You are on page 1of 15

0016-7622/2012-80-5-685/$ 1.00 GEOL. SOC.

INDIA
JOURNAL GEOLOGICAL SOCIETY OF INDIA
Vol.80, November 2012, pp.685-699
Generation of Emissivity and Land Surface Temperature Maps
Using MODIS TIR Data for Lithological Mapping over
the Singhbhum-Orissa Craton
T. J. MAJUMDAR
1*
,

S. K. PAL
2
and AMIT K. BHATTACHARYA
3
1
Space Applications Centre (ISRO), Ahmedabad 380 015
2
Department of Applied Geophysics, Indian School of Mines, Dhanbad 826 004
3
Department of Geology and Geophysics, Indian Institute of Technology, Kharagpur 721 302
Email: tjmajumdar@rediffmail.com
Abstract: The present study was undertaken with four fold objectives, namely, (i) to estimate land surface temperature
using MODIS TIR data; (ii) to calculate relative emissivities from MODIS TIR data; (iii) to identify various lithologies
based on relative emissivity and land surface temperature estimation; and finally, (iv) to carry out comparative assessment
analysis between the prepared lithological map and the published lithological map. The land surface temperatures for
different pixels were estimated using two methods, viz., Reference Channel and Emissivity Normalization; whereas,
relative emissivities were calculated by applying three methods, viz., Reference Channel, Emissivity Normalization and
Alpha Residual. Lithological maps were subsequently prepared based on the estimated land surface temperatures and
relative emissivity values. The present study shows that the Emissivity Normalization method gives the best results for
land surface temperature estimation and also for lithological discrimination based on emissivity estimation. Twenty-four
lithounits demarcated by the present study match with those of the published map, while four lithounits of the published
map could not be identified in the present study. On the other hand, six additional unclassified lithounits could be
demarcated in the present study, which need to be crosschecked by field study.
Keywords: MODIS, Emissivity, Lithological mapping, Singhbhum-Orissa craton.
while it itself, in turn, controls most of the physical, chemical
and biological processes of the earth (Becker and Li, 1990).
Variations in spectral emissivity are particularly useful for
geological mapping since these relate to differences in
compositions (Lyon, 1965). Thus, geological studies are
more concerned with emissivity variations that provide a
means of lithological mapping and less concerned with
temperature. However, surface temperature effects that mask
subtle variations in emissivity dominate the measured
radiance. This leads to the developments of a variety of
techniques for lithological discrimination, which either
enhance or separate the emissivity from the temperature
effects. Number of scientists (Becker, 1987; Kahle, 1987;
Kealy and Hook, 1993; Wan and Dozier, 1996; Wan and
Snyder, 1996; Wan and Li, 1997; Gillespie et al. 1998; Liang,
2001; Dash et al. 2002; Wan et al. 2002; Sobrino et al. 2003;
Wan et al. 2004) have worked extensively for developing
techniques for extracting LST and emissivity values, which
are derived from radiance data separation in TIR regions.
Some of these techniques are: the Reference Channel
INTRODUCTION
The measured radiance from the earth surface in the
thermal infrared region is a function of both emissivity and
temperature information. Emissivity calculations and
subsequent estimation of land surface temperature (LST)
using MODIS thermal infrared (TIR) bands have opened
up new possibilities for satellite based lithological mapping,
as emissivity is controlled by the composition of the surface
rock and is often used for constituent/lithological mapping.
In this context, silicate minerals play important roles, as
emissivity characteristics of silicate minerals are found to
be useful indicators of lithology. In the present study, the
term relative emissivity / emittance is more relevant than
absolute emissivity, as it is related to measurement of
natural surface rather than ideal specimen. LST is known to
be one of the key diagnostic parameters of the physical
processes of land surface, involving both surface and
subsurface geology (Becker and Li, 1990). LST is controlled
by surface energy balance, atmospheric condition, and
thermal properties of surface and subsurface formation;
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
686 T. J. MAJUMDAR AND OTHER
method, the Emissivity Normalization method, the
Temperature-Independent Spectral Indices method, the
Spectral Ratio method, and the Alpha Emissivity (Alpha
Derived Emissivity/Alpha Residual) method (Li et al. 1999).
However, this study assesses the performances of three
techniques, viz., the Reference Channel method, the
Emissivity Normalization method and the Alpha Emissivity
(Alpha Derived Emissivity/ Alpha Residual) method using
Moderate Resolution Imaging Spectrometer (MODIS) data.
MODIS is an earth-viewing sensor that is mounted on
the Earth Observing Systems, namely, Terra and Aqua
satellites, which were launched in 1999 and 2002
respectively. MODIS has a swath of 2330 km (cross-track)
by 10 km (along-track at nadir). It scans the earth from a
polar orbiting sun-synchronous platform at an altitude of
705 km. MODIS provides multispectral image in thirty six
(36) spectral bands ranging between 0.415 m (visible) to
14.235 m (thermal infrared) with spatial resolutions of
250m (first two bands), 500m (next five bands) and 1 km
(remaining twenty nine bands). Details of MODIS band
specifications are available at https://lpdaac.usgs.gov/
products/modis_products_table.
GEOLOGICAL SETTING OF THE STUDY AREA
The present study area lies between latitudes 2050'

N
and 2324'

N and longitudes 8456' E and 875' E, covering
parts of Jharkhand, West Bengal and Orissa States of India.
This area constitutes geologically one of the most complex
and mineralogically rich belts of the Indian subcontinent.
The area has been extensively studied by various geologists
(Dunn, 1929; Sarkar and Saha, 1963; Naha, 1965; Sarkar
and Chakraborty, 1982; Saha, 1994) using ground based
conventional geological methods. The area has a major
tectonic element, known as Singhbhum shear zone (SSZ)
that separates the cratonic block (Singhbhum-Orissa craton)
in the south from the Proterozoic mobile belt (Singhbhum
mobile belt) in the north. The SSZ runs as a northward
dipping zone along a northwardly convex arcuate belt of
length about 160 km from Bhaharagora in the east to
Chakradharpur in the west. Various rocks of the study area
have undergone several phases of deformation and
metamorphism. Rocks south of the Singhbhum shear zone
are relatively less metamorphosed compared to those in the
north. Rocks of Older Metamorphic Group (OMG) form
the basement rocks. They are exposed in the central part of
the basin. OMG mainly consists of schist. The Iron Ore
Group (IOG) rocks overlie the basement rocks and are
exposed over vast areas in the western part and to some
extent in the eastern part. The IOG succession is believed
to have formed a broad NNE plunging synclinorium with
overturned western limb. The published geological map
(Fig. 1) of the study area (after Saha, 1994), prepared by
conventional geological mapping method shows the
occurrence of twenty eight lithologies.
DATA USED
In order to extract emissivity and subsequent estimation
of LST, suitable TIR bands of MODIS data have to be used.
For atmospheric transmission (Wan, 1999) at viewing angle
of 45 from nadir in mid-latitude summer condition (i.e.
column water vapor (cwv) = 2.9 cm and surface visibility
at 0.55 m (vis.) = 23 km, bands 20, 22, and 23 are recorded
in the 3.5 - 4.2 m range, while bands 29 to 32 are recorded
in the 8 -13 m range. All these are within the atmospheric
window. However, bands 33 to 36 (13.34 - 14.24 m) are
recorded at the edge of this atmospheric window. The
atmospheric transmission in band 33 is very low (lower
than 0.35 in many cases) under different atmospheric
conditions (Liang, 2001); while band 30 is strongly affected
by ozone absorption. Hence, retrieval of LST using this
band (30) is not very effective. Major absorbers in bands
20, 22, and 23 are CO
2
, N
2
and water vapor while major
absorbers in bands 29 and 31-33 are water vapor and CO
2
(Wan, 1999). The atmospheric transmission corresponding
to aerosol scat-tering and absorption in these bands is
approximately within the range 0.95-0.98. Therefore, TIR
bands 20, 22, 23, 29, and 31 to 33 can be used for retrieving
surface emissivity and temperature (Wan and Snyder, 1996;
Wan, 1999).
MODIS (Level 1B) bands, 20-36 of MOD021KM of
May 7, 2003 have been chosen for this study. This
MOD021KM contains calibrated radiances generated from
MODIS Level 1A scan of raw radiance. MODIS technical
specifications are available at https://lpdaac.usgs.gov/
products/modis_products_table.
Remotely sensed images include information about the
atmosphere as well as information about the surface. For
quantitative analysis of surface reflectance, removal of the
influence of the atmosphere is important. To compensate
for atmospheric effects, properties such as the amount of
water vapor, distribution of aerosols, and scene visibility
must be known. Because direct measurements of these
atmospheric properties are rarely available, there are
techniques that infer them from their imprint on hyper-
spectral radiance data. These properties are then used to
constrain highly accurate models of atmospheric radiation
transfer to produce an estimate of the true surface reflectance
(Adler-Golden et al. 1999).
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 687
Fig.1. Geological map of the study area (after Saha, 1994). 1-Older Metamorphic Group; 2-Older Metamorphic Tonalite-gneiss; 3-Pala
Lahara Gneiss; 4-Singhbhum Granite-Phase-I; 5-Singhbhum Granite-Phase-II and xenolith-dominated areas of Bonai Granite;
6-Nilgiri Granite; 7-Iron Ore Group lavas, ultramafics; 8-Iron Ore Group shales, tuffs, phyllites; 9-BHJ, BHQ and sandstone-
conglomerate of Iron Ore Group; 10-Singhbhum Granite Phase-III, Bonai Granite, Chakradharpur Granite; 11(a)-Singhbhum
Group pelites, 11(b)-mafic bodies 11(c)-carbon phyllite; 12-Singhbhum Group quartzites; 13-Dhanjori Group(unclassified); 14-
Quarzite-conglomerate-pelite of Dhanjori Group; 15-Dhanjari-Simlipal-Jagannathpur-Malangtoli lavas; 16-Dalma Lavas; 17-
Proterozoic Gabbro-anorthosite-ultramafics; 18-Kolhan Group and equivalents 19-Mayurbanj Granite; 20-Soda granite, Arkasani
Granite, Kuilapal Granite, alkaline granite; 21-Charnockite; 22-Khondalite; 23-Amphibolite enclaves (within CGG) 24-pelitic
enclaves within CGG; 25-Chhotanagpur granite-gneiss(CGG); 26-Porphyritic member of CGG; 27-Gondwana sediments 28-
Alluvium, Tertiaries.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
688 T. J. MAJUMDAR AND OTHER
METHODOLOGY
It is well established that spectral emissivity
characteristics for the terrestrial land cover types are
relatively stable (least variation of emissivity), in the
wavelength range 10.5 to 12.5 m, which are covered by
bands 31 and 32 of MODIS (Wan and Dozier, 1996).
Salisbury and DAria (1992) have recorded spectral
reflectance values of eighty (80) terrestrial material samples,
which include igneous, metamorphic and sedimentary fresh
rocks, varnished rock surfaces, soil, green foliage, water,
ice and oil slicks. Wan (1999) has shown that constant
emissivity approximations within the range of 0.97 to 0.99
in bands 31 and 32 are observed quite good for all natural
land cover types except for exposed rocks and sands.
However, according to Kealy and Hook (1993), most of the
geological earth surfaces have an emissivity of about 0.96.
Accordingly, the present study has been carried out with
two constant emissivitiy approximations of 0.96 and 0.98
for both band 31 and 32 independently. Wan and Snyder
(1996) and Wan (1999) have shown that the band
emissivities in MODIS bands 31 and 32 are relatively stable.
In Plancks equation, the spectral radiance of a black
body at temperature T can be written as,
] 1 ) [exp(
2 5
1

=
T
C
C
L
j
j
BB
ij


(1)
where L
ij
BB
is black body radiance (Wm
-3
), i is the pixel no.
of band j, T is the temperature of black body (K),
j
is
wavelength of band j (in m), first radiance constant, C
1
is
3.74151 x 10
-16
(Wm
2
), and second radiance constant, C
2
is
0.0143879 (mK). The spectral emissivity (
ij
) of a material,
in band j for pixel i, is defined as the ratio of the radiance of
that material (L
ij
) to that of a blackbody (L
ij
BB
) at the same
temperature as the material, which can be mathematically
expressed as,
BB
ij
ij
ij
L
L
= or
BB
ij ij ij
L L =
(2)
The emissivity spectra over various lithounits can be
extracted to evaluate effectiveness of different methods in
geological studies, since lithological discriminations are
dependent on the positions and depths of minima and
maxima in the spectrum.
Reference Channel Method
The Reference Channel method assumes that all the
pixels of a specific band of the thermal infrared data have a
constant emissivity value (Kahle et al. 1980). On this
assumption, in the present study, bands 31 and 32 of MODIS
data are used with constant emissivity of 0.96 and 0.98,
separately to calculate LST and band relative emissivity.
Once the emissivity for a Reference Channel is assumed,
the temperature T can be calculated for that band using the
following equation (Kealy and Hook, 1993):
] 1 ln[
5
1
2
+
=

R R
R
R
R
L
C
C
T (3)
where, subscript R denotes the particular Reference
Channel, and L
R
is the radiance at that band. This
temperature information can then be used to calculate the
emissivity values for all other bands using equations (1)
and (2).
Emissivity Normalization Method
This method is similar to the Reference Channel method
(Gillespie, 1985) with a little difference. In this method,
the temperature is calculated for every pixel in each band
using constant emissivity value (e.g. 0.96 or 0.98 separately).
The highest of these temperatures at each pixel in different
bands is defined as the temperature of that pixel. Thus
temperature image for the entire study area is calculated.
This temperature information is then used to calculate the
emissivity values using equations (1) and (2).
Alpha Residual Method
This method makes use of Wiens approximation of
Plancks equation, which is given as,
)] [exp(
2 5
1
i j
j
BB
ij
T
C
C
L


=
(4)
In Wiens approximation of Plancks equation, the -1
term, in the denominator, is neglected so that the equation
can be linearized with logarithm. The Alpha Residual method
is a two-step procedure: (i) initially, natural logarithms are
formed from the Wiens approximation of Plancks equation
and subsequently wavelength weighted logarithm of Wien
radiance at pixel i in band j (X
ij
) is generated in order to
separate and T ; (ii) the geometric mean over all bands
for pixel j (X
ij
) is then subtracted from the wave-
length weighted log of Wien radiance at pixel i in
band j (X
ij
). The procedure is discussed in details elsewhere
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 689
(Hook et al. 1992). Brief mathematical procedure is as
follows:

(5)
Finally, alpha residual,
ij
, is obtained as given below:
j
n
j
ij j ij j iJ ij ij
K
n
X X + = =

=1
ln
1
ln
(6)
where,
From equation (6), it is clear that K
j
is not a function of
measured radiance but wavelength dependent constant only,
which can be calculated for each band. Thus, the Alpha
Residual spectrum is a function of emissivity only. Kealy
and Hook (1993) have shown that the Alpha Residual spectra
are similar in shape as emissivity spectrum, but have zero
mean. They also concluded that the Alpha Residual spectrum
could be of great help in geological studies since the
lithological discriminations are dependent on the positions
and depths of minima and maxima in the spectrum.
RESULTS
Reference Channel Method
Emissivity values and LST over the study area are
calculated assuming constant emissivities of bands 31 and
32 (as stable bands) for different band combinations, viz.,
(i) 20-33 (i.e., 20, 22, 23, 29, 31, 32 and 33), and (ii) 29-33
(i.e., 29, 31, 32 and 33) of MODIS data using Reference
Channel method. Besides these LST estimations over the
study area, ambient temperatures have been collected from
different available weather stations of Indian Meteorology
Department close to the study area (Table 1, All India
Weather Bulletin, 2003). These stations are: Kolkata,
Jamshedpur, Ranchi, Gaya, Patna and Bhubaneswar. The
temperatures recorded from these stations provided a
comparison of the temperatures estimated by various
methods followed in the present study. Tables 2a and b show
the estimated temperatures using Reference Channel method,
for different band combinations assuming 0.96 and 0.98
emissivity values for bands 31 and 32 respectively, along
with the recorded maximum and minimum temperatures.
The fact that MODIS data acquisition over the study area
has been at 10.30 A.M., the estimated LST using the methods
adopted in the present study will be slightly higher than the
mean of minimum and maximum recorded temperatures at
different weather stations; this fact should be taken into
account for assessment of estimated LST results (Majumdar
and Bhattacharya, 1988).
Histograms of LST over the study area using Reference
Channel method are shown in Fig. 2a for comparative study
and comprehensive assessment of LST calculations. Thus,
from Tables 2 a and b and Fig. 2a, it can be said that LST,
calculated using band combination 29-33 with emissivity
value 0.98 for band 31 (henceforth, denoted as 2933-98-
31) generates a temperature closer to the observed
temperature. Moreover, LST calculation, with the constant
emissivity 0.98, is more accurate in Reference Channel
method than that of the constant emissivity 0.96. Further,
ij
BB
ij ij
L L ln ln ln = + =
i j
j ij
T
C
C


2
1
ln ln 5 ln ln + =
Table 1. All India Weather Bulletin (dated 08.05.2003)
Station Latitude Longitude Temperature (C)
Max. Min. Mean
Patna 25.35N 85.12E 40 22 31
Gaya 24.47N 85.04E 41 24 32.5
Ranchi 23.19N 85.27E 38 22 30
Jamshedpur 22.40N 86.12E 43 26 34.5
Kolkata 22.36N 88.24E 35 18 26.5
Bhubaneswar 20.15N 85.50E 39 28 33.5
Table 2. Comparative study of LST calculation using Reference Channel
Method with the stable band (a) 31, (b) 32 and (c) Normalization
method with constant emissivities 0.96 and 0.98 for different
band combinations for different band combination
(a) Reference Channel method with constant emissivity for band 31
Station 20_33_98_31 20_33_96_31 29_33_96_31 29_33_98_31
Patna 41.69 42.24 39.24 37.69
Gaya 36.49 41.98 37.98 35.58
Ranchi 37.58 42.13 38.13 34.39
Jamshedpur 39.93 43.48 41.48 38.93
Kolkata 32.70 26.09 29.09 26.7
Bhubaneswar 41.98 40.45 36.45 32.98
(b) Reference Channel method with constant emissivity for band 32
Station 20_33_98_32 20_33_96_32 29_33_96_32 29_33_98_32
Patna 43.31 42.00 45 38.31
Gaya 40.74 34.33 34.33 36.74
Ranchi 42.88 41.56 41.56 37.88
Jamshedpur 43.48 43.15 42.15 41.48
Kolkata 23.80 25.30 25.3 23.80
Bhubaneswar 40.50 32.06 32.06 30.50
(c) Normalization method
Station 20_33_98 20_33_96 29_33_96 29_33_98
Patna 43.76 42.34 38 34.31
Gaya 43.67 44.26 36.98 35.49
Ranchi 40.83 42.40 34.56 32.88
Jamshedpur 45.69 45.26 37.48 36.93
Kolkata 38.33 40.87 26.09 25.70
Bhubaneswar 41.27 45.83 34.45 32.98
=
j j j j j
C K
1
ln
ln ln 5 ln

= = =
+ +
n
j
j
n
j
j j
n
j
j
n n n
C
1 1 1
1
ln
ln
5 ln


JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
690 T. J. MAJUMDAR AND OTHER
bands 29 and 31-33, which are in the 8-13 m atmospheric
window are more suitable for LST calculation using
Reference Channel method, than that of the bands 20, 22,
and 23, which are in the atmospheric window in the 3.5-
4.2 m medium wavelength range. Temperatures observed
in bands 20, 22, and 23 are higher than the expected LST
over the area. Detail statistics of calculated LST using
different technique are enlisted in Table 3. LST (Min =
24.59C, Max = 39.16C, Mean = 29.40C, Std. dev =
3.23C) calculated using Reference Channel method from
the bands 29, 31, 32 and 33 with constant emissivity 0.98
for the band 31, is shown in Fig.3. An approach has been
made to discriminate different lithological boundary as per
LST and the inferred lithological boundary map (vector
coverage) has been overlaid on it.
Emissivitiy spectrum is extracted from band emittances
(emissivities) 20, 22, 23, 29, 31, 32 and 33 assuming
constant emissivity values of 0.96 and 0.98, in separate cases
for the stable band 31 and 32, (denoted as Ref_20_33_
96_31, Ref_20_33_98_31, Ref_20_33_96_32 and
Ref_20_33_98_32, respectively). Further relative emissivity
spectra have been collected from ten major distinct lithounits
(1. Iron Ore Group, 2. Chhotanagpur Granite-Gneiss, 3.
Dalma Lavas, 4. Dhanjori-Simlipal-Jagannathpur-
Malangtoli Lavas, 5. Singhbhum Granite, 6. Alluvium, 7.
Alluvium/Gondwana Sediments, 8. Nilgiri Granite,
9. Singhbhum Group Pelites and 10. Granite/Bonigranite,
refer Fig.1) for applicability to demarcate lithology over
the area, separately for four sets of band emittances
(emissivities). However, out of the four, the spectrum
Ref_20-33_96_31 (Fig.4a) has more lithological
distinguishing capability. In the band emissivities, emissivity
calculated using constant emissivity value 0.96 with the
stable band 31 is considered for demarcation of different
lithounits (Kealy and Hook, 1993). In multispectral image,
contiguous bands have very little tonal and textural variation,
therefore, bands are selected for generation of FCC images,











































(a)
(b)
Fig.2. Histograms of LST over the study area for various
combinations using (a) Reference Channel method, and
(b) Emissivity Normalization method.



























0
C
0
C
0
C
0
C
0
C
28
28
30
10
25
25
25
15
15
5
2
8
16
34
27 22
10
11
11
12
19
6
19
31
33
15
35
26
24
19
3
11
16
21 8
9
9
Fig.3. LST calculated using Reference Channel method from bands
29, 31, 32 and 33 assuming constant emissivity 0.98 for
the stable band 31. Inferred lithounits over the study area
are traced out as shown in this figure. Serial numbers 1 to
28 represent similar lithologies as those of Fig. 1. Serial
numbers 30-35 are unclassified A,B,C,D,E,F.
Table 3. Statistics of calculated LSTs using different techniques
Temperature calculation Min Max Mean Stdev
methods C C C
ref_20_33_31_96_T 27.009521 48.787323 38.789538 3.301411
ref_20_33_31_98_T 25.593414 47.167480 37.264544 3.270515
ref_20_33_32_96_T 26.040833 47.446014 38.009202 3.464456
ref_20_33_32_98_T 24.516479 45.707733 36.366500 3.429800
ref_29_33_31_96_T 27.009521 48.787323 38.789538 3.301411
ref_29_33_31_98_T 24.593414 39.167480 29.403814 3.237075
ref_29_33_32_96_T 26.040833 47.446014 38.009202 3.464456
ref_29_33_32_98_T 24.516479 45.707733 36.366500 3.429800
nor_29_31_33_96_T 27.009521 48.787323 38.801684 3.306163
nor_29_31_33_98_T 22.793414 36.932480 25.403361 3.237217
nor_2022232931_33_96_T 32.408447 67.764740 49.040770 4.835214
nor_2022232931_33_98_T 31.901428 67.058929 48.477315 4.818213
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 691
so that the interband relative emissivity difference and
standard deviation are maximum for easy demarcation of
different lithounits based on tonal and textural variations
over surroundings. Figure 5a shows histogram of relative
emissivity of seven bands (20-33) calculated using Reference
Channel method assuming constant emissivity 0.96 for band
31. Figure 5b shows plot of minimum, maximum, and mean
of band emissivities. The detailed statistics of band emittance
are listed in Table 4 a. Bands 23, 20 and 33 have been chosen
to generate FCC image (Fig. 5c) for delineation of various
lithounits based on the Figs. 5a and b, and Table 4a, so that
the interband relative emissivity difference and standard
deviation are maximum.
Emissivity Normalization Method
Emissivity values and LST over the study area have also
been calculated assuming constant emissivities, 0.96 and
0.98, in separate cases for different band combinations,
viz., (i) 20-33 (i.e., 20, 22, 23, 29, 31, 32 and 33), and
(ii) 29-33 (i.e., 29, 31, 32 and 33) of data using Emissivity
Normalization method.
Histograms of LST over the study area using
Normalization method are shown in Fig.2b for comparative
study and comprehensive assessment of LST calculation
using different methods. Table 2c shows a comparative
study of LST calculation using Normalization method
assuming constant emissivities for different band
combinations. Analyses of Fig. 2b and Table 2c indicate
that LST calculation using Normalization method with band
combination 29-33 assuming constant emissivity 0.98
provides closer LST (Table 1, All India Weather Bulletin,
2003) than that by assuming constant emissivity 0.96.






























Wavelength( m)
A
l
p
h
a

r
e
s
i
d
u
a
l


(c) Alpha_20-33
Wavelength( m)
E
m
i
s
s
i
v
i
t
y

(b)Nor_20-33_96
Wavelength( m)
E
m
i
s
s
i
v
i
t
y


(a)Ref_20-33_96_31
Fig.4. Emissivity spectral plots collected from different lithological units (a) obtained from Reference Channel method assuming constant
emissivity 0.96 and band 31 as stable. (b) obtained from Emissivity Normalization method assuming constant emissivity 0.96.
(c) obtained from Alpha Residual method. Lithounits: 1.Iron Ore Group, 2.Chhotanagpur Granite-Gneiss, 3.Dalma Lavas,
4.Dhanjori-Simlipal-Jagannathpur-Malangtoli Lavas, 5.Singhbhum Granite, 6.Alluvium, 7.Alluvium / Gondwana Sediments,
8.Nilgiri Granite, 9.Singhbhum Group Pelites, and 10.Singhbhum Granite/ Bonigranite.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
692 T. J. MAJUMDAR AND OTHER
Further, it is also observed that the LST obtained using
Normalization method is more appropriate than that obtained
using Reference Channel Method. Detailed statistics of
generated LST using different technique are enlisted in
Table 3. The LST (Min= 22.79
0
C, Max= 36.93
0
C, Mean=
25.40
0
C, Std dev= 3.23
0
C) calculated using emissivity
normalization method from the bands 29, 31, 32 & 33
assuming constant emissivity 0.98 is shown in Fig. 6.
Inferred lithounits are similar to them as observed in Fig. 1,
as mentioned in Fig. 6.
Emissivitiy spectra are extracted from band emittance
of the bands 20-33 using constant emissivity values 0.96
and 0.98, (denoted as Nor_20-33_96 and Nor_20-33_98,
respectively). However, out of the two, the spectrum
Nor_20-33_96 (Fig.4b) has more lithological distinguishing
capability. The band emittances (emissivities) estimated by
constant emissivity value 0.96 (Kealy and Hook, 1993) has
been considered for demarcation of different lithounits.
Histograms of relative emissivity of seven bands (20-33)
derived using Emissivity Normalization method with
constant emissivity 0.96, and plots of minimum, maximum
and mean of band emissivities have been generated, similar
in the line of Figs.5a and b used for Reference Channel
method. The detailed statistics of band emitance are listed
in Table 4b. Bands 32, 23 and 33 have been chosen to
generate FCC image (Fig. 7a) for delineation of various
lithounits, after a careful study of the generated histogram
and plots of maximum and minimum emissivities. Another
FCC image (Fig. 7b), generated by selecting bands 29, 32
and 31 from band combination 29, 31, 32 and 33, assuming
constant emissivity 0.98, has been found to be suitable in
discriminating four broad lithologies, namely, (I) Alluvium
fan/sediments, (II) Dhanjari-Simlipal-Jagannathpur-
Malangtoli lavas (III) Singhbhum-Niligiri-Bonai-
Chakradharpur granites etc (IV) Singhbhum group pelites/
metamorphosed argillaceous rock/ Chhotanagpur granite-
gneiss (CGG).
Alpha Residual Method
A spectrum (Fig. 4c) of alpha derived (equation-5) band



F
r
e
q
u
e
n
c
y

Relative emissivity

(a)
(b)
Wavelength ( m)
E
m
i
s
s
i
v
i
t
y






































28
28
30
10
25
25
25
15
15
5
2
8
16
27
22
10
11
11
12
19
6
19
31
33
15
35
26
24
19
3
11
16
21
8
9
9
18
9
34
Fig.5. (a) Histogram of relative emissivity of seven bands (20-33) using Reference Channel method with constant emissivity 0.96 for
band 31. (b) Plot of minimum, maximum and mean of band emissivities. (c) FCC image generated from band emissivities 23, 20
and 33 using Reference Channel method (from bands 20,22,23,29,31,32 and 33 for constant emissivity 0.96 for the band 31 as
stable). Inferred lithounits over the study area are traced out as shown in this figure. Serial numbers 1 to 28 represent similar
lithologies as those of Fig.1. Serial numbers 30-35 are unclassified A,B,C,D,E,F.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 693
emissivities (Alpha Residuals) have been collected from
seven bands, 20, 22, 23, 29, 31, 32 and 33 (denoted as
Alpha_20-33) from different lithounits. Histograms of alpha
residuals of seven bands (20-33) have been generated using
Alpha Residual technique as well as plots of minimum,
maximum and mean of alpha residual for all seven bands,
similar in the line of Figs. 5a and b used for Reference
Channel method. The detailed statistics of band alpha
residuals are listed in Table 4 c. Bands 32, 20 and 33 have
been chosen to generate FCC image (Fig. 8) for delineation
of various lithounits, based on the assessment of the
generated histogram and plots of maximum and minimum
emissivities and Table 4c; so that the interband Alpha
Residual difference and standard deviation are maximum.
DISCUSSION
Reference Channel method has some limitations. First,
it assumes the emissivity in one band is constant and hence,
spectral information in that band is lost. An incorrect
assumption may lead to a decreased or an increased value
of spectrum and also could rotate the spectrum clockwise
or counterclockwise around the assumed stable band,
depending upon whether the assumed emissivity value is an
over or underestimate. Further, if the data in the constant
band contains noise, this noise is introduced into the other
bands. Likewise, if the data of the stable band is highly
correlated with the adjacent bands, the data in the adjacent
bands will also be forced towards the constant emissivity
value and the signal to noise ratio between the two bands
will be decreased. On the other hand, Alpha Residual
method has the disadvantage of assuming initially Wiens
approximation, although for temperature of 300K and a
wavelength of 10 m, the approximation results in less than
1% error (Hook et al. 1992).
The relative emissivity spectra as shown in Fig. 4 have
been collected from ten different lithologies over the study
area, viz., 1. Iron Ore Group 2. Chhotanagpur granite-gneiss
3. Dalma Lavas 4. Dhanjori-Simlipal-Jagannathpur-
Malangtoli Lavas 5. Singhbhum Granite 6. Alluvium 7.
Alluvium/Gondwana Sediments 8. Nilgiri Granite 9.
Singhbhum Group pelites 10. Singhbhum Granite/
Bonigranite. Most of the spectra have two strong emission
minima, near 9 m and 12 m. The strong broad emission
minimum near 9 m would be due to presence of clay
minerals whereas, the strong emission minimum near 12
m would be due to bending modes involving the carbonate
ion occurrences (Salisbury et al. 1987). Further, one
weak emission minimum near 11 m would be due to Si-O
stretching in the silicate lattice, which is generally referred
Table 4. Band-wise statistics of calculated relative emissivities using (a)
Reference Channel method (b) Normalization method (c) Alpha
Residual method
(a) Reference Channel method
Band Min Max Mean Stdev
20 0.339658 0.9956547 0.506918 0.356985
22 0.389469 0.984326 0.497243 0.322132
23 0.365449 0.9836335 0.416184 0.260765
29 0.785645 0.9758229 0.859302 0.387156
31 0.859999 0.960001 0.914099 0.419045
32 0.778927 0.971430 0.857972 0.415532
33 0.486170 0.776168 0.579503 0.282247
(b) Normalization method
Band Min Max Mean Stdev
20 0.459469 0.960001 0.713656 0.419033
22 0.336292 0.960000 0.626716 0.368770
23 0.258977 0.897632 0.529700 0.312483
29 0.374180 0.832647 0.546665 0.321956
31 0.485807 0.904849 0.616616 0.362860
32 0.480991 0.904395 0.618256 0.363931
33 0.374519 0.665441 0.423605 0.249541
(c) Alpha Residual method
Band Min Max Mean Stdev
20 1.334053 6.291672 1.549069 0.938540
22 0.826103 5.102219 1.158558 0.714282
23 0.278305 4.349609 0.637763 0.411774
29 -2.743373 -0.037990 -0.265211 0.193626
31 -2.062538 0.729897 0.265561 0.230183
32 -2.730057 0.781525 0.226415 0.268294
33 -7.244583 -3.357956 -3.572163 2.127594
0
C

35
0
C

0
C

0
C

0
C

28
28
30
10
25
25
25
15
15
5
2
8
16
34
27 22
10
11
11
12
19
6
19
31
33
15
35
26
24
19
3
11
16
21 8
9
9
Fig.6. LST calculated using Emissivity Normalization method from
the bands 29, 31, 32and 33 assuming constant emissivity
0.98. Inferred lithounits over the study area are traced out
as shown in this figure. Serial numbers 1 to 28 represent
similar lithologies as those of Fig.1. Serial numbers 30-35
are unclassified A,B,C,D,E,F.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
694 T. J. MAJUMDAR AND OTHER
as Reststrahlen band (Hunt, 1980). The Reststrahlen band
provides a means for identifying silicate rocks, as it migrates
to longer wavelengths as the material becomes more mafic
with decreasing silica and increasing ferromagnesium
minerals (Hunt, 1980; Salisbury et al. 1988). As MODIS
TIR data have very poor spatial resolution (~1 km) and also
low spectral resolution, spectra collected over the lithologies
are not pure. The potential pitfalls would be due to slight
variations in mineral composition, mineral admixture,
degree of weathering and some uncertainties in sensor
calibration and atmospheric correction, which lead to the
presence of some artifacts and may create difficulty in
correlating the collected spectra with the available library
spectra. Therefore, the spectral classification techniques
based on the spectral values of the extracted emissivity
images, would not be effective for discriminating the various
lithologies.
Hence, in the present study, the different lithounits
have been demarcated on the basis of tonal and textural
variations, textural association and color contrast in the
different FCCs, generated from LST and emissivity images
and corroborated with the published lithological map
(Saha, 1994). However, it is clear that the estimated LST
28
28
30
10
25
25
25
15
15
5
2
8
16
34
27 22
10
11
11
12
19
6
19
31
33
15
35
26
24
19
3
11
16
21 8
9
9
18
9
I
I
IV
II
III II

28
28
30
10
25
25
25
15
15
5
2
8
16
34
27 22
10
11
11
12
19
6
19
31
33
15
35
26
24
19
3
11
16
21 8
9
9
9
18
22
Fig.7. (a) FCC image generated from band emissivities, 32, 23, and 33 using Emissivity Normalization method (from bands
20,22,23,29,31,32 and 33 assuming constant emissivity 0.96). Inferred lithounits over the study area are traced out as shown in
this figure. Serial numbers 1 to 28 represent similar lithologies as those of Fig. 1. Serial numbers 30-35 are unclassified A,B,C,D,E,F.
(b) FCC image generated from band emissivities, 29, 32, and 31 using Emissivity Normalization method (from bands 29,31,32
and 33 assuming constant emissivity 0.98). Four broad comprehensive areas have been identified: (I) Alluvium fan/sediments,
(II) Dhanjari-Simlipal-Jagannathpur-Malangtoli lavas, (III) Singhbhum-Niligiri-Bonai-Chakradharpur granites, and (IV)
Singhbhum group pelites / metamorphosed argillaceous rock / Chhotanagpur granite-gneiss.
Fig.8. FCC image generated from alpha residuals, 31, 20, and 33
using Alpha Residual method (from bands 20,22,23, 29,
31,32 and 33). Inferred lithounits over the study area are
traced out as shown. Serial numbers same as in Fig.7.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 695
are less concerned for discrimination of various lithounit
boundaries. Vector topologies for various identified
lithounits are created to generate closed polygons for all
lithounits. The generated vector topologies are overlaid on
various FCC imageries as shown in Figs. 5c, 7a and 8
obtained using Reference Channel method, Emissivity
Normalization method and Alpha Residual method
respectively. It is clear that the Normalization method is
most helpful in discriminating as observed from the relative
emissivity spectra in Figs. 4 a, b and c and the FCC images
in Figs. 5a, 7a and 8. Figure 9 shows the lithological map
(henceforth denoted as, MOD-MAP) prepared from the FCC
(Fig.7a) of band emitance extracted using Emissivity
Normalization method.
Comparative assessment analysis has been carried out
between the published geological map of Saha (1994) and
the MOD-MAP. Majority of the lithounits, observed in the
published map (Saha, 1994), have been delineated in the
MOD-MAP, as can be seen from Figs. 1, 9 & 10 and
Table 5. The result of this comparative assessment analysis
Table 5. Details of geological interpretation and comparative assessment study between present interpreted (MOD-MAP) and published
lithology maps (Saha, 1994)
SL. Geology/Lithounits Geology of Geology as interpreted Results of comparative
No. Saha (1994) from present study assessment analysis
(Area, km
2
) (Area, km
2
) (Area, km
2
)
1 Older Metamorphic Group 295.6 88.5 72
2 Older Metamorphic Tonalite-gneiss 1541.8 983.6 854.5
3 Pala Lahara Gneiss 1579.7 1872.4 1490.6
4 Singhbhum Granite-Phase-I 218.5 - -
5 Singhbhum Granite-Phase-II and xenolith-dominated
areas of Bonai Granite 1869.4 1285.8 1254.4
6 Nilgiri Granite 989.2 769.4 674.3
7 Iron Ore Group lavas, ultramafics 817.5 174.9 165.5
8 Iron Ore Group shales, tuffs, phyllites 3579.3 4724.6 2133
9 BHJ, BHQ and sandstone-conglomerate of Iron Ore Group 852.6 700.5 344.8
10 Singhbhum Granite -Phase-III, Bonai Granite,
Chakradharpur Granite 8292.4 7099.5 4918.5
11(c) Singhbhum Group pelites 6916.9 3682.8 3573.9
11(b) Mafic bodis 116.3 - -
11(c) Carbon phyllite 40 - -
12 Singhbhum Group quartzites 622.8 416 310.4
13 Dhanjori Group(unclassified) 38
14 Quarzite-conglomerate-pelite of Dhanjori Group 752 161.4 92
15 Dhanjari-Simlipal-Jagannathpur-Malangtoli lavas 2963.8 4640.8 2855
16 Dalma Lavas 1106.4 3006.5 995.6
17 Proterozoic Gabbro-anorthosite-ultramafics 368 69.5 256.4
18 Kolhan Group and equivalents 1348.7 497.7 290
19 Mayurbanj Granite 1695.5 2438 1513.5
20 Soda granite, Arkasani Granite, Kuilapal Granite,
alkaline granite 74.7 - -
21 Charnockite 174 192 152.6
22 Khondalite 358.8 325.4 191.8
23 Amphibolite enclaves (within CGG) 107.8 - -
24 Pelitic enclaves within CGG 124.5 129 103.4
25 Chhotanagpur granite-gneiss(CGG); 4293.5 3488 3438.6
26 Porphyritic member of CGG 309 311.6 234
27 Gondwana sediments 2413 2827.5 2312.7
28 Alluvium, Tertiaries 6160.6 7194 5940.5
30 Unclassified (A) 972.7 -
31 Unclassified (B) 215 -
32 Unclassified (C) 22.5 -
33 Unclassified (D) 480 -
34 Unclassified (E) 1127.4 -
35 Unclassified(F) 123 -
36 Total Unclassified area as par the comparative
assessment [29 =30+31 +32 +33+ 34+35+ overlaps 15852
of some lithounits]
Total 50020 50020 50020
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
696 T. J. MAJUMDAR AND OTHER

Fig.9. Geological map (MOD-MAP) interpreted from the FCC of band emissivities extracted using Emissivity Normalization method
(Fig. 7 a). Serial numbers 1 to 28 represents similar lithologies as those of Fig.1. Serial numbers 30-35 are unclassified A,B,C,D,E,F.
has been highlighted in Fig.10. The total area of the present
study is about 50,000 sq km. It is observed that a total area
of about 30,000 sq km (60%) has been identified correctly
in the present study. However, a few prominent areas, such
as, in Fig. 10, A and B within the CGC, C- near Dalma Lava,
D and E within the Singhbhum Granite, and F near Dhanjori
Lava have been demarcated as different unclassified
lithounits, due to occurrences of color contrasts, tonal
variation and textural variation of these regions from the
surroundings, which do not correlate with the previous
published geological map of Saha (1994). A total of twenty
eight lithounits have been observed in the published
geological map (Saha, 1994; Fig.1). On the other hand,
although, twenty four lithounits (Fig.9) could be identified
in the present study, which matches well with the published
map, the rest four lithounits could not be delineated from
the present study due to lack of color, tonal and textural
variations. These are: Singhbhum 4- Granite - Phase-I, 13-
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 697
Fig.10. Geological map represent comparative assessment between the published geological map of Saha (1994) and the MOD-MAP.
Serial numbers 1 to 28 represent same lithologies as those of Fig. 1. 29-Change in interpreted geology / unclassified geology as
inferred from present study [30-Unclassified-(A); 31-Unclassified-(B); 32-Unclassified-(C); 33-Unclassified-(D); 34-Unclassified-
(E) and 35-Unclassified-(F)].
Dhanjori Group (unclassified), 20-Soda Granite, and 23-
Amphibolite Enclaves (within CGC).
CONCLUSIONS
From the present study, it can be concluded that the
Emissivity Normalization method using MODIS TIR
bands (29, 31, 32 and 33) which lie within the wave-
length region 8-13 m, and assuming constant emissivity
0.98, provides best LST estimation. The next best estimation
of LST is achieved by Reference Channel method
using same band combination and assuming constant
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
698 T. J. MAJUMDAR AND OTHER
emissivity value of 0.98 with band 32 as the stable one.
In the preparation of lithological map of the area using
estimated LST and relative emissivity, it is observed that
LST based method is not suitable. On the other hand, the
FCC image generated using Emissivity Normalization
method, from the band emissivities 32, 23 and 33 selected
from band combination 20,22,23,29,31,32 and 33 with
constant emissivity 0.96, is found to be very appropriate in
demarcating lithologies over the Singhbhum-Orissa Craton,
which agree well with the published lithological map.
Moreover, Emissivity Normalization method with the band
combination 29, 32 and 31 is observed to be suitable in
delineating four major features, such as, Alluvium fan/
sediments, Dhanjari-Simlipal-Jagannathpur-Malangtoli
lavas, Singhbhum-Niligiri-Bonai-Chakradharpur granites,
Singhbhum Group pelites/ metamorphosed argillaceous
rock/Chhotanagpur granite-gneiss), which need to be
crosschecked by field study. The next best technique for
identifying lithologies is Reference Channel; FCC image
generated using this technique, from the band emissivities,
32, 20 and 33, selected from bands 20,22,23,29,31,32
and 33 with constant emissivity 0.96 and band 31 as stable,
is relatively suitable, whereas, other combinations are
noisy.
Finally, while attempting comparative assessment
analysis of the two lithological maps, viz., MOD-MAP
(inferred lithology) and the published map, it is found that
there is complete agreement over sixty percent (60%) of
the study area. Twenty-four lithounits demarcated by the
present study match with those of the published map, while
four lithounits of the published map could not be identified
in the present study. Six additional unclassified lithounits
could be demarcated in the present study, which need to be
crosschecked by the field study.
Acknowledgements: The authors wish to thank two
anonymous reviewers for their valuable suggestions for
improvement of the manuscript. They are thankful to Shri
A.S. Kiran Kumar, Director, SAC, Ahmedabad and Prof.
D.C. Panigrahi, Director, ISM, Dhanbad for their keen
interest in this study. Thanks are also due to Dr. R.
Bhattacharya, Oil India Ltd., Guwahati, Dr. P.K. Srivastava,
Shri Subhranshu Sanyal, Mr. Sreejith A. P., and Mr.
Sabyasachi Maiti, Department of Geology and Geophysics,
IIT, Kharagpur for their help at various stages of
development of this activity. TJM wishes to thank CSIR,
New Delhi for Emeritus Scientist Fellowship since January
2011.
References
ADLER-GOLDEN, S.M., MATTHEW, M.W., BERNSTEIN, L.S., LEVINE, R.Y.,
BERK, A., RICHTSMEIER, S.C., ACHARYA, P.K., ANDERSON, G.P.,
FELDE, G., GARDNER, J., HOKE, M., JEONG, B. PUKALL, L.S.,
RATKOWSKI, A. and BURKE, H.-H. (1999) Atmospheric correc-
tion for short-wave spectral imagery based on MODTRAN4.
Proc. SPIE Imaging Spectrometry, v.3753. pp. 61-69.
ALL INDIA WEATHER BULLETIN (2003) India Meteorological
Department, New Delhi.
BECKER, F. (1987) Impact of spectral emissivity on the measurement
of land surface temperatures from a satellite. Int. Jour. Rem.
Sens., v.8(10), pp.1509-1522.
BECKER, F. and LI, Z.L. (1990) Temperature-independent spectral
indices in thermal infrared bands. Rem. Sens. Environ., v.32,
pp.17-33.
DASH, P., GOTTSCHE, F.M., OLESEN, F.S. and FISHER, H. (2002) Land
surface temperature and emissivity estimation from passive
sensor data: theory and practicecurrent trends. Int. Jour. Rem.
Sens., v.23(13), pp.2563-2594.
DUNN. J.A. (1929) The geology of north Singhbhum. Mem. Geol.
Soc. India, Bangalore.
GILLESPIE, A.R., ROKUGAWA, S., MATSUNAGA, T., COTHERN, J.S.,
HOOK, S. and KAHLE, A.B. (1998) A temperature and emissivity
separation algorithm for Advanced Spaceborne Thermal
Emission and Reflection Radiometer (ASTER) images. IEEE
Trans. on Geosc. and Rem. Sens., v.36, pp.11131126.
GILLESPIE, A.R. (1985) Lithologic mapping of silicate rocks using
TIMS. The TIMS Data Users Workshop, June 18-19, 1985,
JPL Pub. 86-38, pp. 29-44.
HOOK, S.J., GABELL, A.R., GREEN, A.A. and KEALY, P.S. (1992) A
comparison of techniques for extracting emissivity information
from thermal infrared data for geological studies. Rem. Sens.
Environ., v.42, pp.123-135.
HUNT, G.R. (1980) Electromagnetic radiation: communication link
in remote sensing. In: B. S. Siegal and A. Gillespie (Eds.),
Remote Sensing in Geology. Wiley, New York, pp.5-45.
KAHLE, A.B., MADURA, D.P. and SOHA, J.M. (1980) Middle infrared
multispectral aircraft scanner data analysis for geological
application. Appl. Optics. v.19, pp.2279-2290.
KAHLE, A.B. (1987) Surface emittance, temperature, and thermal
inertia derived from Thermal Infrared Multispectral Scanner
(TIMS) data for Death Valley, California. Geophys., v.52(7),
pp.858-874.
KEALY, P.S. and HOOK, S.J. (1993) Separating temperature and
emissivity in thermal infrared multispectral scanner data:
implications for recovering land surface temperature. IEEE
Trans. Geosc. and Rem. Sens., v.31(6), pp.1155-1164.
LI, Z.L., BECKER, F., STOLL, M.P. and WAN, Z. (1999) Evaluation
of six methods for extracting relative emissivity spectra from
thermal infrared images. Rem. Sens. Environ., v.69, pp.197-
214.
JOUR.GEOL.SOC.INDIA, VOL.80, NOV. 2012
EMISSIVITY AND SURFACE TEMPERATURE ESTIMATION OVER SINGHBHUM 699
LIANG, S. (2001) An optimization algorithm for separating land
surface temperature and emissivity from multispectral thermal
infrared imagery. IEEE Trans. Geosc. and Rem. Sens., v.39(2),
pp.264-274.
LYON, R.J.P. (1965) Analysis of rocks by spectral infrared emission
(8 to 25 microns). Econ. Geol., v.60, pp.715-736.
MAJUMDAR, T.J. and BHATTACHARYA, B.B. (1988) Derivation of
surface temperatures on land after corrections due to
atmospheric water vapor a case study with INSAT VHRR
data. Rem. Sens. Environ., v.26, pp.185-191.
NAHA, K. (1965) Metamorphism in relation to stratigraphy,
structure and movements in parts of east Singhbhum, eastern
India. Quart. Jour. Geol., Min. Metall. Soc. India, v.37, pp.41-
88.
SAHA, A.K. (1994) Crustal evolution of Singhbhum - North
Orissa, Eastern India. Mem. Geol. Soc. India, no.27, 341p.
SALISBURY, J.W, WALTER, L.S. and VERGO, N. (1987) Midinfrared
(2.5-25m) spectra of minerals. U.S. Geological Survey Open-
File Report 87-263.
SALISBURY, J.W., WALTER, L.S. and DARIA, D.M. (1988)
Midinfrared (2.5-13.5m) spectra of igneous rocks: first
edition. U.S. Geological Survey Open-File Report 88-686.
SALISBURY, J.W. and DARIA, D.M. (1992) Emissivity of terrestrial
materials in the 8 14 m atmospheric window. Rem. Sens.
Environ., v.42, pp.83106.
SARKAR, S.N. and SAHA, A.K. (1963) On the occurrence of two
intersecting Precambrian orogenic belts in Singhbhum and
adjacent areas, India. Geol. Magz., v.100(1), pp.69-92.
SARKAR, A.N. and CHAKRABORTY, D.K. (1982) One orogenic belt
or two? A structure reinterpretation supported by Landsat data
products of the Precambrian metamorphics of Singhbhum,
Eastern India. Photogrammetria, v.37, pp.185-201.
SOBRINO, J.A., EL KHARRAZ, J. and LI, Z.L. (2003) Surface
temperature and water vapour retrieval from MODIS data. Int.
Jour. Rem. Sens., v.24(24), pp.5161-5182.
WAN, Z. and DOZIER, J. (1996) A generalized split window
algorithm for retrieving land surface temperature from space.
IEEE Trans. on Geosc. and Rem. Sens., v.34(4), pp.892-
905.
WAN, Z. and SNYDER, W. (1996) MODIS land surface temperature
algorithm theoretical basis document, version 2.3, ICESS/
UCSB, December, http://modarch.gsfc.nasa.gov/ MODIS/
ATBD/atbd mod11.pdf.
WAN, Z. and LI, Z.L. (1997) A physics-based algorithm for
retrieving land-surface emissivity and temperature from EOS/
MODIS data. IEEE Trans. on Geosc. and Rem. Sens., v.35,
pp.980996.
WAN, Z. (1999) MODIS Land-Surface Temperature Algorithm
Theoretical Basis Document (LST ATBD), Version 3.3.
Contract Number: NAS5-31370, Institute for Computational
Earth System Science, University of California, Santa Barbara.
WAN, Z., ZHANG, Y., ZHANG, Q. and LI, Z.L. (2002) Validation of
the land surface temperature products retrieved from Terra
moderate resolution imaging spectroradiometer data. Rem.
Sens. Environ., v.83, pp.163-180.
WAN, Z., ZHANG, Y., ZHANG, Q. and LI, Z.L. (2004) Quality
assessment and validation of the MODIS global land surface
temperature. Int. Jou. Rem. Sens., v.25(1), pp.261-274.
(Received: 17 November 2010; Revised form accepted:20 May 2012)

You might also like