|
Abstract
|
Article Information:
Forecast Models for the Yield of Millet and Sorghum in the Semi Arid Region of Northern Nigeria Using Dry Spell Parameters
B.A. Sawa and A.A. Ibrahim
Corresponding Author: B.A. Sawa
Submitted: 2011 February, 12
Accepted: 2011 March, 15
Published: 2011 May, 25 |
Abstract:
|
This study aims at presenting models for the prediction of millet and sorghum yields in the face of
occurrence of dry spells of 5, 7, 10 and equal to or greater than 15 consecutive days during the growing season.
Daily rainfall records and crop yield per hectare for three decades (1976-2005) were used to develop forecast
models for the yield of millet and sorghum in the semi arid region of Northern Nigeria on the basis of dry spell
parameters. Frequencies of dry spells of 5, 7, 10 and equal to or greater than 15 consecutive days were
determined. Bivariate correlation analysis, stepwise regression (forward selection) and double log multiple
regression were then used to develop models for the prediction of the yield of the two crops. Results of the
analyses show that of the twenty-one (21) dry spell parameters, only one (total dry spells in the growing season)
indicated significant correlation with millet at 0.01 levels. Except 7-day dry spells in May and 5-day dry spells
in September, sorghum show significant correlation with the remaining nineteen (19) dry spell parameters.
Stepwise multiple regression identified total dry spells during the growing season as being critical to millet
yield while 5-day dry spells in May and total dry spells in the growing season were picked for sorghum. The
model for millet yield is logY = 0.0715-0.0817 logX2, implying that if the occurrence of total dry spells during
the growing season is unity, the yield of millet would be reduced by about 0.0817 T/ha in the study area. That
for sorghum is logY = 0.170 + 0.0629 logx1 - 0.141 logx2, meaning that the occurrence of a single 5-day dry
spell in May would increase sorghum yield by about 0.0629 T/ha but that of total dry spells would reduce it
by 0.141 T/ha. About 54.5% variation in millet yield is accounted for by occurrences of total dry spells during
the growing season while 63.2% of the variation in sorghum yield is due jointly to the occurrence of 5-day dry
spells in May and total dry spells in the growing season. Both the forecast error and the linear graph of observed
and predicted yields suggest that the models are good enough for the forecast of millet and sorghum yield in
the area.
Key words: Coefficient of determination, double log function, dry spells, forecast model, , ,
|
|
Abstract
|
PDF
|
HTML |
|
Cite this Reference:
B.A. Sawa and A.A. Ibrahim, . Forecast Models for the Yield of Millet and Sorghum in the Semi Arid Region of Northern Nigeria Using Dry Spell Parameters. Asian Journal of Agricultural Sciences, (3): 187-191.
|
|
|
|
 |
ISSN (Online): 2041-3890
ISSN (Print): 2041-3882 |
 |
| Information |
|
|
|
|
| Sales & Services |
|
|
|
|
