INTRODUCTION
Nigeria, for the past decades has been stressing greater emphasis on
irrigation development as a means of increasing food and raw material
production as well as achievement to rural development (Barau et al.,
1999). Irrigation is however a complex system which comprises of different
interacting factors such as water, environment, economic and human factors.
Among these, the human factor plays a most important role (Maskey and
Weber, 1996). It reflects the farmer`s managerial ability as indicated
by his reaction to the dynamic process of decision making and subsequent
implementation that eventually leads to achievements of a set of predetermined
irrigation oriented goals (Rao, 1993). A good knowledge of the impact
of the human factor on the performance of the irrigation systems is very
important for enacting irrigation policies. The understanding of the factors
influencing the farmers` satisfaction based on the human factor is therefore
very important. Unless farmers are satisfied with their irrigation systems,
there would be no incentives or initiatives producible. No irrigation
technology regardless of its ecological and economical soundness will
have any impact on productivity and income unless it is adopted by a significant
proportion of farmers. Consequently, the determination of factors that
influence farmers` satisfaction with irrigation systems is very important
in designing, executing and adjustment of irrigationrelated government
policies in the country. The Nigerian government regards irrigation as
a catalyst to advance farming technology. High priority has been placed
on using large scale irrigation projects known as River Basin Development
Authorities (RBDA) to achieve increased cropping intensities and outputs
particularly in the northern parts of the country. This study therefore
relates the management patterns of the RBDA projects and seeks to identify
factors influencing farmers satisfaction.
MATERIALS AND METHODS
The study was conducted in Kano River Irrigation Project Phase One (KRIP)
which is the major programme of the Hadeja Jama`are River Basin Development
Authority (HJRBDA) with headquarters in Kano, Kano state of Nigeria.
A double stage random sampling was employed in drawing the sample for
this study. Two irrigation communities were selected randomly from each
of the three Local Government Areas that embodies the KRIP and from each
community, 30 farmers were randomly selected. Therefore out of a population
of 4,296 KRIP irrigation farmers (Anonymous, 1989), 180 farmers were sampled
for the study. Structured questionnaire facilitated by oral interview
was administered to the farmers for the 2003/2004 cropping season. Information
collected includes the age, education level and offfarm income of the
farmer, quantity of fertilizer application, use of hybridized seed, water
availability on time and location of farm plots to head section of field
canals.
Kano River Project Phase One (KRIP1)
The KRIP covers an area of 22,000 ha (Anonymous, 1989) and is situated
along the KanoKaduna express way. It spreads over three Local Government
Areas: Bunkure, Garun Mallam and Kura Local Governments. It is administratively
divided into four zones. For efficient water allocation, KRIP is divided
into 29 sectors of varying sizes. A sector is a single unit, with independent
administrative water management and other operations independent of other
sectors.
Water is released to each sector from a main canal through lateral canals.
The canals discharge water to distributor canals and finally to the field
canals. Each field canal relays water to particular fields consisting
of farm plots ranging from 7 to 20 in number. For efficient administrative
management, KRIP is divided into four zones.
Model Specification
A satisfaction model was developed to examine the variables determining
farmers` satisfaction with their irrigation system. Since the dependent
variable (satisfaction) is dichotomous (the farmer stands to be satisfied
or dissatisfied) in nature, the Logit model was used in the analysis instead
of a normal linear regression. The application of the Logit model in the
analysis stands to be the most appropriate because
â€¢ 
The computation of the logistic distribution guarantees
the rate of the probabilities estimated to always lie between 0 and
1. 
â€¢ 
Unlike the Linear Programming Model (LPM), the probability does
not increase linearly with a unit change in the value of the explanatory
variables (Gujarati, 1988). 
â€¢ 
It is easier to compute and interpret than the Probit and Tobit
models (Pindyck and Rubinfeld, 1991). 
The logistic technique makes use of the maximum likelihood estimation
method to analyze the relationship between dichotomous reactions and explanatory
variables. In this model, the satisfaction level of the farmer is assumed
to be based on the objective of the utility maximization. The farmer weighs
up the marginal advantages and disadvantages of the irrigation system
and will therefore be satisfied if the marginal utilities of the irrigation
system outweigh the marginal disadvantages. Since the farmer can either
be in a state of satisfaction or dissatisfaction, let the status of his
satisfaction be represented by j. The underlying utility function for
the farmer can thus be represented as:
j = 0, 1, i = 1, 2, 3...n 
j 
= 
0 for dissatisfaction 
j 
= 
1 for satisfaction. The non observable utility function that ranks
the ith farmers` preference is given by U (H_{ji}, E_{ji},
I_{ji}) 
H 
= 
A vector of human factor as captured by farmer specific characteristics 
E 
= 
A vector of economic factors 
I 
= 
A vector of environmental factor as captured by irrigation system
specific attributes 
Since utilities are random, the farmer will be satisfied in his irrigation
system if the preference comparison is such that U^{a} > U^{b}
or if the non observable (latent) variable Y* = U^{a}  U^{b}
> 0; otherwise the farmer will be dissatisfied. Thus the probability
of satisfaction for the ith farmer can be given by
X_{i} 
= 
Matrix of explanatory variables 
Î² 
= 
Vector of parameters to be estimated 
Î¼_{i} 
= 
Random error term 
Î´(X_{i}Î²) 
= 
The cumulative distribution function for Î¼_{i} estimated
at X_{i}Î² 
The probability that a farmer will be satisfied is thus a function of
the explanatory variables and the unknown error term. If it is assumed
that the error term follows a Logistic distribution, then Î´(.) can
be estimated using a logistic distribution model. Following Eq.
2 and using a logistic distribution, the Logit model that will capture
the above underlying utility maximization is:
The odds ratio which defines the probability of satisfaction relative
to non satisfaction is given by:
In the empirical Logit model, it was assumed that H, E and I vectors
influence farmers` satisfaction. However, due to inadequate data on E
(economic variables such as taxes, subsidies and prices of inputs), it
was difficult to consider these in the analysis. While economic factors
are important in influencing decisions of farmers, such variables are
quite hard to capture in crosssectional surveys. This is because; it
is possible that farmers received free seeds or extension advice on the
technology and inputting the price for such inputs could be quite problematic.
To account for these, the level of disposable income as captured by the
financial status of the farmer was used as proxy for the underlying economic
variable and the irrigation specific attributes were used as proxy for
the environmental factors.
The dependent variable indexes if the farmer is satisfied or not (Table
1). The variable takes the value of 1 if the farmer is currently satisfied
with irrigation system and 0 otherwise. The explanatory variables are
as explained below:
AGE refers to the age of the farmer measured in years. It was hypothesized
that as the age of the farmer increases, the farmer becomes more and more
dissatisfied with their irrigation system. This is because older farmers
may have risk preferences different from those of young farmers.
Table 1: 
Definition of variables 

:EDU measures farmers` educational attainment. This was hypothesized
to have a positive effect on satisfaction. This is due to the ability
of the educated farmers to become aware of improved innovations and to
adopt them in their farming practices.
FERT measures the fertilizer availability on time. It is expected for
farmers who received fertilizer on time and in adequate measure to have
a positive inclination towards satisfaction. With the availability of
fertilizer on time and in adequate quantity, farmers would better combine
irrigation water and fertilizer for higher productivity.
HYV variable measures the high yielding variety of seeds. High yielding
seed variety will result in increase in output which will tilt the farmer
towards satisfaction.
OUTPUT refers to the total harvest of the farmer from his irrigated plot.
It was measured in kilograms. It could affect the farmers` satisfaction
with their irrigation system either positively or negatively. If the output
is high it is expected for satisfaction to be high and low output will
lead to farmers` dissatisfaction.
OFFI refers to the offfarm income of the farmer. It is expected that
farmers with offfarm income source might concentrate less on crop production
and thus be less satisfied with their irrigation system.
PSIZE variable measures the plot sizes. The size of a farmer`s plot may
influence his attitude towards the irrigation system since the impact
of irrigation on small and large plots differ. Farmers are expected to
be likely dissatisfied with their irrigation system the larger their plot
sizes.
WAT is a variable that measures the availability of water on time. The
availability of water on time is expected to make farmers inclined to
be satisfied.
FINSTA refers to the farmers` financial status. It was hypothesized that
the higher the disposable income level of the farmer, the more the farmer
will be able to purchase all necessary inputs on time and also be able
to meet all necessary requirements. This will lead to a good output,
thereby making the farmer to be more satisfied with the irrigation system.
LAC is a variable that measures the distance of the canal to farm plots.
It is expected that farmers with farm plots located near the head of canals
will get more water than those farther away. It was therefore hypothesized
that location of farm plots in the head canal section will be more satisfied
with irrigation system.
RESULTS AND DISCUSSION
The result of the initial estimation of the Logit model on Table
2 shows the estimates of the full Logistic regression model for the
KRIP irrigation system. The result shows that age, education, high yield
variety, offfarm income, financial status and distance of farm plots
to canal were found not to have any statistically significant influence
on he farmers` satisfaction with their irrigation system (Table
2).
Table 3 shows estimates of the reduced Logit model
which included only significant variables. Based on the tstatistic of
the reduced Logit model, the output of the farmers and water availability
on time were the most important variables in determining the farmers`
satisfaction.
The relative importance of the influence of the explanatory variables
is reflected in their coefficients (Î²) which show the magnitude of
change in the log of odds ratio for any change in the explanatory variables.
This however does not explain the change in the probability. In this study,
all the coefficients have positive signs except for plot sizes (PSIZE).
The negative sign of PSIZE implies that increment in the plot sizes of
farmers will cause the log of odds in favour of satisfaction to decrease.
In other words the probability of farmers being satisfied in terms of
the role of the system in increasing their farm income diminished with
the increase in plot sizes. A unit increase in PSIZE variable will lower
the probability of the farmers` satisfaction to 0.28 ceteris paribus.
The farmers were however found to be satisfied if fertilizer was available
on time and at adequate quantity, output is high, water is available on
time and their farm plots located near the head section of field canals.
The 2 Log Likelihood measures the goodness of model employed in the
study. It indicates the difference between the estimated Logistic model
and the perfect model. The values of 2 Log Likelihood for both full and
reduced models showed that there is a significant relationship between
the log of odds ratio, probability of satisfaction and the explanatory
variables included in the model. The Rsquare values and the overall percentage
of correct predictions also suggest that the estimated satisfaction model
had a good explanatory power.
The reduced Logit model (Table 3) shows that if fertilizer
is available on time and at adequate quantity, the farmer has a good output
that can generate a revenue that will improve his social status,the cultivated
area is larger than 0.4 ha and water is available on time, then the probability
of farmers being satisfied is estimated to be 0.728 (73%). Based on this
outcome, it could be predicted that the farmers were likely satisfied.
Nevertheless, if the plot sizes were small and other variables held constant,
then the probability of farmers being satisfied is estimated to higher
than 90%. If measures are put in place such that the farmers` output increases
by a unit, then the odds in favour of his satisfaction will be increased
by a factor close to 7 with the probability of his satisfaction with the
irrigation system increasing to 87.30%. If water is released on time and
at adequate quantity, the farmer`s satisfaction will also increase by
a factor of 6 and the odds of his satisfaction increase to 87.19%.
Table 2: 
Full logistic regression model of socioeconomic factors
of the farmers operating the KRIP irrigation system 

*: p>0.05 
Table 3: 
Reduced logistic regression model of socioeconomic
factors of the farmers operating the KRIP irrigation system 

*: p<0.05 
CONCLUSION
In this study, adequate fertilizer availability on time, output and adequate
releases of water were found to have a great impact on the farmers` satisfaction
with the KRIP irrigation system. Since fertilizer accessibility is largely
dependent on cash availability of the farmer in Nigeria, policy makers
must therefore pay special attention to agricultural credit system in
order to realize the full benefit of the irrigation technology.