IAPRS & SIS, Vol.34, Part 7, "Resource and Environmental Monitoring", Hyderabad, India,2002 
ALL INDIA WHEAT INVENTORY USING MULTI-DATE IRS WiFS 
AND WEATHER DATA 
M. P. Oza *, N. L. Bhagia, D. R. Rajak and V. K. Dadhwal” 
Crop Inventory and Modelling Division, Agricultural Resources Group, (RESA) 
Space Applications Centre, (ISRO), Ahmedabad — 380 015, India 
4 Commission VII, Working Group VII/6 
KEY WORDS: Crop acreage, Multi-date WiFS, Stratification, Pre-harvest Forecast, Wheat production, Meteorological yield 
models. 
ABSTRACT 
This paper describes the methodology and results of forecasting national level wheat acreage/production using multi-date medium 
resolution WiFS data since 1995-96 . Multi-date geometrically registered and radiometrically normalized IRS WiFS data was 
classified using hierarchical decision rules, which exploited differential crop spectral profiles of various crops in winter season. 
Wheat acreage estimates were arrived by aggregation of stratified sample segments. Wheat yields for meteorological sub-divisions 
were predicted by regression models using fortnightly temperatures. Multiple preharvest wheat acreage and production forecasts 
were made since 1997-98 crop season with additional information on crop growth performance in comparison to previous season. 
1. INTRODUCTION 
Use of remote sensing (RS) data in making crop production 
forecast has been actively investigated in India (Navalgund ef 
al., 1991) and many other parts of the world (MacDonald and 
Hall, 1980, Hanuschak et al., 1979, Ulaby, 1986). Monitoring 
Agriculture with Remote Sensing (MARS) program, developed 
at the Joint Research Centre (JRC) for the General Directorate 
for the Agriculture of the Commission of European 
Communities (EEC) and the Statistical Service of the European 
Communities, provides up-to-date agricultural information to 
European agricultural policy makers (Sharman, 1993; De 
Roover et al., 1993). Ortiz et al., (1997) have demonstrated the 
integrated use of RS, Geographical Information System (GIS) 
and historical database for improving cropland classification 
over a test sight of 15 x 15 km. 
In India, Crop Acreage and Production Estimation (CAPE) was 
the first large RS project for crop inventory. It provided the 
acreage and production of major crops at group of district level. 
The study on National Wheat Production Forecasting (NWPF) 
was initiated in 1995-96 for providing multiple forecasts at a 
national level with state level disaggregation. Crop production 
forecasting comprises crop identification, area estimation and 
predicting the yield of the crop. In this study multi-date data 
from Wide field Sensor (WiFS) onboard Indian Remote 
Sensing Satellites (IRS) 1C, 1D and P3 were used for wheat 
area estimation. The WiFS sensor on IRS 1C and ID has 2 
spectral bands (Red & NIR; additional SWIR band in P3) with 
spatial resolution of 188 m and ability to view a given region in 
5-day interval due to large swath of about 810 km. Such a 
configuration is ideal for regional level agricultural inventory. 
Regression based wheat yield models were developed at 
meteorological subdivisions between fortnightly temperatures 
and deviations of wheat yield from technology trend. 
  
" Corresponding author. E-mail address: markandoza@yahoo.com 
2. OBJECTIVES AND STUDY AREA 
The objectives of the study were: 
i) To provide national level multiple pre-harvest forecasts 
for wheat; 
ii) To make spatial analysis of crop growth differences 
across seasons . 
The time schedule for multiple forecasts was: 
(a) apriori production forecast to be made available by 
December end; 
(b) estimates for total rabi crop area to be made available by 
January end; and 
(c) early and final wheat production forecasts to be made 
available by February end and March end, respectively. 
In addition to these forecasts, the areas undergoing major 
change in crop intensity and / or crop type were also to be 
detected and mapped spatially and communicated to the user 
agency. 
The study was confined to six major wheat-producing states 
viz., UP, Punjab, Madhya Pradesh, Haryana, Rajasthan and 
Bihar. The study states contributed 91.78 percent to national 
wheat production (average for years 1995-96 to 1999-2000). 
For the remaining states, the contribution was estimated from 
historical trends. The study states and their contribution of area, 
production and yield to national wheat are given in Table 1. 
3. METHODOLOGY 
3.1 Sampling Approach 
For wheat area estimation, stratified random sampling with two- 
stage stratification was adopted. Within each state, 
meteorological subdivision formed the stratum and three 
substrata were formed within each stratum by allocating the 
segments to A, B, C type based on their rabi crop proportions.