Urban Climate Studies:
applications for weather, air quality,
and climate change
Prof. Robert Bornstein
Dept. of Meteorology
San Jose State University
San Jose, CA USA
pblmodel@hotmail.com
Presented at
Tel Aviv University
8 Jan 2009
Funding sources: USAID-MERC, TCEQ, NASA, NSF, SCU
1
OVERVIEW
• URBAN CLIMATE
– WHY STUDY IT
– ITS CAUSES
– ITS IMPACTS
• CALIFORNIA COASTAL COOLING
– DATA
– ANALYSIS
• URBAN ATMOSPHERIC MODELS
– FORMULATION
– APPLICATIONS (HOUSTON, ATLANTA, ISRAEL)
• FUTURE EFFORTS
2
URBAN WEATHER ELEMENTS:
battles between conflicting effects
•
•
•
•
•
•
•
•
•
VISIBILTITY: decreased
TURBULENCE: increased (mechanical & thermal)
PBL NIGHT STABILITY: neutral
FRONTS (synoptic & sea breeze): slowed
TEMP: increased (UHI) or decreased
PRECIP: increased (UHI) or decreased
WIND SPEED (V): increased or decreased
WIND DIRECTION: con- or divergence
THUNDERSTORMS: triggered or split
3
HUMAN-HEALTH IMPACTS OF
URBAN CLIMATE
• > UHI  THERMAL STRESS
• > PRECIP ENHANCEMENT  FLOODS
• > URBAN INDUCED INVERSIONS 
POLLUTED LAYERS
• > TRANSPORT & DIFF PATTERNS FOR
– POLLUTION EPISODES
– EMERGENCY RESPONSE (i.e., TOXIC
RELEASES)
4
NEW URBAN CLIMATE: CAUSES
• GRASS & SOIL 
CONCRETE & BUILDINGS 
ALTERED SURFACE HEAT FLUXES
• FOSSIL FUEL CONSUMPTION 
ATMOSPHERIC POLLUTION AND HEAT
• ATM POLLUTION
ALTERED SOLAR & IR ENERGY
5
St. Louis nocturnal PBL:
warm near-neutral, polluted urban-plume
vs. rural stable surface-inversion
0F
Tmin
Tmax
urban-plume
inversion
Clark & McElroy (1970):
6
Urban effects on wind speed
• FAST LARGE-SCALE (i.e., SYNOPTIC) SPEEDS 
•
•
SMALL UHI 
URBAN SFC ROUGHNESS (Z0) INDUCED
DECELERATION
SLOW SYNOPTIC SPEEDS
LARGE UHI 
INWARD-DIRECTED ACCELERATION
CRITICAL SPEED ~ 3-4 m/s (FOR NYC & London)
7
NYC DAYTIME ∆V (z)
urban
rural
8
URBAN EFFECTS ON WIND DIR
• FAST SYNOPTIC SPEED WEAK UHI 
URBAN BUILDING-BARRIER EFFECT 
FLOW DIVERGES AROUND CITY
• SLOW SYNOPTIC SPEED  LARGE UHI 
LOW-p  CONVERGENCE INTO CITY
• MODERATE SYNOPTIC SPEED 
CONVERGENCE-ZONE ADVECTED TO
DOWNWIND URBAN-EDGE
9
NOCTURNAL UHI-INDUCED SFC-CONFLUENCE:
otherwise-calm synoptic flow 
confluence-center over urban center of Frankfurt, Germany
10
Weak cold-frontal (N to S) passage over NYC
a. Hourly positions (left)
b. At 0800 EST (right): T, q, & SO2 z-profile-changes
showed lowest 250 m of atm not-replaced, as front
“jumped” over city
See 
11
URBAN IMPACTS ON PRECIP
• INITATION BY THERMODYNAMICS (at SJSU)
– LIFTING FROM
• UHI CONVERGENCE
• THERMAL & MECHANICAL CONVECTION
– DIVERGENCE FROM BUILDING BARRIER EFFECT
• AEROSOL MICROPHYSICS
– SLOWER SECONDARY DOWNWIND ROLE
– METROMEX & PROF. D. ROSENFELD (HUJI)
12
NYC two-summer daytime-average thunderstorm-precip radar-echoes
(σ’s from uniform-distribution) for cases: all, convective, & moving
Formed over city
splitting
case
Split by city
13
Dispersion effects
• Vertical diffusion limited by urban-induced
•
•
elevated inversions (next slide)
Transport: 3-D effects of urban-induced flowmodifications
Convergence-zones effects due to
– Urban effects
– Sea breezes
14
 Urban-induced nocturnal elevated inversion-I traps home-heating emissions
 Power plant plume is trapped b/t urban-induced inversions I & II
 Inversion III is regional inversion  poor estimate of mixing depth
Plume
Home-heating
Sources
15
California Coastal-Cooling
(to appear, J. of Climate, 2009)
• Global & CA observations generally show
– asymmetric warming (more warming for Tmin than
for Tmax) (next graph)
– acceleration since mid-1970s
• CA downscaled global-modeling (next map)
– done onto 10 km grids
– shows summer warming that decreases towards
the coast (but does not show coastal cooling)
16
Not much change from mid40s to mid-70s, when values
started to again rapidly rise
17
Statistically down-scaled (Prof. Maurer, SCU) 1950-2000
Summer (JJA) IPCC temp-changes (0C) show warming rates that
decrease towards coast; red dots are COOP sites used in present
study & boxes are study sub-areas
18
Current Hypothesis
INCREASED GHG-INDUCED
INLAND TEMPS
INCREASED (COAST TO
INLAND) PRESSURE & TEMP
GRADIENTS
INCREASED SEA BREEZE FREQ,
INTENSITY, PENETRATION,
&/OR DURATION 
COASTAL AREAS SHOULD
SHOW COOLING SUMMER
DAYTIME MAX TEMPS (i.e., A
REVERSE REACTION)
NOTE:
NOT A TOTALLY ORIGINAL
IDEA

19
Results 1: SoCAB 1970-2005 summer (JJA) Tmax warming/
cooling trends (0C/decade); solid, crossed, & open circles show
stat p-values < 0.01, 0.05, & not significant, respectively
?
?
?
20
Tmax
Results 2: SFBA & CV 1970-2005 JJA
warming/cooling trends (0C/decade), as in previous figure
?
?
?
21
Results 3: JJA Temp trends; all CA-sites
• LOWER TRENDS
FROM 1950- 70
(EXCEPT FOR TMAX)
• Curve b: TMIN HAD
FASTEST RISE (AS
EXPECTED)
• Curve c: TMAX HAD
SLOWEST RISE; IT IS
A SMALL-∆ B/T BIG
POS VALUE & BIG
NEG-VALUE (AS IN
ABOVE 2 GRAPHS)
• CURVE a: TAVE THUS
ROSE AT MID RATE
• Curve d: DTR (diurnal
temp range) THUS
DECREASED (AS TMAX
FALLS & TMIN RISES)
22
Significance of these all-CA Trends
• HIGHER TRENDS FROM 1970-2005 
•
•
•
•
FOCUS NEEDED ON THIS PERIOD
TMIN HAS FASTER RISE 
ASSYMETRIC WARMING IN LITERATURE
BUT TMAX HAS SLOWER RISE, BECAUSE IT IS A
SMALL DIFFERENCE B/T BIG POS-VALUE & BIG
NEG-VALUE (AS SEEN IN ABOVE SPATIAL PLOTS)
TAVE & DTR ARE ALSO THUS “CONTAMINATED”
NEXT 2 SLIDES THUS SHOW SEPARATE TRENDS
FOR CA COASTAL AND INLAND AREAS
23
Result 4: JJA Tave, Tmin, Tmax, & DTR TRENDS FOR
INLAND-WARMING SITES OF SoCAB & SFBA
a
Curve b: TMIN
INCREASED
(EXPECTED)
b
c
d
Curve c: TMAX HAD
FAST RISE;
(UNEXPECTED),
COULD BE DUE TO
INCREASED UHIs OR
INCREASED DOWNSLOPE FLOWS
CURVE a: TAVE THUS
ROSE AT MID RATE
Curve d: DTR THUS
INCREASED (AS TMAX
ROSE FASTER THAN
24
TMIN ROSE
Result 5: JJA Tave, Tmin, Tmax, & DTR TRENDS FOR
COASTAL-COOLING SITES OF SoCAB & SFBA
a
b
c
d
Curve b: TMIN ROSE
(EXPECTED)
Curve c: TMAX HAD COOLING (UNEXPECTED MAJOR
RESULT OF STUDY)
CURVE a: TAVE THUS
SHOWED ALMOST NO
CHANGE, AS FOUND IN
LIT.), AS RISING Tmin &
FALLING Tmax CHANGES
ALMOST CANCELLED OUT
Curve d: DTR THUS DECREASED, AS TMIN ROSE &
TMAX FELL
25
Note IPCC 2001 does show cooling over
Central California!!
26
Significance of above Coastal-Cooling
and Inland-Warming trends
• CA ASSYMETRIC WARMING IN LITERATURE IS
•
HEREIN SHOWN TO BE DUE TO COOLING TMAX
IN COASTAL AREAS & CONCURRENT WARMING
TMAX IN INLAND AREAS
PREVIOUS CA STUDIES
– DID NOT LOOK SPECIFICALLY AT SUMMER DAYTIME
COASTAL VS. INLAND VALUES HAVE
– THEY THUS REPORTED CONTAMINATED TMAX, TAVE, &
DTR VALUES
– THEY, HOWEVER, ARE NOT INCONSISTENT WITH
CURRENT RESULTS, THEY ARE JUST NOT AS
DETAILED IN THEIR ANALYSES & RESULTS
27
Result 6. JJA 1970-2005 2 m Tmax trends for 4 pairs of
urban (red, solid) & rural (blue, dashed) sites
Notes:
1. All sites are near the
cooling-warming border
2. UHI-TREND (K/DECADE)
= absolute sum b/t
warming-urban &
cooling-rural trends
a. SFBA sites
> Stockton
(0.38 + 0.17 = 0.55)
> Sac. (0.49)
b. SoCAB sites
> Pasadena (0.26)
> S. Ana (0.12)
28
Notes on JJA daytime
UHI-trend results
• Faster growing cities (not shown) had
faster growing UHIs
• As no coastal sites showed warming Tmax
values, calculations could only be done at
these four pairs (at the inland boundary
b/t the warming and cooling areas)
• Coastal sites would have cooled even
more w/o their (assumed) growing UHIs
29
BENEFICIAL IMPLICATIONS OF
COASTAL COOLING
• NAPA WINE AREAS MAY NOT GO EXTINCT
•
•
•
(REALLY GOOD NEWS!) (next map)
ENERGY FOR COOLING MAY NOT INCREASE AS
RAPIDLY AS POPULATION (next graph)
LOWER HUMAN HEAT-STRESS RATES
OZONE CONCENTRATIONS MIGHT CONTINUE TO
DECREASE, AS LOWER MAX-TEMPS MEAN
REDUCED
– ANTHROPOGENIC EMISSIONS
– BIOGENIC EMISSIONS
– PHOTOLYSIS RATES
30
NAPA WINE AREAS MAY NOT GO EXTINCT DUE TO ALLEGED
RISING TMAX VALUES, AS PREDICTED IN NAS STUDY
31
Result 7: Peak-Summer Per-capita Electricity-Trends
 Down-trend at cooling
Coastal: LA (blue) & Pasadena
(pink, 8.5%/decade)
> Up-trend at warming
inland Riverside (green)
Up-trend at warming Sac
& Santa Clara
Need:
 detailed energy-use
data for more sites
 to consider changed
energy efficiency
32
Future Coastal-Cooling Efforts (PART 1 OF 2)
• EXPAND (TO ALL OF CA & ISRAEL?)
– ANALYSIS OF OBS (IN-SITU & GIS)
– URBANIZED MESO-MET (MM5, RAMS, WRF) MODELING
• SEPARATE INFLUENCES OF CHANGING:
– LAND-USE PATTERNS RE
• AGRICULTURAL IRRIGATION
• URBANIZATION & UHI-MAGNITUDE
– SEA BREEZE:
INTENSITY, FREQ, DURATION, &/OR PENETRATION
• DETERMINE POSSIBLE “SATURATION” OF SEABREEZE EFFECTS FROM
• FLOW-VELOCITY & COLD-AIR TRANSPORT
• AND/OR STRATUS-CLOUD EFFECTS ON LONG- & SHORT-WAVE
RADIATION
33
POSSIBLE FUTURE EFFORTS (PART 2 OF 2)
• DETERMINE CUMULATIVE FREQ DISTRIBUTIONS
OF TMAX VALUES, AS
– EVEN IF AVERAGE TMAX DECREASES,
– EXTREME VALUES TMAX MAY STILL INCREASE (IN
INTENSITY &/OR FREQUENCY)
• DETERMINE CHANGES IN LARGE-SCALE ATM
FLOWS:
– HOW DO GLOBAL CLIMATE-CHANGE EFFECT POSITION
& STRENGTH OF: PACIFIC HIGH & THERMAL LOW?
– THESE TYPES OF CLIMATE-CHANGES ARE THE
ULTIMATE CAUSES OF TEMP AND PRECIP CHANGES
34
OUR GROUP’S MESO-MODELING EXPERIENCE
• SJSU (MM5 & uMM5)
–
–
–
–
–
–
–
Lozej (1999) MS: SFBA winter wave cyclone
Craig (2002) MS: Atlanta UHI-initiated thunderstorm (NASA)
Lebassi (2005) MS: Monterey sea breeze (LBNL)
Ghidey (2005) MS: SFBA/CV CCOS episode (LBNL)
Boucouvula (2006a,b) Ph.D.: SCOS96 episode (CARB)
Balmori (2006) MS: Tx2000 Houston UHI (TECQ)
Weinroth (2009) PostDoc: NYC-ER UDS urban-barrier effects (DHS)
• SCU (uRAMS)
– Lebassi (2005): Sacramento UHI (SCU)
– Lebassi (2009) Ph.D.: SFBA & SoCAB coastal-cooling (SCU)
– Comarazamy (2009) Ph.D.: San Juan climate-change & UHI (NASA)
• Altostratus (uMM5 & CAMx)
–
–
–
–
–
SoCAB (1996, 2008): UHI & ozone (CEC)
Houston (2008): UHI & ozone (TECQ)
Central CA (2008): UHI & ozone (CEC)
Portland (current): UHI & ozone (NSF)
Sacramento (current): UHI & ozone (SMAQMD)
35
SJSU IDEAS ON GOOD MESO-MET
MODELING
MUST CORRECTLY REPRODUCE:
– UPPER-LEVEL Synoptic/GC FORCING FIRST:
pressure (“the” GC/Synoptic driver) 
Synoptic/GC winds
– TOPOGRAPHY NEXT:
min horiz grid-spacing 
flow-channeling
– MESO SFC-CONDITIONS LAST:
temp (“the” meso-driver) & roughness 
meso-winds
36
Case 1: ATLANTA UHI-INITIATED STORM: OBS GOES &
PRECIP (UPPER) & MM5 w’s & precip (LOWER)
37
Recent Meso-met Model Urbanizations
• Need to urbanize momentum, thermo, & TKE
•
•
– Surface & SfcBL Diagnostic-Eqs.
– PBL Prognostic-Eqs. (not done in NCAR uWRF)
Start: veg-canopy model (Yamada 1982)
Veg-param replaced with GIS/RS urban-param/data
– Brown and Williams (1998)
– Masson (2000)
– Martilli et al. (2001) in TVM/URBMET
– Dupont, Ching, et al. (2003) in EPA/MM5
– Taha et al. (‘05, ‘08a,b,c) [& Balmori et al. (‘06)]: his uMM5
uses improved urban dynamics, physics, parameterizations,
& inputs
38
From EPA uMM5:
Within Gayno-Seaman
Mason + Martilli (by
Dupont)
PBL/TKE scheme
New
Roughness
approach
Sensible
heat flux
Rn pav Hsens pav LEpav
Drag-Force
approach
Net radiation
Latent
heat flux
Storage
heat flux
Anthropogenic
heat flux
Precipitation
Ts roof
roof
natural
soil
Qwall
Ts pav
Gs pav
Tint
Paved
surface
bare
soil
Surface layer
water
Drainage outside
the system
Infiltration
Drainage
network
Drainage
Root zone layer
Diffusion
Deep soil layer
39
But, uMM5 needs extra GIS-derived inputs
as f (x, y, z, t)
 land-use (38 categories)
 roughness heights z0 (see next slide)
 anthropogenic heat
 building heights
 paved-surface 2-D fractions
 building H to W, wall-plan, & impervious-area
2-D ratios
 building frontal, plan, & rooftop 3-D area densities
40
S. Stetson: Houston GIS/RS zo inputs
But, values are too
large, as they were
f(h) & not f(ơh)
h = building height
Values up to 413 m
uMM5 for Houston: Balmori (2006)
Goal: Accurate urban/rural temps & winds for
Aug 2000 O3 episode via
– uMM5
– Houston LU/LC & urban morphology
parameters
– TexAQS2000 field-study data
– USFS urban-reforestation scenarios 
UHI & O3 changes
42
uMM5 Simulation period: 22-26 August 2000
• Model configuration
•
•
– 5 domains: 108, 36, 12, 4, 1 km
– (x, y) grid points:
(43x53, 55x55, 100x100, 136x151, 133x141
– full-s levels: 29 in D 1-4 & 49 in D-5; lowest ½ s-level=7 m
– 2-way feedback in D 1-4
Parameterizations/physics options
> Grell cumulus (D 1-2)
> ETA or MRF PBL (D 1-4)
> Gayno-Seaman PBL (D-5) > Simple ice moisture,
> urbanization module NOAH LSM > RRTM radiative cooling
Inputs
> NNRP Reanalysis fields, ADP obs data
> Burian morphology from LIDAR building-data in D-5
> LU/LC modifications (from Byun)
43
1-km grid, uMM5 Houston UHI: 8 PM, 21 Aug
UHI
UHI
Bay
Gulf
MM5 UHI (2.0 K)
uMM5 UHI (3.5 K)44
UHI-Induced Convergence: obs vs. uMM5
C
C
Krieged Obs
uMM5 output
45
min
max
increase
Base-case (current)
veg-cover (0.1’s)
 urban min (red)
 rural max (green)
Modeled changes of
veg-cover (0.01’s)
 Urban-reforestation
(green)
Rural-deforestation
(purple)
46
Run 12 (urban-max reforestation) minus Run 10 (base case) 
near-sfc ∆T at 4 PM shows that:
reforested central urban-area cools &
surrounding deforested rural-areas warm
warmer
cooler
warmer
47
DUHI(t): Base-case minus Runs 15-18
Urban temp difference between runs
0.4
run14-run13s
run15-run13s
run16-run13s
0.2
run17-run13s
run18-run13s
RURAL
Tdiff (K)
0.0
-0.2
-0.4
URBAN
Max-impact of –0.9 K on
a 3.5 K noon-UHI, of which
1.5 K was from uMM5
-0.6
-0.8
-1.0
20
0
4
8
12
16
20
0
4
8
12
16
LST
• UHI = Urban-Box minus Rural-Box
• Runs 15-18: Urban re-forestation scenarios
• DUHI = Run-17 UHI minus Run-13 UHI 
max effect (green line)
• Reduced UHI  lower max-O3 (not shown) 
EPA emission-reduction credits  $ saved
48
RAMS, MM5, & CAMx SIMULATIONS OF
MIDDLE-EAST O3 TRANSBOUNDARY
TRANSPORT
E. Weinroth1,2, S. Kasakseh1,3
M. Luria2, R. Bornstein1
1San
Jose State Univ.
2Hebrew Univ. Jerusalem, Israel
3Applied Research Institute Jerusalem (ARIJ),
Bethlehem, West Bank
In Atmos. Environ. (2008)
49
USAID-MERC project (2000-)
• Involves scientists from Palestinian Territories, Israel,
•
USA (& now Jordan and Lebanon)
Objectives accomplished:
– Installation of environmental stations in West Bank & Gaza
(and now Jordan & Beirut)
– Preparation of environmental databases (SJSU web page)
– Field campaigns during periods of poor air quality (Prof. Luria)
– Application of numerical models for planning
• RAMS & MM5 (Kasakseh 2007) meso-met
• CAMx photochemical air-quality (Weinroth et al. 2007 in
Atmos. Environ.)
50
10 m obs speed (m/s) & O3 at 0300 LST or 0000 UTC on 1 Aug
33
50 m
250 m
750 m
1000 m
Night obs of sfc flow:
3-AM LST (00 UTC)
3 m/s
H
H
32.4
Flow Dir: weak down-slope off
coastal-mountains at
 Coastal plain: offshore (to W)
from W-facing slopes
 Haifa Pen. (square): offshore
(to E ) from E-facing slopes
 Inland sites: directed inland (to
E) from E-facing slopes
Med
iterr
anea
n Se
a
LATITUDE (Deg N)
32.7
32.1
31.8
31.5
34.5
L
Low-O3
generally <40 ppb)
Haifa still at 51 ppb
L
34.8
35.1
LONGITUDE (Deg E)
35.4
51
10 m obs speed (m/s) at 1200 LST or 0900 UTC on 1 Aug
33
Day Obs: 1200 NOON LST
50 m
250 m
750 m
1000 m
L
Winds:
Reversed
Stronger: up 6 m s-1
3 m/s
32.4
Med
iterr
anea
n
LATITUDE (Deg N)
Sea
32.7
 Coastal plain: Onshore/upwind,
from SW
 Inland sites: Channeling (from
W) in corridor (box; focus of
modeling) from Tel-Aviv to
J. area (at Modiin site).
H
H
32.1
H
L
Higher daytime O3
 max at Mappil, 66 ppb
 2nd max at Modiin, 63 ppb
31.8
31.5
34.5
34.8
35.1
LONGITUDE (Deg E)
35.4
52
MM5 Configuration
 Version 3.7
 3 domains







– 15, 5, 1.67 km Grid Spacings
– 59 x 61, 55 x 76, 58 x 85 Grid
Points
32 σ-levels
– up to 100 mb
– first full σ-level at 19 m
Lambert-conformal map projection
(suitable for mid lat regions)
Two-way nesting
5-layer soil-model
Gayno-Seaman PBL
Simulations
– End: 00 UTC, 3 Aug
– Start: 00 UTC, 29 July
Single CPU , LINUX
53
MM5 Domain-3 winds (m/s) at 1100 LST on 1 Aug ‘97
red lines = topo heights (m); yellow line = sea breeze
front; note reverse upslope-flow & channeling to J.
Sea
Max
J.
Max
54
Same, but at 2300 LST; where yellow line = land
breeze front; note down-slope flow; still inland
directed flow in inland areas & still channeling to J.
Sea
Max
J.
Max
55
Mid-east Obs vs. MM5: 2 m temp (Kasakech ’06 AMS)
First 2 days show GC/Syn trend not in MM5,
as MM5-runs had no analysis nudging
Obs
Run 1
July 29
August 1
obs
July 31
Aug 1
Run 4:
AugustReduced
2
Seep-soil
MM5:Run
4 T
Aug2
Standard-MM5 summer night-time min-T,
56
But lower input deep-soil temp  better 2-m T results  better winds  better
O3
Obs vs. MM5: wind speed (m/s)
Run 3
OBS
July 31
August 1
August 2
57
RAMS/CAMx (left) O3 vs. airborne obs (right) at 300 m:
> Secondary-max: over J. in obs; due to coastal N-S highway
> Primary-max: in Jordan (no obs); due to Hadera
Airborne obs
Jerusalem
O3 ppb
Hadera
Power 
0
.
Irbid,
Jordan
0-20
20-40
Plant
40-60
60-70
70-80
80-90
90-95
95-105
105-120
1 Aug, 1500 LST
58
Overall Modeling Lessons
• > Models can’t be
– assumed to be perfect (i.e., model user vs. modeler)
– used as black boxes
• > Need good large-scale forcing model-fields
• > If obs are not available, OK to make reasonable
educated estimates, e.g., for rural
– deep-soil temp
– soil moisture
• > Need data to compare with simulated-fields
• > Need good urban
– morphological data
– urbanization schemes
59
Thanks for listening!
Questions?
60