Better Simulations of Hurricane Winds in Urban Canopies

BACKGROUND

The myriad ways that weather and climate affect society motivate research that increases our understanding of natural hazards, and that provides opportunities and options to inform society’s responses to those hazards.  For a large fraction of the world, hurricanes are a perennial threat.  When they make landfall over or near cities, extreme wind in the urban canopy can damage infrastructure and endanger lives over large areas, including well inland.

The goal of this project sponsored by the National Science Foundation is to use numerical simulations from Cloud Model 1 (CM1) at the large-eddy-scale and from the Weather Research and Forecasting (WRF) Model at the mesoscale to improve our fundamental understanding of, and ability to forecast, how buildings affect the intensity and heterogeneity of hurricane winds in the urban canopies of coastal cities.  Results from these numerical simulations will contribute to technology that provides authorities with more skillful short-term predictions of landfalling hurricanes; more detailed, accurate projections of wind damage; and more reliable guidance on long-term vulnerability from hurricane-force winds in the future.  This project focuses only on wind damage, notwithstanding the dangers of storm surge, flooding rainfall, and so forth.

Models

Weather Research and Forecasting (WRF) Model

The WRF Model is a long-established industry standard for numerical weather analysis and prediction in operations and research.   The model code is open source.  It was developed by a group of partners including NCAR, the National Oceanic and Atmospheric Administration, the Air Force Weather Agency, the Federal Aviation Administration, and the university community.  The model is used across many scales, from global to microscale.  The WRF Model includes a Building Effect Parameterization (BEP), which estimates the aggregate influences from horizontal (roofs and roads) and vertical (walls) surfaces on momentum, heat, and turbulence kinetic energy in the atmosphere.  BEP was optimized for wind speeds far lower than are found in hurricanes.  Improving the BEP’s performance for hurricane-force wind is a key focus of the project.

Cloud Model 1 (CM1)

CM1 is an open-source, non-hydrostatic, three-dimensional, prognostic atmospheric model.  It has been used extensively for idealized simulations, at the scales of large eddies to the fine mesoscale (grid intervals of orders 1–1000 m), of moist atmospheric phenomena, especially individual thunderstorms, complexes of thunderstorms, and hurricanes.  High-fidelity simulations with CM1 are being used as a standard of reference for the more course simulations being conducted with the WRF Model.

SELECTED ACCOMPLISHMENTS IN FY2018

Buildings in CM1

Figure 1.  CM1 simulations and observations of flow around a solid cube.  In the panels on the left (grid marked in km), horizontal wind speed (top) and vertical wind speed (bottom) from CM1 are shaded according to the color bar (ms-1).  The cube is in black.  In the panels on the right, profiles of normalized horizontal wind speed from CM1 (black) are plotted against experimental results from a wind tunnel (red) up wind of the cube (left profiles) and over the cube (right profiles).
Figure 1.  CM1 simulations and observations of flow around a solid cube.  In the panels on the left (grid marked in km), horizontal wind speed (top) and vertical wind speed (bottom) from CM1 are shaded according to the color bar (ms-1).  The cube is in black.  In the panels on the right, profiles of normalized horizontal wind speed from CM1 (black) are plotted against experimental results from a wind tunnel (red) up wind of the cube (left profiles) and over the cube (right profiles).

In FY2018, members of the project team at NCAR introduced into CM1 an immersed boundary method (IBM) for representing vertical faces (e.g., walls of buildings) in the model.  IBM is often used for studying detailed flow patterns in urban environments and over steep terrain.  To validate the results, NCAR compared CM1 simulations to results from wind-tunnel measurements of three-dimensional wind around solid objects in fully developed channel flow.  After adjusting surface drag coefficients in CM1, the simulations and results from the tunnel compare reasonably well, although further optimization is still possible (Figure 1). 

Observations of hurricane winds around buildings

NCAR has located at the Center for Severe Weather Research a dataset of observations from mobile Doppler radar that captured high-velocity wind around and over buildings as Hurricane Frances made landfall at Fort Pierce, FL in September 2004.  The team is using those data as another source of validation for CM1.

WRF Model simulations of real hurricanes

Members of the project team at the University of Miami are conducting real-case simulations of tropical cyclones affecting urban environments.  The work began with a review of all US hurricane landfalls in the last 20 years, as well as possible cases of hurricanes or typhoons making landfalls in other countries.  Based on this review, it was decided that Hurricane Wilma (2005) would be the first storm to be simulated; Wilma brought category 1 winds to a long stretch of developed land from Miami to West Palm Beach.  Other storms that remain under consideration are Hurricane Ike (2008) and Typhoon Ma-on (2004), which hit Tokyo, Japan.  Preliminary simulations were fairly successful in reproducing the track and intensity of Hurricane Wilma as it passed over southern Florida.  To make the landfall structure as realistic as possible, a vortex bogussing technique was applied to modify the initial location, strength, and size of the vortex so that the storm.

SELECTED KEY PLANS FOR FY2019

Improving the WRF Model’s simulated boundary layer

In FY2019 the project team will improve the BEP urban physics scheme in the WRF Model to see if our changes increase the accuracy of the simulated hurricane wind field over developed land.  To evaluate local effects, the team will compare the model output to observations such as time series from airport data (1-min intervals) and tower data (0.1-sec intervals).

Simulations of more historical hurricanes

With the improved BEP (described earlier) the team will execute more simulations of historical hurricanes, focusing on those for which we can obtain datasets of observations of wind amid and near buildings.

Archetypes of coastal cities and hurricanes

In the next step toward improving our understanding of the ranges of likely wind conditions in the urban canopy within different types of hurricanes making landfall over different types of coastal cities, the project team will begin defining idealized archetypal buildings and cities, and archetypal hurricanes.  The latter will comprise different sizes (small vs. large), strengths (weak vs. strong), and translational speeds (slow vs. fast).  We then will simulated a suite of idealized hurricane strikes on cities to find patterns in the results.