142
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
England et al. 1992; Xue and Cracknell 1995). The negative correlation between
LST and VIs was found with various remote sensing data at different spatial scales
and temporal resolution for microclimate studies in developed regions. Moron et al.
(1994) thought that the scatterplot-combined NDVI and LST data are trapezoidal
from a theoretical point of view. Carlson et al. (1994) and Goetz (1997) analyzed
NDVI and LST data derived from different resolutions of the sensors and found that
a significant negative correlation exists between LST and NDVI.
Soil moisture plays a key role in surface–subsurface water and heat exchanges
through infiltration, percolation, and capillary processes. When the range of vegetation cover and soil moisture in the study area was large, the scatterplot-combined
NDVI and LST remote sensing data resulted in a triangle, which can be verified
using the soil–vegetation–atmosphere transfer model (Price 1990; Carlson et al.
1995a; Gillies et al. 1997; Sandholt et al. 2002). Weng et al. (2004) investigated the
applicability of using a vegetation fraction derived from a spectral mixture model as
an alternative indicator of vegetation abundance and found that LST had a slightly
stronger negative correlation with the unmixed vegetation fraction than with NDVI
for all land cover types across the spatial resolution from 30 to 960 m. This may be
further linked to the effect of urban heat island (UHI). Chen et al. (2006) studied the
spatial and temporal relationships between LST and VIs based on the analysis of UHI
effects due to urbanization impacts. Martha et al. (2008) further described the spatial
relationship among satellite-derived LST, circumpolar arctic vegetation, and NDVI.
In comparison, spatial VITT has been applied widely in many studies reflecting the potential impact of LST on NDVI. Several studies that monitored ET and
soil moisture with spatial VITT have illuminated this correlation between LST and
NDVI (Goward and Hope 1989; Price 1990; Ridd 1995; Gillies et al. 1997; Gillies
and Carlson 1995; Sandholt et al. 2002; Wang and Moran 2004; Han et al. 2006).
Moron et al. (1994) explained the algorithm of the crop water stress index (CWSI),
which avoids measurements of leaf temperature when studying vegetation cover.
The slope of scatterplot-combined LST and VI represents the degree of crop water
stress gradient based on the negative relationship between LST and VI (Carlson et al.
1995b; Moran et al. 1996; Fensholt and Sandholt 2003; Venturini et al. 2004; Wang
et al. 2007). Such findings lead to a more accurate evaluation of the spatial and temporal variations of drought. The water stress index method is the ratio of actual ET
and potential ET, which is a kind of CWSI. With this ratio, Jackson and Idso (1981)
put forward the CWSI concept, and Moron et al. (1994) proposed the water deficit
index (WDI). In addition, the moisture index method is an approach for monitoring
regional drought with water characteristics of strong absorption in shortwave infrared band (Xu 2006; Fensholt and Sandholt 2003; Chen et al. 2005). For example,
Kogan (1995) proposed the VCI, and McFeeters (1996) proposed the normalized
difference water index (NDWI) by combining the Landsat TM green band and the
TM NIR band; both VCI and NDWI are variations of the moisture index method.
7.1.3 need to develoP a coMPoSite dRought indicatoR
A number of specific indices for drought monitoring and assessment have been
recently developed. Wang and Takahashi (1999) developed the WDI and applied this
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
England et al. 1992; Xue and Cracknell 1995). The negative correlation between
LST and VIs was found with various remote sensing data at different spatial scales
and temporal resolution for microclimate studies in developed regions. Moron et al.
(1994) thought that the scatterplot-combined NDVI and LST data are trapezoidal
from a theoretical point of view. Carlson et al. (1994) and Goetz (1997) analyzed
NDVI and LST data derived from different resolutions of the sensors and found that
a significant negative correlation exists between LST and NDVI.
Soil moisture plays a key role in surface–subsurface water and heat exchanges
through infiltration, percolation, and capillary processes. When the range of vegetation cover and soil moisture in the study area was large, the scatterplot-combined
NDVI and LST remote sensing data resulted in a triangle, which can be verified
using the soil–vegetation–atmosphere transfer model (Price 1990; Carlson et al.
1995a; Gillies et al. 1997; Sandholt et al. 2002). Weng et al. (2004) investigated the
applicability of using a vegetation fraction derived from a spectral mixture model as
an alternative indicator of vegetation abundance and found that LST had a slightly
stronger negative correlation with the unmixed vegetation fraction than with NDVI
for all land cover types across the spatial resolution from 30 to 960 m. This may be
further linked to the effect of urban heat island (UHI). Chen et al. (2006) studied the
spatial and temporal relationships between LST and VIs based on the analysis of UHI
effects due to urbanization impacts. Martha et al. (2008) further described the spatial
relationship among satellite-derived LST, circumpolar arctic vegetation, and NDVI.
In comparison, spatial VITT has been applied widely in many studies reflecting the potential impact of LST on NDVI. Several studies that monitored ET and
soil moisture with spatial VITT have illuminated this correlation between LST and
NDVI (Goward and Hope 1989; Price 1990; Ridd 1995; Gillies et al. 1997; Gillies
and Carlson 1995; Sandholt et al. 2002; Wang and Moran 2004; Han et al. 2006).
Moron et al. (1994) explained the algorithm of the crop water stress index (CWSI),
which avoids measurements of leaf temperature when studying vegetation cover.
The slope of scatterplot-combined LST and VI represents the degree of crop water
stress gradient based on the negative relationship between LST and VI (Carlson et al.
1995b; Moran et al. 1996; Fensholt and Sandholt 2003; Venturini et al. 2004; Wang
et al. 2007). Such findings lead to a more accurate evaluation of the spatial and temporal variations of drought. The water stress index method is the ratio of actual ET
and potential ET, which is a kind of CWSI. With this ratio, Jackson and Idso (1981)
put forward the CWSI concept, and Moron et al. (1994) proposed the water deficit
index (WDI). In addition, the moisture index method is an approach for monitoring
regional drought with water characteristics of strong absorption in shortwave infrared band (Xu 2006; Fensholt and Sandholt 2003; Chen et al. 2005). For example,
Kogan (1995) proposed the VCI, and McFeeters (1996) proposed the normalized
difference water index (NDWI) by combining the Landsat TM green band and the
TM NIR band; both VCI and NDWI are variations of the moisture index method.
7.1.3 need to develoP a coMPoSite dRought indicatoR
A number of specific indices for drought monitoring and assessment have been
recently developed. Wang and Takahashi (1999) developed the WDI and applied this