Remote Sensing of the Urban Heat

Island Effect Across Biomes in the

Continental USA

Marc Imhoff , Lahouari Bounoua, Ping Zhang, and Robert Wolfe

NASA’s Goddard Space Flight Center

Land Cover Land Use Change - 2012

Studying the UHI Phenomenon

Focus on surface temperature

- influenced by albedo, emissivity, and surface fluxes.

Bioclimatic context

- setting is important as UHI is a relative measure

(Temp of urban core compared to nearby non-

urban cover).

- Important to climate change as urbanization of

specific biomes represents a departure from

previous flux characteristics.

- Impact may be disproportionate to simple area

based assessments. 2

Urban Occupation of Fertile Soils

0

1

2

3

4

5

6

7

8

0 1 2 3 4 5 6 7 8

Percent of Soils (UN/FAO) Covered by LightsGrouped by Number of FCC Limiting Factors

Per

cent

of S

oil A

rea

Cov

ered

by

Ligh

ts

Number of Factors Limiting Agriculture

CHINA

Consequences of Urbanization

NPP Lost or Gained (annually) going from a

pre-urban to a post urban world

Annual loss of food web energy 400 Trillion kilocalories

(roughly equal to food energy requirement for 448 million people).

Reduction of actual food products for 16.5 million persons annually.

Disproportionate impact: Urbanization (3% surface area) has reduced

Continental NPP by 41.5 Million tons C/year

Roughly equal to the gains made by agriculture (40% surface area).

Imhoff et al. 2004, RSE, Vol. 89, Issue 4, pp. 434-443.

Consequences of

Urbanization on NPP

DMSP/OLS Urban Map

Urban, Peri-urban, Non-urban

AVHRR Monthly NPP (g Cm-2)

NPP and Local Climate: Urban Heating Extends

Satellite Observations

Winter NPP

gain negated

in peak

season by

reduced

vegetation

and heat

stress.

Seasonal Offset

diminishes in

tropics

Length of growing season

locally in cold climates.

In semi-arid regions cities

enhance NPP relative to

surrounding areas

North East

Mid-Atlantic

South East South West

Atlanta Urban Heat Island and Air Quality Modeling Study

D. Quattrochi, MSFC

120oF 49oC

Surface Temperature and Albedo

83oF 28oC

7

Surface composition (parameters)

in and outside urban area

8

Schematic functioning of the Simple Biosphere Model (SiB2)

showing the pathway for the latent heat flux calculation.

Surface Layer

Root Zone

Recharge Zone

etc r c

2r b

r d

r soil

l E c

e a

e m

l E c + l E g

l E g

r a

S L

R n

P

+W leaf

W c

+ W thru W drip

-W run W g

W 1

W 3

W 2

-W drain

U m

T m

eci

T c

egi

egs

z 2

z 1

d 1

d 2

d 3

Canopy

Air Space

SiB2 Transfer pathway for Latent Heat Radiation

Precipitation

Interception Loss

interception

Ground evaporation

Transpiration

More than Albedo Biome Level Surface Parameters Influence Heat

9

Datasets and compilations

Domains: cities in U.S.

Scales: cities with area size from 10km2 to more than 4000 km2

Datasets:

- Impervious surface area (ISA) from the Landsat TM-based NLCD 2001 dataset [Yang et al., 2002 ]

- Land surface temperature (LST) from MODIS [Wan et al., 2004]

- Terrestrial ecoregion map developed by Olson et al [2001]

- Topographic data from SRTM30 [Farr and Kobrick, 2000]

- NDVI from MODIS [Huete et al., 1997]

10

Geospatial urban area definition using Landsat

Impervious Surface Area

Identifies boundary between urban and low intensity residential area

Provide spatially coherent urban groups

11

Classification of urban density

12

Classification of urban density

Four broad biomes: forest, grass/shrub, arid (desert and

xeric shrubland), and semi-arid (Mediterranean)

323 settlements are evenly distributed over USA ranging

from around 10 km2 to more than 4,000 km2

Five urban zones are defined based on ISA density

- Urban core pixels: 75% > ISA ≥ 50%

- Medium density urban pixels: 50% > ISA ≥ 25%

- Suburban pixels: 0 to 5 km outside the 25% ISA contour

- Rural pixels: 15 to 20 km outside the 25% ISA contour

Sampling is constrained by biome and elevation

UHI=LSTurban core- LSTRural

13

UHI and ecological context

Average UHI for US cities grouped by biomes

-2

0

2

4

6

8

Forest Grass Mediterranean Desert

UH

I (d

egre

e)

• UHI=7.1 oC in cities built in forest

• UHI=4.2 oC in cities built in grass/shrubs

• UHI=6.0 oC in cities built in Mediterranean

• UHI=0.25 oC in cities built in desert

• The average city size of each group is 159/206/1055/160 km2

14

UHI Baltimore vs. Las Vegas

• Baltimore, Maryland, in Notheastern temperate broadleaf and mixed forest

• Las Vegas, Nevada, in desert and xeric shurbland

Baltimore Las Vegas

0

20

40

60

80

Avera

ge I

SA

(%

)

Urban Core (ISA>=75%)

Urban1 (75%>ISA>=50%)

Urban2 (50%>ISA>=25%)

Suburban (0-5Km buffer)

Rural (45-50Km buffer)

A

Baltimore Las Vegas

0.0

0.2

0.4

0.6

0.8

B

Jun-Aug LST

Jun-Aug NDVI

Avera

ge N

DV

I

Fractional ISA

Baltimore Las Vegas

25

30

35

40

45

50

55

C

Avera

ge L

ST

(C

)

15

UHI and ecological context: Baltimore vs. Las Vegas

Composite averaged LST during summer (Jun-Jul-Aug) day time

across all urban zones at Baltimore and Las Vegas.

25

30

35

40

45

50

55

60

Las Vegas

Baltimore

Avera

ge s

um

mer

daytim

e L

ST

(C)

50km 20km 15km 10km 5km 25% core 25% 5km 10km 15km 20km 50km

The composite summer daytime LST profile

25

30

35

40

45

50

55

60

A Biome-Centric View of Remotely Sensed Surface Parameters in US cities

Imhoff, Zhang, Wolfe and Bounoua, 2009, RSE 114 (2010)

Urban Heat Islands and Ecological Context

MODIS - Aqua LST (2003-2005)

Summer = Jun-Aug (average LST)

Winter = Dec - Feb (average LST) LST at 1:30 pm local time

Day = 1:30PM Local time

Night = 1:30AM Local time

Imhoff, Zhang, Wolfe and Bounoua, 2009, RSE 114 (2010)

FE FA GN DE MS GS GT

FW

18

Quantifying the relationship between LST and ISA

Variations in ISA explain 88% of the variation in LST for urban areas in forested biomes

The rate of change in LST as a function of ISA is 11.5% for urban areas in forests, while only 6.9% for those characterized by short vegetations.

In desert environments, the LST’s response to ISA presents is a “U-shaped” horizontal gradient decreasing from the urban core to the outskirts of the city and then increasing again in the suburban to the rural zones

A: Cities in forests

LST = 0.115 X ISA

R2 = 0.88

-8

-6

-4

-2

0

2

4

6

8

-40 -20 0 20 40

NLCD2001 ISA anomaly (%)

Su

mm

er

da

y L

ST

an

om

aly

(oC

) B: Cities in desert

LST = 0.001 X ISA2 + 0.010 X ISA - 0.54

R2 = 0.14

-8

-6

-4

-2

0

2

4

6

8

-40 -20 0 20 40

NLCD2001 ISA anomaly (%)

Su

mm

er

da

y L

ST

an

om

aly

(oC

)

19

The magnitude of UHI and size of urban area

Average UHI for US cities grouped by size

0

2

4

6

8

10

10-100 100-1000 >1000 (km2)

UH

I (d

eg

ree)

Forest Grass

• US urban settlements are grouped based on the total contiguous area for

each urban polygon defined by 25% ISA threshold

• Average UHI during summer day time is affected by urban size

20

UHI and urban extent for discrete urban

centers in temperate forest (continental US)

Urban extent defined by contiguous area with 25%+ ISA

21

Conclusions and future directions

This research highlights significant positive relationships

between the UHI magnitude, the ISA, and ecological setting

estimated exclusively from remotely sensed observations

The use of ISA as an estimator of the extent and intensity of

urbanization is more objective than population density based

methods and can be consistently applied across large areas for

inter-comparison of impacts on biophysical processes.

We will use a combination of satellite and ecological map data

to extend the characterization of the UHI response to global

urban settlements.

The uncertainties in ISA outside USA are challenges for future

studies as the concept of ISA in the USA does not necessarily

transfer to many developing country cities where infrastructure

characteristics and properties are often different