Progress in Urban LandUse Modeling for MM5 and WRF Models

1
Progress in Urban Land-Use Modeling for MM5 and
WRF Models

• Fei Chen1, Yubao Liu1, Hiroyuki Kusaka1 4, Mukul
  Tewari1, Jian-Wen Bao2, Chun Fung Lo3, and Kai
  Hon Lau3
• 1 National Center for Atmospheric Research,
  Boulder, CO
• 2 National Oceanic and Atmospheric Administration
  (NOAA), Boulder, CO
• 3 Hong Kong University of Science and Technology,
  Hong Kong
• 4 Central Research Institute of Electric Power
  Industry (CRIEPI), Abiko, Japan
• _at_ AHPCRC Workshop on Mesoscale and Microscale
  Meteorological Modeling for Military Applications
  
•   26 May 2004, Jackson State University

2
Urban Landuse Modeling for High-Resolution (1-5
km) NWP Models

• Goals to provide
• more accurate weather forecasts (near surface and
  PBL structures) for urban regions
• meteorological fields (initial and boundary
  conditions) for air quality and dispersion
  models, and CFD models
• Challenge
• Specification of urban landuse
• Degree of complexity of urban modeling
• Initializing state variables of urban models

3
Land surface and urban modeling and assimilation
system
snow
Leaf area index
Vegetation type
Urban type
Vegetation cover
Soil texture
Terrain
Obs. Precipitation Radiation, T, Q, U, V
High resolution land data assimilation system
(HRLDAS)
Soil moisture, soil temperature, snow cover,
canopy water, wall/roof/road temperature
Noah land surface model, Urban canopy model
Boundary layer parameterization
Coupled mode
4
Urban Modeling Approaches

• In-building scale modeling (typical grid
  resolution lt 1 meter forecast time seconds to
  minutes)
• Single to many building scale modeling (typical
  grid resolution 1-100 meter forecast time
  minutes to a few hours)
• Urban-canopy model parameterization (gt 100
  meters forecast time many hours)
• Simple bulk parameterization
• Urban canopy model

5
Simple Parameterization of Urban Effects in Noah
LSM for JUT 2003 (OKC) RTFDDA

• Large roughness length
• turbulence generated by roughness elements
• drag due to buildings
• Low surface albedo
• radiation trapping
• Large thermal capacity and thermal conductivity
• heat storage in soil
• Low evaporation

6
JUT 2003 (OKC) RTFDDA URBAN SIMULATIONAverage of
9 clear-sky days in July 2003, at 06Z, 1.5-km grid
2-m Temperature and 10-m Winds
Surface Sensible flux ( W/m2 )
Negative Flux
29.5
Positive Flux
7
JUT 2003 (OKC) RTFDDA URBAN SIMULATIONCase
examples Nocturnal PBL 06Z June 24
PBL Height (m)
Wind Speed (m/s LLJ)
B
H(km)
A
URBAN
B
A
10 km
8
New MM5 Landuse in Pearl River Delta, China
Domain 1 40.5km Domain 2 13.5km Domain 3 4.5km
9
New MM5 Landuse (1-km) in Pearl River Delta, China

• Based on 30-m local landuse map
• More urban areas
• Better river network and
• water bodies

Original USGS Landuse
New Landuse
10
(Lowest sigma level)
Magenta Observation Blue MM5/LSM
forecast YellowMM5 forecast
LSM (with urban modification) seems to better
reproduce the land sea breeze circulation in HK
11
Deficiencies of the simple approach

• Poor representation of radiative exchange in
  complex urban canyon geometries
• Lack of representation of heat storage in roofs
  and walls and their exchange with atmosphere
• Lack of detail with respect to urban landuse
  differentiation
• Lack of wind variation within canopy layer

12
Our Effort to Develop a Coupled
Land-Surface/Urban-Canopy Model

• Couple Noah LSM in WRF with a single layer
  urban-canopy model (UCM), based on Kusaka et al,
  2001 similar to Masson 2000 Martilli and Rotach
  2002 Brown et al. 2000, including
• 2-D urban geometry (orientation, diurnal cycle of
  solar azimuth), symmetrical street canyons with
  infinite length
• Shadowing from buildings and reflection of
  radiation
• Multi-layer roof, wall and road models

13
Single-Layer Urban Canopy Model Shadow and
Radiation Trapping
14
30-meter resolution Landuse for the Houston Area
15
UCM Produce More Pronounced Nocturnal Heat Island
Lowest model level ? at 12 UTC 26 Aug 2000
(2-km WRF)
With Urban Canopy Model and new urban landuse
Simple urban treatment Old urban landuse map
16
Wall surface temperature responsible for a large
part of nocturnal urban heterogeneity
Road surface
Roof surface
Wall surface
Surface temperature at 06 UTC 26 Aug 2000
17
Simulations with UCM Enhance Strength of Sea
Breezelowest-model level wind speed at 21 UTC 25
Aug 2000
With Urban Canopy Model and new urban landuse
Simple urban treatment Old urban landuse map
18
Reflection on the Current Work

• The new capability of urban-canopy modeling
• Provide detailed urban heat island
• Enhance mesoscale model forecasted wind and
  thermal structure over urban area
• Improve input for air quality and dispersion
  model
• Specification of urban landuse is critical
• Initialization will remain a major challenge
• High-resolution land data assimilation (HRLDAS)
  needs to incorporate UCM component

19
Urban Effects on Airflow and Meteorology in
Context of Mesoscale NWP Modeling

• Dynamic effects
• Intense shear layer at the top of the canopy
• Development of wake diffusion generated by
  roughness elements
• Drag due to buildings (pressure differences
  across individual roughness elements)
• Thermal effects
• Differential heating/cooling of sunlit/shaded
  surfaces, radiation trapping, heat storage in
  buildings ? urban heat island effect
• CCN and cloud formation

20
Single-Layer Urban Canopy Model Temperatures
and Thermal Transfer
Temperature are explicitly calculated for roof,
wall, and road surface, and for urban canyon
21
WRF Sensible heat (Wm-2) on 2-kmat 18 UTC 25 Aug
2000
With Urban Canopy Model new urban landuse
Simple urban treatment Old urban landuse map
200 km
22
Sensible heat (Wm-2) at 06 UTC 26 Aug 2000
With Urban Canopy Model and new urban landuse
Simple urban treatment Old urban landuse map
23
JUT 2003 (OKC) RTFDDA URBAN SIMULATIONCase
examples Day-time PBL 18Z July 5, 2003
W ( m / s )
PBL Height ( m )
H(km)
updraft
A
B
10 km
24
JUT 2003 (OKC) RTFDDA URBAN SIMULATIONLand use
and water body on Domain 4 (1.5-km) defined by
USGS (1994) Terra MODIS (2002) landuse data
Aerial picture
7 larger urban area based on MODIS landuse
25
Dynamic Effects
Thermal effects