Surface-Atmosphere Interactions (page 1 of 2)

• Page 1 (you are here)
• Land-surface interactions
• Ocean-atmosphere interaction at turbulence- and mesoscales

• Page 2
• Chemical transports and transformations
• Wildfire research

land-surface interactions (top)

International H₂O Project (IH₂OP-2002) surface characteristics

Peggy LeMone, Diane Strassberg (MMM student visitor), and Joe Alfieri (RAP student assistant) compared the normalized differential vegetation index (NDVI) from hand-held sensors at the ten IH2OP surface flux stations to corresponding values from the University of Wyoming King Air Aircraft. They used aircraft videos to verify that the aircraft was over the correct land cover for comparison. This procedure proved robust, with early-season difficulties in distinguishing winter wheat from grasses resolved through use of post-harvest videos. In addition, landmark identification confirmed that corrected aircraft location was accurate to within the researchers’ ability to relate video images to aircraft location.

Atmospheric response of spatial variations of soil moisture

Jielun Sun, in collaboration with Larry Mahrt and Dean Vickers (both of Oregon State University), Thomas Jackson (USDA/ARS Hydrology), Ian MacPherson (National Research Council, Canada), Paul Houser (NASA Goddard Space Flight Center), and Eleanor Burke (University of Arizona), investigated the response of atmospheric moisture flux to temporal and spatial variations in soil moisture and spatial vegetation heterogeneity (http://www.mmm.ucar.edu/science/abl/sgp/sgp.html). Sun generalized the commonly used parameterization for estimating evapotranspiration with consideration of soil moisture and surface type. Using the Southern Great Plains (SGP) dataset, Sun found that evapotranspiration depends not only on soil moisture and surface type, but also on atmospheric conditions for ventilation.

Response of the PBL to land-surface heterogeneity

Edward Patton (MMM visitor, Pennsylvania State University), Peter Sullivan, and Chin-Hoh Moeng used their coupled PBL/land-surface large-eddy simulation code to study the planetary boundary layer (PBL) response to large-scale soil moisture heterogeneity (with scales ranging from 1-18 times the PBL height). In the presence of heterogeneity, the atmospheric transport of moisture varies with the initial atmospheric state, wet versus dry. In both situations, land-surface heterogeneity induces organized motions that scale with the heterogeneity. Surface fluxes respond asymmetrically in that they are minimum near the center of the heterogeneity and maximum at about one-fifth the way into the dry soil region. Phase-averaged statistics also reveal this skewed distribution; intense vertical motion is found in a narrow band centered over the dry soil, while the descending leg of the circulation is weaker and extends over a area greater than the wet soil region (see Figure 34).

Figure 34: Normalized phase-averaged horizontal velocity (up/w*, top) and vertical velocity (wp/w*, bottom) as a function of x/l and z/zi for the case with l/zi = 4. The dashed-line is the phase-averaged boundary layer depth, zip; dotted contours represent negative phase-correlated values; and the hatches demarcate the x/l extent of the wet soil.

The horizontal motions (associated with the circulations) also show an asymmetric response. Depending on the moisture state of the overlying atmosphere, time-averaged measurements at a point can incorrectly estimate the total vertical moisture flux by up to 60%. The error varies with the height and location of the measurement station in the region of heterogeneity (see Figure 35).

Figure 35: Vertical profiles of normalized vertical water vapor mixingratio flux for a case with a dry PBL over top of heterogeneitythat is about four times the height of the PBL. The greenprofile represents the the total flux. The blue line representscontribution to the total flux by the background turbulence (or the non-organized component of the turbulence), and the hatchmarks demarcate plus and minus one standard deviation of thisbackground component from the y-average. The red line isthe time-average at each x/l and z/zi location at y =Ly/2. The (top-left, top-right, lower-left, lower-right) panels represent locations x/\lambda = (0 or 1, 0.25, 0.5, 0.75).

Investigation of impacts of cold land transition for global climate change

Sun participated in the Fluxes over Snow Surfaces (FLOSS) experiment that focused on energy transfer over snow-covered cold land (http://www.mmm.ucar.edu/science/abl/floss/). This effort was organized by Larry Mahrt (Oregon State University) and was sponsored by the NASA Hydrology program. Sun designed an observation plan for measuring radiative flux divergence close to the ground in the presence of a snow layer. She contributed thermocouples to the observational study and measured temperature distributions around sage shrubs, in collaboration with Steve Oncley, Tony Delany, and Steve Semmer (all of ATD) and Larry Mahrt, Dean Vickers, and Richard Cuenca (all of Oregon State University). These results will shed light on the role of sage shrubs in energy transfer during snow-covered freezing periods and also lead to better understanding of scenarios related to the global warming associated with snow and frozen-land melting. This research effort is an interdisciplinary study relating ecology, micrometeorology, and hydrology.

CO2 transport over complex terrain

Jeilun Sun and Sean Burns, in collaboration with Steve Oncley, Tony Delany, Britt Stephens, and Teresa Campos (all of ATD), and Alex Guenther and Andrew Turnipseed (both of ACD), Russell Monson (University of Colorado), and Dean Anderson (USGS) led an investigation of CO2 transport over complex terrain at Niwot Ridge, Colorado in September 2002 (http://www.mmm.ucar.edu/science/abl/forest/). The research was sponsored by the NCAR Director's Opportunity Fund and is now a research focus of the NCAR Biogeoscience Initiative. In addition, this research concept has led to a successful NSF proposal for a full-scale study of CO2 transport over complex terrain. Using data from the 2002 field campaign, Sun and Burns found that the spatial distribution of CO2 is sensitive to major steep slopes, and also to small gullies embedded in steep slopes (Figure 36).

Figure 36: Investigation of CO2 transport over complex terrain at Niwot Ridge showed that spatial distribution of CO2 depends on radiative energy transfer and atmospheric wind gusts. The results provide critical information on the many ongoing long-term observational programs, which currently ignore horizontal transport of CO2 in their CO2 budgets.

The spatial distribution of CO2 depends on radiative energy transfer and atmospheric wind gusts. The results provide critical information on the many ongoing long-term observational programs, which currently ignore horizontal transport of CO2 in their CO2 budgets.

LES of stable boundary layers

Sullivan and Moeng participated in the first intercomparison of LES solutions of stable planetary boundary layers organized by the GEWEX atmospheric boundary layer study (GABLS). The long-term goal of this international working group is to improve the parameterization of stable (nocturnal) PBLS in large-scale numerical weather prediction codes (see the May issue of GEWEX news posted on http://www.gewex.org/gewex_nwsltr.html). This first intercomparison case focused on the idealized Arctic boundary layer described by Kosovic and Curry (JAS 2000). Nine-hour simulations with resolutions varying from 643 to 2003 were generated using the two-part subgrid scale model described in Sullivan et al. (1994). Sullivan and Moeng's results, along with those from the nine groups that participated in the intercomparison, are described at http://metresearch.net/gabls/index.shtml. The findings show that LES is capable of simulating stratified surface layers, but the results from the different groups are highly variable depending on the grid resolution and subgrid-scale model. High resolution LES, run with two-m spacing (2003 grid points see figure 36) tend to converge, but still exhibit surprising differences because of varying subgrid scale parameterizations used in the surface layer. Sullivan and Moeng plan to utilize results from HATS to develop an improved subgrid parameterization and thus enhance the fidelity of LES predictions for stable surface layers.

Figure 37: Snapshot of the instantaneous vertical velocity (w) field in x-z, y-z and x-y planes from an LES of a stable PBL with 2m resolution illustrating how w is dominated by small scale, intermittent motions. In this simulation a persistent low-level jet develops at approximately z = 200m above the surface. Color contours are in m/s and the size of the computational domain is 400x400x400m.

Energy transfer under nocturnal stable conditions

Sun, in collaboration with Sean Burns, Donald Lenschow, Tom Horst (ATD), Tony Delany (ATD), and Steve Oncley (ATD) continued analyzing data collected from the Cooperative Atmosphere-Surface Exchange Study-1999 (CASES-99, http://www.mmm.ucar.edu/science/abl/cases/cases.html). Sun focused on the nocturnal heat balance. This effort is the first attempt to measure precisely radiative flux divergence over a relatively deep stable layer with consideration of surface heterogeneity. She found that the radiative flux divergence decreases with height and is large during the early evening when the ground is still warm. The research is fundamental since it sheds light on the applicability of Monin-Obukhov similarity theory, which is used in all numerical models to relate surface fluxes and mean flowfields. This research is currently being supported by the Army Research Office, and additional funding has been awarded to support further data analysis.

Numerical simulation of sand dune evolution in severe winds

Predicting sediment transport and bed evolution in severe wind conditions depends on accurate prediction of flow past a complex boundary evolving with the flow itself. The geometrical complexity of the evolving interfaces, which either accommodate dynamically to the external/internal boundaries of the domain or convolute in response to internal flows, is per se a challenge to numerical modeling. Pablo Ortiz (University of Granada, Spain) and Piotr Smolarkiewicz developed a variant of the nonhydrostatic model EULAG that couples the internal flow with a lower boundary evolving in response to the sand saltation and/or dust storms (Fig. 38).

Figure 38: Two numerical simulations of a sand dune evolution. First three frames show potential-flow initial condition (including perspective view of the initial sand pile placed on a hard bed and isolines of vertical velocity in the vertical center plane and following the surface). Next three frames show the developed solution, whereas the subsequent three frames show a similar solution, but for the initial sand pile placed on a layer of sand.

The key prerequisite facilitating this development is the use of a time-dependent curvilinear coordinate transformation that accommodates rapid changes in the boundary shape. The nonoscillatory forward-in-time numerical technology of EULAG enables novel numerical designs that improve the accuracy, stability, and robustness of the traditional saltation models that govern the evolution of dunes.

Improvements in subgrid-scale modeling of plant canopy environments

Roger Shaw (University of California, Davis) and Patton (visitor, Pennsylvania State University) developed a subgrid-scale (SGS) model for large-eddy simulations of canopy flows that incorporates the clumpiness of leaf matter on the breakdown of resolved- and subgrid-scale motions into wake-scale (WS) motions in the lee of plants.

Figure 39: Schematic of the processes that convert resolved-scale kineticenergy to subgrid scale, wake scale, and ultimately to internalenergy. The numbers identify the various physical processesof energy exchange.

Figure 39 is a schematic of the processes responsible for energy transfer from the mean flow to internal energy and depicts the processes included in the new SGS model. In this study, it was found that inclusion of the breakdown from SGS motions into WS motions (Process 6 in the figure) is vital to the proper simulation of canopy flows as WS motions act to enhance dissipation of SGS energy. Ongoing collaborations with Monique Leclerc (University of Georgia) and Stathis Michaelides (Tulane University) are taking advantage of this new code. Collaborations with John Finnigan (CSIRO, Australia) and Shaw are further improving this SGS model by deriving the canopy influence on dissipation from first principles.

Subfilter scale motions in LES

Sullivan, Lenschow, Moeng, Thomas Horst (ATD), and Jeffrey Weil (MMM visitor, University of Colorado/CIRES) published a manuscript (JFM 482) describing their analysis of the field data collected from the Horizontal Array Turbulence Study (HATS). One of the major highlights from this work exposed the shortcomings of current subgrid scale modeling in situations where large amounts of momentum and scalar flux are transmitted by unresolved motions in LES. New anisotropic subgrid scale models for LES are needed near boundaries and in cases with strong stable stratification. The HATS datasets were shared with John Wyngaard (Pennsylvania State University), Chenning Tong (Clemson University), and Ronald Adrian (University Illinois, Urbana-Champagne).

Mechanisms of up-valley winds

Richard Rotunno in collaboration with Gabriele Rampanelli and Dino Zardi (both of the University of Trento, Italy) investigated the basic physical mechanisms governing the daytime evolution of up-valley winds in mountain valleys using a series of numerical simulations of thermally driven flow over idealized three-dimensional topography. As in previous studies of two-dimensional circulations in valleys, the heated valley side walls require a compensating subsidence in the valley core which brings potentially warmer air from the stable free atmosphere into the valley core. In the context of the three-dimensional valley-plain simulations, they found that the subsidence heating in the valley core is the main contributor to the valley-plain temperature contrast along the valley axis and, under the hydrostatic approximation, the pressure difference that accelerates the up-valley wind (see Figure 40, below).

Figure 40: The Physical mechanisms governing the daytime evolution of up-valley winds in mountain valleys using a series of numerical simulations of thermally driven flow over idealized three-dimensional topography.

The foregoing theory improves on the standard textbook explanation that assumes zero heat transfer between the free atmosphere and the air within the valley.

Moist neutral flow over a ridge

Although moist neutral flow over a ridge is a fairly common atmospheric condition in orographic-rain scenarios, relatively little is known about it from theory and modeling. Marcello Miglietta (Consiglio Nazionale della Ricerca, Italy) and Rotunno carried out numerical simulations of the orographic-flow modification occurring for a two-dimensional moist neutral flow over a ridge. If an initially saturated moist neutral flow were to remain everywhere saturated as it flows over an obstacle, then the expected solution would be the linear solution because the moist stability is small. However, Miglietta and Rotunno found that the numerical solutions indicate the development of areas of unsaturated air, with correspondingly larger values of local static stability. For certain parameter settings, the subsaturated zone may even propagate upwind (see Figure 41, below).

Figure 41: Numerical simulations of the orographic-flow modification occurring for a two-dimensional moist neutral flow over a ridge.

This internal switching from small to large values of static stability is an inherent nonlinearity of moist airflow, which has far-reaching consequences for understanding the orographic-flow modification and predictability of rainfall in this regime.

Lee vortex formation related to upstream blocking

Craig Epifanio (Texas A&M University) and Rotunno explored the basic fluid mechanics of orographic wake formation using simple scaling arguments and diagnostic vorticity inversion techniques. Their work pinpoints upstream blocking as the key process associated with lee-side mountain wake formation at the surface. Upstream blocking causes the descent of warm air over the lee slope to replace the colder surface air blocked upstream. The result is a temperature gradient along the lee slope that quickly collapses into a front. The colder surface air downstream of the front tends to propagate upstream into the lee-side warm anomaly, and for sufficiently large temperature difference, the flow behind the front stagnates and reverses to form a wake at the surface. In three dimensions the flow wraps around the lateral edges of the wake as suggested by the tilting of horizontal vortex lines to produce vertical vorticity (see Figure 42, below).

Figure 42: Basic fluid mechanics of orographic wake formation using simple scaling arguments and diagnostic vorticity inversion techniques pinpoints upstream blocking as the key process associated with lee-side mountain wake formation at the surface.

This work provides a conceptual model to describe surface wake formation downstream of mountains and clarifies the previously observed association between orographic wake formation and upstream blocking.

An all-scale anelastic model for geophysical flows: dynamic grid deformation

Joseph Prusa (Iowa State University) and Smolarkiewicz continued the development of an adaptive grid-refinement approach, embedded in the framework of a nonhydrostatic anelastic model for simulating a broad range of geophysical flows using nonoscillatory forward-in-time (NFT) numerical methods. The focus of the past year was the extension of the generalized coordinate technique into the viscous portions of the model. This required a rigorous development of the higher order differential operators for scalar and vector diffusion, as well as development of the accompanying boundary conditions along deformable surfaces. These developments have been coded into the numerical model, and are currently being tested in preliminary simulations. Additional refinements were also developed regarding the use of analytically specified transformations for grid adaptation, and are currently being tested in the context of idealized Held-Suarez climates with orographic forcing as well as oceanic flows in channels with irregular coastlines (Figure 43). It is expected that the former experiments will clarify to what extent local grid adaptation can enhance the accuracy of climate statistics. The experiments with irregular channel geometry open new opportunities for studies of valley and canyon flows.

Figure 43: Continued development of an adaptive grid-refinement approach, embedded in the framework of a nonhydrostatic anelastic model for simulating a broad range of geophysical flows using nonoscillatory forward-in-time (NFT) numerical methods, have been coded into the numerical model and are currently being tested in the context of idealized Held-Suarez climates with orographic forcing as well as oceanic flows in channels with irregular coastlines.

Multidimensional Positive Definite Advection Transport Algorithm (MPDATA): an edge-based unstructured-data formulation

Joanna Szmelter (Cranfield University, Srivenham, UK) and Smolarkiewicz developed the iterative upwind scheme MPDATA in a Finite Volume framework with an edge-based data structure and arbitrary hybrid mesh. MPDATA has proven successful in simulations of geophysical flows using single-block, structured, topologically rectangular meshes employing continuous time-dependent curvilinear mapping approach (see Smolarkiewicz and Margolin, J. Comput. Phys., 1998 for a comprehensive review). The motivation for the finite-volume formulation and the choice of unstructured meshes is to facilitate the use of MPDATA schemes for a wider range of applications involving complex geometries and/or inhomogeneous anisotropic flows where mesh adaptivity is advantageous. Their new development preserves the signature benefits of the standard, Cartesian-mesh MPDATA scheme; i.e., second-order accuracy, sign preservation, and a full multidimensionality free of directional-splitting errors (see Fig. 44).

Figure 44: Development of an iterative upwind scheme MPDATA in a Finite Volume framework with an edge-based data structure and arbitrary hybrid mesh has proven successful in simulations of geophysical flows using single-block, structured, topologically rectangular meshes employing continuous time-dependent curvilinear mapping approach.

Laboratory for internal-gravity-wave dynamics: the numericale equivalent to the quasi-biennial oscillation (QBO) analogue

Nils Wedi (European Center for Medium-range Weather Forecasting) and Smolarkiewicz extended the classical terrain-following coordinate transformation of Gal-Chen and Somerville (J. Comput. Phys., 1974) to a broad class of time-dependent curvilinear vertical domains. In particular, their development allows the simulation of stratified flows with intricate geometric, time-dependent boundary forcings, either at the top or bottom of the domain. They applied their mathematical/numerical framework to the direct numerical simulation of the celebrated laboratory experiment of Plumb and McEwan (J. Atmos. Sci, 1978) thereby creating the numerical equivalent to the laboratory quasi-biennial oscillation (QBO) analogue. The QBO represents a conspicuous example of a fundamental dynamical mechanism with challenging detail, but it is difficult to identify all the physical mechanisms from experimental evidence alone. A series of 2D and 3D simulations demonstrate their ability to reproduce the laboratory experiment (Fig. 45) and to associate reversing flow patterns as an entirely wave-driven phenomena.

Figure 45: A series of 2D and 3D simulations demonstrate the ability to reproduce the laboratory experiment of classical terrain-following coordinate transformation of Gal-Chen and Somerville, and to extend it to a broad class of time-dependent curvilinear vertical domains.

Their results enhance the confidence in the numerical approach and further elevate the importance of the laboratory setup for its fundamental similarity to the atmosphere, while allowing study of the principal atmospheric mechanisms and their numerical realization in a well-suited environment.

Ocean-atmosphere interaction at turbulence- and mesoscales (TOP)

A new method to measure oceanic waves from a moving platform

Jeilun Sun, in collaboration with Sean Burns, Douglas Vandemark (NASA Goddard Space Flight Center), and Mark Donelan (University of Miami), continued efforts at retrieving two-dimensional wave spectra using three laser altimeters on board the LongEZ aircraft. Sun found that a wavelet analysis method is able to retrieve wave propagation direction and wavenumber using datasets collected from the Shoaling Waves Experiment (SHOWEX, http://www.mmm.ucar.edu/science/abl/showex/showex.html) and Coupled Boundary-Layers/Air-Sea Transfer (CBLAST-low, http://www.mmm.ucar.edu/science/abl/cblast/). This wave analysis method provides a new capability to efficiently and economically measure waves and atmospheric turbulence gathered from a moving platform.

Air-sea interactions under weak wind conditions

Sun, in collaboration with Haflidi H. Jonsson (Navy Postgraduate School), Djamal Khelif (University of California, Irvine), Larry Mahrt and Dean Vickers (both of Oregon State University), participated in the ONR-sponsored CBLAST-low experiment off the coast of Martha's Vineyard, Massachusetts. Sun and colleagues planned the scientific missions and detailed flight plans for the Pelican research aircraft from the Navy Postgraduate School. Preliminary data analysis showed the existence of complex internal boundary layer structures associated with spatial variations in sea surface temperature (Figure 46).

Figure 46: Preliminary data analysis showing the existence of complex internal boundary layer structures associated with spatial variations in sea surface temperature, as studied in teh CBLAST-low experiment off the coast of Martha's Vineyard, Massachusetts.

LES of marine boundary layers with swell

Recent observations in the marine surface layer provide evidence that swell (fast moving waves) can generate spectacular behavior in the low-wind neutral planetary boundary layer (PBL). Low-level jets, positive (upward) vertical momentum flux, and negative mean gradient profiles are often observed. To model the impact of surface waves on the PBL, Peter Sullivan, James McWilliams (University of California, Los Angeles), and Chin-Hoh Moeng recently developed a new large-eddy simulation (LES) code with the capability of imposing a moving sinusoidal wave at its lower boundary. Preliminary simulations focused on PBLs with light winds, neutral stratification, and different surface conditions.

Figure 47: Snapshot of the instantaneous velocities over swell in neutrally stratified winds of 5ms-1 from LES. The wave age c/U10 > 2.2, where U10 is the wind at z=10m and c is the wave phase speed. The properties of the imposed wave are amplitude = 1.6m, waveslope = 0.1, wavelength = 100m, and wave phase speed c = 12.5m/s. Upper panel u and lower panel w. Note the strong positive correlation between u and w on the downwind side of each wave crest which leads to a positive upward momentum flux for z > 10m. The color bar, located in the lower right portion of each figure, shows the variation in units of m/s.

Flow visualization of the LES solutions (see Figure xx, above) shows that in the case of fast moving swell a coherent pattern of accelerated winds occurs downwind of each wave crest. At the same time, the vertical velocity is biased towards negative (positive) values upstream (downstream) of the wave crest, respectively. This organization induces positive vertical momentum flux that accelerates the near surface winds. The LES is thus able to replicate important features of a wave-driven boundary layer. The appearance of a low-level jet and vertically varying vertical momentum flux makes surface layer measurements dependent on wave state and vertical distance above the surface, and invalidates the use of Monin-Obukhov similarity theory most often used to predict air-sea fluxes.

Turbulence simulations of ocean boundary layers

Sullivan, James McWilliams (University of California, Los Angeles), and Ken Melville (Scripps Institute of Oceanography) are developing a stochastic model for the effects of breaking waves for use in turbulence-resolving simulations of the ocean boundary layer (OBL). The breaking parameterization, based on laboratory and field data, is being evaluated in direct numerical simulation (DNS) with large-eddy simulation (LES) implementations expected in the next fiscal year. They found that DNS with wave breaking generates long-lived vortices close to the water surface that are effective in energizing the surface region of oceanic boundary layers. A comparison of idealized oceanic boundary layers driven by constant current, constant stress, or a mixture of constant stress plus stochastic breakers provides evidence that intermittent stress transmission from breaking waves significantly alters the instantaneous flow patterns (see Movie 5, below) as well as the ensemble statistics. Analysis of the mean current profiles shows that breaking effectively increases the surface roughness zo by more than a factor of 30 (see also Figure 48, below).

Movie 5: Animation of the spanwise vorticity for an OBL simulation with 100% breaking waves. Note the difference in flow structure near the no-slip lower boundary and the uppoer boundary where breaking waves are present. Near the top of the water the usual turbulence production mechanism associated with low-speed streaks and hairpin vortices is disrupted by intermittent breaking. The highly energized ocean surface layer enhances turbulence mixing which leads to an increase in surface roughness. In the above animation, time is made dimensionless by a large scale turnover time.

Mouse over image to begin movie. Alternately, you may download the animation.

Figure 48: Vertical profiles of the mean current in wall coordinates with wave breaking. The dotted line is the linear curve u+ = -z+; the thick solid line is the log-linear law u+ = ln(-z+)/k + b valid for smooth walls with (k, b) = (0.41,5); and the thin solid line is the log-linear law for fully rough walls with b = (0.41,-2.1). The different simulations, indicated by dots, are: Couette flow (red); stress driven OBL with no breaking (pink); 25% breaking (green); 80% breaking (blue); 100% breaking (black). Notice how the effective surface roughness z0 deduced from the current profiles, increases with the amount of breaking.

Compared to a flow driven by a constant current, the extra mixing from breakers increases the mean eddy viscosity by more than a factor of ten near the water surface. Breaking waves alter the usual balance of production equals dissipation in the turbulent kinetic energy (TKE) budget. Turbulent boundary layers driven by constant current and constant stress (i.e. with no breaking) are also found to differ in fundamental ways.