Search

The accuracy and reliability of 3D steady RANS CFD simulations of wind flow in urban environments can be affected by numerical settings including the turbulence model and the imposed roughness heights. In that regard, various k-ε and k-ω turbulence models and roughness height (ks) values are commonly used when predicting wind flow in urban environments. However, it is insufficiently known to which extent the CFD results may be influenced by these settings when simulating wind flows in complex urban environments with large changes in surface roughness. This is the scope of the present paper, for which wind-tunnel (WT) measurements and CFD simulations were performed on a reduced-scale model (1:300) of a district of Livorno (Italy). Mean wind speed (U), turbulent kinetic energy (k) and turbulence dissipation rate (ε) profiles from WT measurements and CFD simulations were compared at 25 positions and deviations between experimental and numerical results were quantified by three metrics: fractional bias, correlation coefficient and fraction of data within a factor of 1.3. The turbulence model selection had a larger impact compared to the surface roughness selection on U, k and ε values. The best and worst performing turbulence models (e.g. for α = 240° at 0.02 m above the bottom) showed a deviation in terms of correlation (0.89 and 0.61, respectively) of about 0.28. Conversely, the best and worst performing roughness set, (e.g. for α = 240° at 0.02 m above the bottom), showed a deviation in terms of correlation (0.77 and 0.78, respectively) of only 0.01.

The surface-averaged forced Convective Heat Transfer Coefficient (CHTCavg) at a windward building facade is influenced by the complex interaction between a wide range of parameters. Existing CHTC expressions, however, consider the impact of these parameters either incompletely or not at all. Earlier studies have shown that this shortcoming can lead to significant errors in Building Energy Simulations. In this paper, therefore, the combined impacts of wind speed (U10), building height (H) and width (W), and wind direction (θ) on the CHTCavg for the windward facade of buildings are systematically investigated. High-resolution CFD simulations of wind flow and forced convective heat transfer, validated with wind-tunnel measurements, are performed for 64 building geometries (10 m ≤ H and W ≤ 80 m), 8 wind directions (0° ≤ θ ≤ 78.75°) and 4 reference wind speeds (1 m/s ≤ U10 ≤ 4 m/s). The 3D steady RANS equations with the realizable k-ε turbulence model and the low-Re number Wolfshtein model are used. The results show that for a given building geometry and U10, the CHTCavg decreases as θ increases from 0° to 78.75°. The maximum reduction of about 42% occurs for the building with H = 8W = 80 m. In addition, for a given θ and U10, by increasing H, the CHTCavg increases, while increasing W has the opposite impact on the CHTCavg. Finally, a new generalized CHTC expression is presented as a function of U10, H, W and θ and its accuracy is confirmed by detailed in-sample and out-of-sample evaluations.

The importance of vertical wind velocities (in particular positive vertical wind velocities or updrafts) in atmospheric science has motivated the need to deploy multi-hole probes developed for manned aircraft in small remotely piloted aircraft (RPA). In atmospheric research, lightweight RPAs (

A comparative analysis of seven commonly-used eddy-viscosity turbulence models for CFD simulations of VAWTs is presented. The models include one- to four-equations, namely the Spalart-Allmaras (SA), RNG k-ε, realizable k-ε, SST k-ω, SST k-ω with an additional intermittency transition model (SSTI), k-kl-ω and transition SST (TSST) k-ω models. In addition, the inviscid modeling is included in the comparison. The evaluation is based on validation with three sets of experiments for three VAWTs with different geometrical characteristics operating in a wide range of operational conditions, from dynamic stall to optimal regime and to highly-rotational flow regime. The focus is on the turbine wake, the turbine power performance, and the blade aerodynamics. High-fidelity incompressible unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations are employed. The extensive analysis reveals high sensitivity of the simulation results to the turbulence model. This is especially the case for the turbine power coefficient CP. The results show that the inviscid, SA, RNG k-ε, realizable k-ε and k-kl-ω models clearly fail in reproducing the aerodynamic performance of VAWTs. Only the SST model variants (SST k-ω, SSTI and TSST) are capable of exhibiting reasonable agreement with all the experimental data sets, where the transitional SST k-ω versions (SSTI and TSST) are recommended as the models of choice especially in the transitional flow regime.

© 2019, Tsinghua University Press and Springer-Verlag GmbH Germany, part of Springer Nature. Windexchangers are relatively small structures located on the building rooftop to promote natural ventilation. This paper presents a Computational Fluid Dynamics (CFD) validation study, sensitivity analysis and performance comparison of three windexchanger (WE) configurations applied to a generic isolated building with a windward window. The study is limited to wind-driven (isothermal) ventilation, for wind perpendicular to the windward facade. The CFD simulations are based on the 3D-steady Reynolds-Averaged Navier–Stokes (RANS) equations. The validation study is performed with experimental results from a previously published water channel test. The sensitivity analysis focuses on the domain size, grid resolution and turbulence model. The performance evaluation of the three WE configurations is based on the mean velocity and mean static pressure coefficients in the vertical centerplane, the ventilation volume flow rate and the volume percentage of the living zone with air speed ratio equal to or above 0.10. The WE configuration with four openings and one duct shows the highest ventilation flow rate (0.232 m 3 /s) and the highest volume percentage (21%). This study shows that the assessment and selection of WE configurations should not only be based on volume flow rate or ACH but should consider the living zone air speed ratio as well, specifically concerning the flow distribution in the living zone. ispartof: BUILDING SIMULATION vol:12 issue:3 pages:475-488 status: published

Coastal areas and seaport areas are exposed to high wind speeds which may involve risks for the ships and people working in the area. Therefore, knowledge of the microscale wind conditions is essential for safe maneuvering and mooring of ships and optimizing harbor design. In the present study, 3D steady RANS CFD simulations with the realizable k-e turbulence model are performed for the new configuration of the “IJmuiden sea lock” in Amsterdam, the largest sea lock in the world at the time of writing this article. The computed wind speed and turbulence intensity amplification factors and the local wind directions are validated with on-site measurements for the old configuration of the sea lock. For the wind speed amplification factor and local wind direction, a satisfactory agreement is obtained with 90% of CFD data within ±30% from the measured data. Conversely, for the turbulence intensity amplification factor, less satisfactory agreement is found with 74% of CFD data within ±30% from the measured data. Overall, the 3D steady RANS approach shows a sufficiently high reliability for predicting the wind conditions in the seaport area under neutral atmospheric conditions.

Most natural landscapes are characterized by multiscale (often multifractal) topography with well-known scale-invariance properties. For example, the spectral density of landscape elevation fields is often found to have a power-law scaling behaviour (with a -2 slope on a log-log scale) over a wide span of spatial scales, typically ranging from tens of kilometres down to a few metres. Even though the effect of topography on the atmospheric boundary layer (ABL) has been the subject of numerous studies, few have focussed on multiscale topography. In this study, large-eddy simulation (LES) is used to investigate boundary-layer flow over multiscale topography, and guide the development of parametrizations needed to represent the effects of subgrid-scale (SGS) topography in numerical models of ABL flow. Particular emphasis is placed on the formulation of an effective roughness used to account for the increased aerodynamic roughness associated with SGS topography. The LES code uses the scale-dependent Lagrangian dynamic SGS model for the turbulent stresses and a terrain-following coordinate transformation to explicitly resolve the effects of the topography at scales larger than the LES resolution. The terrain used in the simulations is generated using a restricted solid-on-solid landscape evolution model, and it is characterized by a -2 slope of the elevation power spectrum. Results from simulations performed using elevation fields band-pass filtered at different spatial resolutions indicate a clear linear relation between the square of the effective roughness and the variance of elevation.

Single-column models (SCMs) are widely employed to evaluate boundary-layer parametrizations under well-controlled conditions. To compare SCM results to observations, these models must be driven by realistic forcings of the three-dimensional (3D) atmospheric state. However, these forcings are inherently uncertain. The central research question is therefore: can observations be used to distinguish between different parametrization schemes in SCM simulations or is the spread due to uncertainties in the forcings too large? This study investigates this question for the nocturnal low-level jet (LLJ) at Cabauw. First, we analyse a single LLJ case that has been used for the third GEWEX Atmospheric Boundary Layer Studies (GABLS3) intercomparison study. To estimate the forcings, for this case a blend of local observations and 3D model output has been used. A sensitivity study to both the forcings and the turbulence formulation is performed by using the SCM version of the ECMWF model. The sensitivity to the turbulence parametrization is largest in the bulk of the stable boundary layer (SBL). The influence of the forcings manifests itself mainly in the upper part of the SBL and above. Second, an ensemble of eight comparable LLJ cases is considered. Using forcings derived from 3D model integrations, SCM results of the separate cases show significant deviations from the observations. However, the mean of the SCM simulations agrees well with the mean of the observations. Based on the eight selected LLJ cases, a composite SCM case is defined from the mean forcings of the 3D model. This improves the signal-to-noise ratio, which enables a better judgement of the quality of boundary-layer parametrizations in a comparison with the observations

Thesis entitled: Modelling Stable Atmospheric Boundary Layers over Snow H.A.M. Sterk Wageningen, 29th of April, 2015 Summary The emphasis of this thesis is on the understanding and forecasting of the Stable Boundary Layer (SBL) over snow-covered surfaces. SBLs typically form at night and in polar regions (especially in winter), when radiative cooling at the surface causes a cooler surface than the overlying atmosphere and a stable stratification develops. This means that potential temperature increases with height and buoyancy effects suppress turbulence. Turbulence is then dominated by mechanical origin. If sufficient wind shear can be maintained, turbulence remains active, otherwise it will cease. A proper representation of SBLs in numerical weather prediction models is critical, since many parties rely on these forecasts. For example, weather prediction is needed for wind energy resources, agricultural purposes, air-quality studies, and aviation and road traffic. Knowledge on SBLs is also essential for climate modelling. In the Arctic regions, climate change is most pronounced due to stronger changes in near-surface temperature compared to other latitudes. Though this `Arctic amplification' is not yet fully understood, possible responsible processes are the ice-albedo feedback, alterations in cloud cover and water vapour, different atmospheric and oceanic circulations, and the weak vertical mixing in the lower atmosphere. However, many interactions exist between these processes. With positive feedbacks, changes are even further enhanced. This could have worldwide consequences, i.e. due to affected atmospheric circulations and sea level rise with Greenland's melting ice-sheets. Scientists try to explain the observed climate changes, as well as provide outlooks for future changes in climate and weather. However, the understanding is hampered by the fact that many model output variables (e.g. regarding the 2 m temperature) vary substantially between models on the one hand, and from observations on the other hand. Modelling the SBL remains difficult, because the physical processes at hand are represented in a simplified way, and the understanding of the processes may be incomplete. Furthermore, since processes can play a role at a very small scale, the resolutions in models may be too poor to represent the SBL correctly. Additionally, there are many different archetypes of the SBL. Turbulence can be continuous, practically absent, or intermittent, and vary in strength which affects the efficiency of the exchange of quantities horizontally and vertically. Processes that are considered critical for the SBL evolution are e.g. turbulent mixing, radiative effects, the coupling between the atmosphere and the underlying surface, the presence of clouds or fog, subsidence, advection, gravity waves, and drainage and katabatic flows. In this thesis, the focus is on the first three processes, as these are most dominant for the evolution of the SBL (e.g. Bosveld et al., 2014b). In Chapter 3 an idealized clear-sky case over sea-ice was studied based on the GABLS1 benchmark study (e.g. Cuxart et al., 2006), but extended by including radiative effects and thermal coupling with the surface. Hence the following research questions were posed: Question 1: What is the variety in model outcome regarding potential temperature and wind speed profiles that can be simulated with one model by using different parametrization schemes? Question 2: Which of the three governing processes is most critical in determining the SBL state in various wind regimes? Question 3: Can we identify compensation mechanisms between schemes, and thus identify where possible compensating model errors may be concealed? From the analysis with different parametrization schemes performed with the WRF single-column model (SCM, Question 1), it followed that quite different types of SBLs were found. Some schemes forecasted a somewhat better mixed potential temperature profile where stratification increased with height, while another scheme produced profiles with the strongest stratification close to the surface and stratification decreased with height. After only 9 h of simulation time, a difference in temperature of almost 2 K was found near the surface. Regarding the wind speed profile, some variation was found in the simulated low-level jet speed and height. Mainly the difference in atmospheric boundary-layer (BL) schemes which parametrize the turbulent mixing are responsible for these model output variations. A variation in long wave parametrization schemes hardly affected the model results. Question 2 addresses the problem whether other processes than turbulent mixing may be responsible for a similar spread in model results. A sensitivity analysis was performed where for one set of reference parametrization schemes the intensity of the processes was adjusted. The relative sensitivity of the three processes for different wind regimes was analysed using `process diagrams'. In a process diagram, two physically related variables are plotted against each other, which in this case represent either a time average or a difference over time of the variable. A line connects the reference state with the state for which the process intensity is modified. By comparing the length and orientation of the lines, the relative significance of the individual processes for the different wind speeds can be studied. Overlapping line directions identify possible compensating errors. Geostrophic wind speeds of 3, 8 and 20 m s-1 were selected representing low, medium and high wind speeds, capturing the range of wind speeds frequently occurring in the Arctic north of 75oN according to the ERA-Interim reanalysis dataset. Overall, a shift in relative importance was detected for the various wind regimes. With high geostrophic wind speeds, the model output is most sensitive to turbulent mixing. On the contrary, with low geostrophic wind speeds the model is most sensitive to the radiation and especially the snow-surface coupling. The impact of turbulent mixing is then minor, unless when mixing in both boundary layer and surface layer is adjusted. This stresses that proper linking between these two layers is essential. Also with one set of parametrization schemes different SBL types were simulated. Potential temperature profiles were better mixed (increasing stratification with height) for high geostrophic wind speeds, and this tended to develop to profiles with the strongest stratification near the surface (decreasing stratification with height) for low geostrophic wind speeds. However, a variety in types was also found when keeping the same wind regime, but by varying the mixing strength. With enhanced mixing, the profile became better mixed, also when the reference profile showed the strongest stratification near the surface. With decreased mixing, profiles with a stronger stratification were found, again shaped with the strongest stratification near the surface. Thus a different mixing formulation has a strong impact on the vertical profiles, even when it may not necessarily strongly affect the surface variables. Therefore, it is recommended that when a model is evaluated and optimized, the vertical structure is also regarded in this process, since near-surface variables may be well represented, strong deviations aloft are still possible. Furthermore, the process diagrams showed overlap in sensitivity to some processes. Therefore errors within the parametrizations of these processes could compensate each other and thus remain hidden (Question 3), making the model formulation possibly physically less realistic. This study did not reveal an unambiguous indication for the compensating processes regarding the various sets of variables, though overlap for single variables is seen. This study also revealed a non-linear behaviour regarding the 2 m temperature, which is also found in observations (e.g. Lüpkes et al., 2008) and in a model study by McNider et al. (2012). Here the 2 m temperature decreased with enhanced mixing strength and increased with a lower mixing intensity. This counter-intuitive behaviour is explained by that mixing only occurs in a shallow layer close to the surface. Cold air that is mixed up by the enhanced mixing, is insufficiently compensated by the downward mixed relatively warm air. This behaviour was found mostly for low wind speeds or with decreased mixing at the medium wind regime, when the potential temperature profile showed the strongest stratification near the surface. The study proceeds with a model evaluation against observations in low wind speed regimes. Three stably stratified cloud-free study cases with near-surface wind speeds below 5 m s-1 were selected with each a different surface: Cabauw in the Netherlands with snow over grass, Sodankylä in northern Finland with snow in a needle-leaf forest, and Halley in Antarctica with snow on an ice shelf. Chapter 4 presents the evaluation of the WRF-3D and SCM for these cases. In this study, the WRF-3D model was used to determine the forcings, as often not all the required observations at high resolution in time and space are available. Hence the following questions were formulated: Question 4: What is the performance of WRF in stable conditions with low wind speeds for three contrasting snow-covered sites? Question 5: How should we prescribe the single-column model forcings, using WRF-3D? The standard WRF-3D simulation had an incorrect representation of the snow-cover and vegetation fraction, which deteriorates the conductive heat flux, the surface temperatures and the SBL evolution. Indeed, Chapter 3 highlighted the critical role of the land-surface coupling representation. Adjusting the settings with site specific information, improved model simulations compared with the observations. In general, the performance of WRF-3D was quite good for the selected cases, especially regarding the wind speed simulations. The temperature forecast proved to be more challenging. For Cabauw and Sodankylä, 2 m temperatures were strongly overestimated, though a better simulation was seen at higher tower levels. For Halley a better representation of the 2 m temperature was found, though aloft potential temperatures were underestimated. Hence, the three cases had an underestimated modelled temperature gradient in common. This study also investigated how the forcing fields for the SCM should be prescribed. Model results for the three study cases all showed a significant deviation from the observed wind field without lateral forcings and time-invariant geostrophic wind speed. Including only a time-varying geostrophic wind speed did not improve the results. Prescribing additional momentum advection did have a positive impact on the modelled wind speed. The results regarding temperature, specific humidity and their stratification improved when temperature and humidity advection was also taken into account. Forcing the SCM field towards a prescribed 3D atmospheric state is not recommended, since unrealistic profiles were found below the threshold forcing height. Having established the optimal model set-up, the SCM can be used as a tool to further study the small-scale processes for the three study cases, addressing the following questions: Question 6: How do the model results with various process intensities compare with observations? Question 7: Are any differences in relative process impacts found for the three contrasting sites? Question 8: Does the model sensitivity vary between two different BL schemes? The sensitivity analysis was performed with the WRF-SCM and repeated for two BL schemes. In general, the temperature and humidity stratifications intensified by decreasing the process strengths and hence were in better agreement with observations than the reference cases. The wind field was most sensitive to turbulent mixing, with a weaker low-level jet at a higher altitude for enhanced mixing and the opposite for less mixing, while the impact of the other processes was small. Contrary to the temperature profiles, a better agreement with wind observations was found with amplified mixing, except for Halley where results improved with reduced mixing. Regarding the surface energy budget, the conductive heat flux was greatly overestimated at Cabauw due to an overestimated snow conductivity, while better agreements were found for the other sites. A revision of the definition for snow conductivity in the model is recommended, because rather large values were assumed for fresh snow, and indeed results improved when the coupling strength was reduced for Cabauw and Sodankylä. For Halley almost the same snow conductivity was modelled as was used to determine the observed conductive heat flux, however, then the temperature gradient through the first soil/snow layer was underestimated leaving the flux too small. The net radiation was strongly too negative for the Cabauw and Halley case-studies. This is likely due to an underestimation of the incoming long wave radiation as part of a deficiency in the long wave radiation scheme. For all sites the sensible heat flux was overestimated, and decreased mixing improved the results. However, the eddy covariance measurements may have been made outside the constant flux layer, which hampers the model evaluation. Though Question 6 aims to obtain understanding in which processes are most responsible for simulating model results that are in closer agreement with observations, measuring in these cold and dry circumstances is especially challenging. Furthermore, the measurements are mostly point measurements while the model grid represents a larger area, such that the measurements may be influenced by local features which are not captured in the model. These issues hinders a clear comparison of the model results with observations, and the observation uncertainties may be greater than what was represented in the process diagrams. When comparing the process sensitivity for the different sites (Question 7), we found some distinct variations in relative process significance. The radiation impact was relatively large at Cabauw and Sodankylä where the specific humidity was higher such that a larger impact on the incoming long wave radiation can be obtained. The snow-surface coupling is more important at Halley. This is related to the higher snow cover at Halley compared to the other sites. Additionally, the conductivity of the underlying medium at Halley is set equal to that of snow. These two factors ensure that the impact of an altered snow conductivity is greater. From the comparison of the sensitivity analyses for the two BL schemes (MYJ and YSU, Question 8), it followed that the overall direction of the sensitivity orientation is similar. However, stronger BL temperature stratifications were found with YSU, though between the surface and the first model level stronger stratifications were simulated with MYJ. This is related to the relatively high ratio of mixing in the boundary layer versus the surface layer with MYJ. Therefore the mixing in the BL is relatively more efficient and the surface layer cannot keep up the mixing to keep a smooth profile at the surface-layer / boundary-layer interface. This indicates the importance of a consistent transition between the BL and surface layer, as also pointed out by Svensson and Holtslag (2009). Furthermore, the non-linearity concerning the 2 m temperature behaviour discussed earlier is most profound with YSU, and not as obvious with MYJ due to a stronger implemented minimum diffusivity. The results point towards the direction of focus for future research. This could be achieved by e.g. re-evaluating the snow representation, as well as investigating the long-standing problem of the underestimated long wave radiation. Additionally, the mixing seems to be too high in some of the simulations. As such, care should be taken in choosing the BL scheme and its constraints on the mixing, as these may hamper the development of the observed behaviour on non-linear near-surface temperature evolution for example.

Thesis entitled: Modelling Stable Atmospheric Boundary Layers over Snow H.A.M. Sterk Wageningen, 29th of April, 2015 Summary The emphasis of this thesis is on the understanding and forecasting of the Stable Boundary Layer (SBL) over snow-covered surfaces. SBLs typically form at night and in polar regions (especially in winter), when radiative cooling at the surface causes a cooler surface than the overlying atmosphere and a stable stratification develops. This means that potential temperature increases with height and buoyancy effects suppress turbulence. Turbulence is then dominated by mechanical origin. If sufficient wind shear can be maintained, turbulence remains active, otherwise it will cease. A proper representation of SBLs in numerical weather prediction models is critical, since many parties rely on these forecasts. For example, weather prediction is needed for wind energy resources, agricultural purposes, air-quality studies, and aviation and road traffic. Knowledge on SBLs is also essential for climate modelling. In the Arctic regions, climate change is most pronounced due to stronger changes in near-surface temperature compared to other latitudes. Though this `Arctic amplification' is not yet fully understood, possible responsible processes are the ice-albedo feedback, alterations in cloud cover and water vapour, different atmospheric and oceanic circulations, and the weak vertical mixing in the lower atmosphere. However, many interactions exist between these processes. With positive feedbacks, changes are even further enhanced. This could have worldwide consequences, i.e. due to affected atmospheric circulations and sea level rise with Greenland's melting ice-sheets. Scientists try to explain the observed climate changes, as well as provide outlooks for future changes in climate and weather. However, the understanding is hampered by the fact that many model output variables (e.g. regarding the 2 m temperature) vary substantially between models on the one hand, and from observations on the other hand. Modelling the SBL remains difficult, because the physical processes at hand are represented in a simplified way, and the understanding of the processes may be incomplete. Furthermore, since processes can play a role at a very small scale, the resolutions in models may be too poor to represent the SBL correctly. Additionally, there are many different archetypes of the SBL. Turbulence can be continuous, practically absent, or intermittent, and vary in strength which affects the efficiency of the exchange of quantities horizontally and vertically. Processes that are considered critical for the SBL evolution are e.g. turbulent mixing, radiative effects, the coupling between the atmosphere and the underlying surface, the presence of clouds or fog, subsidence, advection, gravity waves, and drainage and katabatic flows. In this thesis, the focus is on the first three processes, as these are most dominant for the evolution of the SBL (e.g. Bosveld et al., 2014b). In Chapter 3 an idealized clear-sky case over sea-ice was studied based on the GABLS1 benchmark study (e.g. Cuxart et al., 2006), but extended by including radiative effects and thermal coupling with the surface. Hence the following research questions were posed: Question 1: What is the variety in model outcome regarding potential temperature and wind speed profiles that can be simulated with one model by using different parametrization schemes? Question 2: Which of the three governing processes is most critical in determining the SBL state in various wind regimes? Question 3: Can we identify compensation mechanisms between schemes, and thus identify where possible compensating model errors may be concealed? From the analysis with different parametrization schemes performed with the WRF single-column model (SCM, Question 1), it followed that quite different types of SBLs were found. Some schemes forecasted a somewhat better mixed potential temperature profile where stratification increased with height, while another scheme produced profiles with the strongest stratification close to the surface and stratification decreased with height. After only 9 h of simulation time, a difference in temperature of almost 2 K was found near the surface. Regarding the wind speed profile, some variation was found in the simulated low-level jet speed and height. Mainly the difference in atmospheric boundary-layer (BL) schemes which parametrize the turbulent mixing are responsible for these model output variations. A variation in long wave parametrization schemes hardly affected the model results. Question 2 addresses the problem whether other processes than turbulent mixing may be responsible for a similar spread in model results. A sensitivity analysis was performed where for one set of reference parametrization schemes the intensity of the processes was adjusted. The relative sensitivity of the three processes for different wind regimes was analysed using `process diagrams'. In a process diagram, two physically related variables are plotted against each other, which in this case represent either a time average or a difference over time of the variable. A line connects the reference state with the state for which the process intensity is modified. By comparing the length and orientation of the lines, the relative significance of the individual processes for the different wind speeds can be studied. Overlapping line directions identify possible compensating errors. Geostrophic wind speeds of 3, 8 and 20 m s-1 were selected representing low, medium and high wind speeds, capturing the range of wind speeds frequently occurring in the Arctic north of 75oN according to the ERA-Interim reanalysis dataset. Overall, a shift in relative importance was detected for the various wind regimes. With high geostrophic wind speeds, the model output is most sensitive to turbulent mixing. On the contrary, with low geostrophic wind speeds the model is most sensitive to the radiation and especially the snow-surface coupling. The impact of turbulent mixing is then minor, unless when mixing in both boundary layer and surface layer is adjusted. This stresses that proper linking between these two layers is essential. Also with one set of parametrization schemes different SBL types were simulated. Potential temperature profiles were better mixed (increasing stratification with height) for high geostrophic wind speeds, and this tended to develop to profiles with the strongest stratification near the surface (decreasing stratification with height) for low geostrophic wind speeds. However, a variety in types was also found when keeping the same wind regime, but by varying the mixing strength. With enhanced mixing, the profile became better mixed, also when the reference profile showed the strongest stratification near the surface. With decreased mixing, profiles with a stronger stratification were found, again shaped with the strongest stratification near the surface. Thus a different mixing formulation has a strong impact on the vertical profiles, even when it may not necessarily strongly affect the surface variables. Therefore, it is recommended that when a model is evaluated and optimized, the vertical structure is also regarded in this process, since near-surface variables may be well represented, strong deviations aloft are still possible. Furthermore, the process diagrams showed overlap in sensitivity to some processes. Therefore errors within the parametrizations of these processes could compensate each other and thus remain hidden (Question 3), making the model formulation possibly physically less realistic. This study did not reveal an unambiguous indication for the compensating processes regarding the various sets of variables, though overlap for single variables is seen. This study also revealed a non-linear behaviour regarding the 2 m temperature, which is also found in observations (e.g. Lüpkes et al., 2008) and in a model study by McNider et al. (2012). Here the 2 m temperature decreased with enhanced mixing strength and increased with a lower mixing intensity. This counter-intuitive behaviour is explained by that mixing only occurs in a shallow layer close to the surface. Cold air that is mixed up by the enhanced mixing, is insufficiently compensated by the downward mixed relatively warm air. This behaviour was found mostly for low wind speeds or with decreased mixing at the medium wind regime, when the potential temperature profile showed the strongest stratification near the surface. The study proceeds with a model evaluation against observations in low wind speed regimes. Three stably stratified cloud-free study cases with near-surface wind speeds below 5 m s-1 were selected with each a different surface: Cabauw in the Netherlands with snow over grass, Sodankylä in northern Finland with snow in a needle-leaf forest, and Halley in Antarctica with snow on an ice shelf. Chapter 4 presents the evaluation of the WRF-3D and SCM for these cases. In this study, the WRF-3D model was used to determine the forcings, as often not all the required observations at high resolution in time and space are available. Hence the following questions were formulated: Question 4: What is the performance of WRF in stable conditions with low wind speeds for three contrasting snow-covered sites? Question 5: How should we prescribe the single-column model forcings, using WRF-3D? The standard WRF-3D simulation had an incorrect representation of the snow-cover and vegetation fraction, which deteriorates the conductive heat flux, the surface temperatures and the SBL evolution. Indeed, Chapter 3 highlighted the critical role of the land-surface coupling representation. Adjusting the settings with site specific information, improved model simulations compared with the observations. In general, the performance of WRF-3D was quite good for the selected cases, especially regarding the wind speed simulations. The temperature forecast proved to be more challenging. For Cabauw and Sodankylä, 2 m temperatures were strongly overestimated, though a better simulation was seen at higher tower levels. For Halley a better representation of the 2 m temperature was found, though aloft potential temperatures were underestimated. Hence, the three cases had an underestimated modelled temperature gradient in common. This study also investigated how the forcing fields for the SCM should be prescribed. Model results for the three study cases all showed a significant deviation from the observed wind field without lateral forcings and time-invariant geostrophic wind speed. Including only a time-varying geostrophic wind speed did not improve the results. Prescribing additional momentum advection did have a positive impact on the modelled wind speed. The results regarding temperature, specific humidity and their stratification improved when temperature and humidity advection was also taken into account. Forcing the SCM field towards a prescribed 3D atmospheric state is not recommended, since unrealistic profiles were found below the threshold forcing height. Having established the optimal model set-up, the SCM can be used as a tool to further study the small-scale processes for the three study cases, addressing the following questions: Question 6: How do the model results with various process intensities compare with observations? Question 7: Are any differences in relative process impacts found for the three contrasting sites? Question 8: Does the model sensitivity vary between two different BL schemes? The sensitivity analysis was performed with the WRF-SCM and repeated for two BL schemes. In general, the temperature and humidity stratifications intensified by decreasing the process strengths and hence were in better agreement with observations than the reference cases. The wind field was most sensitive to turbulent mixing, with a weaker low-level jet at a higher altitude for enhanced mixing and the opposite for less mixing, while the impact of the other processes was small. Contrary to the temperature profiles, a better agreement with wind observations was found with amplified mixing, except for Halley where results improved with reduced mixing. Regarding the surface energy budget, the conductive heat flux was greatly overestimated at Cabauw due to an overestimated snow conductivity, while better agreements were found for the other sites. A revision of the definition for snow conductivity in the model is recommended, because rather large values were assumed for fresh snow, and indeed results improved when the coupling strength was reduced for Cabauw and Sodankylä. For Halley almost the same snow conductivity was modelled as was used to determine the observed conductive heat flux, however, then the temperature gradient through the first soil/snow layer was underestimated leaving the flux too small. The net radiation was strongly too negative for the Cabauw and Halley case-studies. This is likely due to an underestimation of the incoming long wave radiation as part of a deficiency in the long wave radiation scheme. For all sites the sensible heat flux was overestimated, and decreased mixing improved the results. However, the eddy covariance measurements may have been made outside the constant flux layer, which hampers the model evaluation. Though Question 6 aims to obtain understanding in which processes are most responsible for simulating model results that are in closer agreement with observations, measuring in these cold and dry circumstances is especially challenging. Furthermore, the measurements are mostly point measurements while the model grid represents a larger area, such that the measurements may be influenced by local features which are not captured in the model. These issues hinders a clear comparison of the model results with observations, and the observation uncertainties may be greater than what was represented in the process diagrams. When comparing the process sensitivity for the different sites (Question 7), we found some distinct variations in relative process significance. The radiation impact was relatively large at Cabauw and Sodankylä where the specific humidity was higher such that a larger impact on the incoming long wave radiation can be obtained. The snow-surface coupling is more important at Halley. This is related to the higher snow cover at Halley compared to the other sites. Additionally, the conductivity of the underlying medium at Halley is set equal to that of snow. These two factors ensure that the impact of an altered snow conductivity is greater. From the comparison of the sensitivity analyses for the two BL schemes (MYJ and YSU, Question 8), it followed that the overall direction of the sensitivity orientation is similar. However, stronger BL temperature stratifications were found with YSU, though between the surface and the first model level stronger stratifications were simulated with MYJ. This is related to the relatively high ratio of mixing in the boundary layer versus the surface layer with MYJ. Therefore the mixing in the BL is relatively more efficient and the surface layer cannot keep up the mixing to keep a smooth profile at the surface-layer / boundary-layer interface. This indicates the importance of a consistent transition between the BL and surface layer, as also pointed out by Svensson and Holtslag (2009). Furthermore, the non-linearity concerning the 2 m temperature behaviour discussed earlier is most profound with YSU, and not as obvious with MYJ due to a stronger implemented minimum diffusivity. The results point towards the direction of focus for future research. This could be achieved by e.g. re-evaluating the snow representation, as well as investigating the long-standing problem of the underestimated long wave radiation. Additionally, the mixing seems to be too high in some of the simulations. As such, care should be taken in choosing the BL scheme and its constraints on the mixing, as these may hamper the development of the observed behaviour on non-linear near-surface temperature evolution for example.

This paper deals with the utility of Monin-Obukhov similarity theory for wind speed simulation in stable conditions over flat terrain in northern Croatia. Wind speed profiles observed through a 150-m deep layer are compared with those obtained from two simple wind speed models. For this purpose, the similarity functions for two stable classes based on Monin-Obukhov theory are locally determined. They are tested in wind speed simulations using a model based on the Monin-Obukhov theory (MO) and a model which combines the surface-layer model of Monin-Obukhov theory and the Ekman-Taylor type model (MOE). Under less stable conditions the combined model (MOE) produces better skill, whereas the model based on the Monin-Obukhov theory (MO) performed better under more stable conditions. Copyright © 2002 Royal Meteorological Society.

Air temperature and specific humidity inversions and low-level jets were studied over two Svalbard fjords, Isfjorden and Kongsfjorden, applying three tethersonde systems. Tethersonde operation practices notably affected observations on inversion and jet properties. The inversion strength and depth were strongly affected by weather conditions at the 850 hPa level. Strong inversions were deep with a highly elevated base, and the strongest ones occurred in warm air mass. Unexpectedly, downward longwave radiation measured at the sounding site did not correlate with the inversion properties. Temperature inversions had lower base and top heights than humidity inversions, the former due to surface cooling and the latter due to adiabatic cooling with height. Most low-level jets were related to katabatic winds. Over the ice-covered Kongsfjorden, jets were lifted above a cold-air pool on the fjord; the jet core was located highest when the snow surface was coldest. At the ice-free Isfjorden, jets were located lower.

We give a new derivation of the familiar linear relation for the dimensionless velocity gradient in the stably stratified surface layer and provide physical and empirical grounds for its universal applicability in stationary homogeneous turbulence over the whole range of static stabilities from Ri = 0 to very large Ri. Combining this relation with the budget equation for the turbulent kinetic energy we obtain the “equilibrium formulation” of the turbulent dissipation length scale, and recommend it for use in turbulence closure models.

Single-column models (SCMs) are widely employed to evaluate boundary-layer parametrizations under well-controlled conditions. To compare SCM results to observations, these models must be driven by realistic forcings of the three-dimensional (3D) atmospheric state. However, these forcings are inherently uncertain. The central research question is therefore: can observations be used to distinguish between different parametrization schemes in SCM simulations or is the spread due to uncertainties in the forcings too large? This study investigates this question for the nocturnal low-level jet (LLJ) at Cabauw. First, we analyse a single LLJ case that has been used for the third GEWEX Atmospheric Boundary Layer Studies (GABLS3) intercomparison study. To estimate the forcings, for this case a blend of local observations and 3D model output has been used. A sensitivity study to both the forcings and the turbulence formulation is performed by using the SCM version of the ECMWF model. The sensitivity to the turbulence parametrization is largest in the bulk of the stable boundary layer (SBL). The influence of the forcings manifests itself mainly in the upper part of the SBL and above. Second, an ensemble of eight comparable LLJ cases is considered. Using forcings derived from 3D model integrations, SCM results of the separate cases show significant deviations from the observations. However, the mean of the SCM simulations agrees well with the mean of the observations. Based on the eight selected LLJ cases, a composite SCM case is defined from the mean forcings of the 3D model. This improves the signal-to-noise ratio, which enables a better judgement of the quality of boundary-layer parametrizations in a comparison with the observations

A climatology of nocturnal low-level jets (LLJs) is presented for the topographically flat measurement site at Cabauw, the Netherlands. LLJ characteristics are derived from a 7-yr half-hourly database of wind speed profiles, obtained from the 200-m mast and a wind profiler. Many LLJs at Cabauw originate from an inertial oscillation, which develops after sunset in a layer decoupled from the surface by stable stratification. The data are classified to different types of stable boundary layers by using the geostrophic wind speed and the isothermal net radiative cooling as classification parameters. For each of these classes, LLJ characteristics like frequency of occurrence, height above ground level, and the turning of the wind vector across the boundary layer are determined. It is found that LLJs occur in about 20% of the nights, are typically situated at 140–260 m above ground level, and have a speed of 6–10 m s−1. Development of a substantial LLJ is most likely to occur for moderate geostrophic forcing and a high radiative cooling. A comparison with the 40-yr ECMWF Re-Analysis (ERA-40) is added to illustrate how the results can be used to evaluate the performance of atmospheric models.

It is common knowledge that in the evening the wind close to the ground drops. However, it is less well known that at the same time the wind at a few hundreds of meters height can strongly accelerate. In this case, the vertical wind-profile shows a distinct maximum. This phenomenon is called nocturnal low-level jet (LLJ). Knowledge of the LLJ is relevant for aviation, the generation of wind energy and the distribution of pollutants. This thesis presents a climatology of LLJ characteristics (like the frequency of occurrence and the typical height at which the phenomenon occurs) for observations from the KNMI measurement site at Cabauw. The development of nocturnal LLJs is very sensitive to the amount of turbulent mixing. In this thesis different ways on which turbulence is represented in atmospheric models are tested. Various strategies for model evaluation are discussed. This thesis aims to contribute to the understanding of physical processes and their representation in weather and climate models. Het is algemeen bekend dat dichtbij de grond de wind in de avond wegvalt. Wat minder bekend is, is dat op enkele honderden meters hoogte op hetzelfde ogenblik de wind sterk kan toenemen. Dit gebeurt vooral bij mooi weer. In dit geval vertoont het verticale profiel van de windsnelheid een duidelijk maximum. Dit verschijnsel wordt low-level jet (LLJ) genoemd. Kennis van de LLJ is belangrijk voor de luchtvaart, het opwekken van windenergie en de verspreiding van vervuilende stoffen. Dit proefschrift presenteert een klimatologie van LLJ karakteristieken (zoals de frequentie van voorkomen en de typische hoogte waarop het verschijnsel optreedt) voor de meetlocatie van het KNMI in Cabauw. De ontwikkeling van LLJs is erg afhankelijk van de hoeveelheid turbulente menging. In dit proefschrift worden verschillende manieren getest waarop turbulentie in modellen gerepresenteerd is. Ook worden verscheidene strategieën voor modelevaluatie besproken. Dit proefschrift heeft tot doel om bij te dragen aan het begrijpen van fysische processen en het verbeteren van hun representatie in weer- en klimaatmodellen.

Blowing snow causes disturbances that affect both tourism and transportation. Moreover, snowdrifts and accelerated snow metamorphosis dramatically increase the risk of avalanches. These observations call for more scientific and technical researches like those realized in Cemagref, Grenoble, France, where this Ph-D was achieved. Past works contributed to the development of physical and numerical models about blowing snow but the difficulties encountered to match the results get in the field and the general complexity involved in Aeolian particle transport required more studies. That's why new experimental measurements were recorded either in a wind-tunnel or in the field. On one hand, image processing and spectral analysis were used to investigate the fluctuations of particle concentration blown in a stationary turbulent flow. On the other hand, the variability of concentration profiles during natural events of blowing snow was studied thanks to flux measurements. These latter required some dedicated sensors with specific calibration., Le transport de neige par le vent occasionne des nuisances qui entravent certaines activités du secteur des transports et du tourisme. Il déplace les dépôts de neige et modifie la structure du manteau neigeux. A ce titre, il participe à l'augmentation du risque d'avalanche. Ces différents aspects justifient l'activité de recherche menée depuis plusieurs années au Cemagref de Grenoble et dans laquelle cette étude trouve naturellement sa place. Ainsi les travaux précédents ont contribué à la modélisation physique et numérique de prévision du transport de neige par le vent. Cela dit, le transport éolien de particules met en jeu différents phénomènes complexes que les connaissances actuelles ne suffisent pas à décrire avec précision. Dans ce contexte, des travaux expérimentaux ont été conduits à la fois en soufflerie et sur le terrain. Ces derniers ont permis des premières investigations sur le comportement des fluctuations de concentration en régime permanent, en particulier par le biais de traitements d'images et d'analyses spectrales. Cette thèse a également porté sur la variabilité des profils de concentration en conditions naturelles, c'est-à-dire dans un écoulement turbulent instationnaire. Tous les résultats in-situ ont nécessité l'usage d'instruments adaptés dont certains ont du être réétalonnés de manière spécifique.

We describe the results of an experiment focusing on wind speed and momentum fluxes in the atmospheric boundary layer up to 200 m. The measurements were conducted in 1996 at the Cabauw site in the Netherlands. Momentum fluxes are measured using the K-Gill Propeller Vane. Estimates of the roughness length are derived using various techniques from the wind speed and flux measurements, and the observed differences are explained by considering the source area of the meteorological parameters. A clear rough-to-smooth transition is found in the wind speed profiles at Cabauw. The internal boundary layer reaches the lowest k-vane (20 m) only in the south-west direction where the obstacle-free fetch is about 2 km. The internal boundary layer is also reflected in the roughness lengths derived from the wind speed profiles. The lower part of the profile (

In this paper, the degree of scatter in flux–gradient relationships for stably stratified conditions is analyzed. It is generally found that scatter in the dimensionless lapse rate ϕh is larger than in the dimensionless shear ϕm when plotted versus the stability parameter z/Λ (where Λ is the local Obukhov length). Here, this phenomenon is explained to be a result of self-correlation due to the occurrence of the momentum and the heat flux on both axes, measurement uncertainties, and other possibly relevant physical processes left aside. It is shown that the ratio between relative errors in the turbulent fluxes influences the orientation of self-correlation in the flux–gradient relationships. In stable conditions, the scatter in ϕm is largely suppressed by self-correlation while for ϕh this is not the case (vice versa for unstable stratification). An alternative way of plotting is discussed for determining the slope of the linear ϕm function.

The aim of this work is to present experimentally evaluated effective roughnesses (z(oe)) of a partly forested landscape. Although the ratio of boundary-layer height to obstacle size was only of the order of 50, there still seemed to exist a height range of 75-200 m where surface-layer similarity was approximately valid. Attempts were made to use conventional wind profile analysis to evaluate z(oe), but the small height range and the large number of variables initially led to unacceptable uncertainties. Fixing the displacement height z(d), rather than fitting it, reduced the data scatter to an acceptable level. The profile-derived roughness lengths z(op) obtained in this way were in good agreement with previous work, and with an alternative roughness length estimate z(of) for which flux-derived profile parameters u* and theta* were used. This implies that the profile-derived roughnesses were consistent with the measured surface-layer momentum flux. Comparison of both roughness estimates also yielded an improved estimate of the displacement height. Besides this, the authors tested a landscape roughness evaluation method which makes use of the gustiness parameter T-u = sigma(u)/U in the surface layer. The results obtained by this method were in fair agreement with the profile-derived data. In previous work, the gustiness method was advocated because it could be used at relatively low levels, perhaps even within the roughness sub-layer. At the present measuring Site, this was not the case as the gustiness method was only valid in an approximate way, and for a limited height range.

Observations of the dependence of the dimensionless wind speed gradient φm as a function of the Monin–Obukhov stability parameter z/Lo under strong stability diverge from results of large-eddy simulation (LES) modelling. A kinetic energy budget analysis indicates that it is likely caused by violations of the assumptions of stationarity and/or homogeneity of turbulence in the field experiments rather than in imperfections of the LES. This confirms the validity of the widely used linear approximation for φm not only at weak to moderate stability, but also under strong stability. The new interpretation of the linear approximation of φm is given in terms of turbulent scales, which gives hope for its applicability to the free atmosphere as well. publishedVersion

Large-eddy simulation (LES), coupled with a wind-turbine model, is used to investigate the characteristics of a wind-turbine wake in a neutral turbulent boundary-layer flow. The tuning-free Lagrangian scale-dependent dynamic subgrid-scale (SGS) model is used for the parametrisation of the SGS stresses. The turbine-induced forces (e.g., thrust, lift and drag) are parametrised using two models: (a) the ‘standard’ actuator-disk model (ADM-NR), which calculates only the thrust force and distributes it uniformly over the rotor area; and (b) the actuator-disk model with rotation (ADM-R), which uses the blade-element theory to calculate the lift and drag forces (that produce both thrust and rotation), and distribute them over the rotor disk based on the local blade and flow characteristics. Simulation results are compared to high-resolution measurements collected with hot-wire anemometry in the wake of a miniature wind turbine at the St. Anthony Falls Laboratory atmospheric boundary-layer wind tunnel. In general, the characteristics of the wakes simulated with the proposed LES framework are in good agreement with the measurements in the far-wake region. The ADM-R yields improved predictions compared with the ADM-NR in the near-wake region, where including turbine-induced flow rotation and accounting for the non-uniformity of the turbine-induced forces appear to be important. Our results also show that the Lagrangian scale-dependent dynamic SGS model is able to account, without any tuning, for the effects of local shear and flow anisotropy on the distribution of the SGS model coefficient.