Tina Katopodes Chow

My current research interests are in performing large-eddy simulations (LES) of the atmospheric boundary layer, with a focus on the development and testing of new turbulence models and improved boundary conditions for flow over complex terrain (such as mountainous and urban areas).

Current projects include studies of boundary layer processes in steep valleys, turbulence in urban boundary layer flow, source inversion for contaminant dispersion in urban areas, and development of coupled models for improving the representation of land-atmosphere interactions, among others. The Environmental Fluid Mechanics at Berkeley page describes further research in our group.

Rica Enriquez, Megan Bela and Megan Daniels, installing soil moisture sensors in Owens Valley, CA.

Owens Valley, California is famous for many reasons, from battles over its water resources (the Los Angeles Aqueduct starts here), to the "Sierra Wave" and the daring flights of the Sierra Wave Project in the 1950s. Owens Valley was recently the site of the Terrain-induced Rotor Experiment (T-REX) field campaign in Spring 2006, which focused on understanding lee waves and atmospheric rotors generated by strong winds over the Sierra-Nevada mountains. The T-REX field campaign built upon the goals of the 1950's Sierra Wave Project but now using sophisticated remote sensing platforms and aircraft, combined with many detailed numerical models. We are performing numerical simulations of flow in Owens Valley to understand valley wind dynamics under quiescent and strongly forced wind conditions. Land-surface forcing is a primary focus of this work because valley winds exhibit strong sensitivity to land cover and soil moisture. Simulations are being compared to data collected in the field (including soil moisture data) to improve understanding of atmospheric flow over complex terrain. See Daniels et al. 2006 and 2008, Schmidli et al. 2009, and Grubisic et al. 2008 for details.

Wind farm at Altamont Pass, California.

Wind turbine micrositing, for operational wind power forecasting and for turbine design, require high-resolution simulations of atmospheric flow over complex terrain. We are developing large-eddy simulations (LES) for wind energy applications using the Weather Research and Forecasting (WRF) model, a community mesoscale model developed by NCAR. Grid nesting is used to refine the grid from mesoscale to LES scales which can adequately resolve terrain and turbulence in the atmospheric boundary layer. Grid nesting provides time-dependent lateral boundary conditions which incorporate changing weather conditions even in the smallest domain. Our group has improved WRF's LES capability by implementing an explicit filtering and reconstruction turbulence model (described below, following Chow et al. 2005) and an immersed boundary method (IBM) to accommodate complex terrain. This project is in collaboration with Julie Lundquist at Lawrence Livermore National Laboratory. See Lundquist JK et al. 2007 and 2009, Mirocha et al. 2007, Lundquist KA et al. 2007 and 2008 for details.

Composite plume showing 90% confidence intervals for concentration levels for flow in Oklahoma City. Observed concentrations are also shown as small colored squares.

The ability to determine the source of a contaminant plume in urban environments is crucial for emergency response applications. Locating the source based on downwind concentration measurements, however, is complicated by the presence of buildings which can divert flow in unexpected directions. High-resolution flow simulations are now possible for predicting plume evolution in complex urban geometries, where contaminant dispersion is affected by the flow around individual buildings. Using stochastic sampling algorithms and Bayesian inference together with a high-resolution CFD model, we have developed methods to reconstruct an atmospheric release event to determine the plume source and release rate in urban environments based on point measurements of concentration. Event reconstruction algorithms are applied first for flow around a prototype isolated building (a cube), and then using observations and flow conditions from Oklahoma City during the Joint URBAN 2003 field campaign. Stochastic sampling methods (Markov Chain Monte Carlo) are used to extract likely source term parameters, taking into consideration measurement and forward model errors. See Chow et al. 2006 and 2008 for details.

Vertical profiles of normalized mean horizontal wind speed (<U>) for neutral boundary layer simulations using an eddy-viscosity closure (Smagorinsky) and the dynamic reconstruction model (DRM) with a near-wall stress model. The traditional (Smagorinsky) closure model overestimates the flow by 10% at heights z/H ~ 0.1.

The equations for large-eddy simulation (LES) are obtained by applying a low-pass filter to the Navier–Stokes equations. This filtering operator divides the flow into so-called resolved and subfilter-scale (SFS) motions. When the equations are solved on a discrete grid, a discretization operator is applied to the equations as well, which further divides the turbulent flow field; the subfilter scales are divided into resolved SFS and unresolved SFS regions. The RSFS contribution can be theoretically reconstructed (e.g. using series expansions), and the SGS stress must be modeled. In this investigation, we study the influence of numerical errors on LES of turbulent channel flow, as well as the influence of the filtering approach and the reconstruction level on the turbulence models. The results demonstrate that improvements can be obtained for a given resolution and code by using explicit filtering and the dynamic reconstruction model (DRM). A similar approach is taken for LES of atmospheric boundary layer flow, again with significant improvement in comparisons with similarity theory (log velocity profile). In the atmospheric boundary layer, surface roughness becomes important, and its contribution must be included in the total stress distribution. These ideas are also extended for simulations of flow over an isolated hill in Scotland. See Gullbrand and Chow 2003, Chow et al. 2005, and Chow and Street 2004 and 2009 for details.

Water table depth in the Little Washita watershed in Oklahoma obtained from offline spin-up and used to initialize coupled groundwater-atmosphere simulations.

Both field observations and simulations indicate strong sensitivity of atmospheric dynamics to land-surface conditions, in particular surface soil moisture. We are developing a coupled model which connects subsurface, surface, and atmospheric dynamics to allow a complete representation of the hydrologic cycle. We can therefore capture feedbacks in the land-atmosphere system that occur through precipitation events, evapotranspiration, surface runoff, and infiltration. Use of a three-dimensional variably saturated/unsaturated groundwater model explicitly accounts for lateral moisture transport in the subsurface. The soil moisture distributions naturally correspond to variations in topography, soil types, and vegetation. A land-surface model provides the interface between the atmosphere and the subsurface, passing moisture and heat fluxes between the models. The results of our coupled simulations show the importance of the groundwater-atmosphere connection in determining soil moisture distributions and land-surface fluxes. The effects of realistic, spatially-varying soil moisture forcing on boundary layer development can be equal or greater than the effects from heterogeneous land-cover (soil and vegetation types), thus pointing to the need for improved soil moisture representations in current mesoscale atmospheric models. See Chow et al. 2006 and Maxwell et al. 2007 for details.

Results from simulations of flow in the Riviera Valley at 150 m horizontal resolution. Contours of along-valley flow in meters per second (red: up-valley, into the page, blue: down-valley, out of the page). Vectors show flow in the cross-valley direction (not all vectors are shown).

Increases in available computational power now allow high-resolution simulations of flow over complex terrain, but the appropriate numerical and physical parameters required by such simulations are not generally known. The influence of parameterizations such as those used for turbulence, soil moisture, solar radiation, surface roughness, the configuration of initial conditions, lateral boundary conditions, and the choice of numerical grids is highly situation dependent. This project investigates the steps necessary to achieve accurate large- eddy simulations of flow in highly complex terrain. Specifically, we examine the flow and temperature fields in the Riviera Valley, located in the Alps in southern Switzerland. The simulation results are verified through comparisons to surface and radiosonde observations in the Riviera Valley, obtained during the Mesoscale Alpine Programme (MAP) Riviera Project. We evaluate the model sensitivity to changes in parameterizations such as those listed above. The boundary layer processes in our simulated valley are also studied, including comparisons to aircraft flight data, descriptions of along-valley wind transitions and secondary cross-valley circulations, and a heat budget analysis. See Chow et al. 2005 and Weigel et al. 2005 for details.

Fotini Katopodes Chow

Research interests

*Environmental Fluid Mechanics at UC Berkeley*

Home Research Teaching Publications Students Awards Department website		Fotini Katopodes Chow Research interests My current research interests are in performing large-eddy simulations (LES) of the atmospheric boundary layer, with a focus on the development and testing of new turbulence models and improved boundary conditions for flow over complex terrain (such as mountainous and urban areas). Current projects include studies of boundary layer processes in steep valleys, turbulence in urban boundary layer flow, source inversion for contaminant dispersion in urban areas, and development of coupled models for improving the representation of land-atmosphere interactions, among others. The Environmental Fluid Mechanics at Berkeley page describes further research in our group. See the environmental engineering alumni newsletter or the College of Engineering Lab Notes for a non-technical overview of some aspects of my work. Selected projects (see this page too) The Terrain-induced Rotor Experiment - Owens Valley Large-eddy simulation for wind energy applications Source inversion for urban contaminant dispersion Explicit filtering and turbulence modeling for LES Coupled modeling of land-atmosphere interactions Flow over complex terrain - Riviera Valley *The Terrain-induced Rotor Experiment - Owens Valley* Rica Enriquez, Megan Bela and Megan Daniels, installing soil moisture sensors in Owens Valley, CA. Owens Valley, California is famous for many reasons, from battles over its water resources (the Los Angeles Aqueduct starts here), to the "Sierra Wave" and the daring flights of the Sierra Wave Project in the 1950s. Owens Valley was recently the site of the Terrain-induced Rotor Experiment (T-REX) field campaign in Spring 2006, which focused on understanding lee waves and atmospheric rotors generated by strong winds over the Sierra-Nevada mountains. The T-REX field campaign built upon the goals of the 1950's Sierra Wave Project but now using sophisticated remote sensing platforms and aircraft, combined with many detailed numerical models. We are performing numerical simulations of flow in Owens Valley to understand valley wind dynamics under quiescent and strongly forced wind conditions. Land-surface forcing is a primary focus of this work because valley winds exhibit strong sensitivity to land cover and soil moisture. Simulations are being compared to data collected in the field (including soil moisture data) to improve understanding of atmospheric flow over complex terrain. See Daniels et al. 2006 and 2008, Schmidli et al. 2009, and Grubisic et al. 2008 for details. *Large-eddy simulation for wind energy applications* Wind farm at Altamont Pass, California. Wind turbine micrositing, for operational wind power forecasting and for turbine design, require high-resolution simulations of atmospheric flow over complex terrain. We are developing large-eddy simulations (LES) for wind energy applications using the Weather Research and Forecasting (WRF) model, a community mesoscale model developed by NCAR. Grid nesting is used to refine the grid from mesoscale to LES scales which can adequately resolve terrain and turbulence in the atmospheric boundary layer. Grid nesting provides time-dependent lateral boundary conditions which incorporate changing weather conditions even in the smallest domain. Our group has improved WRF's LES capability by implementing an explicit filtering and reconstruction turbulence model (described below, following Chow et al. 2005) and an immersed boundary method (IBM) to accommodate complex terrain. This project is in collaboration with Julie Lundquist at Lawrence Livermore National Laboratory. See Lundquist JK et al. 2007 and 2009, Mirocha et al. 2007, Lundquist KA et al. 2007 and 2008 for details. *Source inversion for urban contaminant dispersion* Composite plume showing 90% confidence intervals for concentration levels for flow in Oklahoma City. Observed concentrations are also shown as small colored squares. The ability to determine the source of a contaminant plume in urban environments is crucial for emergency response applications. Locating the source based on downwind concentration measurements, however, is complicated by the presence of buildings which can divert flow in unexpected directions. High-resolution flow simulations are now possible for predicting plume evolution in complex urban geometries, where contaminant dispersion is affected by the flow around individual buildings. Using stochastic sampling algorithms and Bayesian inference together with a high-resolution CFD model, we have developed methods to reconstruct an atmospheric release event to determine the plume source and release rate in urban environments based on point measurements of concentration. Event reconstruction algorithms are applied first for flow around a prototype isolated building (a cube), and then using observations and flow conditions from Oklahoma City during the Joint URBAN 2003 field campaign. Stochastic sampling methods (Markov Chain Monte Carlo) are used to extract likely source term parameters, taking into consideration measurement and forward model errors. See Chow et al. 2006 and 2008 for details. *Explicit filtering and turbulence modeling for LES* Vertical profiles of normalized mean horizontal wind speed (<U>) for neutral boundary layer simulations using an eddy-viscosity closure (Smagorinsky) and the dynamic reconstruction model (DRM) with a near-wall stress model. The traditional (Smagorinsky) closure model overestimates the flow by 10% at heights z/H ~ 0.1. The equations for large-eddy simulation (LES) are obtained by applying a low-pass filter to the Navier–Stokes equations. This filtering operator divides the flow into so-called resolved and subfilter-scale (SFS) motions. When the equations are solved on a discrete grid, a discretization operator is applied to the equations as well, which further divides the turbulent flow field; the subfilter scales are divided into resolved SFS and unresolved SFS regions. The RSFS contribution can be theoretically reconstructed (e.g. using series expansions), and the SGS stress must be modeled. In this investigation, we study the influence of numerical errors on LES of turbulent channel flow, as well as the influence of the filtering approach and the reconstruction level on the turbulence models. The results demonstrate that improvements can be obtained for a given resolution and code by using explicit filtering and the dynamic reconstruction model (DRM). A similar approach is taken for LES of atmospheric boundary layer flow, again with significant improvement in comparisons with similarity theory (log velocity profile). In the atmospheric boundary layer, surface roughness becomes important, and its contribution must be included in the total stress distribution. These ideas are also extended for simulations of flow over an isolated hill in Scotland. See Gullbrand and Chow 2003, Chow et al. 2005, and Chow and Street 2004 and 2009 for details. *Coupled modeling of land-atmosphere interactions* Water table depth in the Little Washita watershed in Oklahoma obtained from offline spin-up and used to initialize coupled groundwater-atmosphere simulations. Both field observations and simulations indicate strong sensitivity of atmospheric dynamics to land-surface conditions, in particular surface soil moisture. We are developing a coupled model which connects subsurface, surface, and atmospheric dynamics to allow a complete representation of the hydrologic cycle. We can therefore capture feedbacks in the land-atmosphere system that occur through precipitation events, evapotranspiration, surface runoff, and infiltration. Use of a three-dimensional variably saturated/unsaturated groundwater model explicitly accounts for lateral moisture transport in the subsurface. The soil moisture distributions naturally correspond to variations in topography, soil types, and vegetation. A land-surface model provides the interface between the atmosphere and the subsurface, passing moisture and heat fluxes between the models. The results of our coupled simulations show the importance of the groundwater-atmosphere connection in determining soil moisture distributions and land-surface fluxes. The effects of realistic, spatially-varying soil moisture forcing on boundary layer development can be equal or greater than the effects from heterogeneous land-cover (soil and vegetation types), thus pointing to the need for improved soil moisture representations in current mesoscale atmospheric models. See Chow et al. 2006 and Maxwell et al. 2007 for details. *Flow over complex terrain - Riviera Valley* Results from simulations of flow in the Riviera Valley at 150 m horizontal resolution. Contours of along-valley flow in meters per second (red: up-valley, into the page, blue: down-valley, out of the page). Vectors show flow in the cross-valley direction (not all vectors are shown). Increases in available computational power now allow high-resolution simulations of flow over complex terrain, but the appropriate numerical and physical parameters required by such simulations are not generally known. The influence of parameterizations such as those used for turbulence, soil moisture, solar radiation, surface roughness, the configuration of initial conditions, lateral boundary conditions, and the choice of numerical grids is highly situation dependent. This project investigates the steps necessary to achieve accurate large- eddy simulations of flow in highly complex terrain. Specifically, we examine the flow and temperature fields in the Riviera Valley, located in the Alps in southern Switzerland. The simulation results are verified through comparisons to surface and radiosonde observations in the Riviera Valley, obtained during the Mesoscale Alpine Programme (MAP) Riviera Project. We evaluate the model sensitivity to changes in parameterizations such as those listed above. The boundary layer processes in our simulated valley are also studied, including comparisons to aircraft flight data, descriptions of along-valley wind transitions and secondary cross-valley circulations, and a heat budget analysis. See Chow et al. 2005 and Weigel et al. 2005 for details.