Title: Remote Sensing of Urban Landscapes and contributions of remote sensing to the Social Sciences
1Remote Sensing of Urban Landscapes and
contributions of remote sensing to the Social
Sciences
2Urban-Suburban Land Use
- Urban and suburban expansion
- almost 1/2 the Earths population lives in cities
- rapid expansion of urban centers and their
peripheries - impacts on land cover, societal structure of the
cities, population distribution, land use
characteristics - interconnectivity of cites at large scales
3Urban remote sensing
- High spatial resolution data are needed
- Temporal and spectral resolution are typically
not a significant requirement for most
applications - Ancillary data typically used (census data)
- Can measure variables such as urban extent,
housing density, structure type, urban vegetation
cover, air quality, change detection
4- Temporal and spatial resolution requirements vary
depending on applications - short term (event-scale, sub-annual) vs. long
term (interannual) - high spatial resolution (lt 1m) vs. medium spatial
resolution (15-30m) - (see Jensen Fig. 12-1)
5High Spatial Resolution Sensors
- QuickBird (65cm B/W, 4m multispectral)
- IKONOS (1m B/W, 4m multispectral)
- SPOT (2.5m - 20m multispectral)
- ASTER (15 -30m multispectral)
- Landsat ETM (15m B/W, 4m multispectral)
6Delineation of Urban Areas
- Difficult to do because urban areas are diverse
and complex - Boundaries between urban and suburban are not
always clear - Lack of a consistent definition of what is
urban - administrative boundaries
- population density, etc.
7Balitmore, MD Well-developed city
center Diffuse boundary between urban and
natural environment
Landsat TM multi-spectral image
8Las Vegas, NV Indistinct city center Distinct
boundary between urban and natural environments
Landsat TM multi-spectral image
9Riyadh, Saudi Arabia Intermediate case
10Demographic/Socioeconomic Patterns
- Census data lack spatial details and are
infrequently updated (not globally available) - Remote sensing is useful for monitoring urban
growth in developing countries - Need ancillary data plus repeat temporal coverage
from remote sensing - Important to integrate physical and socioeconomic
variables
11Example
- Pozzi and Small (2002) produced a study of
relationship between population density (from US
census) and vegetation cover (from Landsat TM)
12NYC Population Density
NYC Vegetation Fraction
(source US Census)
(source Landsat TM)
Linear inverse correlation between population and
vegetation fraction
13Urban Heat Island Monitoring
- Project ATLANTA (Atlanta Land-use Analysis
Temperature and Air-quality) - Uses remote sensing to observe, measure, and
monitor impacts of rapid urban growth
14ATLAS Thermal Images of Atlanta, Georgia
Atlanta - Daytime Image
Atlanta - Nighttime Image
15City Lights Imagery
- Uses visible band of the Operational Linescan
System (on board the DMSP satellite) - Useful for making global inventories of human
settlements - Spatial resolution of 1km
- Relationships between city lights and
socioeconomic variables such as population
density, economic activity, electric power
consumption, etc.
16Earth Lights from OLS
17Measurements of Pollution in the Troposphere
(MOPITT)
Carbon monoxide plumes from China 22 km spatial
resolution, 640 km FOV
18Disaster Monitoring
- Volcanic eruptions
- Tornados
- Hurricanes
- Oil spills
- Earthquakes
- War/terrorism
- Floods
19ASTER image of Maryland tornado path
before
after
20AVHRR image of Hurricane Floyd September 1999
21RADARSAT image of oil spill
22ERS-2 Interferometric SAR Mapping of Ground
Displacement
23Bam, Iran Earthquake destruction IKONOS image
from 12/27/2003
24Flooding in Dresden, Germany August 22,
2002
QuickBird Satellite Image 65 cm spatial
resolution
25Epidemiology
- Cholera virus attaches to zooplankton (copepods)
and phytoplankton. Plankton plumes emanating from
the Ganges are being monitored - SST and plankton can be monitored in Bay of
Bengal to track this - Hanta virus (carried by mice) correlates to
changes in precipitation (El Nino) and vegetation
cover, especially grasses - NDVI can be used to track these changes
- Townshend et al. found that Ebola outbreaks
corresponded to changes in land use and seasonal
climate patterns
26Landsat TM map of land cover near Kikwit, Zaire
(location where Ebola outbreaks were first
reported in 1995) pinkcleared areas greenjungle