Abstract

Water adequacy is central to maximised agricultural production in irrigation schemes. Smallholder Irrigation Schemes (SISs) are designed to distribute water efficiently, adequately and equitably. Water governance, defined as the institutions, processes, procedures, rules and regulations involved in water management, plays an important role in water allocation and subsequently water adequacy. The intersectoral institutions involved in water governance in SISs, i.e., government, Water User Associations (WUAs), Irrigation Management Committees (IMCs) and traditional authorities, interact to formulate and design policies for running SISs. However, multilevel interaction amongst the active stakeholders at multiple levels shapes policy and underlies SIS performance. This research aimed to investigate the impacts water governance had on adequacy of water in irrigation schemes and was premised on the hypothesis that governance had no effect on water adequacy. Water adequacy describes water supply relative to demand. Adequacy indicates whether the water delivery system supplies the required amount to a section in the irrigation scheme over a period of time (daily, monthly or seasonally). Two irrigation schemes, the Mooi-River Irrigation Scheme (MRIS) and Tugela Ferry Irrigation Scheme (TFIS) were used as case studies. A descriptive analysis showed that 86% of the farmers in the TFIS had adequate water, whereas 24% had water adequacy in the MRIS. A Binary Logit model was employed to investigate the factors that influence water adequacy among irrigators. The regression model identified eight statistically significant factors that influenced water adequacy: the irrigation scheme, location of plot within the scheme, training in water management, training in irrigation, SIS irrigators' knowledge about the government's aims, availability of water licences, payment of water fees and satisfaction with the irrigation schedule. The study concluded that governance factors had influence on water adequacy in the selected SISs. The implication is that stakeholders should make irrigators aware of government Irrigation Management Transfer (IMT) policy and strategies. The study recommends that the SISs introduce rules, procedures and protocols to support irrigators to enhance scheme governance and lead to the realisation of government policies.

Introduction

Agricultural water withdrawals account for 72% of global surface and groundwater, with abstraction being c. 90% in developing countries (Wisser et al., 2008). The erratic rainfall patterns bedeviling South Africa affect crop production, such that irrigation is the best alternative to augment the water deficit (Van Averbeke et al., 2011). To reduce dependence on rain-fed agriculture, developing countries have invested heavily in irrigation infrastructure as a means to improving water adequacy and subsequently mitigating rural poverty. Such investments coupled with improved crop production technologies and fertiliser application processes, plant breeding and the like has contributed to countries attaining food security (Gorantiwar & Smout, 2005). An irrigation system is a physical water delivery network comprised of subsystems that supply water to land by means of artificial canals and ditches. The system also has a socio-economic dimension that operates and derives benefits from the infrastructure. As such, it is set to attain farm level agricultural production goals such as maximising net economic and social welfare benefits (Molden & Gates, 1990).

The government of South Africa's efforts to invest in SIS infrastructure are aligned with the global Sustainable Development Goals (SDGs) of eliminating poverty, achieving zero hunger and reducing inequality. To achieve these goals it is paramount to identify and characterise the nature of the population at risk (Hope, 2006), which in this case is made up of vulnerable rural farmers occupying the previous poverty nodes of the apartheid era, hence the South African government's allocation of more resources in SIS infrastructure. However, despite massive investments, the performance of SISs has remained below expectations (Shah et al., 2002; Van Averbeke et al., 2011; Sinyolo et al., 2014; van Koppen et al., 2018). Indicators (both static and dynamic) such as water resource use, irrigation area, irrigation technology, agricultural productivity, poverty and food security have been used to assess the performance of SISs and are centered on water adequacy criterion (Svendsen et al., 2009).

Water adequacy

Water adequacy is the capacity of a supply source to meet demand (Frederick, 2013). According to Mehta et al. (2017), adequacy can be defined as the availability of water for food, energy, domestic supply and irrigation. Water adequacy is dependent on dynamic parameters such as water resource use and institutional frameworks, and on both technical/physical and socio-economic factors (institutional, gender, age and so on). Water adequacy entails delivery of the required amount of water at the right time and is a function of the condition of the hydraulic infrastructure and the governance structures in place (Figure 1). Irrigation scheme performance is judged by its adequacy characteristics; adequate supply and good governance will yield the desired irrigation scheme goals (Abernethy, 1990). The human dimension encompasses design of delivery schedules, operation and maintenance of the water control infrastructure (Molden & Gates, 1990; Horst, 1998). Water User Associations (WUAs) and irrigation management committees (IMCs) distribute and manage water resources to improve adequacy and equity (Kuşçu et al., 2008).

Fig. 1.

Physical and governance attributes that influence water adequacy (adopted from Taylor, 2002).

Fig. 1.

Physical and governance attributes that influence water adequacy (adopted from Taylor, 2002).

Water governance and its relationship with water adequacy

Water adequacy is a function of effective water policies determined by the intertwined socio-economic and physical/technical factors in SISs (Prathapar et al., 2002). Policies inform water governance strategies that are adopted in SISs and their design and implementation is a complex process that takes place in a dynamic social setting. The complexities arise from the varying degrees of representation of diverse interest groups that are clustered into organisations/institutions termed WUAs and IMCs. The diversity of SISs results in the formation of hybrid governance models where a combination of informal and formal institutions, processes, procedures, rules and regulations merge to create a framework that influences water adequacy (Colona & Jaffe, 2016). The creation of an ‘ideal state’ can only be met when water governance factors merge (Figure 2), i.e. national and local level institutions work together to promote an Irrigation Management Transfer (IMT) policy that ensures rule enforcement and promotes WUAs and IMCs to aid them in executing their legal mandates.

Fig. 2.

Water governance factors and their interaction within SISs.

Fig. 2.

Water governance factors and their interaction within SISs.

Formal and informal arrangements more often than not shape rule enforcement mechanisms and regulation arrangements, and dictate the policy paths that SISs take (Meagher et al., 2014). For instance, Prathapar et al. (2002) highlighted how scheme governance in south Asia adopted a two-way system of rule enforcement whereby rules in SISs could either be sanction- or compliance-based. Under such circumstances, a lack of command control policy has negative effects on water adequacy for the diverse irrigators. Owing to hybridised arrangements, water policies are not aligned to each other, as evidenced by perpetual spatial water conflicts (user upstream vs downstream users). Such governance-related failure of schemes result in the derailment and reshaping of agreed policies (Esteban et al., 2018), which inadvertently influences the availability of water and subsequently its perceived adequacy.

According to Movik et al. (2016), the 98.8% of South Africa's rural poor who inhabit SISs access only 5% of the water. This creates a welfare disparity, considering that the huge population depends on small-scale agricultural practices. As such, water reform policy and strategies were adopted in the form of integrated water resource management (IWRM), which sought to address the inequalities and ensure water adequacy at local level. Local level SIS politics require participation to promote alignment of goals between institutions, i.e., with WUAs and IMCs and the government on one hand, and the National Water Act (NWA), which promotes a water licensing system facilitating access to adequate water, on the other.

Various studies have been undertaken that investigate farmer perception of water security and water use efficiency (Gomo et al., 2014; Sharaunga & Mudhara, 2016; Alcon et al., 2017). The management of common pool resources (CPRs) such as adequate canal irrigation water and maintenance of the hydraulic infrastructure is key to improved crop production and poverty alleviation; however, SISs still remain poverty nodes despite the potential benefits of irrigation water and the conveyance system (Van Averbeke et al., 2011). This study investigated the infrastructural and socio-economic factors that influenced water adequacy for the irrigators in the Mooi-River Irrigation Scheme (MRIS) and Tugela Ferry Irrigation Scheme (TFIS). Xie (2006) argued that scheme governance has a bearing on availing adequate water for cropping requirements, i.e., methods employed in water distribution and water allocation influence water adequacy. The current literature looks at how socio-economic and institutionally related factors influence water security (Sharaunga & Mudhara, 2016) without necessarily looking at water adequacy and the underlying governance factors that influence it. This paper differs from Sharaunga & Mudhara (2016) as it focuses on water adequacy and the underlying water governance factors that influence it (see Figure 2).

Materials and methods

Study sites

The study was carried out in the Mooi-River Irrigation Scheme and Tugela Ferry Irrigation Scheme. The two irrigation schemes are located in the Msinga District of KwaZulu-Natal, South Africa (Figure 3). The irrigation schemes exhibit both similar and dissimilar governance-related characteristics, as shown in Table 1.

Table 1.

Governance-related characteristics of the MRIS and TFIS.

Irrigation scheme
 
Governance aspect MRIS TFIS 
Land allocation Tribal authority Tribal authority 
Water allocation and access Scheduled irrigation Subject to fee payment 
Conflict management Reported to the scheme committee or tribal authority Executive committee or tribal authority 
Election of committee Irrigators Irrigators 
Penalties for non-compliance Pay fines Not enforced 
Irrigation scheme
 
Governance aspect MRIS TFIS 
Land allocation Tribal authority Tribal authority 
Water allocation and access Scheduled irrigation Subject to fee payment 
Conflict management Reported to the scheme committee or tribal authority Executive committee or tribal authority 
Election of committee Irrigators Irrigators 
Penalties for non-compliance Pay fines Not enforced 
Fig. 3.

Study site locations within KwaZulu-Natal.

Fig. 3.

Study site locations within KwaZulu-Natal.

Mooi-River Irrigation Scheme (MRIS)

The MRIS is located along the floodplains of the Mooi-River in the Midlands region of KwaZulu-Natal province (Gomo, 2012). The scheme accommodates 842 farmers over 600 hectares of land. Land units are 0.1 ha in size but farmers tend to have more than one plot (Sharaunga & Mudhara, 2016). For ease of management and water distribution the scheme is divided into 15 blocks and each block has its own water management and allocation committee (Gomo, 2012). The upper section has blocks 1–5, the mid-section has blocks 6–11 and the tail section is comprised of irrigators in blocks 12–15 (Muchara et al., 2014). Water is diverted from a weir constructed across the Mooi-River into a parabolic canal which runs for 20.8 km from the diversion point to the end of the scheme (DAEA, 2001). The concrete-lined canal with a top width of 2.0 m and a depth of 1 m was designed to convey approximately 0.36 m3.s−1 (DAEA, 2001). The canal either feeds directly into the field or into night-storage/balancing dams. Distribution canals are used to convey water from the balancing dams to the fields as per the irrigation schedule. A diesel pump is used by tail-end user to abstract water from the Mooi-River to the canal. The diesel pump augments the irrigation canal water (Muchara et al., 2014).

The scheme (see Figure 4) is said to have been constructed by the South African National Army in the early 20th Century as a goodwill gesture to the locals. Initially the scheme had earthen-lined canals which were later upgrade to concrete lining (Gomo, 2012).

Fig. 4.

Mooi-River Irrigation Scheme layout.

Fig. 4.

Mooi-River Irrigation Scheme layout.

Tugela Ferry Irrigation Scheme (TFIS)

The TFIS is located on both banks of the Thukela River near the town of Tugela Ferry, in the Msinga Local Municipality. The approximate co-ordinates of the scheme entrance are: 28° 45″ 46.5′ S, 30° 26″ 33.2′ E (Mnkeni et al., 2010; UWPConsulting, 2012). With a land area of 800 hectares, an estimated area of 500 hectares is cultivated by approximately 1,500 producers (Cousins, 2012; Sharaunga & Mudhara, 2016).

Blocks 1–3 are on the right bank of the Thukela River (Figure 5). The scheme abstracts water from the Thukela River at a weir approximately 4 km upstream of Block 1. Water is transported under gravity to the scheme through a parabolic canal pipeline to Block 1. Block 1 is the first block to receive water from the main canal. The main canal at this point is 2.1 m across the top and has a capacity of 0.4 m3.s−1 (UWPConsulting, 2012).

Fig. 5.

Tugela Ferry Irrigation Scheme layout.

Fig. 5.

Tugela Ferry Irrigation Scheme layout.

Block 4 is on the right bank of the Thukela River downstream of Block 3. It is separated from Block 3 by a headland that protrudes towards the river. Due to failure issues of the connecting pipeline between Block 3 and Block 4, it is no longer connected to the main canal supply. The installed pump station provides water for irrigation; the water is pumped into a night storage dam where it is conveyed to the fields through existing concrete-lined canals. The canal section is an estimated 1.1 m across the top and the average carrying capacity is 0.1 m3.s−1, reducing to 0.014 m3.s−1 towards the tail end.

Block 5 lies on the left bank of the river and water supply is from the main canal via an inverted siphon across the Thukela River. Block 6 is not functional due to tribal politics and, as a result, there is no farming activity. Block 7 has been cut off from the main supply canal due to the failure of the Sampofu River aqueduct and the siphons through Tugela Ferry. It has two sub-divisions: Blocks 7A and 7B. Both blocks abstract from the Thukela River via an electricity powered pump that feeds a reservoir. The water is conveyed from reservoirs to farmlands via gravity (UWPConsulting, 2012).

Model specification

There is not much literature on the use of empirical models in trying to investigate governance issues that influence water adequacy. Speelman et al. (2008) applied a Tobit regression model from first step data generated by Data Envelopment Analysis (DEA) to investigate water use efficiency in smallholder irrigation schemes in North-West Province, South Africa. Various statistical tools allow data processing and the most common are generalised linear models and the Logit and Probit models (Cakmakyapan & Goktas, 2013). Both the Logit and Probit models can be used to investigate the influence of governance on adequacy because the dependent variable is a binary variable; careful consideration has to be made to select the best performing model of the two. In the literature, the Logit model has been used because it is stable for a large population size (Amemiya, 1981; Maddala, 1986; Cakmakyapan & Goktas, 2013; Sharaunga & Mudhara, 2016). According to Damisa et al. (2008) and Gomo et al. (2014), the advantages of using a Logit model are that they are easier to compute and interpret than Probit models, and eliminate heteroskedasticity, i.e., the probability does not increase linearly with a unit change in the value of the independent variable and the computation of the logistic distribution guarantees the rate of the probabilities to always lie between the interval 0 and 1.

The Binomial Logit (BNL) model was applied and run on STATA 15 to analyse the governance factors assumed to influence adequacy. Analysis was made using STATA statistical software. According to Pindyck & Rubinfeld (1981), the Binary Logit model can be defined as: 
formula
(1)
where:
  • is the probability of the farmer asserting water adequacy;

  • is the observed variable for adequacy (i.e., 1 = adequate and 0 = otherwise);

  • are factors determining water adequacy status for farmer i, with j parameters to be estimated.

For ease of presentation, Sharaunga & Mudhara (2016) substituted the variable with Z and thus Equation (1) becomes: 
formula
(2)
From Equation (1) the probability of a farmer citing water adequacy is given by which gives Equation (3) as stated by Sharaunga & Mudhara (2016): 
formula
(3)
The odds ratio for asserting water adequacy is given by: 
formula
(4)
Taking natural logarithms for Equation (4) gives: 
formula
(5)
Taking into consideration the disturbance term , the Binary Logit model (BNL) finally becomes: 
formula
(6)
where:
  • = the fixed component of the log odds

  • = explanatory (independent) variable(s)

  • = error term

Similar empirical models have been applied in various water cases. Gomo et al. (2014) used the Probit model, which is in the same family as the Logit generalised linear models, to investigate farmer satisfaction with SIS performance. Sharaunga & Mudhara (2016) used a BNL model run on STATA 15 to investigate factors that influence water use security in smallholder irrigation farming.

Data collection

Data was collected from the two irrigation schemes (the TFIS and MRIS). Table 2 shows the detailed stratification of the irrigation schemes.

Table 2.

Stratification of the TFIS and MRIS.

  Irrigation scheme
 
  MRIS
 
TFIS
 
Stratum Water availability Blocks SP* Blocks SP* 
Head blocks Always available 1–4 40 1–3 37 
Mid blocks Intermittently available 5–11 52 5 & 4A 52 
Tail-end blocks Limited availability 12–15 28 4B & 7 31 
  Irrigation scheme
 
  MRIS
 
TFIS
 
Stratum Water availability Blocks SP* Blocks SP* 
Head blocks Always available 1–4 40 1–3 37 
Mid blocks Intermittently available 5–11 52 5 & 4A 52 
Tail-end blocks Limited availability 12–15 28 4B & 7 31 

*SP = Sampled population.

Each scheme was divided into three sections, the blocks within each section exhibiting similar characteristics in terms of water availability and irrigation days. These sections were further divided into three strata (upper, middle and tail-end blocks) from which a sample was taken, for farmers located closest to the water source, mid-way and furthest away, respectively. The sample size was 240 for both schemes and was determined by a Raosoft® sample calculator at 95% confidence level (Incorporation, 2010).

A structured questionnaire was administered by five local isiZulu speaking enumerators to 240 randomly selected farmers from both schemes. Questionnaire experts established a face validity test; furthermore, a pilot test was carried out in each strata of the selected study sites.

Age, gender, irrigation training and adequacy were among the variables recorded by the questionnaire. The key dependent variable, adequacy, was an independent question which captured the farmers' perception of adequacy of water, in any dimension they perceived it. The variable had a binary response, i.e., 1 and 0. The description and coding for the variables assumed to influence adequacy of water are shown in Table 3. The model processed 15 selected independent variables.

Table 3.

Variables defining water adequacy status for the various farmers in the schemes and their respective coding as used in STATA 15.

Variable Code Description Hypothesis 
Location in block BLCK Location of plot in relation to main supply line
Dummy; 1 if located at tail and 0 otherwise 
Farmers closer to the main supply line have adequate water 
Irrigation scheme (MRIS vs TFIS) IRRG_SCHEME Location of scheme with in the district
Area dummy: 1 if household is from Mooi-River and 0 Tugela Ferry 
Revitalisation improves water adequacy 
Water management training WAMTR Determination of farmers who have received water management training
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Irrigation training IRRTR Determination of farmers who have received irrigation training
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Agricultural training AGRTR Determination of farmers who have received training in agriculture
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Knowledge of government aims in SISs KGOVAIMS Measures farmers awareness of governments agenda for the SIS
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Awareness of the government's agenda for SISs improves water adequacy 
Election process fair FAIRELECTPRCSS Allows determination of farmers in election participation
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Representation of all interest groups improves water adequacy 
Available water licence AVWALIC Determination of farmers who have water licences
Dummy; 1 if farmer cites yes and 0 otherwise 
Farmers with water permits have water adequacy 
Available water rights AVWATRGHT Measure farmer perception on water rights
Dummy; 1 if farmer cites yes and 0 otherwise 
Framers citing availability of water rights have adequate water 
Water conflicts between farmers WACONFLFRM Allows for the determination of conflicts at farm/village level
Dummy; 1 yes and 0 otherwise 
Less conflicts improves water adequacy 
Water conflicts between blocks WATCONFLBLK Allows for the determination of conflicts block level
Dummy; 1 yes and 0 otherwise 
Less conflicts improves water adequacy 
Water payments PAYMNTWAT Determination of farmers that pay for water
Dummy; 1 if farmer cites yes and 0 otherwise 
Farmers that pay water fees have access to water hence are water adequate 
Satisfied with irrigation schedule SATIRRSCHDL Measures timeliness in water delivery
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Farmers receiving water on time and irrigating for a long time are likely to cite water adequacy 
Attending water-related training ATTWTRLDTRA Determination of farmers who participate actively in water-related training exercises
Likert scale; 1 = sometimes, 2 = always, 0 = otherwise 
Farmers who actively participate in such exercises are more likely to report water adequacy 
Attending irrigation meetings ATTIRRMTNGS Determination of farmers who participate actively in irrigation meetings
Likert scale; 1 = sometimes, 2 = always, 0 = otherwise 
Farmers who actively participate in such meetings are more likely to report water adequacy 
Variable Code Description Hypothesis 
Location in block BLCK Location of plot in relation to main supply line
Dummy; 1 if located at tail and 0 otherwise 
Farmers closer to the main supply line have adequate water 
Irrigation scheme (MRIS vs TFIS) IRRG_SCHEME Location of scheme with in the district
Area dummy: 1 if household is from Mooi-River and 0 Tugela Ferry 
Revitalisation improves water adequacy 
Water management training WAMTR Determination of farmers who have received water management training
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Irrigation training IRRTR Determination of farmers who have received irrigation training
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Agricultural training AGRTR Determination of farmers who have received training in agriculture
Dummy; 1 if farmer had training and 0 otherwise 
Training improves water adequacy 
Knowledge of government aims in SISs KGOVAIMS Measures farmers awareness of governments agenda for the SIS
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Awareness of the government's agenda for SISs improves water adequacy 
Election process fair FAIRELECTPRCSS Allows determination of farmers in election participation
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Representation of all interest groups improves water adequacy 
Available water licence AVWALIC Determination of farmers who have water licences
Dummy; 1 if farmer cites yes and 0 otherwise 
Farmers with water permits have water adequacy 
Available water rights AVWATRGHT Measure farmer perception on water rights
Dummy; 1 if farmer cites yes and 0 otherwise 
Framers citing availability of water rights have adequate water 
Water conflicts between farmers WACONFLFRM Allows for the determination of conflicts at farm/village level
Dummy; 1 yes and 0 otherwise 
Less conflicts improves water adequacy 
Water conflicts between blocks WATCONFLBLK Allows for the determination of conflicts block level
Dummy; 1 yes and 0 otherwise 
Less conflicts improves water adequacy 
Water payments PAYMNTWAT Determination of farmers that pay for water
Dummy; 1 if farmer cites yes and 0 otherwise 
Farmers that pay water fees have access to water hence are water adequate 
Satisfied with irrigation schedule SATIRRSCHDL Measures timeliness in water delivery
Likert scale; 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree 
Farmers receiving water on time and irrigating for a long time are likely to cite water adequacy 
Attending water-related training ATTWTRLDTRA Determination of farmers who participate actively in water-related training exercises
Likert scale; 1 = sometimes, 2 = always, 0 = otherwise 
Farmers who actively participate in such exercises are more likely to report water adequacy 
Attending irrigation meetings ATTIRRMTNGS Determination of farmers who participate actively in irrigation meetings
Likert scale; 1 = sometimes, 2 = always, 0 = otherwise 
Farmers who actively participate in such meetings are more likely to report water adequacy 

Variable selection

For this study 15 governance-related factors were identified and run in the regression model. Location in the block (BLCK) is a factor determining water access. The literature indicates that farmers located at upper reaches of the scheme have unlimited access to water which subsequently impacts on adequacy (Bruns, 2007; Sharaunga & Mudhara, 2016). Another governance-related factor, Irrigation scheme (IRRG_SCHEME), was identified as a contributor to water adequacy. The methods used to abstract, convey and apply water determines how successful scheme goals can be met. The literature suggests that farmers using advanced methods (e.g. electric and diesel pumps) for water abstraction, conveyance and application were more likely to have adequate water supplied to their plots (Sinyolo et al. (2014).

Water management training (WAMTR), irrigation training (IRRTR) and agricultural training (AGRTR) were also identified as factors that influence water adequacy in irrigation schemes. Water adequacy perception is centred on the farmers ability to utilise acquired skills. van Koppen et al. (2018) stated that investing in soft skills (i.e., farmer training) yields improved production stemming from improved water use efficiency and subsequently improved water adequacy. Allouche (2016) stated that farmers with knowledge of government's aims for SISs (KGOVAIMS) were most likely to have adequate water. Government designs policy, which is diffused and translated by the decentralised institutions (WUAs and IMCs) to the farmers. Farmers with a clear understanding of the policies and strategies are more inclined to effectively manage the water resource.

Fair election process (FAIRELECTPRCSS), attending water-related training and irrigation meetings (ATTWTRLDTRA and ATTIRRMTNGS), water conflicts amongst blocks and farmers (WATCONFLBLK and WACONFLFRM) were each identified as water governance-related factors that could potentially influence water adequacy. A fair election process is a participatory approach that ensures all interest groups are properly represented, averting conflict. Studies have revealed that a participatory approach harmonises irrigation scheme dynamics and hence water equity, delivery and adequacy are met (Groenfeldt & Svendsen, 2000; Yildirim & Çakmak, 2004; Uysal & Atış, 2010). Water licensing (AVWALIC) and water fees (PAYMNTWAT) were also identified as water governance-related factors that potentially influenced water adequacy. The literature reveals that for a farmer to have ‘wet water’ they have to make the necessary payment to acquire permits (‘paper water’) (Denby et al., 2017). The variable available water rights (AVWATRGHT) were considered to be a potentially water adequacy influencing governance factor. Studies have shown that water rights contribute immensely to improved water access and subsequently to crop production. Water rights allow farmers to protect their water from being stealthily taken away by other farmers which subsequently impacts on water adequacy (Bruns, 2007). Lastly, satisfaction with irrigation schedule (SATIRRSCHDL) was also deemed as influencing water adequacy.

Results and discussion

Descriptive statistics for the MRIS and TFIS

The variable IRRWTADQCY was a dummy variable, with 1 indicating water adequacy and 0 otherwise (Table 4). The table shows significant differences in water adequacy for some variables across farmers from the TFIS and MRIS. It reveals that 87% of the TFIS farmers are water adequate compared to 23% in the MRIS. A high proportion of farmers did not receive training in agriculture, irrigation and water management, which is contrary to expectations. Table 4 reveals that 54% of farmers who do not pay for water (PAYMNTWAT) are not water adequate. Some 94% of farmers who were satisfied with the irrigation schedule (SATIRRSCHDL) were water adequate whilst all those who expressed dissatisfaction did not have adequate water. Farmers who did not experience water conflict were more likely to be water adequate compared to their counterparts. Farmers' participation in irrigation and water management training increased the chances of experiencing water adequacy. This was evident as 64% who participated in water-related training indicated water adequacy. Farmers who paid water fees indicate water adequacy: this is evident as 79% who paid water fees expressed water adequacy. By contrast, 54% of those who did not pay water fees indicated water inadequacy. Farmer location across the canal reach was a significant factor: 65% of farmers at the head end indicated water adequacy, whereas 24% of those at the tail end indicated water adequacy. Finally, 58% of farmers who agreed to a fair election process within the scheme indicated water adequacy whilst 54% of water adequate farmers neither agreed nor disagreed.

Table 4.

Behaviour of independent variables with respect to water adequacy.

  Water adequacy
 
 
Variable name All sample (n = 239) No (%) Yes (%) -value 
BLCK Tail (76 24 0.000*** 
Head (35 65 
IRRG_SCHEME TFIS (13 87 0.000*** 
MRIS (77% 23% 
WAMTR Yes (43 57 0.689 
No (46 54 
IRRTR Yes (36 64 0.104 
No (48 52 
AGRTR Yes (43 57 0.496 
No (47 53 
KGOVAIMS Agree (36 64 0.394 
Neutral (34 66 
Disagree (47 53 
FAIRELECTPRCSS Agree (42 58 0.091* 
Neutral (46 54 
Disagree (80 20 
AVWALIC Agree (50 50 0.985 
Neutral (46 54 
Disagree (47 53 
AVWATRGHT Yes (13 87 0.954 
Not sure (47 53 
No (12 88 
WACONFLFRM Yes (70 30 0.000*** 
No (36 64 
WATCONFLBLK Yes (91 0.000*** 
No (35 65 
PAYMNTWAT Yes (21 79 0.000*** 
No (54 46 
SATIRRSCHDL Agree (94 0.000*** 
Neutral (68 32 
Disagree (100 
ATTWTRLDTRA Always (39 61 0.091* 
Sometimes (53 47 
Never (52 48 
ATTIRRMTNGS Always (42 58 0.300 
Sometimes (53 47 
Never (50 50 
  Water adequacy
 
 
Variable name All sample (n = 239) No (%) Yes (%) -value 
BLCK Tail (76 24 0.000*** 
Head (35 65 
IRRG_SCHEME TFIS (13 87 0.000*** 
MRIS (77% 23% 
WAMTR Yes (43 57 0.689 
No (46 54 
IRRTR Yes (36 64 0.104 
No (48 52 
AGRTR Yes (43 57 0.496 
No (47 53 
KGOVAIMS Agree (36 64 0.394 
Neutral (34 66 
Disagree (47 53 
FAIRELECTPRCSS Agree (42 58 0.091* 
Neutral (46 54 
Disagree (80 20 
AVWALIC Agree (50 50 0.985 
Neutral (46 54 
Disagree (47 53 
AVWATRGHT Yes (13 87 0.954 
Not sure (47 53 
No (12 88 
WACONFLFRM Yes (70 30 0.000*** 
No (36 64 
WATCONFLBLK Yes (91 0.000*** 
No (35 65 
PAYMNTWAT Yes (21 79 0.000*** 
No (54 46 
SATIRRSCHDL Agree (94 0.000*** 
Neutral (68 32 
Disagree (100 
ATTWTRLDTRA Always (39 61 0.091* 
Sometimes (53 47 
Never (52 48 
ATTIRRMTNGS Always (42 58 0.300 
Sometimes (53 47 
Never (50 50 

Source: Survey data (2017).

***, **, *indicate statistically significant at 1%, 5% and 10% levels, respectively.

Binary logit analysis

This section presents the results of the model. The model had two levels of responses based on water adequacy (IRRWTADQCY). A frequency table was generated which indicated that 28 irrigators in the MRIS had adequate water and 92 had inadequate water; in the TFIS, 104 irrigators had adequate water and 16 had inadequate water (Table 5).

Table 5.

Frequency of water adequacy responses in the TFIS and MRIS.

Response TFIS
 
MRIS
 
Frequency Percentage Frequency Percentage 
Yes 104 87 28 23 
No 16 13 92 77 
Total 120 100 120 100 
Response TFIS
 
MRIS
 
Frequency Percentage Frequency Percentage 
Yes 104 87 28 23 
No 16 13 92 77 
Total 120 100 120 100 

Regression model diagnostics

The model results are presented in Table 6. Multicollinearity diagnostics showed no significant correlation between the explanatory variables. The variance inflation factors (VIF) averaged 1.698 which is within the acceptable range of below 10 (Rogerson, 2001; Hounsome et al., 2006). The standard error estimates were low (i.e., below 10) which indicated that micro-numericity was not a problem. The model revealed an overall classification accuracy of 89.12%, which illustrates that it successfully predicted irrigation water adequacy. The Chi-square test was significant ( indicating a good predictive capacity of the model (Table 6).

Table 6.

Estimates of the Binary Logit model for the TFIS and MRIS.

IRRWTADQCY Coefficient Standard Error  Marginal effects  
BLCK −2.383 0.656 0.001 −0.176 
IRRG_SCHEME −4.581*** 1.141 0.000 −0.338 
WAMTR 2.190** 0.874 0.012 0.162 
IRRTR −1.440* 0.843 0.088 −0.106 
AGRTR −0.634 0.634 0.317 −0.047 
KGOVAIMS 0.517** 0.232 0.026 0.038 
FAIRELECTPRCSS 0.204 0.269 0.447 0.015 
AVWALIC 0.799*** 0.252 0.002 0.059 
AVWATRGHT 0.309 0.532 0.561 0.023 
WATCONFLBLK −1.055 1.025 0.303 −0.052 
WATCONFLFRM −0.709 0.774 0.360 −0.078 
PAYMNTWAT −1.728** 1.028 0.093 −0.128 
SATIRRSCHDL 1.601*** 0.297 0.000 0.118 
ATTWTRLDTRA −0.296 0.404 0.465 −0.022 
ATTIRRMTNGS 0.795 0.551 0.149 0.059 
_cons −6.658 2.325 0.004 − 
Number of observations 239    
LR Chi-square (Prob > Chi-square)     
Pseudo  0.6518    
correct prediction     
IRRWTADQCY Coefficient Standard Error  Marginal effects  
BLCK −2.383 0.656 0.001 −0.176 
IRRG_SCHEME −4.581*** 1.141 0.000 −0.338 
WAMTR 2.190** 0.874 0.012 0.162 
IRRTR −1.440* 0.843 0.088 −0.106 
AGRTR −0.634 0.634 0.317 −0.047 
KGOVAIMS 0.517** 0.232 0.026 0.038 
FAIRELECTPRCSS 0.204 0.269 0.447 0.015 
AVWALIC 0.799*** 0.252 0.002 0.059 
AVWATRGHT 0.309 0.532 0.561 0.023 
WATCONFLBLK −1.055 1.025 0.303 −0.052 
WATCONFLFRM −0.709 0.774 0.360 −0.078 
PAYMNTWAT −1.728** 1.028 0.093 −0.128 
SATIRRSCHDL 1.601*** 0.297 0.000 0.118 
ATTWTRLDTRA −0.296 0.404 0.465 −0.022 
ATTIRRMTNGS 0.795 0.551 0.149 0.059 
_cons −6.658 2.325 0.004 − 
Number of observations 239    
LR Chi-square (Prob > Chi-square)     
Pseudo  0.6518    
correct prediction     

Source: Survey data (2017).

***, **, *indicate statistically significant at 1%, 5% and 10%, respectively.

The results show that irrigation scheme infrastructure characteristics (IRRG_SCHEME) have a bearing on water adequacy (). In addition, other statistically significant variables are location of plot in scheme (BLCK) (), water management training (WAMTR) (), irrigation training (IRRTR) (), knowledge of government aims in the SISs (KGOVAIMS) (), availability of water licence (AVWALIC) (), paying water fees (PAYMNTWAT) (), and satisfaction with irrigation schedule (SATIRRSCHDL) ().

Irrigation scheme and farmer location in scheme

The variable for location of farmer within the district (IRR_SCHEME) was negative and statistically significantly influenced water adequacy (). This revealed that being located in the MRIS is associated with an increase in probability of citing water inadequacy by 0.3. The farmers in the TFIS are more likely to have adequate water than those in the MRIS. The scheme infrastructure characteristics contribute to the positive attributes of the TFIS. The TFIS adopted an underground-piped water conveyance system, whereas the MRIS has the canal system that is old and needs repair. The upgraded technology in the TFIS has minimal human interaction and interference requirements (i.e., operability of the water control infrastructure), whilst Mooi River farmers use a manually operated canal and gated systems that can be tampered with. Horst (1998) highlighted that a system that is user friendly and requires minimal human interaction during operation tends to serve the purpose better and offer farmers/irrigators a correct irrigation function. Gomo (2012) revealed that the average conveyance efficiency for the canal systems in the MRIS was 40% instead of the designed 85% for the concrete-lined canals. This was attributed to leakages along the canals where maintenance was overdue. Illegal water abstractions (Figure 6), amplified by leaking canals and silt accumulation (Figure 7), are amongst the major factors that impact on water conveyance and subsequently influence water adequacy negatively in the MRIS.

Fig. 6.

A comparison of the rate of illegal water abstraction between the MRIS and TFIS.

Fig. 6.

A comparison of the rate of illegal water abstraction between the MRIS and TFIS.

Fig. 7.

(a) Silted canal and (b) cracks and failed concrete lining.

Fig. 7.

(a) Silted canal and (b) cracks and failed concrete lining.

The variable which signifies the farmers' location within the scheme (BLCK) had a statistically significant bearing on water adequacy (). Farmers located at the tail end were approximately 18% more likely to indicate water inadequacy than those at the head. Sharaunga & Mudhara (2016) established that farmers at the tail end in the MRIS were water deficient.

According to Letsoalo & Van Averbeke (2006), poor infrastructural maintenance leads to weeds growing in canals so that when water carries silt it enhances sedimentation, subsequently reducing the cross-sectional area and thus the canal's carrying capacity. In addition, sediment accumulation increases Manning's roughness coefficient which consequently reduces flow velocities and hence causes a reduction in canal carrying capacity.

Water management training

The variable for ‘training in water management’ (WAMTR) was statistically significant () suggesting that it influenced water adequacy. Training in water management is associated with an increase in the probability of indicating water adequacy by 16%. Water management is defined as determining and controlling the rate, amount and timing of irrigation water in a planned and efficient manner. Farmers with knowledge of Crop Water Requirement (CWR) and the irrigation schedule were most likely to have adequate water for cropping requirements. Pereira et al. (2002) also argued that for water supplied on a demand-scheduled basis, farmers with knowledge of irrigation scheduling and CWR (which stems from training) were more likely to be water efficient, which subsequently means they had adequate water to meet their cropping needs. Langyintuo & Mekuria (2005) revealed that educating farmers in adopting technologies increased their adaptive capacities, which improves natural resource usage. In addition, Balasubramanya et al. (2017) argued that training farmers facilitated proper agronomic planning that could ensure improved access to water and subsequently water adequacy in SISs. Extension workers (agronomists offering advice to farmers) indicated that water inadequacy can be attributed to lack of knowledge as farmers deliberately absconded training.

Irrigation training

The negative and statistically significant variable ‘irrigation training’ (IRRTR showed that its lack increased the probability of indicating water adequacy by 11%. This finding was not expected. The anticipated outcome was for exposure to irrigation training to increase the probability of a farmer citing water adequacy. The finding resonates with the notion that farmers are comfortable with their inherited irrigation knowledge (Mehta et al., 2017). Furthermore, Mehta et al. (2017) pointed out the complexity of the African matrix of water resource management, as being one that has a web of customary beliefs.

Knowledge of government aims

The variable for ‘farmers had knowledge of the government's aims’ (KGOVAIMS) was positively statistically significant. Farmers who had knowledge of the government's aims in SIS were likely to experience water adequacy, i.e., a knowledgeable farmer was more likely to increase the probability of water adequacy by 4%. Government overseers of the irrigation schemes formulate policy and strategies that trickle down from basin scale to village level irrigation schemes; thus, proper policy articulation and implementation of strategies improve the handling of a water resource, which subsequently increases water adequacy. Decentralisation, which saw the creation of WUAs and IMCs, has been the driver for improved policy and information dissemination. Local institutions and organisations have promoted participation. This finding concurs with Denby et al. (2017) who argued that the 1998 National Water Act facilitated decentralising participatory institutions that articulate and reconcile the divergent goals between government and farmers. In addition, Alba & Bolding (2016) also argued that the creation of local level institutions that articulate a policy and implement government strategy allow policy to move through various channels and, in the process, it gains more transformative perspectives and thus generates new ideas.

Availability of water licence

The positive and statistically significant parameter estimate for possessing a water licence (AVWALIC) means that farmers who possessed water licences were more likely to experience water adequacy. The positive impact for possessing a water licence increased the probability of citing water adequacy by 59%. Water permits allow equitable water appropriation. Denby et al. (2017) argued that in order to obtain ‘wet water’ there has to be ‘paper water’ first, hence institutions that carry out their legal mandates promote compliance and subsequently ensure water adequacy for the irrigators. In addition, Alba et al. (2016) argued that water licensing presented a robust accountability system that improved water allocation amongst users, with the underlying concept being that proper volumes are specifically allocated to a user, thus preventing conflicts.

Water fees

The coefficient estimate for concurring to paying water fees (PAYMNTWAT) was negative and statistically significant. The finding did not concur with the anticipated outcome. It was expected that farmers who pay water fees would be more likely to indicate water adequacy since water payments secure water abstraction rights. Access to water can be blocked when no payments have been made. In addition, water payments facilitate infrastructure maintenance, which subsequently improves water conveyance, and water adequacy. The negative coefficient could be indicating that those paying for water would have expected better water adequacy than they are getting.

Satisfaction with the irrigation schedule

Irrigation scheduling is primarily related to technological and operational factors (Horst, 1998). The variable called ‘satisfaction with irrigation scheduling’ (SATIRRSCHDL), was statistically significant () and positively influenced water adequacy. According to Table 6, farmers who were content with the irrigation schedule were more likely to have adequate water for their operations. A farmer satisfied with the irrigation schedule increased the probability of indicating water adequacy by 0.12. This phenomenon may be attributable to effective information dissemination on the delivery schedule to the farmers. A satisfactory irrigation schedule is one that is flexible in supply, timing and rate. A flexible irrigation schedule facilitates easy agronomic planning and decision making. Palmer et al. (1989) cites that a flexible irrigation schedule promotes satisfaction amongst irrigators thus giving them more freedom in economic and agronomic decision making. Balasubramanya et al. (2017) argued that when farmers participated in training that incorporates long-term planning and irrigation scheduling, the chances of having adequate water are increased.

Conclusion and recommendations

This study assessed the factors that influence water adequacy in smallholder irrigation schemes in KwaZulu-Natal, South Africa. Eight factors were statistically significant. In addition, poor infrastructure maintenance procedures have had a negative impact on the carrying capacity of the concrete lined canals. Results showed that investing in water management training improved access to water. Results also revealed that conflicts impacted negatively on water adequacy for crop production. A flexible irrigation schedule, i.e., one designed to operate in a timely and accurate manner, empowers farmers to make informed agronomic and economic decisions.

Evidence established that strong water governance mechanisms, i.e., those that provide farmer training and those that are participatory, improve water adequacy. Furthermore, proper policy interventions empower WUAs and IMCs to carry out their mandate such as collecting water fees and promoting participation in training, which subsequently facilitates a participatory design of irrigation scheduling, positively influencing water adequacy and promoting routine canal maintenance.

The study also assessed the implication of technology upgrade on water adequacy. the TFIS farmers were more likely to be water adequate compared to the MRIS farmers, partly due to the scheme upgrade undertaken in the TFIS. The TIFS moved from a wholly canal system to piped networks across the blocks. Through this, it was established that the human dimension, as effected by operability and canal structure handling, has a bearing on water adequacy. Sensitive points are less likely to be identified in a canal system.

Farmers who received water management training were highly likely to have adequate water to meet their cropping needs. Illegal abstractions taxed compliant farmers. The rule breakers seem to reap the benefits of accessing water at any given time at the expense of the other farmers. Water conflicts between blocks were also another factor that negatively influenced water adequacy. Policy incoherence and lack of coordination amongst blocks were the primary cause for water conflicts. In addition, institutions showed poor capacity in rule enforcement and reining in delinquent water users in the block.

It is recommended that farmers properly organise themselves and re-align thoughts and attitudes towards their scheme's infrastructure and be aware of the consequences of their actions for other tail-end users. Additionally, there is no need to formalise every aspect of SISs. SIS frameworks should incorporate ‘working’ informal mechanisms to bridge the gaps created by ad hoc legal frameworks. A flexible policy arrangement should prevail so that a balance is created between structured water governance regimes and informal/traditional arrangements, which will create an equilibrium and harmony between the ‘situation on the ground’ and the designed structured frameworks. In addition, IMT policy arrangements have to be implemented to motivate farmers to attend farmer training sessions and discussion groups.

Acknowledgments

The study was undertaken as part of project (K5/2556/4) initiated, managed and funded by the Water Research Commission (WRC) South Africa, entitled, ‘Assessment of Policies and Strategies for the Governance of Smallholder Irrigation Farming in KwaZulu-Natal Province, South Africa’. The authors would also like to acknowledge the support from the Department of Agriculture, Forestry and Fisheries (DAFF), South Africa, and the University of KwaZulu-Natal, Pietermaritzburg, KwaZulu-Natal, South Africa.

References

References
Abernethy
C. L.
, (
1990
).
Indicators of the Performance of Irrigation Water Distribution Systems
.
International Water Management Institute
,
Colombo
,
Sri Lanka
.
Alba
R.
&
Bolding
A.
, (
2016
).
IWRM avant la lettre? Four key episodes in the policy articulation of IWRM in downstream Mozambique
.
Water Alternatives
9
(
3
),
549
.
Alba
R.
,
Bolding
A.
&
Ducrot
R.
, (
2016
).
The politics of water payments and stakeholder participation in the Limpopo River Basin, Mozambique
.
Water Alternatives
9
(
3
),
569
.
Alcon
F.
,
García-Bastida
P.
,
Soto-García
M.
,
Martínez-Alvarez
V.
,
Martin-Gorriz
B.
&
Baille
A.
, (
2017
).
Explaining the performance of irrigation communities in a water-scarce region
.
Irrigation Science
35
(
3
),
193
203
.
Allouche
J.
, (
2016
).
The birth and spread of IWRM – a case study of global policy diffusion and translation
.
Water Alternatives
9
(
3
),
412
.
Amemiya
T.
, (
1981
).
Qualitative response models: a survey
.
Journal of Economic Literature
19
(
4
),
1483
1536
.
Balasubramanya
S.
,
Price
J. P.
&
Horbulyk
T. M.
, (
2017
).
Impacts assessments without true baselines: assessing the relative effects of training on the performance of water user associations in Southern Tajikistan
.
Water Economics and Policy
DOI: 10.1142/S2382624X18500078
.
Bruns
B.
, (
2007
).
Irrigation water rights: options for pro-poor reform
.
Irrigation and Drainage
56
(
2–3
),
237
246
.
Cakmakyapan
S.
&
Goktas
A.
, (
2013
).
A comparison of binary logit and probit models with a simulation study
.
Journal of Social and Economic Statistics
2
(
1
),
1
17
.
Colona
F.
&
Jaffe
R.
, (
2016
).
Hybrid governance arrangements
.
The European Journal of Development Research
28
(
2
),
175
183
.
Cousins
B.
, (
2012
).
Smallholder irrigation schemes, agrarian reform and ‘accumulation from below’: Evidence from Tugela Ferry, KwaZulu-Natal
. In:
Paper for A Conference on ‘Strategies to Overcome Poverty and Inequality: Towards Carnegie III’
,
3–7 September 2012
.
University of Cape Town
,
South Africa
.
DAEA (Department of Agriculture and Environmental Affairs)
(
2001
).
Mooi-River Irrigation Scheme Records, Tugela Ferry Offices. Msinga, KwaZulu-Natal, South Africa
.
Damisa
M.
,
Abdulsalam
Z.
&
Kehinde
A.
, (
2008
).
Determinants of farmers’ satisfaction with their irrigation system in Nigeria
.
Trends in Agricultural Economics
1
(
1
),
8
13
.
Denby
K.
,
Movik
S.
,
Mehta
L.
,
van Koppen
B.
, (
2017
).
The ‘trickle down' of integrated water resources management: a case study of local-level realities in the Inkomati water management area, South Africa
. In:
Flows and Practices: The Politics of Integrated Water Resources Management in Eastern and Southern Africa
.
Mehta
L.
,
Derman
B.
&
Manzungu
E.
(eds).
Weaver Press
,
Harare
,
Zimbabwe
, pp.
107
.
Esteban
E.
,
Dinar
A.
,
Albiac
J.
,
Calera
A.
,
García-Mollá
M.
&
Avellá
L.
, (
2018
).
Interest group perceptions on water policy reforms: insight from a water stressed basin
.
Water Policy
20
(
4
),
794
810
.
Frederick
K. D.
, (
2013
).
Scarce Water and Institutional Change
.
Earthscan
,
New York
,
USA
.
Gomo
T.
, (
2012
).
Assessing the Performance of Smallholder Irrigation in South Africa and Opportunities for Deriving Best Management Practices
.
Unpublished MSc Thesis
,
Bioresources Engineering, University of KwaZulu-Natal, Pietermaritzburg, Pietermaritzburg South Africa
.
Gorantiwar
S.
&
Smout
I. K.
, (
2005
).
Performance assessment of irrigation water management of heterogeneous irrigation schemes: 1. A framework for evaluation
.
Irrigation and Drainage Systems
19
(
1
),
1
36
.
Groenfeldt
D.
&
Svendsen
M.
, (
2000
).
Case Studies in Participatory Irrigation Management
.
World Bank Publications
,
Washington, DC
,
USA
.
Hope
R.
, (
2006
).
Water, workfare and poverty: the impact of the working for water programme on rural poverty reduction
.
Environment, Development and Sustainability
8
(
1
),
139
156
.
Horst
L.
, (
1998
).
The Dilemmas of Water Division: Considerations and Criteria for Irrigation System Design
.
International Water Management Institute
,
Colombo
,
Sri Lanka
.
Hounsome
B.
,
Edwards
R. T.
&
Edwards-Jones
G.
, (
2006
).
A note on the effect of farmer mental health on adoption: the case of agri-environment schemes
.
Agricultural Systems
91
(
3
),
229
241
.
Incorporation
R.
, (
2010
).
Sample size calculator (online). Available from: http://www.raosoft.com/samplesize.html [Accessed: 10 May 2017]
.
Kuşçu
H.
,
Demir
A. O.
&
Korukçu
A.
, (
2008
).
An assessment of the irrigation management transfer programme: case study in the Mustafakemalpaşa irrigation scheme in Turkey
.
Irrigation and Drainage
57
(
1
),
15
22
.
Langyintuo
A.
&
Mekuria
M.
, (
2005
).
Modeling agricultural technology adoption using the software STATA
. In:
CIMMYT-ALP Training Manual
,
21–25 February 2005
,
Harare, Zimbabwe
.
Letsoalo
S.
&
Van Averbeke
W.
, (
2006
).
Infrastructural maintenance on smallholder canal irrigation schemes in the north of South Africa
. In:
Proc. International Symposium on Water and Land Management for Sustainable Irrigated Agriculture
.
Cukurova University
,
Adana
,
Turkey
, pp.
4
8
.
Maddala
G. S.
, (
1986
).
Limited-dependent and Qualitative Variables in Econometrics
.
Cambridge University Press
,
Cambridge
,
UK
.
Meagher
K.
,
De Herdt
T.
&
Titeca
K.
, (
2014
).
Unravelling Public Authority: Paths of Hybrid Governance in Africa
.
IS Academy
,
London
,
UK
.
Mehta
L.
,
Movik
S.
,
Bolding
A.
,
Derman
B.
&
Manzungu
E.
, (
2017
).
Introduction-Flows and practices
. In:
Flows and Practices
,
Vol. 1
.
Weaver Press
,
Harare
,
Zimbabwe
, pp.
1
29
Mnkeni
P.
,
Chiduza
C.
,
Modi
A.
,
Stevens
J.
,
Monde
N.
,
Van der Stoep
I.
&
Dladla
R.
, (
2010
).
Best Management Practices for Smallholder Farming on two Irrigation Schemes in the Eastern Cape and KwaZulu-Natal Through Participatory Adaptive Research
.
WRC Report No. TT 478/10
.
Pretoria
,
South Africa
.
Molden
D. J.
&
Gates
T. K.
, (
1990
).
Performance measures for evaluation of irrigation-water-delivery systems
.
Journal of Irrigation and Drainage Engineering
116
(
6
),
804
823
.
Movik
S.
,
Mehta
L.
,
van Koppen
B.
&
Denby
K.
, (
2016
).
Emergence, interpretations and translations of IWRM in South Africa
.
Water Alternatives
9
(
3
),
456
.
Palmer
J. D.
,
Clemmens
A. J.
&
Dedrick
A. R.
, (
1989
).
Several sources of nonuniformity in irrigation delivery flows
.
Journal of Irrigation and Drainage Engineering
115
(
6
),
920
937
.
Pereira
L. S.
,
Oweis
T.
&
Zairi
A.
, (
2002
).
Irrigation management under water scarcity
.
Agricultural Water Management
57
(
3
),
175
206
.
Pindyck
R. S.
&
Rubinfeld
D. L.
, (
1981
).
Econometric Models and Economic Forecasts
.
New York: McGraw-Hill
,
New York
,
USA
.
Prathapar
S.
,
Hassan
M.
,
Mirza
Z. I.
&
Tahir
Z.
, (
2002
).
Constraints on enforcement of water policies: selected cases from South Asia
. In:
ACIAR Proceedings
.
Brennan
D.
(ed.).
ACIAR
,
Canberra
,
Australia
, pp.
171
176
.
Rogerson
P.
, (
2001
).
Statistical Methods for Geography
.
Sage
,
London
,
UK
.
Shah
T.
,
Van Koppen
B.
,
de Lange
D. M. M.
&
Samad
M.
, (
2002
).
Institutional Alternatives in African Smallholder Irrigation: Lessons From International Experience with Irrigation Management Transfer
.
International Water Management Institute
,
Colombo
,
Sri Lanka
.
Svendsen
M.
,
Ewing
M.
&
Msangi
S.
, (
2009
).
Measuring Irrigation Performance in Africa
.
International Food Policy Research Institute (IFPRI)
,
Washington
,
USA
.
Taylor
P.
, (
2002
).
Some principles for development and implementation of water-allocation schemes
. In:
ACIAR Proceedings
.
Brennan
D.
(ed.).
ACIAR; 1998
,
Canberra
,
Australia
, pp.
62
74
.
UWPConsulting
(
2012
).
Tugela Ferry Irrigation Scheme Refurbishmnet
.
Cascades Pietermaritzburg
,
South Africa
.
Uysal
Ö. K.
&
Atış
E.
, (
2010
).
Assessing the performance of participatory irrigation management over time: a case study from Turkey
.
Agricultural Water Management
97
(
7
),
1017
1025
.
Van Averbeke
W.
,
Denison
J.
&
Mnkeni
P.
, (
2011
).
Smallholder irrigation schemes in South Africa: a review of knowledge generated by the water research commission
.
Water SA
37
(
5
),
797
808
.
van Koppen
B.
,
Tapela
B.
&
Mapedza
E.
, (
2018
).
Joint Ventures in the Flag Boshielo Irrigation Scheme, South Africa: A History of Smallholders, States and Business
.
International Water Managenment Institute
,
Pretoria
,
South Africa
.
Wisser
D.
,
Frolking
S.
,
Douglas
E. M.
,
Fekete
B. M.
,
Vörösmarty
C. J.
&
Schumann
A. H.
, (
2008
).
Global irrigation water demand: variability and uncertainties arising from agricultural and climate data sets
.
Geophysical Research Letters
35
(
24
),
1
5
.
Xie
M.
, (
2006
).
Integrated water resources management (IWRM) – introduction to principles and practices
. In:
Africa Regional Workshop on IWRM
.
World Bank Institute
,
Nairobi
,
Kenya
.
Yildirim
Y. E.
&
Çakmak
B.
, (
2004
).
Participatory irrigation management in Turkey
.
International Journal of Water Resources Development
20
(
2
),
219
228
.