Models in Irrigation andWater Management - Practices of Irrigation and On-farm Water Management - page 415

10.5.4 Discussion

Output of a model depends on the accuracy of the input data and the framework of
the model itself. The yield and weather data should be taken from reliable source
(or it should be primary source data). The model should be tested on independent
weather and crop data. If the model predicts yield with reasonable accuracy, it may
be used with confidence.
Better climatic data, particularly measurements of solar radiation, sunshine
hour, and a greater number of rainfall stations, and yield data for homogeneous
sub-areas or micro-zones might improve the yield prediction. The sub-area
could be constructed on the basis of soil-climate zones or using spatial anal-
ysis and interpolation tool available with GIS software. Forecasted yields from
sub-areas could then be aggregated to obtain the forecast for the whole coun-
try. In order for the concept of the sub-areas to succeed for improving the yield
prediction, the availability of yield data for such small arial units will also be
required.
The trend component of simulation may be updated from time to time to
incorporate the non-linear development in the future.
Since the crop yield depends on many types of variables viz. weather factors,
performance of plant during crop growth stages, agricultural inputs, it will be better
if all these aspects can be considered while forecasting crop yield. For this, compos-
ite index (weather, input, management; and weighting them) for forecasting may be
tried.
In case of large area (say, for a country), the same date of crop transplanting or
swing and consequently the same date of physiological maturity may not appro-
priate. This factor can be more realistically included if prediction can be done for
micro-zones.

10.5.5 Sample Example of Formulating Weather-Based
Yield-Prediction Model

For example, for a particular region (or crop district), the yield rate and weather
variables for the effective crop growing period are given as follows:

Year Yield rate (t/ha) Degree-day Rainfall (cm)

1980 0.8970 1,270 37.5

1981 0.9725 1,260 28.2

1982 0.8069 1,292 8

1983 0.9285 1,239 15.4

1984 0.8512 1,276 40.8

1985 0.8797 1,330 16.5

1986 1.0856 1,217 45.2

1987 1.0334 1,320 32