Image Processing Reference
determines the nature for an area. Urban heat island intensity, DT(u − r), of up to
12°C has been recorded under ideal conditions (Roth et al. 1989 ).
One of the earliest applications of spaceborne measurements was for surface
temperature and its relationship to the urban heat island effect and urban climate.
Rao ( 1972 ) is credited with the first study of urban heat islands from an environmental
satellite. Since then, several other studies have utilized the thermal infrared data from
the AVHRR, Heat Capacity Mapping Mission (HCMM), Landsat TM, TIROS
Operational Vertical Sounder (TOVS). Carlson et al. ( 1977 ) used satellite-derived
measurements of surface temperature to investigate the relationship between urban
land use and heating patterns. Other studies have used satellite measurements of both
reflected and emitted infrared radiation to quantify the relationship between urban
land use and the urban heat island effect (e.g., Owen et al. 1998 ; Roth et al., 1989 ; Gallo
et al. 1993 ; Gillies et al. 1997 ). At finer scales, Nichol ( 1994 ) used Landsat TM thermal
imagery to quantify the effect of solar radiation on the microclimate in Singapore.
Comprehensive summaries of remote sensing application to urban heat island are
presented in Gallo et al. ( 1995 ), Roth et al. ( 1989 ), and Voogt and Oke ( 2003 ).
Within the urban environment, the abundance and distribution of vegetation plays
an important role in controlling temperatures and air quality (Akbari et al. 1996,
2001 ; Nowak et al 2000 ). To ameliorate the effect from urban heat island, several
urban areas have recommended the planting of trees and grasses and the use of
reflective building and roofing material and pavement surfaces. Taha et al. ( 2000 )
observed that through the impact of vegetation and reflected roofs and pavement,
the daytime ambient temperature could be reduced between 1 and 2 K.
Relations Between Satellite Thermal Measurements
and In Situ Air Temperatures Observations
A major objective of researchers is to derive measurements that are meaningful or
familiar to human interpreters or city planners, and thus, the tendency is to com-
pare satellite-derived measurements to those acquired using in situ techniques.
However, the fundamental differences between the two datasets cannot be over-
looked and any attempt to establish direct correlations is
not always straightforward (Price 1979 ; Vukovich 1983 ;
Roth et al. 1989 ). Satellite measurements are unique and
have some attractive features, which include dense grid
synoptic view that eliminates synchronizing of data. The
differences between traditional and satellite-derived ther-
mal measurements of urban areas have been discussed
extensively in the literature (Roth et al. 1989 ; Voogt and
Oke 2003 ; Nichol 2003 ). While attempts have been made
by some researchers to compare the two datasets, others
have warned against using the two datasets as surrogates
due the inherent differences and the measured surfaces
in situ and