Details
| Original language | English |
|---|---|
| Qualification | Doctor rerum naturalium |
| Awarding Institution | |
| Supervised by |
|
| Date of Award | 6 Dec 2023 |
| Place of Publication | Hannover |
| Publication status | Published - 10 Apr 2024 |
Abstract
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Hannover, 2024. 205 p.
Research output: Thesis › Doctoral thesis
}
TY - BOOK
T1 - High resolution numerical simulations of dust devils in the convective boundary layer
T2 - Effects of detailed process representation on Vortex development and dust release
AU - Giersch, Sebastian
PY - 2024/4/10
Y1 - 2024/4/10
N2 - Dust Devils (DDs) are convective whirlwinds that frequently occur in the atmospheric Convective Boundary Layer (CBL) of arid regions during daytime. If strong enough, they transport soil particles into the atmosphere. These aerosols alter cloud microphysics and the Earth’s radiation budget. Existing quantifications of the particle release and estimates of the contribution of DDs to the local, regional, and global dust cycle are highly uncertain. Good knowledge about the DD statistics and dynamics would help to improve them. A better understanding of DDs is also beneficial for investigations of other vortex types like tornadoes or waterspouts. This thesis presents results from numerical investigations of DD-like vortices that occur in CBLs. The simulations are carried out with the turbulence-resolving PALM model system. The first study investigates the effects of the grid spacing, background wind, and surface heat flux heterogeneities on the DD characteristics with a focus on the vortex strength. It is shown that grid spacings of 2 m, a striped pattern of heat flux heterogeneities, and moderate background winds significantly increase the vortex strength in Large-Eddy Simulation (LES) of the atmospheric CBL. These are the first numerical simulations that have ever produced DD-like vortices of observed intensity. In the second study, DD-like vortices are investigated in Direct Numerical Simulation (DNS) of Rayleigh-Bénard convection for Rayleigh numbers up to 10 to the power of 11 for the first time. The model domain’s aspect ratio, the velocity boundary condition, and the Rayleigh number are the main control parameters that are varied. It is shown that a minimum Rayleigh number of 10 to the power of 7 is necessary for the development of DD-like vortices, which is much less than a typical atmospheric value of 10 to the power of 18. While the aspect ratio shows only minor effects on the vortices, the Rayleigh number and surface friction are critical parameters for nearly all vortex properties. The results also reveal that the three-dimensional structure of the DDs is very similar to the one in LES of the atmospheric boundary layer, indicating that subfilter-scale models and parameterizations of the surface-atmosphere exchange in LES have a negligible influence on the vortices. A grid convergence study of the DD statistics, where the grid spacing is gradually decreased from 10 to 0.625 m, is conducted in the third study of this thesis. Grid spacings of 1 m or less have never been used before for the analysis of DDs that develop in atmospheric LES while capturing the large-scale cellular pattern of the CBL. It is demonstrated that the derivation of meaningful quantitative vortex features requires a resolution of approximately 1 m or less. At this resolution, mean quantities averaged over all detected DDs are converged. However, maxima show no convergence, which is mainly attributed to the detachment of the vertically thin super-adiabatic layer in the vortex core from the near-surface region. This layer needs very fine grid spacings of less than 1 m to be sufficiently resolved. A comparison with measurements indicates that a grid spacing of just less than 0.5 m might be adequate for a convergence of the maximum values. For a qualitative investigation of the DD flow structure, a resolution of 2.5 m or smaller is recommended. In the final study, the surface particle emission, the near-surface particle transport in vertical direction, and the particle concentration of DDs are investigated in LES of the atmospheric CBL. The focus is on dust-sized particles. It is found that DDs cause peak dust emission fluxes and dust mass concentrations of up to 50 mg m^−2 s^−1 and 10 mg m^-3, respectively, which is 1–2 orders of magnitude larger than previous numerical estimates. On average, the dust transport at 10 m height is five times larger than the surface dust emission, making DDs an important phenomenon for air quality and visibility. The DD’s contribution to the total dust emission in desert-like regions reveals a mean of 5 %, which is almost one order of magnitude less compared to other measurement-based studies. The small contribution is attributed to large-scale patterns of relatively high dust emission, which follow the cellular flow pattern of the CBL and which are the dominant mechanism for the saltation-based dust release.
AB - Dust Devils (DDs) are convective whirlwinds that frequently occur in the atmospheric Convective Boundary Layer (CBL) of arid regions during daytime. If strong enough, they transport soil particles into the atmosphere. These aerosols alter cloud microphysics and the Earth’s radiation budget. Existing quantifications of the particle release and estimates of the contribution of DDs to the local, regional, and global dust cycle are highly uncertain. Good knowledge about the DD statistics and dynamics would help to improve them. A better understanding of DDs is also beneficial for investigations of other vortex types like tornadoes or waterspouts. This thesis presents results from numerical investigations of DD-like vortices that occur in CBLs. The simulations are carried out with the turbulence-resolving PALM model system. The first study investigates the effects of the grid spacing, background wind, and surface heat flux heterogeneities on the DD characteristics with a focus on the vortex strength. It is shown that grid spacings of 2 m, a striped pattern of heat flux heterogeneities, and moderate background winds significantly increase the vortex strength in Large-Eddy Simulation (LES) of the atmospheric CBL. These are the first numerical simulations that have ever produced DD-like vortices of observed intensity. In the second study, DD-like vortices are investigated in Direct Numerical Simulation (DNS) of Rayleigh-Bénard convection for Rayleigh numbers up to 10 to the power of 11 for the first time. The model domain’s aspect ratio, the velocity boundary condition, and the Rayleigh number are the main control parameters that are varied. It is shown that a minimum Rayleigh number of 10 to the power of 7 is necessary for the development of DD-like vortices, which is much less than a typical atmospheric value of 10 to the power of 18. While the aspect ratio shows only minor effects on the vortices, the Rayleigh number and surface friction are critical parameters for nearly all vortex properties. The results also reveal that the three-dimensional structure of the DDs is very similar to the one in LES of the atmospheric boundary layer, indicating that subfilter-scale models and parameterizations of the surface-atmosphere exchange in LES have a negligible influence on the vortices. A grid convergence study of the DD statistics, where the grid spacing is gradually decreased from 10 to 0.625 m, is conducted in the third study of this thesis. Grid spacings of 1 m or less have never been used before for the analysis of DDs that develop in atmospheric LES while capturing the large-scale cellular pattern of the CBL. It is demonstrated that the derivation of meaningful quantitative vortex features requires a resolution of approximately 1 m or less. At this resolution, mean quantities averaged over all detected DDs are converged. However, maxima show no convergence, which is mainly attributed to the detachment of the vertically thin super-adiabatic layer in the vortex core from the near-surface region. This layer needs very fine grid spacings of less than 1 m to be sufficiently resolved. A comparison with measurements indicates that a grid spacing of just less than 0.5 m might be adequate for a convergence of the maximum values. For a qualitative investigation of the DD flow structure, a resolution of 2.5 m or smaller is recommended. In the final study, the surface particle emission, the near-surface particle transport in vertical direction, and the particle concentration of DDs are investigated in LES of the atmospheric CBL. The focus is on dust-sized particles. It is found that DDs cause peak dust emission fluxes and dust mass concentrations of up to 50 mg m^−2 s^−1 and 10 mg m^-3, respectively, which is 1–2 orders of magnitude larger than previous numerical estimates. On average, the dust transport at 10 m height is five times larger than the surface dust emission, making DDs an important phenomenon for air quality and visibility. The DD’s contribution to the total dust emission in desert-like regions reveals a mean of 5 %, which is almost one order of magnitude less compared to other measurement-based studies. The small contribution is attributed to large-scale patterns of relatively high dust emission, which follow the cellular flow pattern of the CBL and which are the dominant mechanism for the saltation-based dust release.
U2 - 10.15488/16927
DO - 10.15488/16927
M3 - Doctoral thesis
CY - Hannover
ER -