of October and first week of November clearly showed
the areas affected due to flood inundation caused by
super cyclone in Orissa in last week of October 1999.
IAPRS & SIS, Vol.34, Part 7, "Resource and Environmental Monitoring", Hyderabad, India,2002
temperatures estimated from present coefficients and
McFarland study is shown in figure 4.
Fig.3: Surface wetness as observed in two distinct
irrigated (Punjab) and unirrigated ( Rajasthan)
region of India showing the early high
wetness condition in punjab as compared to
high wetness after onset of monsoon in
Rajasthan.
Descending pass of DMSP- F13 satellite has early
morning equatorial crossing, which matches with daily
minimum temperature of the diurnal cycle. Screen air
temperature was estimated by developing empirical
relationship between the screen air temperature and the
brightness temperature of 19H, 22V, 37V and 85V of
SSM/I sensor. Comparison of McFarland et al. (1990)
and independently developed coefficients were carried
out. It was found that although the McFarland estimates
gave high correlation of 0.86 but most of the values were
under estimated. The root mean square error (RMSE) of
3.5 °C was found from IMD observations. The values of
new coefficients i.e. Co, Cj, C2, C4 and C, derived with
our dataset as described in the equation 6 were 91.39,
0.006, 0.547, -0.010 and 0.177, respectively. The
correlation coefficient of 0.82 was found between the
estimated and observed surface temperature with RMSE
of 1.4 °C. The relative comparison between screen air
25 - 30-4 ———ttline
; € McFarland et al. 1990
Rajasthan e = Oo
= O presentstudy -
20 + * 2 o
3 9 e c 25 o O
€ 45 - > S $
2 10 - e e +
© : o *? $6
= e e 34204 e
5 * * © + $ +
*. + o
* 2 =
0 "T e T T T T T FT 1 g
B R88 23 BRI LS I 8 15
Xr ee CONE ONE IONE EON ONE NET 0:0
15 20 25 30
Julian Day Temperature (°c) from IMD
ap Fig. 4: Estimation of screen air temperature using SSM/I
HT . Punjab data using McFarland Coefficient and
491€ developed coefficients from the present study
x
is
a 254 e Results indicate the useful of the SSM/I data
s nd ° for temperature estimation and point the need of
3 15 ? e development of regional coefficients. This was an
10 + e ve ° à exploratory analysis carried out over representative sites
s. 2.3. % e in first week of June without stratifying land surface into
. 2. vo different land cover class. A further analysis is needed
mE i um over large time interval over different land stratum such
i rM Oo x e rel M Oo + e A as cro 1 il 1 1
I No Ts oe ps, moist soils, dry soils etc for more precise and
Julian Day generalize model. The interpolated/blended temperature
from SSM/I and meteorological observations can be used
for interfacing crop simulation models with Geographical
Information System.
5. CONCLUSION
This study provides examples over India of rapidly
accumulating literature on application of passive
microwave radiometry for large area agriculture-
hydrology domain. Multi frequency and multi
polarization information of daily observations allowed
estimation of vegetation growth, surface wetness and
screen air temperature, which is unique in comparison to
information collected from optical data. As the spatial
resolution of current sensors like TMI, AMSR-E are finer
and will further improve with development of synthetic
aperture radiometers passive radiometry through its
complementary information will become an important
data source for agriculture monitoring. Use of low
frequency channels (« 10 GHz) is expected to provide
better retrieval of soil moisture without influence of
vegetation and atmosphere.
