Environmental And Agronomic Factors Influencing Variations In Smallholder Cotton Yields In The Kano Plain, Kisumu District, Western Kenya
Abstract
The primary objective of this study was to examine the
interrelationships between cotton yields, environmental and
agronomic factors in the Kano Plain. Since these factors are
highly variable in space and time, there is a spatio-temporal
variation in cotton yields as well. This study is an attempt
to test the hypotheses that: (1) spatial variation in smallholder
cotton yields is a function of environmental and agronomic
factors; and (2) the interrelationships between these
explanatory variables are significant. To verify these
hypotheses, simple regression and correlation analyses, and
factor and step-wise multiple regression analyses were used.
The latter techniques have been used in various geographical
studies and were found to be profitable in this particular
study undertaken in the Kano Plain. They provided sufficient
explanations for some of the major environmental and agronomic
factors influencing spatial variations in cotton yields. The
step-wise multiple regression analysis corroborated the results
of simple regression and correlation analyses, providing in
particular support for the hypothesis that agronomic practices
are the major determinants of the spatial variations in yields.
It is clear from the analyses that when correctly applied
by the smallholders, improved agronomic practices will
provide significant improvements in yields, provided that
the socio-economic constraints are alleviated. Theoretically,
the farmers are applying recommended agronomic practices,
but practically they have not been fully implemented at the
farm level largely because of inadequate education and
extension services; lack of improved marketing and ginning
facilities; diversion of insecticides to other uses; low
producer prices; and crop diversification, leading to
competition for available labour and capital resources.
Resulting data indicated that most soil properties,
agronomic and socio-economic factors were considerably varied
and significantly intercorrelated. The highest interrelationships
occurred between labour and field size (r=0.92),
carbon and nitrogen (r=0.85), clay and sand (r=-0.S4), and
labour and monetary return (r=0.7S). Such significant interactions
showed that in a smallholder cotton farming system,
a change in anyone part of the system or sub-systems directly
or indirectly affects yields. Results of the laboratory
analyses of 100 soil samples revealed that: (1) montmorillonite
soils are suitable for cotton production, confirmed by
soil reaction (pH) and nutrient levels; (2) carbon to nitrogen
ratio, cation exchange capacity, carbon and moisture
contents, are relatively low; (3) percentage base saturation
is generally high; (4) manganese is deficient in 3%, phosphorus
12%, copper 73%, iron 79%, zinc SO% and nitrogen 100%,
of the cotton fields; (5) low levels of manganese are associated
with pH, nitrogen, phosphorus, calcium and iron
contents; (6) soil texture is predominated by clay; (7) montmorillonite
soils tend to deflocculate with a consequent
decreasing permeability and increasing waterlogging and
anaerobic conditions Consequently, interactions between
soil properties suggested that fertilizer application to
the cotton crop requires proper management.
Statistical analyses of the data collected led to
the conclusion that cotton is the least profitable crop,
partly because of the low producer prices, relative to the
cost of cotton production and vis-a-vis other competing
enterprises: maize and sugarcane. Furthermore, very low
average smallholder cotton yields (about 350 kg/ha), in
contrast to those obtained from experimental plots (about
1500kg/ha), suggested that there is considerable scope to
quadruple cotton yields at the farm level. This potential
for increased cotton productivity cannot be realized unless
extension services and research related to cotton growing
are intensified and cotton marketing and development are
streamlined. The low cotton yields were significantly
associated with agronomic and manganese dimensions of the
cotton farming system.
The final step-wise multiple regression model
demonstrated that agronomic dimension, which was related
to weeding, planting density, spraying and picking regimes
and monetary returns, accounted for 51.60% of the spatial
variations in yields; whilst manganese dimension, which
was associated with manganese and cation exchange capacity
contents, planting density and farmer's age, contributed
to 8.53% of the variations in yields. Both factors explained
60.13% of the geographical variations in yields. The
interpretation of the residuals indicated that the step-wise
multiple regression model excluded some environmental factors,
such as bulk density, soil temperature and perception of
pests. The inclusion of these factors would improve
prediction of yields. Finally, an analysis of the farmers'
repertory-grids revealed that biological strategies for pest
control include proper disposal of old cotton stalks, and
presence of predators and parasites in the cotton fields.
Recommendations on policy implications for future
decision-making, planning and research on increasing cotton
yields in the study area have been made. Some of these are:
(1) soil analyses before the next cotton crop is sown should
be performed to indicate whether nutrients have remained
constant or dropped lower than desired to facilitate appropriate
fertilizer management; (2) the micro-nutrient deficiencies
should be solved by applications of sulphates of copper,
iron, manganese and zinc; whilst application of ammonium
sulphate should alleviate nitrogen deficiency; (3) there is
urgent need for correct applications of improved agronomic practices;
increased number of extension officers; participation
of younger people in cotton growing; distribution of seeds,
insecticides and storage facilities on time; (4) there is need
to provide credit and transportation facilities; increase
producer prices; and adhere to buying and paying calendar
schedules; and (5) research on soil nutrients' utilization
by the cotton crop; more sophisticated development of "highyielding
varieties" resistant to weeds, pests and diseases;
bionomics and biological control of pests, should be intensified.
The observations made above demonstrate the necessity
to approach cotton development on an inter-disciplinary basis.
This is the only way of maximizing the use of limited human,
technical and capital resources to increase cotton yields
Citation
Doctor of Philosophy, Agricultural Geography, University of Nairobi (1983)Publisher
University of Nairobi. Faculty of Science