IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India,2002 
    
  
  
  
   
    
  
  
   
   
area sown for sorghum crop in each year is taken from 
Dept. of Agriculture. The production forecast of 
sorghum is given by multiplying the district level 
acreage with the estimated yield. 
3. MATERIALS & METHOD 
SPI is determined by calculating a probability density 
function that describes the long term time series of 
precipitation. Cumulative probabilities are assigned to 
empirical amounts recorded for every district and month 
of the year by using the gamma distribution (X. Lana et 
al. 2001). After that, an equiprobabale transformation is 
used going from the gamma cumulative distribution to 
the standardized normal distribution, by means of the 
relationship 
fg (m B, Ddu 7 1/2 2)? | e7#* du 
Once the probability density function is determined, the 
cumulative probability of an observed precipitation 
amount is computed. The inverse normal (Gaussian) 
function, with mean zero and variance one, is then 
applied to the cumulative probability for computing SPI 
(Gauttman 1999). The scatter plot of SPI with respect to 
rainfall follows a skewed distribution as shown in fig.1. 
For a given time scale, SPI values are positive for 
greater than median precipitation, or negative for less 
than median precipitation. A value of zero corresponds 
to the median precipitation. The magnitude of the 
departure from zero is a probabilistic measure of the 
severity of a wet or dry event. The SPI is equally 
effective as a measure of both wetness and aridity. SPI 
value higher (lower) than 2.0 —2.0) can be considered to 
represent extreme wet (dry) events. Dry events. are 
defined as the sequence of months in which the SI is 
continuously negative with the condition that the SPI is 
less than or equal to —1 in at least one of the months in 
sequence. 
Since sorghum yield was available from 1980 to 1996 
only, therefore SPI from 1980 to 1996 was taken into 
account for estimating yield. of Raichur, Bijapur, 
Dharwad and Gulbarga districts. The precipitation at 
early and mid season showed positive response to yield. 
  
  
  
  
Rainfall 
  
  
  
  
« RF. june 4 RF july - RF aug x RF. sept 
Figl. Scatter plot between monthly rainfall and SPI for 
the period from June-Sept. for Bijapur district. 
  
  
The relationship developed by using monthly SPI with 
yield showed to have good fit with the reported yield. 
The predicted yield for the year 1997 to 2001 using SPI 
model for sorghum crop have shown good accuracy, 
even though it being an abnormal (drought) years. 
4. RESULTS & DISCUSSION 
The amount and pattern of rainfall are among the most 
important weather characteristics affecting agriculture. 
The districts are manly rain defecit areas, the daily 
rainfall varied from 0 to 20 mm. Rain, in addition to 
direct effect on soil moisture balance, they are strongly 
related to other meteorological variables like solar 
radiation, humidity, etc. the correlation of yield with 
monthly rainfall and monthly SPI have been almost 
similar. For eg. in Bijapur district, correlation of yield 
with monthly rainfall of August, September and October 
are 0.18, 0.43 and 0.56 respectively, whereas the 
correlation of yield with monthly SPI of August, 
September and. October are 0.19, 0.56 and 0.50 
respectively. Still SPI is considered better indicator of 
crop response where the distribution pattern of rainfall is 
considered by fitting it to a probability distribution (Oza 
et al. 2002). 
  
  
  
  
  
  
  
  
—3— REPORTED YIELD —4— ESTIMATED Y LD — -x- - PREDICETD Y IE-D 
  
  
  
Fig. 2. Yield forecast of Sorghum in Bijapur district of 
Karnataka state using SPI index method. 
The sorghum yield forecasts of 4 districts namely 
Raichur, Bijapur, Dharwad and Gulbarga using 
standardised precipitation index model are shown in 
fig.2 to 5. The significant SPI months were selected 
using stepwise regression method at 90% confidence 
interval. The models developed are: 
Bijapur Yield = -128261.4246 +76.095 x SPI 1_10 
+ 65.689 x SPI 1_7 + 65.106 x Year 
r’=0.96 n= 17 df 13 
Raichur Yield = 777.948 + 621.356 SPI 1_6 + 733.842 
x SPI 1_7 — 856.328 x SPI 2_7 + 
60.58 x SPI3_10 
r°=044 n=l7" di=12 
Gulberga Yield ^ 1007.139 - 247.8 x SPI1 10