Network efficiency is its capacity to transport a defined quantity of water from point A to B. Flow measurements is therefore necessary to determine this efficiency. In the literature, several methods have been developed to estimate the flow rates of free surface water [20]. From methods by velocity field exploration to the approach by the general equations of hydraulics, through volumetric (gravimetric) methods, tracer dilution, gauging channels and calibrated weirs, the choice of approach to use is an important exercise [20] [21]. Many factors can guide the choice of appropriate method: flow rates, channel size, type of flow, flow velocity, required accuracy, safety considerations… [20] [21]. For this study, the current meter gauging and the float gauging methods with two types of floats (weighted bottles and stick) thrown on the two banks and the middle of the channel was used. Indeed, these methods, which are simpler in their implementation, are adapted to this study and have already been used for flow measurements on several irrigated perimeters in Burkina Faso (perimeter of the Kou Valley, Bagré, Mogtedo, etc.) [13] [22]. To study the functioning of the irrigation network in a wet campaign, the flow rates were measured twice during the wet campaign of the year 2020: at the beginning (July) and the middle of the campaign (August). A dozen measurement sections were retained in total on the channels. 1) For the feeder, a section at the start and the end were retained. 2) For the primary channel, the measurements concerned the start of the primary channel and that of the secondary channels. 3) For the secondary channels, the measurements concerned only the S2 at its entrance and at the level of the tertiary outlets. In this study, the section is defined as any portion of the network where there is no water intake, no service on the way and no water supply. According to the continuity equation, the sum of the incoming flows must be equal to the sum of the outgoing flows plus or minus the losses [23].

Thus, for a section:

Inflow = outflow + ∑ losses (1)

eff [ % ] = outflow Inflow × 100 = Qdownstream Qupstream (2)

This method, based on general hydraulic equations, has permitted us to compare the current roughness of the channels with that of new channels and assess the current quality of the channel. In this study, the Manning and Strickler equation was used to determine the channel roughness parameter, using the measured flow rates as input data. For greater reliability, on the same reach, the calculation was made on the basis of two flow measurements.

U = K S R H 2 / 3 I ⇒ K S = Q R H 2 / 3 S I (3)

R H = S wetted P wetted (4)

I: slope of the canal;

S: wetted section [m²];

P: wet perimeter [m];

R_H: hydraulic radius [m];

Q: flow rate [m³/s];

K_s: Manning’s roughness coefficient [m^1/3/s].

From the first phases of this study, the various actors involved in irrigated perimeter management were consulted in the development of methodological protocols and their applications. In this dynamic, surveys were carried out to collect quantitative and qualitative data on the operation and management of water and structures in the perimeter. The investigation mainly targeted four entities, including institutional managers through the Provincial Directorate of Agriculture and Hydro-agricultural Development (DPAAH) and the Cascades Water Agency (AEC), local cooperatives and operators. Thus, three types of surveys were administered according to the target group in French and in the local language (Dioula) according to the preferences of the interviewee: 1) semi-structured interviews with the chief of the perimeter and the operation’s technicians, 2) structured interviews with each of the farmers’ cooperatives at a meeting organized for this purpose and 3) questionnaires with eighteen (18) farmers of all sex, age and position on the perimeter. The questions addressed revolve about several points, namely the administrative and institutional situation of the perimeter, the historical and physical description of the irrigation system, the dynamics of villages in irrigation, the functioning of cooperatives, the water and infrastructure management mode, and the difficulties encountered by management, technicians, cooperatives and operators in the efficient operation of the perimeter.

Validated mapping (Figure 3) determined that the irrigation canals have varying lengths (Table 1) depending on the geometry of the area they serve. The irrigated perimeter of Karfiguéla has a gross area of 400.4 ha and a net area of 360 ha. It’s served by a gravity network supplied by a feeder located on the Comoé River far of 2.230 km from the perimeter.

The hydraulic district, called “block” on the perimeter, generally has an area of 15 ha which may vary depending on the configuration of the network (Table 1). The block has 30 plots of 0.5 ha. Each farmer generally divides his plot into ten boxes of 25 m by 20 m each. The hydraulic districts are grouped into four large blocks representing the areas dominated by the secondary canals as presented in Figure 3.

Karfiguéla irrigation network is made of the feeder, primary, secondary and tertiary channels of trapezoidal shape with an ordinary concrete coating and by quaternaries in clay. The supply channel has two stairs and a reading scale for the water level. The primary channel has four valves and stairs installed at the entrance of each secondary and three coupled control valves to lateral weirs, installed after the intakes of the secondary channels S1, S2 and S3. The secondary channels are each supplied by an intake valve located on the right bank of the main channel. Right at their intersections with the tertiary sectors they supply, tertiary intake valves and flow control valves are installed. The tertiary sectors are equipped with quaternary intake and flow regulation valves. The quaternaries directly supply the plots and are scoured and traced in an archaic way in each campaign by the farmers who depend on them. Their shape and number are undetermined.

The perimeter sanitation network consists of two main drains and an outfall, all of clay and which drain the excess water from the irrigation canals and the various blocks of plots through secondary drains towards Karfiguéla Lake (Figure 4).

The first primary drain collects water from tertiary T1-i and from the blocks irrigated by these tertials. The second main drain drains the water from T2-i and the plots are located to its right through secondary drains. The outfall collects drainage water from the tertiary sectors and blocks served by the secondary canal S3 and the upper part of the secondary S4. As presented in Figure 4, the drainage of the lower part of the perimeter isn’t really assured. The secondary drains aren’t connected to any collector, thus making drainage non-existent.

Also, during the mapping, interruptions on several secondary drains were noticed. The results of the visual inspection, presented below, explain these remarks and validate the map.

Infrastructure is said to be functional when it achieves the objective for which it was installed [24]. It’s therefore qualified as non-functional when it’s no longer suitable or used for its primary function. However, the functionality doesn’t attest to the quality of the work because it can be functional, but qualified as poor when it doesn’t fully play its role [24]. Inventory results in Table 2 show a rate of functionality of the infrastructure of 79%. The main malfunctions encountered are the breaks or landslides of the stairs and weirs (ii) the valve mechanism problems or the total lack of valves at the level of the tertiary and quaternary canals and the total lack of scuppers on the entire perimeter.

In general, the feeder and the primary channel suffer from some minor slight pathologies ranging from grass cover to cracks in the panels, including silting up. As for the secondary and tertiary canals, the damage is greater (Figure 4). The eight (8) main impairments observed on the different channels can be summarized as follows (Figure 5): 1) transverse, vertical, oblique or any other cracking of the majority of the panels; 2) complete degradation of the panels joints on all the channels and grassing of these joints; 3) settlements and overturning of certain canals; 4) advanced erosion of the canals lining, especially in the tertiary canals; 5) erosion of the backfill of canals and weirs (intake infrastructure) certainly due to overflows caused by improper handling of valves; 6) presence of shrubs and large trees in close proximity of the canals leading to serious cracks and even overturning of the concrete modules over a length of several meters for the secondary and tertiary canals; 7) significant siltation and grassing of the various canals and 8) intentional filling of the last sections of certain tertiary canals with a view to redirecting all the water towards the plots upstream. The state of deterioration of the channels can be explained by the old age of the network, built more than 45 years ago.

The drainage network is traced in each campaign by the producers, but in an anarchic and disordered. Several malfunctions hamper its proper functioning (Figure 6). These are: 1) the interruption of the drainage network in certain places, by some producers who by this sabotage hope to gain a few additional square meters of irrigable area; 2) voluntary widening of drains in order to engage in fishing activities; 3) rejecting the rice straw directly inside the drains after threshing; 4) the diversion of drainage water for the irrigation of plots outside the consensus irrigation schedules.

Current meter and float gauging were carried out in the feeder, in various sections of the primary canal and secondary canals (Table 3) during the July and August months (peak rainfall) of the 2020 year. Between July and August in the feeder, there is very little variation in water flow. In other words, the water quantities entering the network are practically constant. In addition, these values of water flow in the feeder are consistent with the flow measurements carried out on the feeder during 2018 dry season [22]. In which the efficiency of the network was estimated at 52%. On the other hand, the water flow values in the primary, secondary and tertiary networks vary enormously depending on the period. Indeed, this is explained by the variation in the crop’s water requirement in the rainy season, due to rainfall variability. The distribution of flows in the network depends on rainfall episodes and the crops vegetative stage. The specific case of the large variation in the flow of the secondary channel S4, whose flow goes from 0.053 m³/s in July to 0.24 m³/s, is explained in particular by the downstream position of this channel in the network. Indeed, S4 being a continuity of the primary channel (Figure 3); it therefore constitutes a drainage channel for unused flow of the primary channel. During August month, the month of high rainfall, the irrigation water needs are very low or even non-existent, S4 drains almost all water entering the network.

For the 2020 wet season, at first glance, the efficiency during the month of July in the feeder and primary channel seem quite good for an old gravity network. The efficiency during August months (all greater than 100%) suggests the water unforeseen inflows into the canals from 1) surface water runoff in the secondary and tertiary coming from unplanned plots, and the strong degradation of canals, 2) the drainage of plots which are flooded or caused by the presence of large cracks in the canals serving as drainage corridors during irrigation, 3) the drainage of surface water. The general observation is that during the wet season, the role of irrigation network is reversed; it receives more water into the different channels than it supplies. Instead of participating in the water distribution, it turns into a drainage network. This situation was also noted during surveys with operators and the visual inspection.

The aim of estimating the roughness coefficients is to assess, with relative magnitudes, the level the channel degradations. The values K_s1 and K_s2 (Table 4) represent the Strickler roughness coefficients, calculated using two flow measurements in July and August 2020. An average K_s is calculated when the variation between K_s1 and K_s2 is acceptable and if not, the measurements are repeated. Thus, the variation values between the K_s1 calculated with the July 2020 flow measurements and K_s2 taken from the August 2020 measurements also permitted us to validate the values of the measured flows. In this study, the maximum variation in roughness is 8 m^1/3/s, the flow values are therefore good. With regard to the roughness values, the general observation is that the roughness of channels increases in the network and that all the values of Strickler coefficients of channels are lower than that of the estimated new concrete at about 70 m^1/3/s [25] except the feeder which has a K_s of 68 m^1/3/s. This situation is explained, on the one hand, by the increase in the solid load and water pollution (organic and chemical) as one as moves forward in the network and, on the other hand, by human intervention level in the network which creates mechanical and chemical erosion of channels which are closest to the plots. Indeed, the operators use clay mixed with herbs to plug leaks in the valves and to replace the non-existent valves at the tertiary level. Thus, the channel roughness results illustrate very well the level of degradation of channels on the perimeter. They’re consistent with the results obtained during the visual diagnosis.

The perimeter is used by about 1200 producers coming from eight (8) villages: Karfiguéla, Banfora, Tingréla, Lémouroudougou, Kiribina, Sibiéna, Tiékouna and Nafona [14]. They’re grouped into five (5) cooperatives each managing one block with the exception of cooperatives n˚4 and 5 which share only block IV (Figure 7).

Each cooperative is managed by an office made of a president, a treasurer, a secretary all unpaid except for the water worker who has a monthly salary between 15.19 Euros and 22.78 Euros. The cooperatives are grouped into a union called the Union of Karfiguéla’s Perimeter Operators (UCEPAK). It benefits from the regular support of technicians from the DPAAH for better management of water, fertilizers, infrastructure and marketing. In theory, UCEPAK has internal regulations and includes three bodies: the general assembly, the management board and the control committee. Unfortunately, despite the relatively high level of education of cooperative members, namely 75% of which 60% are at the primary level and 15% at the secondary level and 25% are uneducated members (Figure 8), it’s clear that the organization of cooperatives and the union leaves something to be desired.

While the internal regulations are almost non-existent, for some social considerations, no punitive measure is taken in the event of non-compliance with the rules. No clear and formal task or responsibility is assigned to each member of the union and cooperatives, thus leading to poor management and organization of cropping campaigns. Also, it should be noted that the ignorance of principles and regulations of cooperatives and the UCEPAK by the farmers create a climate of mistrust, individualism and even conflicts within the perimeter. The most explicit situation encountered during field trips is non-compliance with the crop calendar proposed by the DPAAH, leading to crops cycle differences of up to 2 months.

In the dry campaign, the irrigated perimeter of Karfiguéla is divided into two parts: the first part is made up of blocks I, II and III and the second is made up of blocks IV and V [4]. Water problems are such that a system of alternating seeding of each part of the perimeter has been adopted. Thus, only one part of the perimeter is exploited in dry campaign. Even under these conditions, the water deficit is still present on the perimeter and also persists in wet seasons.

In theory, the entire perimeter is seeded during the wet campaign. However, for several years, some plots have received other crops than rice, which is the main vocation of the perimeter, or have been abandoned for hydraulic, topographical or pedological reasons. Already in 2002, there was a land loss rate of 8.2% and 7.3% in 2003. In 2018, the abandoned area was about 25 ha [14]. In the middle of a wet season in the year 2020 (Figure 9) speculations other than rice were recorded over an area of 10.8 ha, the rest of the plots seeded with rice. These speculations were put in place from the first rains and mainly concern corn, okra, beans, sorrel, chili and bananas. At the end of the wet season, before the harvest period, the abandoned area was about 10.6 h (Figure 9). For the majority of abandoned plots, it was noticed that crops from nearby plots have wilted stems and empty panicles thereby indicating a water problem in this area.

On Karfiguéla irrigated perimeter, collective activities are programmed in advance on cropping calendars by a collaboration between farmers, technicians who support them and the DPAAH. Unfortunately, the schedule is rarely respected especially during the wet season because of rainfall delays. In 2020, for example (Table 5), cycle deviations of up to two (2) months were noted on the perimeter, complicating the management of water, fertilizer inputs and crop diseases on the perimeter.

• Management of water turns and cultural calendar

The distribution of water in the plots during dry campaign is done by water turns on the secondary canals. As for the application of water to the plot, it’s mainly done by submersion basins for rice, and by short lines for vegetables. The water release program on the perimeter during dry campaign provides a water volume per cycle of 8000 m³/ha for the part of the perimeter that will be developed and considering the efficiency of 0.6 for the irrigation system. However, the problems linked to climatic hazards (increased evapotranspiration), topography, degradation of canals and possibly the diversity of speculations make this resource insufficient and don’t always facilitate the strict application of water turns.

In the wet campaign, the water in the irrigation network comes exclusively from the natural flow of the Comoé River. There is no water release operated by SN-SOSUCO, the main controller of valve openings at the dam. There is no precise watering schedule in the wet season, because the irrigation water supply depends on the quantities of rain that has fallen. This situation causes the same inconvenience as in the dry campaign, especially during periods of delayed rainfall or pockets of drought. To these problems are added anarchist and disorderly manipulations of flow control valves by farmers which lead to water deficits on certain plots and floods on others, consequently generating latent and real conflicts between farmers.

• Upkeep and Maintenance of Infrastructures

The maintenance and upkeep of hydraulic infrastructure remain a central component in the management of irrigated areas [13]. These operations allow better management of water distribution while strengthening the sustainability and efficiency of the infrastructure. On the perimeter, channel maintenance operations generally consist of weeding and cleaning (Figure 10(d)). For a synergy

and staggering of maintenance tasks, the roles are distributed as follows: the feeder and the primary channel are maintained by all the cooperatives, the portion of the canal which directly feeds its block and the secondary canals are maintained by each cooperative, the tertiary canals are maintained by the irrigators concerned and finally, the quaternary canals are traced at each campaign. The weeding and cleaning works are carried out with archaic materials such as hoes and manual shovels.

Rehabilitation such as the rebuilding of degraded, cracked or overturned canals (a) and b), are often carried out by farmers on the irrigation canals. These maintenance or rehabilitation works are made possible through water charges (10 Euros/ha) paid at the end of each season by the operators. The rehabilitation is controlled by the office of the cooperative concerned with the repairs. Unfortunately, it’s clear that this work is often botched and the rehabilitated infrastructure fails quite frequently (Figure 10(c)). As for the drainage network, it is not frequently rehabilitated and the malfunctions only accumulate.