season.
r were
ve rabi
ade by
) were
im and
15 at
incated
erature
st was
t starts
wheat
Iti-date
:b end.
ven in
| based
2]
96 are
dates
'evious
during
(Rajak
es and
time to
ulture,
sments
major
n parts
aryana
1 East
shift in
wheat
1. The
es.
Table 5: The highlights since 1997-98 are summarized in the
following table.
Year Highlights
1997-98 | Use of weather data to make multiple pre-
harvest production forecasts
Multi-year RS based comparison of wheat
growth demonstrated
Delayed sowing and later recovery of wheat in
Bihar picked up
1998-99 | Crop shift from mustard to wheat over large area
in Kota-Baran identified & mapped
1999-00 | Spatial & temporal changes in crop VI applied
to identify crop shifts
2000-01 | Drop in mustard acreage in MP & UP detected
in December ;
Reduction in cropped area in Western MP &
South Rajasthan picked up as early as Mid Jan.
2001-02 | Increased wheat area in Madhya Pradesh with
respect to 2000-01
6. CONCLUSIONS
The study clearly demonstrates that it is possible to give
regional-level multiple pre-harvest production forecast for
wheat, using multi-date IRS WiFS and meteorological data. The
pre-harvest forecasts for last four seasons (1997-98 to 2000-01)
deviated from post-harvest estimates by +0.9, +2.1, -5.4 and
+0.1 Mt. The extension of this methodology for other crops
require availability of RS data at finer spatial resolution and
more spectral bands which would be available from Advanced
WiFS (AWiFS) sensor onboard IRS P6. There is a plan to
extend such a study for other crops, mainly cotton, mustard and
sugarcane.
7. PLAN FOR FUTURE IMPROVEMENTS
Use of monitoring temperature through agromet indicator such
as Growing Degree Days (GDD) has been recently introduced.
Such studies bring out temperature anomalies, which could
influence yields. The deviations from normal for crop season
are calculated. Month wise deviations in GDD from its normal
explain behavior of weather and its likely effect on crop yield.
Realising the limitations of the regression based yield models in
abnormal season, there is a need to look for an alternate class of
models. Crop simulation model (CSM) is a useful tool, which
mimics the growth of crop to its surrounding atmosphere and
management inputs. CSM calculates dry matter production and
its partitioning on daily basis. In order to use CSM for yield
forecast, actual weather data upto February end, and normal
weather data thereafter, have been used. Use of this approach
has been already demonstrated at district level (Nain et al.,
2002). It was applied to met-subdivisions. Forecasts are in
agreement except for UP. As a further improvement,
development of Crop Growth Simulation Model (CGMS) with
RS and weather data over Haryana has been initiated (Sehgal et
al, 2002). Use of predicted weather data from ERMP
(Extended Range Monsoon Prediction) model for early crop
yield forecasting is also planned in future. :
Food and Agriculture Organisation (FAO), uses ten-day
composite SPOT- VEGETATION derived vegetation indices as
a part of global food security assessment. An analysis
methodology was developed for using multi date FAO SPOT-
VEGETATION derived ten day (14 class) normalized
IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India,2002
difference vegetation index for monitoring near real time crop
condition at meteorological sub division level. The inter-annual
comparison of Vegetation Index Number (VIN) temporal
profile indicates the condition of crop with respect to a previous
reference year. Such an assessment could provide RS-based
continuous monitoring during the crop growth season.
At present, pre-harvest forecasts from RS data are available
from two approaches, one uses multi-date IRS WiFS and
weather data and the other using single date IRS LISS III data
with weather and RS derived vegetation indices. A unified
approach, which combines strengths of these two approaches, is
planned to be developed.
The extension of this methodology for other crops require
availability of RS data at finer spatial resolution and more
spectral bands which would be available from Advanced WiFS
(AWIFS) sensor onboard IRS P6. There is a plan to extend such
a study for other crops, mainly cotton, mustard and sugarcane.
ACKNOWLEDGMENT
The work was sponsored by Department of Agriculture and Co-
operation (DAC), Ministry of Agriculture (MOA), Government
of India. The suggestions and feedback by Secretary DAC,
Econ. & Stat. Adv. - MOA and Dr Rajeev Mehta, Addl. Stat.
Adv. - MOA are thankfully acknowledged. The project team is
grateful to Dr. R R Navalgund, Director, NRSA and Shri J S
Parihar, Group Director, ARG and Mission Director RSAM, for
taking keen interest in the study and providing encouragement.
The contribution of CAPE team form State Remote Sensing
Centres by way of providing ground information and other
collateral information and scientists from SAC in providing
critical inputs through discussions and suggestions is gratefully
acknowledged. Dr. Manab Chakraborty, System Manager and
Shri H P Bhatt helped the team by way of providing computer
facilities.
REFERENCES
Appa Rao, G., 1983. Estimation of wheat yields over Punjab
using district and state models. Mausam, 34, 275-280.
Bhagia N., Oza M.P., Patel J.H., Dadhwal V.K., 1996. An
approach for all India wheat production forecasting using
remotely sensed data. Scientific Note. RSAM/SAC/CAPE-
II/SN/53/96 April 1996, Space Applications Centre Ahmedabad
380 015.
Bhagia N., Oza M.P., Rajak D.R., Singh R.P., Sehgal V.K.,
Ravi N., Srivastava H.S., Patel J.H., Ray S.S. and Dadhwal
V.K.,1997. An attempt to make national wheat production
forecast using multi-date WiFS data for 1996-97 season. Bull.
National Natural Resources Management System, NNRMS(B)-
21, 54-58.
Box G.E.P. and Jenkins J.R.,1976. Time series analysis :
forecasting and control. Holden - Day. San Francisco.
Dadhwal, V.K., 1989. Effect of temperature on wheat in India.
Climate and food security, IRRI, Philippines, 137-144.
De Roover, B., De Mudker, S. C., Goossens, R., 1993. The
regional inventory (MARS) in Belgium. Proc. Int. Symp.
