IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring". Hyderabad, India,2002
temperature T,, was formulated by McFarland et al.
(1990) as
Ty, 7 Aj * Tgyz y. * A?* (Tgszy- Tp22y)
* 45* (Tgyzy. — Tgog) + As* Tggsy (5)
Where A;, A;, As, A4 are regression coefficients and
where V and H refer to the polarization of the channel.
The brightness temperature difference between 37 and
22V is measure of the atmospheric water vapour. The
polarization difference between the 37 and 19 GHz
brightness temperature is a function of water present in
the land surface scene. The screen air temperature was
retrieved by rearrangement of these above terms as linear
combination of four channels as follows
T, 7 Co * C; * Tgigu. tC? * Tay
* C; * Tgyzy, * C4 * Tgasy (6)
Where C, is the constant of regression and C,, C2, C3, C4
are the respective coefficients of brightness temperature
(in °Kelvin) of 19, 22, 37 and 85 GHz channels. The 22V
and 37V channel correct for the influence of atmospheric
water vapour. The average atmosphere attenuation, in
degree Kelvin equivalent, is incorporated into the
constant of regression (Co) and departure from the
average is included in the 85V coefficient. A regression
analysis was carried out between above SSM/I channels
and minimum screen air temperature of first week of June
over representative meteorological stations in India.
4. RESULTS
Maximum value composite of NDVI image and MPDI
images of 19, 37 and 85 GHz of June 1-10, is shown in
figure 1. It can be seen that the vegetated areas are seen
in brighter tone in the NDVI image (figure 1a) whereas
the non-vegetated soil areas, particularly with moist
condition (in Punjab) is seen bright in MPDI images
(figure 1, b, c, d). Vegetation shows high NDVI due to
large reflectance difference in Near Infra Red (NIR) and
Red bands and low MPDI from un-polarized microwave
emission. To the contrary, exposed soil with high soil
moisture emits very high polarization difference signal in
microwave region and is associated with low response of
NDVI in optical region. It can also be observed from
figure 1 that the higher frequency channels of 37 and 85
GHz have higher spatial information as compared to low
frequency 19 GHz image. A temporal profile of both
MPDI and NDVI of different land cover classes (figure
2) showed that, at lower NDVI in early stages, the MPDI
values were high with large variation due to soil moisture
variability in rice grown area. The sensitivity of the
MPDI at low NDVI varied for different crops as well as
over the region. The decrease in the MPDI values was
found with crop growth associated with increase in
NDVI. Overall, a nonlinear inverse relationship was
found between the MPDI and NDVI of different crops
over various regions (figure 2).
The surface wetness index was estimated
during kharif season over different parts of India. Two
distinct temporal behaviour of wetness as observed in
Fig. 1:Observations of (a) NDVI, (b) 19 GHZ, (c) 37
GHz MPDI and (d) 85 GHz Microwave Polarization
Difference Index (MPDI) of June, 1999
8
] Rice
7 1
A Soya, Cotton, Scane
6 + X N.Veg, Forest
m
©
4
z s
3 Sc ^ A
eX X 1
2 C ^ ; X
£84 ©, a
4 X X X A^ X X X X
1 EXX x x SM X XXX
0 T
0 0.2 0.4 0.6 0.8
NDVI
Fig 2: Relationship between the NDVI and MPDI of 37
GHz of different land cover classes viz. rice, soyabean,
cotton, sugarcane, natural vegetation and forest during
the Monsoon season of 1999 in India.
irrigated as well as unirrigated regions, represented by
Punjab and Rajasthan, respectively are shown in figure 3.
The irrigated region of Punjab (figure 3) with rice as
major crops showed very high wetness before the onset
of monsoon. The wetness in unirrigated region of
Rajasthan was low during June, which increased after the
onset of southwest monsoon. It can be seen from the
figure 3 that maximum wetness index in irrigated
condition was in range of 40 —50 while the same for
unirrigated region varied from 25- 30. Analysis was also
carried out to detect the flooding condition based on
wetness index. The difference in the wetness in first week
