Detect%20and%20Simulate%20Urban%20Climate%20Change

1
Detect and Simulate Urban Climate Change Via
EOS Observations and Land Surface Model
Dr. Menglin Jin, Department of
Meteorology University of Maryland, College
Park Dr. Christa D. Peters-Lidard NASA GSFC
April 20, 2004,
Acknowledgements Funded by NASA EOSIDS and NASA
DDF
2

• Outline
• Rationale and Objectives
• Observed Modifications of Urban Regions
• Skin Temperature
• Surface Albedo
• Surface Emissivity
• Aerosol and Clouds
• Description of urban model
• Model Results
• Summary and future direction

Dr. Menglin Jin Univ. of Maryland, College Park
3

1. Rationale and Objectives
Current Problem Land surface schemes in the
leading GCM/regional models do not include
urban landscape. For example, NCAR
Community Land Model (CLM), NASA land model,
NCEP land surface model, etc
Objectives Develop urban scheme in land surface
model
Needs Need to know what is urban how to
simulate urban
Basic idea Optimally combine satellite data into
urban model. Satellite observations can help
(a) better identify urban features and (b)
improve models surface parameters
Dr. Menglin Jin Univ. of Maryland, College Park
4
Question 1 Is urban region important enough for
us to simulate it specifically in a land surface
scheme?
(a) Is urban region big enough? (b) Are urban
physical processes unique enough?
Question 2 How to simulate urbanization?
Dr. Menglin Jin Univ. of Maryland, College Park
5
Human Density of 1700
(Source Ame. Association for the Advancement of
Science)
Dr. Menglin Jin Univ. of Maryland, College Park
6
Human Density of 1800
(Source Ame. Association for the Advancement of
Science)
Dr. Menglin Jin Univ. of Maryland, College Park
7
Human Density of 1900
(Source Ame. Association for the Advancement of
Science)
Dr. Menglin Jin Univ. of Maryland, College Park
8
Human Density of 1998
(Source Ame. Association for the Advancement of
Science)
Dr. Menglin Jin Univ. of Maryland, College Park
9
MODIS Observed Urban and Built-up

• 1000 household can make Tair higher about 2ºC
  than surroundings
• (Oke, 1976, Torok et al. 2002)

10
Question 1. Is Urban region important enough for
us to simulate it a GCM?
As an extreme case of land cover and land use
changes, urban region

1. modifies surface and atmospheric properties, and
   thus

changes heat, water, and momentum
transports 2. Adds new surface physical
processes into original existing, natural
physical processes
Land Surface Energy Budget
(1-a)Sd LWd-esTskin4 SHLE G 0
Dr. Menglin Jin Univ. of Maryland, College Park
11
2. How to Simulate Urban?
Urbanization Modifies Surface Energy Budget
(1-a)Sd LWd-esTskin4 SHLE G 0
Urban adds new physical processes
Storage term, Anthropogenic flux Modified
roughness length Canyon effect
Dr. Menglin Jin Univ. of Maryland, College Park
12
Radiation attenuation
Canyon effect
Turbulence production

Radiation trapping
Canopy heating cooling
Urban thermal properties
Dr. Menglin Jin Univ. of Maryland, College Park
13
50km
MODIS
Nighttime
50km
50km
Jin et al. 2004, submitted to J. of Climate
14
3.2 MODIS Observed Global urban heat island effect
Dr. Menglin Jin Univ. of Maryland, College Park
15
Comparison of skin temperature for urban and
nearby forests
MODIS
Cities have higher Tskin than forests
16
3.3 Urbanization changes surface albedo (MODIS)
Dr. Menglin Jin Univ. of Maryland, College Park
17
Urban region
NIR
Albedo
VIS
The decrease of urban albedo is mainly caused by
the decrease of reflectance at NIR
18
Urban albedo lower than cropland
urban
Spectral Albedo, January 2001, courtesy M. King
19
July
Seasonality of urban Albedo, courtesy M.
King
January
Evident seasonality is observed on urban albedo
20
3.4 Urbanization changes surface emissivity
(MODIS)
50km
50km
21
Zonal Averages from MODIS
Urban albedo is lower than that of cropland
Urban emissivity is lower than that of cropland
22
3.5. Urbanization changes atmospheric
conditions
MODIS Aerosol Optical Depth
Dr. Menglin Jin Univ. of Maryland, College Park
23
Interannual variation of Aerosol optical depth
24
Diurnal variation of aerosol from EPA PM2.5
Dr. Menglin Jin Univ. of Maryland, College Park
25
MODIS Seasonal variation of urban aerosol
Houston
Jin, Shepherd, and King 2004
26
Houston
Jin, Shepherd, and King 2004
27
Aerosol decreases surface insolation
Total solar radiation decreased by aerosol
20Wm-2
(Based on model of Ming-Dah Chou of NASA GSFC)
28
(1-a)Sd LWd-esTskin4 SHLE G 0
SH, LE, and G cannot be directly observed from
satellite. Need to use model framework to
examine their changes.
Dr. Menglin Jin Univ. of Maryland, College Park
29
Conceptual Urban Model
Is land cover urban
Urban model type
n
y
Urban Canyon Suburban Human-grass Roads
Existing CLM
Urban scheme
Dr. Menglin Jin Univ. of Maryland, College Park
30

• Model Design--- Two Steps
• Modify Physical Parameters which are changed for
  urban regions
• 1). Albedo,
• 2). Emissivity
• 3). LAI,
• 4). Heat capacity
• 5). Thermal conductivity
• 6). Hydro-conductivity

2. Modify physical process 1). Anthropogenic
heat flux (follow Voogt and Grimmond 2000) 2).
Storage heat flux 3). Roughness length 4).
Momentum changes by buildings
Dr. Menglin Jin Univ. of Maryland, College Park
31

• Model Design ---Two Steps
• Modify Physical Parameters which are changed for
  urban regions
• 1). Albedo,
• 2). Emissivity
• 3). LAI,
• 4). Heat capacity
• 5). Thermal conductivity
• 6). Hydro-conductivity

MODIS
2. Modify physical process 1). Anthropogenic
heat flux (follow Voogt and Grimmond 2000) 2).
Storage heat flux 3). Roughness length (Follow
Grimmond and Oke 2002) 4). Momentum changes by
buildings (not implemented yet)
Dr. Menglin Jin Univ. of Maryland, College Park
32
Use MODIS observed surface properties into model
Dr. Menglin Jin Univ. of Maryland, College Park
33
MODIS15_A2 Leaf Area Index (LAI) over Houston
regions
Dr. Menglin Jin Univ. of Maryland, College Park
34
MODIS11_L2 Emissivity_BAND 32 over Houston regions
Dr. Menglin Jin Univ. of Maryland, College Park
35
Model Design Two steps
2. Modified physical processes 1).
Anthropogenic heat flux (Qf)
Following Voogt and Grimmond 2000
Qf(hour) Qf0 1-0.6cos(p(hour-3)/12) Qf0
62.5W/m2 in January Qf0 37.5W/m2 in July
2.) Storage term (?Qs)
Adopted from Arnfield and Grimmond 1998
?Qs S(aiQbiQsci)Ai a,b,c are experical
coefficients corresponding to surface type Ai
surface areas for each urban land surface
type Q - net radiation radiation Qs absorbed
surface solar radiation
36
New Physical Processes (cont.)
3). Roughness Length (rub), Building impact on
momentum calculation
37
Complex Surface Parameters -Some Background
Literature

• Earliest study U Wisc. MS thesis of John
  Kutzback, 1961 studies effect on surface
  roughness of density of objects that were bushel
  baskets placed on frozen Lake Mendota
• Mike Raupach1992,1994 formulates theoretically
  the issue of partitioning of stress between
  smooth and such roughness objects- relates
  roughness length and displacement height to
  fractional area of obstacles normal to wind
  direction
• Linroth, 1993, identifies LAI as contributing to
  tree effect
• MacDonald et al, 1998, formulates modified
  version for buildings

(Copied from R. Dickinson Fall AGU 2003 urban
session invited talk)
38
Key Elements

• Uh wind at the top of the canopy
• Cd drag coefficient ,such that the overlying
  atmosphere loses t l Cd Uh 2 of momentum to the
  complex canopy
• where l effective frontal area density
• But sqrt(l Cd ) 0.4 / log(( h d)/ z0 )
  from the simple relationship between logarithmic
  wind profile and momentum transfer, which is
  inverted to get z0 / (h d )

(Copied from R. Dickinson Fall AGU 2003 urban
session invited talk)
39
Resolution of Roughness Issues

• McDonald scaling likely ok if modified to include
  reduction by leaf area factor for vegetation
• Raupach expression for d/h should be modified to
  allow for eddy sheltering (i.e. wake vortices of
  building only interact over the area not covered
  by building so that d should become h as
  structures become a single building - however
  for closed canopy, this is between 0.7 and 0.85 h
  depending on crown-aspect ratio

(Copied from R. Dickinson Fall AGU 2003 urban
session invited talk)
40
Table for properties modified for Case 1 run
0.5
1.5
Control run 0.25
Control run 0.15
0.92
0.96
1.5control run
Set as zero at first layer
Dr. Menglin Jin Univ. of Maryland, College Park
41
4. CLM-urban model results
Ground Temperature
Urban increase ground temperature by 1-3ºC, with
the largest increase occurring at local daytime
42
4.2 CLM-urban model results
Surface air temperature

• Urban increases land surface 2m surface air
  temperature,
• at a lower rate than its effects on ground
  temperature/skin temperature
• maximum at nighttime!

43
4.3 Urban Model Results
Absorbed Solar Radiation
Urban absorbs more Solar radiation
44
4.3 CLM-Urban Model Results
Urban increase of SH can be as high as 15Wm-2,
with maximum at local afternoon.
45
4.3 CLM-Urban Model Results
Urban increase upward longwave radiation
46
4.3 CLM-Urban Model Results
Urban reduces ground flux
47
4.4 Single Column NCAR SCAM-CLM-Urban Model
Beijing
48
4.4 SCAM/CLM2/urban-scheme Results Beijing
Sep. 1998
LE
49
4.4 Single Column NCAR CAM-CLM-Urban Model
Houston
50
4.4 SCAM/CLM2/urban-scheme Results
LE
51
Summary

• Satellite observations are extremely useful for
  understanding and
• simulating urbanization in climate models.

2. Urbanization scheme is needed in GCM/RCMs
land surface model, in order to accurately
reflect human impacts on global land climate
system. 3. We need more accurate urban land
cover, building density, and population
information for simulating urban in global and
regional scales.
Dr. Menglin Jin Univ. of Maryland, College Park