for southern areas which show an increase in NDVI. LST averages stay stable or
slightly decrease over the whole continent, except southern areas, which show an
increase. These results provide evidence that arid and semiarid areas of southern
Europe have become more arid, while the remainder of Europe has seen an in increase
in vegetated lands. Wloczyk et al. (2011) used ETM+ data in conjunction with a
temperature–vegetation index method (TVX) for areawide mapping of instantaneous
air temperature. The TVX method was applied to a multitemporal data set of nine
ETM+ scenes covering large parts of northeastern Germany. These satellite-derived
measurements were compared with in situ measurements showing an average error of
about 3 K whereas the mean error in LST estimation was about 2 K.
4 These results are
comparable with previously reported results for the TVX method.
The MODIS and ASTER sensors have been a critical tool for providing regional
estimates of LSTs. LSTs are warmer in the early afternoon than in the morning
because this is a peak time for solar insolation. MODIS data from the Aqua mission,
therefore, are more likely to be closer to the maximum daily LST than that acquired
earlier in the day by Terra. Coops et al. (2007) investigated the differences in LST
between Aqua and Terra to get an assessment of how large these differences are across
Canada. Using MODIS Aqua and Terra data for 2000 through mid-2002, they found
there are statistically significant differences between AM and PM LSTs ranging from
1.2 and 5°C, depending on the time of year. On the average, over 90% of the variation
observed in the PM data can be explained by the AM LST land cover type and
location.
Yang et al. (2011) employed several land cover indices, the soil-adjusted vegetation index (SAVI), the normalized multiband drought index (NMDI), the normalized
difference built-up index (NSBI), and the normalized difference water index (NDWI),
to investigate four land cover types (vegetation, bare soil, impervious, and water) in a
suburban area of Beijing, China. They applied these indices to MODIS and ASTER
data acquired in May 2001. The study was designed to evaluate differences in LST as
a function of spatial scale differences between the 1-km MODIS TIR and the 90-m
ASTER TIR data. They applied a disaggregation method for subpixel temperature
analysis using a remote sensing end-member index-based technique to derive land
surface temperature.
5 It was found that there was good agreement in LSTs between
the two spatial resolutions. Another scaling study by Liu et al. (2006) used different
scaling approaches to compare LSTs for MODIS and ASTER data over a part of the
Loess Plateau in China. ASTER 90m TIR data were scaled up to match the 1km
spatial resolution of the MODIS sensor to compare LST values between the two
4 Kelvin is a measurement of heat energy or temperature which advances in the same increments as does
Celsius. Its principal difference is that Kelvin measurements are written as K and have a much lower starting
point; 0 K or 0 Kelvin is measured at −273.15°C, which is the point at which no heat energy exists in a
substance (called absolute zero). Celsius converts to K by adding 273.15 to the Celsius number.
5 Spectral mixture analysis provides an efficient mechanism for the interpretation and classification of
remotely sensed multidimensional imagery. It aims to identify a set of reference signatures (also known as
“end members”) that can be used to model the reflectance spectrum at each pixel of the original of a finite
number of ground components (Plaza et al., 2002).
ESTIMATING LAND SURFACE ENERGY BUDGETS USING REMOTE SENSING DATA
39
slightly decrease over the whole continent, except southern areas, which show an
increase. These results provide evidence that arid and semiarid areas of southern
Europe have become more arid, while the remainder of Europe has seen an in increase
in vegetated lands. Wloczyk et al. (2011) used ETM+ data in conjunction with a
temperature–vegetation index method (TVX) for areawide mapping of instantaneous
air temperature. The TVX method was applied to a multitemporal data set of nine
ETM+ scenes covering large parts of northeastern Germany. These satellite-derived
measurements were compared with in situ measurements showing an average error of
about 3 K whereas the mean error in LST estimation was about 2 K.
4 These results are
comparable with previously reported results for the TVX method.
The MODIS and ASTER sensors have been a critical tool for providing regional
estimates of LSTs. LSTs are warmer in the early afternoon than in the morning
because this is a peak time for solar insolation. MODIS data from the Aqua mission,
therefore, are more likely to be closer to the maximum daily LST than that acquired
earlier in the day by Terra. Coops et al. (2007) investigated the differences in LST
between Aqua and Terra to get an assessment of how large these differences are across
Canada. Using MODIS Aqua and Terra data for 2000 through mid-2002, they found
there are statistically significant differences between AM and PM LSTs ranging from
1.2 and 5°C, depending on the time of year. On the average, over 90% of the variation
observed in the PM data can be explained by the AM LST land cover type and
location.
Yang et al. (2011) employed several land cover indices, the soil-adjusted vegetation index (SAVI), the normalized multiband drought index (NMDI), the normalized
difference built-up index (NSBI), and the normalized difference water index (NDWI),
to investigate four land cover types (vegetation, bare soil, impervious, and water) in a
suburban area of Beijing, China. They applied these indices to MODIS and ASTER
data acquired in May 2001. The study was designed to evaluate differences in LST as
a function of spatial scale differences between the 1-km MODIS TIR and the 90-m
ASTER TIR data. They applied a disaggregation method for subpixel temperature
analysis using a remote sensing end-member index-based technique to derive land
surface temperature.
5 It was found that there was good agreement in LSTs between
the two spatial resolutions. Another scaling study by Liu et al. (2006) used different
scaling approaches to compare LSTs for MODIS and ASTER data over a part of the
Loess Plateau in China. ASTER 90m TIR data were scaled up to match the 1km
spatial resolution of the MODIS sensor to compare LST values between the two
4 Kelvin is a measurement of heat energy or temperature which advances in the same increments as does
Celsius. Its principal difference is that Kelvin measurements are written as K and have a much lower starting
point; 0 K or 0 Kelvin is measured at −273.15°C, which is the point at which no heat energy exists in a
substance (called absolute zero). Celsius converts to K by adding 273.15 to the Celsius number.
5 Spectral mixture analysis provides an efficient mechanism for the interpretation and classification of
remotely sensed multidimensional imagery. It aims to identify a set of reference signatures (also known as
“end members”) that can be used to model the reflectance spectrum at each pixel of the original of a finite
number of ground components (Plaza et al., 2002).
ESTIMATING LAND SURFACE ENERGY BUDGETS USING REMOTE SENSING DATA
39