Session MG: GFD: Atmospheric Flows III

A new anemometer for 2D atmospheric flow measurements in rough environments

One major downside of cup anemometry is the different response time for increasing and decreasing wind speeds, causing a systematic over-estimation of the mean wind speed under turbulent conditions. Especially under harsh environmental conditions like in offshore operation, the measuring principle leads to a wear of bearings causing a de-calibration over time and the requirement of regular maintenance. Therefore, we propose the newly developed sphere anemometer as a simple and robust alternative without any moving parts. The sphere anemometer consists of a flexible tube with a sphere mounted on top of it. The drag force acting on the sphere and its support causes a deflection, which is measured by means of a light pointer. Via calibration, this allows for simultaneous determination of wind speed and direction using only one sensor. In our contribution, we introduce the anemometer's setup and it's optimization towards offshore application. Additionally, experimental results obtained from wind tunnel measurements of turbulent flows are presented. Measurements under real wind conditions are compared to those of state-of-the-art wind speed sensors, such as cup and ultrasonic anemometers.   

Interaction of pollution plumes and discontinuous fields in atmospheric chemistry models

Atmospheric pollutants originate from concentrated sources such as cities, power plants, and biomass fires. They are injected in the troposphere where eddies and convective motions of various scales act to shear and dilute the pollution plumes as they are advected downwind. Despite this shear and dilution, observations from aircraft, sondes, and satellites show that pollution plumes in the remote free troposphere can preserve their identity as well-defined layers for a week or more as they are transported on intercontinental scales. This structure cannot be reproduced in the standard Eulerian chemical transport models used for global modeling of tropospheric composition, instead, the plumes dissipate far too quickly. In this work, we study how the structure of plumes is modified when they cross discontinuities arising for example: from the moving day-night boundaries or from abrupt unresolved horizontal temperature changes (for example in horizontal ocean-land or ocean-ice transitions). Chemical reactions within the plumes depend strongly on photon availability and temperature, and thus, discontinuities in these variables lead to discontinuous changes in reaction rate constants.   

Punctuated changes in plant pathogen populations associated with passage of atmospheric Lagrangian coherent structures

The atmospheric transport of airborne microorganisms (e.g., plant pathogens) is poorly understood, yet necessary to assess their ecological roles in agricultural ecosystems and to evaluate risks posed by invasive species. The atmospheric transport of plant pathogens can be roughly divided into three phases: liberation of pathogen spores, drift (transport in the atmosphere) and deposition. If liberated spores escape into the planetary boundary layer, they could be transported over thousands of kilometers before being deposited. The drift phase is poorly understood, due to the complex nature of atmospheric transport and relative lack of observational data. In this talk, we present a framework of Lagrangian coherent structures to determine the important atmospheric transport barriers (ATBs) that partition the atmosphere and systematically organize the mesoscale transport problem. Using autonomous unmanned aerial vehicles, we measure the concentration of spores of a plant pathogenic fungus (\textit{Fusarium}) sampled in the atmosphere above Virginia Tech's Kentland Farm. We report correlations between concentrations of \textit{Fusarium} with the local movement of ATBs determined from archived meteorological data.   

Statistical scale invariance in satellite observations of water vapor mixing ratio from the Atmospheric Infrared Sounder

Statistical scale invariance appears almost ubiquitously in fluid dynamical systems and often characterizes universal aspects of particular classes of flows. Perhaps the most famous instances of statistical scale-invariance in the atmospheric sciences are the Kolmogorov's -5/3 and Charney's -3 variance spectra for passive scalars and velocity in 3D and quasi-geostrophic turbulence respectively. Parameterizations of radiative transfer and clouds in global climate models (GCMs) depend on proper characterization of the spatial statistics of water vapor, which is not a passive scalar. Empirical investigations of the scale dependence of water vapor statistics have largely depended on aircraft observations, which are limited in spatial and temporal extent. We will present results from a structure function analysis of statistical scale invariance of water vapor mixing ratio fields as observed by the Atmospheric Infrared Sounder (AIRS) onboard NASA's Aqua satellite and discuss the application of these results to the GCM parameterization problem.   

Analysis and numerical simulation of a laboratory analog of radiatively induced cloud-top entrainment

Numerical simulations using the One-Dimensional-Turbulence model are compared to water-tank measurements\footnote{B. J. Sayler and R. E. Breidenthal, J. Geophys. Res. \textbf{103} (D8), 8827 (1998).} emulating convection and entrainment in stratiform clouds driven by cloud-top cooling. Measured dependences of the entrainment rate on Richardson number, molecular transport coefficients, and other experimental parameters are reproduced. Additional parameter variations suggest more complicated dependences of the entrainment rate than previously anticipated. A simple algebraic model indicates the ways in which laboratory and cloud entrainment behaviors might be similar and different.   

Laboratory simulations of cumulus cloud flows explain the entrainment anomaly

In the present laboratory experiments, cumulus cloud flows are simulated by starting plumes and jets subjected to off-source heat addition in amounts that are dynamically similar to latent heat release due to condensation in real clouds. The setup permits incorporation of features like atmospheric inversion layers and the active control of off-source heat addition. Herein we report, for the first time, simulation of five different cumulus cloud types (and many shapes), including three genera and three species (WMO Atlas 1987), which show striking resemblance to real clouds. It is known that the rate of entrainment in cumulus cloud flows is much less than that in classical plumes - the main reason for the failure of early entrainment models. Some of the previous studies on steady-state jets and plumes (done in a similar setup) have attributed this anomaly to the disruption of the large-scale turbulent structures upon the addition of off-source heat. We present estimates of entrainment coefficients from these measurements which show a qualitatively consistent variation with height. We propose that this explains the observed entrainment anomaly in cumulus clouds; further experiments are planned to address this question in the context of starting jets and plumes.   

The evaporatively driven cloud-top mixing layer

Turbulent mixing caused by the local evaporative cooling at the top cloud-boundary of stratocumuli will be discussed. This research is motivated by the lack of a complete understanding of several phenomena in that important region, which translates into an unacceptable variability of order one in current models, including those employed in climate research. The cloud-top mixing layer is a simplified surrogate to investigate, locally, particular aspects of the fluid dynamics at the boundary between the stratocumulus clouds and the upper cloud-free air. In this work, direct numerical simulations have been used to study latent heat effects. The problem is the following: When the cloud mixes with the upper cloud-free layer, relatively warm and dry, evaporation tends to cool the mixture and, if strong enough, the buoyancy reversal instability develops. This instability leads to a turbulent convection layer growing next to the upper boundary of the cloud, which is, in several aspects, similar to free convection below a cold horizontal surface. In particular, results show an approximately self-preserving behavior that is characterized by the molecular buoyancy flux at the inversion base, fact that helps to explain the difficulties found when doing large-eddy simulations of this problem using classical subgrid closures.   

A study of the interactions between turbulence and small inertial droplets

Understanding the dynamics of particles in turbulent flows is important to many engineering and environmental problems including spray atomization and cloud droplet growth and precipitation. Specifically, we have studied the effect of turbulence on droplet collision-coalescence in an effort to clarify its role in the process of warm rain formation. The hypothesis that turbulence-induced-collisions can explain the size gap between the limit of condensational growth and the onset of gravitational collisions and sedimentation is supported by our measurements and analysis. Wind tunnel experiments were used to study the evolution of water droplets in homogeneous, isotropic, slowly decaying grid turbulence. Droplets between 1 and 120 $\mu m$ were injected into the wind tunnel and their diameter, position and velocity were measured at different distances downstream by Phase Doppler Particle Analysis (PDPA). Statistics of the radial distribution function (RDF), relative velocity distribution and settling velocity have been produced and analyzed. They will be compared to the same statistics computed from 3D hybrid direct numerical simulations (DNS) at similar $Re$. High-speed visualizations of the droplet dynamics will be explored in an effort to understand and quantify coalescence efficiency.   

Enhancement of coalescence due to droplet inertia in turbulent clouds

In the Explicit Mixing Parcel Model of mixing effects on cloud-droplet evolution, turbulent advection of fluid is implemented by permutations (``triplet maps'') of the fluid cells in chosen segments of the 1D domain, each representing an individual eddy. This captures motions as small as the smallest turbulent eddies (Kolmogorov microscale), but there is important droplet-inertia phenomenology, such as droplet clustering that increases droplet collision rates, at much smaller scales. We have developed and demonstrated a 3D triplet map for droplets (and an associated drag-law representation) that captures clustering behaviors at small Stokes numbers St (such as those of cloud droplets). There is excellent agreement between our results (for radial distribution functions and collision kernels) at small St and direct-numerical-simulation (DNS) results that omit gravity, and good agreement with DNS results that include gravity. We are currently testing an extension of our model that is intended to broaden its applicability to higher St, and we are using a collision-detection algorithm to simulate coalescence.   

Experimental study of droplet condensational growth in a wet/dry turbulent mixing layer

Droplet condensational growth has been experimentally studied as a function of supersaturation and turbulence. The experimental setup consists of two coaxial round jets, with a moist, warm inner round jet, surrounded by an annular sheath of cold air. Supersaturation is controlled by the relative humidity and temperature of the two streams, while the turbulence levels are determined by the shear at the interface between the two jets. Turbulent mixing of the supersaturation field, as well as particle clustering in regions of low vorticity, are expected to lead to large inhomogeneities in the growth rate and resulting discrepancies in the particle size distribution. Analysis of large data sets of droplet behavior at different locations along the mixing layer and under different conditions of water vapor concentration, temperature and mixing intensity are used to understand the dynamics of the interaction between the turbulent eddies and condensational droplet growth. A Phase Doppler Particle Analyzer (PDPA) is used to collect statistics of droplet growth, and to characterize the turbulent velocity field. The goal of this study is to improve the understanding of the competition between turbulent mixing replenishing the water content around the droplet and depletion of the vapor surrounding the droplet by condensation, and to develop quantitative models that can be applied to the problem of edge mixing in clouds.