errors are computed for different tracer displacements, noise powers, pixel particle sizes, image bit-depths and particle concentrations.The accuracy of the OpF methods is mainly affected by the magnitude of velocity gradients and convective accelerations. The inner region of the Poiseuille flow and the saddle point in the Rankine vortex combined with uniform flow pose significant difficulties to all methods. The accuracy of the Lucas-Kanade/Liu-Shen combination is high for all flow types, image conditions image bit depths. The Farneback/Liu-Shen combination has a similar high performance but only for images of 10 bit or higher. The Horn-Schunck algorithm is the worst performing method, due to high sensitiveness to particle concentration variations or particle sizes.

Wildfires are responsible for feeding ash to watercourses that are downstream of the burned area. The presence of riparian vegetation promotes the retention of ash from wildfires during the first rainfalls. Thus, the study of the transport of wildfire ash suspended in flows in rivers with vegetation is relevant for the management of watersheds and the conservation of biodiversity. In this scope, the present study focuses on the vertical profiles and turbulence intensities in the free region of a flow with suspended ashes in a partially vegetated channel. An experimental campaign was carried out in the Laboratory of Fluvial Hydraulics and Structures of Universidade da Beira Interior. The emergent vegetation stems were inserted on the right side of the channel in an extension of 6.00 m long and 0.30 m wide. The matrix was made with rigid cylinder tubes with two areal number densities, 500 stems/m(2) and 1034 stems/m(2), in a staggered array. The ashes adopted has a specific gravity of both the whole ash (Gs=2.64) and its fine fraction (Gs=2.54). The present work allows to conclude that: the streamwise velocities are higher at the interface for the sparser matrix and, moving away from the vegetation, the differences between the two vegetation areal number densities becomes increasingly smaller, existing an intermediate region where both areal number densities' velocities overlap, and, in the outermost region of the channel, velocities are higher for the denser matrix; the spanwise velocities present similar behavior for both matrices; the turbulent intensity is not affected by matrix areal number density.

Gravity currents propagating over and within porous layers occurs in natural environments and in industrial processes. The particular modes by which the dense fluid flows into the porous layer is a subject that is not sufficiently understood. To overcome this research gap, we conducted laboratory experiments aimed at describing experimentally the dynamics of the drainage flow. The experiments were conducted in a horizontal channel with a rectangular cross-section. The channel is 3.0 m long, 0.05 m wide. The porous bottom was composed of 5 cm and 10 cm layers of 3 mm borosilicate spheres – unimodal bed – and of a mixture of 3 mm (50% in weight) and 5 mm spheres (50%) – bi-modal bed. The porosity of the unimodal bed ranged between 0.60 and 0.64 (compatible with loose packing). The porosity of the bi-modal bed ranged between 0.61 and 0.65. All gravity currents were generated by releasing suddenly denser fluid locked by a thin vertical barrier placed at 0.2 m from the channel end. The dense fluid consists in a mixture of freshwater and salt (coloured with Rhodamine) while the ambient fluid is a solution of freshwater and ethanol. The density difference between the ambient fluid and the current, and the need to maintain the same refractive index, determine the amount of salt and alcohol added in each mixture. Here we report the findings of currents with a reduced gravity of 0.06 ms-2. Each experiment was recorded by an high-speed camera with a frame-rate of 386 Hz and a resolution of 2320 x 1726 pxxpx. Measurements were based on light absorption techniques: a LED light panel 0.3 m high and 0.61 m long was used as back illumination. All images were calibrated to ascribe, pixel by pixel, a concentration value from a 8 bit gray level. Different calibrations were performed for the porous layer and for the surface current. Results show that, in the slumping phase, the gravity current flows with velocities compatible with those over rough beds. As the current progresses further attenuation of momentum is noticed owing to mass loss to the porous bed. The flow in the porous bed reveals plume instability akin to a Saffman-Taylor instability. The growth of the plumes seems independent from the initial fluid height in both types of porous beds. The wavelength and the growth rate of the plumes depends on the bed material. Plumes grow faster in the case of the bi-modal bed and the wavelength of the bi-modal bed is about 1.5 as that of the unimodal bed. It is hypothesised that the gravity-induced porous flow is best parameterized by a Péclet number defined as a ratio of dispersive (mechanical diffusion) and advective modes of transport. Smaller wavelengths and slower growths are attained for stronger dispersion, characterisitic of the unimodal bed. For bimodal beds, permeability is larger, and thus also advection. This causes the flow to concentrate in faster growing but farther apart plumes. This research was funded by national funds through Portuguese Foundation for Science and Technology (FCT) project PTDC/CTA-OHR/30561/2017 (WinTherface).

We present a benchmark study of Optical Flow (OpF) methods for fluid mechanics applications. It is aimed at assessing the performance of three OpF methods, Lucas and Kanade (in: Proceedings of the 7th 1519 international joint conference on artificial intelligence, 1981), Horn and Schunck (AI 17:185–203, 1981) and Farnebäck (Two-frame motion estimation based on polynomial expansion, in: Bigun, Gustavsson (eds) Image analysis, Springer, 2003), combined or not with the Liu and Shen (JFM 614:253–291, 2008) algorithm. The performance of the OpF methods, evaluated exclusively as the difference between the values of the methods and of ground-truth (reference values), is benchmarked for deformation dominated, rotation-dominated and uniform flows. For each flow type, relative and absolute errors are computed for different tracer displacements, noise levels, pixel particle sizes, image bit-depths and particle concentrations. The accuracy of the OpF methods seems mainly affected by the magnitude of velocity gradients and convective accelerations. It does not seem to be affected by the relative preponderance of rotational and deformation components. The inner region of the Poiseuille flow and the saddle point in the Rankine vortex combined with uniform flow pose significant difficulties to all methods. The performance of the Lucas-Kanade/Liu-Shen combination is the best for all flow types, image conditions and image bit depths. The Farnebäck/Liu-Shen combination has a similar high performance but only for image depths of 10 bit or higher. These OpF methods maintain a high performance for tracer sizes and concentrations outside the PIV optimal range. Horn-Schunck is the worst performing method, due to high sensitiveness to particle concentration variations or particle sizes. These results can be used to plan new particle-based velocimetry experiments or to retrieve further information from existing PIV databases.

Multiple-cylinder configurations present complex flow fields generated by interactions between shear layers, vortexes and wakes. The present work is aimed at the characterization of vortex interaction within arrays of randomly placed cylinders through detection of Lagrangian Coherent Structures (LCS) in a numerical database from 2D Smoothed Particle Hydrodynamics. LCS are bounded by ridges in the field of the finite-time Lyapunov exponents (FTLE), which characterizes the rate of separation of neighboring trajectories over a fi-nite time interval. The results show that attracting LCS demonstrated ability to detect vortexes paths and that repelling LCS allowed the identification of the locus of the flow stagnation in the stoss side of each cylinder, which was found to vary in time. Insights on the mechanism of generation of background turbulence are pre-sented. Concerning time-averaged flow, it is proposed that the density of unstable manifolds detected within the array can be seen as a measure of the flow spatial complexity, and therefore correlated with spatial variance.