Urban impacts on climate - The Global Climate System: Patterns, Processes, and Teleconnections - page 245

Table 7.4 Important considerations in the accurate use of satellite
thermal imagery for urban climate analysis (after Roth et al. 1989 )
Resolution of satellite sensor (and image)
Noise equivalent associated with wavelengths used
Angle of satellite image collection (angle of scan, geometric distortion)
Weather (cloud, wind, etc.)
Impact of intervening atmosphere (scattering, absorption, refraction,
re-radiation)
Image enhancement needs and procedures
Emittance or emissivity of surface
Geometry of surface (three-dimensional roughness)
Surface land use
Surface directional effects (anisotropy)
Size and orientation of urban canyons
Surface layering (how much of the actual surface is obscured)
resolution of 120m, to find a good relationship between SUHI and AUHI in
Singapore. While surface temperatures were higher and had greater variability
compared to air temperatures, the diurnal air temperature pattern did follow the
surface temperature changes. Figure 7.11 (Plate 6 ) demonstrates the resolution
detail a high-quality sensor can provide.
Voogt and Oke ( 1997 ) attempted to overcome the problems created by the use
of SUHI to define AUHI, by defining a ''complete surface temperature''
between the surface and the air. Using a combination of ground and airborne
thermal sensors in Vancouver, Canada, they estimated the impacts of emissions
from all surfaces and all directions, removing the inherent satellite viewing
biases described earlier. The result was an area-weighted overall temperature
(Tc) that included inputs from rooftops, sides of buildings, under trees, etc. Tc
must be calculated, and is not directly observable, and depending on the urban
environment, can be very complicated to determine. Voogt and Oke suggested
that hemispheric measurements of outgoing longwave radiation might be a
reasonable surrogate for Tc. While Tc did provide a significant improvement
compared to SUHI in the description of AUHI, especially around noon, there
were still important differences between the two values.
Both Roth et al.( 1989 ) and Voogt and Oke ( 2003 ) suggest that, as a result of
the complications mentioned above, many different types of UHI can be defined
(Figure 7.6 ). While Voogt and Oke in 1997 stated that the linkages between air
and surface temperatures are far too complex to define properly, six years later,
their viewpoint was a little more positive (Voogt and Oke 2003 ). While quali-
tative descriptions and simple correlations based on remote sensing are still very
much the norm, there have been some improvements. Advances in satellite