HIPPO InfoService

perspectives4 Feb 16, 2003

Van 't Hof, S. & Maurice, L. (2002) Efficiency, cost, optimization and spread of spray irrigation in West Africa. HIPPO Perspectives 4, 14
pp., 4 figs.
perspectives4.pdf, 180k

Abstract: The cost and efficiency of spray irrigation are analysed and the prospects for its optimization are examined. Spray irrigation is a low-cost, overhead system that consists of a small petrol pump (Honda, Yamaha, Robin) with a lay-flat hose of 40 or 50 mm diameter at the end of which is a hand-held spraying head of the type normally found on watering cans. Its use by professional market gardeners was observed in Togo, Niger and Mali in 2000 and 2001. Spray irrigation is energy efficient, water efficient, adapts well to the yield limitations of low-cost, hand-dug wells, is readily available, and poses no problems in terms of technology transfer. There are various options for optimization: (1) reducing the price of equipment, including that of accessories, such as layflat hose; (2) improving spraying head efficiency; (3) improving pump efficiency; (4) improving engine efficiency; and (5) replacing the petrol pump with a treadle pump. It is recommended to collect information on existing spray systems to enable their cost and efficiency to be compared with those of fully developed technologies, such as sprinkler and trickle irrigation. Contact: HIPPO Foundation, De Verwondering 27, 3823HA Amersfoort, the Netherlands, e-mail hippomp@net.hcc.nl, www.hipponet.nl

xxxxxxxxxxxx

Key words: irrigation methods, costs, water use, efficiency, appropriate technology, innovation adoption, irrigations schemes, small farms, private farms Mots cls: mthode d'irrigation, cout, utilisation de l'eau, efficience, technologie approprie, adoption de l'innovation, primtre irrigu, petit agriculteur, exploitation agricole prive Annotation : (S. van t Hof, hippomp@net.hcc.nl) The technology of spray irrigation was pioneered by African farmers to suit their conditions. It has a number of advantages that make it probably make it a better choice than trickle or sprinkler solutions. As a result the technology spread like wildfire throughout West Africa (Benin, Niger, Mali) without the assistance of development organizations.
1

xxxxxxxxxxxxx

The annotation serves to highlight aspects that are related to the mission or work of the HIPPO Foundation (http://www.hipponet.nl).

SPRAY IRRIGATION

HIPPO Foundation

Efficiency, cost, optimization and spread of spray irrigation in West Africa

Sjon van t Hof and Laurent Maurice
2002

HIPPO Perspectives 4
Prepublication

Secretariat
De Verwondering 27 3823 HA Amersfoort The Netherlands www.hipponet.nl hippomp@net.hcc.nl xxxxxxxxxxxxxx

Efficiency, cost, optimization and spread of spray irrigation in West Africa

Sjon van t Hof1 and Laurent Maurice2
Abstract: The cost and efficiency of spray irrigation are analysed and the prospects for its optimization are examined. Spray irrigation is a low-cost, overhead system that consists of a small petrol pump (Honda, Yamaha, Robin) with a lay-flat hose of 40 or 50 mm diameter at the end of which is a hand-held spraying head of the type normally found on watering cans. Its use by professional market gardeners was observed in Togo, Niger and Mali in 2000 and 2001. Spray irrigation is energy efficient, water efficient, adapts well to the yield limitations of low-cost, hand-dug wells, is readily available, and poses no problems in terms of technology transfer. There are various options for optimization: (1) reducing the price of equipment, including that of accessories, such as layflat hose; (2) improving spraying head efficiency; (3) improving pump efficiency; (4) improving engine efficiency; and (5) replacing the petrol pump with a treadle pump. It is recommended to collect information on existing spray systems to enable their cost and efficiency to be compared with those of fully developed technologies, such as sprinkler and trickle irrigation.

INTRODUCTION
Spray irrigation is a low-cost, overhead system that consists of a small petrol pump with a lay-flat hose of 40 or 50 mm diameter at the end of which is a hand-held spraying head of the type normally found on watering cans. Its use by professional market gardeners was observed in Lom, Togo
1 HIPPO (High-efficiency Irrigation Pumps, Procurement & Organization) Foundation, De Verwondering 27, 3823 HA Amersfoort, The Netherlands, tel/fax +31.33.4553623, HIPPOMP@net.hcc.nl, www.hipponet.nl 2 SahelConsult, Bamako, consult@afribone.net.ml

Spray Irrigation

(ICRA, 2000), but also Niamey in Niger and Bamako, Mali (Van t Hof, 2001a and b). It resembles stationary spraylines as described by Kay (1983). Over the past few years, spray irrigation technology spontaneously spread across West Africa, i.e. from farmer to farmer and without the assistance of development organizations. A recent survey in Bamako among 80 market gardeners with small petrol or diesel pumps showed that more than 60% use spray irrigation (Maurice, 2002). Farmers in Bamako did not know of its use elsewhere in West Africa and considered it a local invention that dates back to about 1998. Similar remarks were recorded in Niamey, so it can be assumed that the centre of origin is unknown. Spray irrigation is an intermediate technology between completely manual bucket or watering can systems and more sophisticated sprinkler or drip irrigation technologies. The investment cost is intermediate, water use efficiency is high, energy efficiency is very good, and labour productivity is high by West African standards. Moreover, it has an excellent record of spontaneous technology transfer. In this article, spray irrigation is described, its cost is determined, and the prospects for its optimization are assessed. By documenting the technology it will be possible to compare its cost and efficiency with the fully developed technologies of sprinkler and trickle irrigation. The focus of the article is on spray irrigation using groundwater from non-reinforced wells as observed in Bamako (S. van t Hof, 2001b), which seems to be more common than spray irrigation from a surface source, such as the Niger river, as observed both in Niamey and in Bamako (S. van t Hof, 2001a and b).

SYSTEM DESCRIPTION
The source of water is either a shallow aquifer or a surface source, such as the Niger river. In Bamako, intensive market gardening has been going on for decades. Most of it takes place in relatively small orchard-like zones between built-on areas. Non-reinforced wells are dug by hand in the alluvial plain between the Mandingue hills and the Niger river. Because of

Spray Irrigation

the clayey subsoil, these wells are remarkably stable3 but there yield is very limited, hence the necessity to construct up to 10 wells/ha. Prior to the introduction of small petrol pumps, water was lifted by hand from these wells and distributed on sunken or raised4 vegetable beds with watering cans. In some cases the task of water lifting was eased by pumping water to small concrete or masonry basins5. Where the water is pumped from a surface source, there is often a considerable distance from the river or reservoir to the vegetable garden to avoid the garden from being submerged during high water. Here pumping is used for transportation rather than lifting to reduce the drudgery of carrying water over distances of 100 m or more. In the Bamako-type of spray irrigation, the length6 of the lay-flat hose is 25-50 m and its diameter is 40 or 50 mm. The spraying heads are locally made. So far, no measurements have been made of their dimension, number and size of holes, or discharge/pressure relationship. In a 0.5 ha vegetable garden, there may be 5 wells. A portable, selfaspiring pumpset is carried from one well to the other to empty them one by one. With a discharge of 1-2 l/sec, it takes 15 to 30 minutes to empty each well, a bit less during the dry season when the groundwater level reaches a seasonal low with a static head of 2.5 m below the surface. The static water level in the well is between 0.5 and 2 m, depending on the season. Typically, the well diameter is 2 m at the surface and 1.5 m at the bottom. Fuel consumption is in the order of 0.5 l/hour.

SPRAYING HEAD HYDRAULICS

Measurements were made on a 10 litre industrial watering can of Geli Thermoplastics to provide the basic data for hydraulic calculations. The punctured disc of the spraying head had a diameter of about 10 cm with 221 holes in it of 1-1,5 mm diameter. The holes are situated in 9 concentric
We were informed by an owner that they can last 10 years with normal maintenance. Sunk beds are used during the dry season, raised beds during the wet season. 5 The dimensions of these basins are not standardized. In some cases their width x length x depth could be about 2 x 2 x 1 m. Some are provided with steps to facilitate taking the last remaining water.
4 6 3

Based on 10 wells/ha, its minimum length is Lmin =

1 1000 4 2

= 17,8 m.

Spray Irrigation

circles. It took 25 seconds to empty a full 10 litre can. The average head was estimated at 25 cm. The discharge of a spraying head can be calculated using Torricellis theorem ( v = 2gh ) corrected for flow through a sharp-edged orifice where the proper area to use for the jet cross-section is that at the smallest diameter at a point, called the vena contracta, slightly outside the orifice, as follows: Q = 0.62 A 2 9.81 H (eq. 1), where Q is the discharge in m3/sec, A is the area of the orifice in m2 and H is the head in metres (Mott, 2000). With an average orifice diameter of 1.25 mm, the total orifice area A of the Deli watering can is 0.0002711 m2. Using Eq. 1, Q = 0.37 l/sec and the volume emptied in 25 seconds is estimated at 9.3 litres, which is reasonably close to the maximum volume of water that can be contained in a full Deli watering can. This shows that equation 1 can be used to vary the head-discharge relationship of the spraying head by changing the diameter and number of spraying holes.

EFFICIENCY
The following main types of efficiency will be discussed below: energy efficiency and irrigation efficiency. Together with the procurement cost of the equipment, the prevailing interest rate, the fuel price and the lifespan of the equipment they are the main factors that determine the cost of operation of an irrigation pump system. Energy efficiency is mainly determined by pump efficiency, engine efficiency, and water-lifting efficiency. Water-lifting efficiency is defined here as the ratio of static head to dynamic head, where the static head is expressed as the difference in height between the surface or depth of the source of water and the point of application (e.g. the soil surface) to dynamic head. The dynamic head is the sum of the static head, head losses and operation pressure. To simplify matters, we will start by considering a Robin EY15 7 /Koshin SEH-50X 8 pumpset, with a 50 mm, 6 m suction hose and a 50
7
8

http://www.robinamerica.com/engines/robinsv.html

http://www.contractorstools.com/koshin.html. The SEH-50X pump can also be found with Honda engines, such as the GX120 or the G150.

Spray Irrigation

mm, 30 m lay-flat discharge hose. The EY15 is the smallest engine of the SV series of Robin-Subaru (Japan) and the SEH-50X is the smallest centrifugal (non-trash) pump of Koshin Ltd. (Japan). They are closecoupled, so the transmission is considered 100% efficient. The EY15/SHE-50X pumpset is described in ANPIP/PPIP (2000). The data were entered in pumpselectSMALL.xls, a Microsoft Excel spreadsheet on pump selection and configuration (Van t Hof, 2002).
Table 1: pump data
Koshin SHE-50X Manufacturer Model Maximum speed 3600 rpm suc./disch. 50/50 mm At a speed of 2925 rpm the pump has the following characteristics Q - m3/h P - kW H-m -% 0 0,89 15,8 0% 7,2 0,89 14,2 31% 14,4 0,89 10,2 45% 20,16 0,89 5,2 32% 25,2 0,89 0 0%

NPSHR m N/A N/A N/A N/A N/A

In ANPIP/PPIP (2000) only the combined pump-engine performance was recorded, i.e. discharge, total dynamic head and fuel consumption. Therefore the data for pump efficiency and power requirement had to be estimated. This was done by assuming that the best efficiency of this small, self-aspiring pump is 45% and that its power curve is perfectly horizontal. Again, most of the values had to be estimated. We assumed a linear relationship between engine speed and power. The correction of actual consumption in relation to engine load was varied until a fairly satisfactory match with the test bench data was obtained. The same applies to specific fuel consumption. A fuel consumption of 0.5 l/h suggests an engine speed of about 2500 rpm. Using the spraying head of the Geli watering can (221 holes of 1,25 mm), a 30 m discharge hose of 40 mm diameter, a 6 m suction hose of 50 mm and a static head of 4 m, pump efficiency is 30%, pump power is 0,56 kW, engine efficiency is 11.7%., the total dynamic head becomes 10.48 m, while the operating pressure of the spraying head is 4.79 m for a discharge of 1.63 l/sec. Such an operating pressure is well below the operating head required for sprinkler systems (10 to 30 m). Keeping the discharge

Spray Irrigation

constant and replacing the spraying head with a 10 m or a 20 m head rotary sprinkler we find fuel consumptions of 0.63 l/hour at 3010 rpm and 0.9 l/hour at 3810 rpm, respectively9. In other words, fuel consumption of a 10 m or a 20 m head rotary sprinkler is 26% to 80% higher compared to a standard watering can spraying head that was not optimized (see the section on optimization below).
Table 2 Engine data
Manufacturer Robin Max. continuous power Max. continuous speed Minimum speed Specific fuel consumption Fuel Engine characteristics rpm 1500 2000 2500 3000 Model 2.0 kW (2.7 hp) 3600 tours/min 1500 rpm g/kWh 340 petrol (1 l = 760 g) Engine load (%) 100 80 60 40 EY15

P (kW) 0.83 1.11 1.39 1.67

Fuel consumption (%)

100 112 130 200

Irrigation efficiency is the product of application, distribution and storage efficiency. Storage is not normally considered when groundwater is used. Distribution efficiency is 100% because impervious plastic suction and delivery hoses are used. As to application efficiency, watering with a hose may not provide good uniformity or control of the amount of water applied (ASA, 1987). On the other hand, much depends on the operator holding the hose and the cumulative effect of successive irrigations may well improve uniformity, similar to the accumulative effect for a handmove sprinkler lateral reported by Pair (1968. In ASAE, 1983). In Bamako, spray irrigators usually apply water during the last hours before sunset. The walking paths between the beds are not irrigated, which represents a considerable saving compared with sprinkler irrigation (2001,

Use is made of PUMPSELECT.xls, a pump selection programme that attempts to simulate variable pump efficiency, engine efficiency and friction losses at variable speed.

Spray Irrigation

personal observation). A typical application for hand watering is 1 to 2 cm of water (ASA, 1987). In May-June, at the end of the dry season, the depth of water in the wells drops to 0.5-1.0 m. Many of the vegetable beds are not in use, because the ground water supply from 10 non-reinforced wells per hectare is inadequate. The following preliminary calculation could serve as a basis for further investigations. To keep things simple, it is hypothesized that: (1) the depth of water in the well is 0.8 m; (2) well diameter is 1.5 m; (3) there is no recharge during pumping; (4) the average irrigation depth supplied per day is equal to evapotranspiration (about 8 mm/day); (5) the area covered by vegetable beds (i.e. the net irrigated area ) is 60% of the total area10; and (6) application efficiency is 100%. This suggests that of the 600 m2 of vegetable beds only 177 m2 or close to 30% are in use if irrigation is carried out only once a day. Supposed that full recharge takes place in about 12 hours, double that area or close to 60% of the beds available can be irrigated if water is applied twice a day. In September, at the beginning of the dry season, the situation differs mainly with regard to the depth of water in the well. It is supposed here that during the rainy season it increases from 0.8 m to 2 m. As the wells are conical (i.e. wider at the surface than at the bottom), this means that the amount of water available is increased by a factor 3. This suggests that pumping once a day could provide enough water to irrigate 90% of the beds and pumping twice a day could supply more than enough water to irrigate all the beds. Maurice (2002) found that the average daily irrigation period on 0.5 ha is 4 hours. Supposed that the average discharge is 1.63 l/sec (see above), this corresponds to 23.47 m3/day or about 2934 m2, which is a bit more than 58% of 0.5 ha. This is close enough to the hypothesized cropping intensities to lend credibility to the above calculation. Provided the results of above calculations are a close match to reality, it can be concluded that the spray system has high irrigation efficiency that is probably comparable to that of drip irrigation. It is not excluded that other advantages of trickle irrigation mentioned by the ASAE (1983),

See also Tindall (1975) where a typical raised beds is described as having a width of 4 ft, any convenient length up to 30 ft, with pathways of 12 to 18 inch. This works out at 60-75% net cropping area.

10

Spray Irrigation

including enhanced plant growth and yield but with the exception of weed growth reduction, apply to spray irrigation.

COST
The 2.7 hp (continuous power at 3600 rpm) EY15/SHE-50X pumpset was bought new in 1999 at 180,000 CFA. It is used 4 hours/day and 11 months/year. This corresponds to 1340 hours/year. In 2001, petrol cost 425 CFA/l. Layflat hose of 40 mm costs about 3500 CFA/m. Suction tube of 50 mm diameter costs about twice that amount. The watering can head may cost about 2500 CFA. A complete system with 6 m of suction tube, 30 m of discharge hose and a spraying head will cost 180,000 + 147,000 + 2500 = 329500 CFA. With an estimated lifespan of 4 years11, the annual cost can be calculated as follows:
Table 3: irrigation cost (in CFA francs)
Depreciation Interest (12%) Maintenance & repair (10%) Fuel Labour (100 CFA/person/hour) Total (7863 m3/year) CFA/m3 * see section below on optimization Current system 82375 39540 32950 284750 134000 573615 73 Optimized system* 50000 18000 15000 142375 134000 359375 46 Treadle system* 13000 15600 13000 0 349500 391100 50

In table 3 we note that fuel costs are almost 50% of all cost, followed by labour at less than half of the fuel costs. This is all the more noteworthy, when we consider that the Robin EY15/SHE-50X had the best fuel efficiency of all petrol pumpsets tested by Abric et al. (ANPIP/PPIP, 2000), while pump efficiency is fairly low at 30% and engine efficiency an estimated 11.7% only.

11

For the time being there are no data for the average lifespan of small petrol-powered equipment, such as pumpsets.

Spray Irrigation

ADVANTAGES AND DISADVANTAGES

The main advantages of spray irrigation are: (1) its relatively low investment cost per unit area12; (2) widespread availability of equipment; (3) well adapted to situations with multiple low-yielding wells, requiring the equipment to be shifted every 15 to 30 minutes; (4) high application efficiency; (5) probably enhanced plant growth and yield; (6) negligible problems of rodent damage or theft; (7) no emitters, so no emitter clogging; and (7) African farmers quickly take up the technology. Possibly, the main disadvantage is its labour requirement, but this will require further study for the case of multiple low-yielding wells. Compared to a situation where water is lifted and distributed by hand, spray irrigation represents an enormous saving in labour requirement. Assuming an average discharge of 0.33 l/sec, the irrigation cost of completely hand-operated watering can irrigation can be estimated at 83 CFA/m3, which is 14% higher than the current of spray irrigation and 80% higher than the irrigation cost of optimized spray irrigation.

OPTIMIZATION
There are various options for optimization: (1) reducing the price of equipment, including that of accessories, such as layflat hose; (2) improving spraying head efficiency; (3) improving pump efficiency; (4) improving engine efficiency; and (5) replacing the petrol pump with a treadle pump. Ad 1.1: small pumpsets with 2-stroke engines are sold at very low cost in China. Bulk cost is about 45 US$/unit. It is expected (ANPIP/PPIP, 2000) that they can be sold to individual farmers at 80,000 CFA/unit in West Africa. Ad 1.2: the production cost of low-pressure layflat hose is perhaps about 5 US$/kg. With a wall thickness of 1 mm and a diameter of 40 or 50 mm, the ex-factory cost in China is probably about 0.60 US$ or 400 CFA/m. Ad 2.1: using Torricellis theorem it can be shown that larger nozzles of 1.75 mm would reduce head loss from almost 5 m to 1.25 m.
12

For crops with high plant densities requiring large amounts of pipe per land unit, drip irrigation may not be economical.

Spray Irrigation

10

Alternatively the number of nozzles could be doubled to reach the same result. Ad 3 & 4: pump and engine efficiency have to be improved simultaneously by selecting the most appropriate equipment for a particular job. Using PumpSelect it can be shown that with an improved spraying head a discharge of 1.63 l/sec can be achieved at a speed of 2080 rpm. Pump efficiency is improved from 30 to 35% and fuel consumption is decreased to 0.37 l/hour, although engine efficiency is further decreased to about 8%. If engine efficiency could be improved by a factor 2, fuel consumption could become a third of actual field values. The combined optimization potential is estimated at 27 CFA/m3. This suggests that the current irrigation costs are 59% higher than the level obtainable after optimisation (see table 3). Ad 5: another possible approach for optimization would be the replacement of the petrol pump by a treadle pump. A treadle pump with a total dynamic head of 6 m, can supply 4,5 m3/h13 . A bit over twice as much labour is required than with a petrol pump. This is almost compensated by the absence of a fuel bill and lower other costs (see table 3).

CONCLUSION AND RECOMMENDATIONS

In countries with low labour costs, spray irrigation is an interesting alternative for sprinkler irrigation, trickle irrigation and various types of manual irrigation. Spray irrigation is energy efficient, water efficient, adapts well to the yield limitations of low-cost, hand-dug wells, is readily available, and poses no problems in terms of technology transfer, since farmers in at least 3 West African countries adopted the system. The question is not whether spray irrigation is interesting, but rather whether its performance can be improved and how it can be compared with more classical system, such as sprinkler irrigation, to allow a rational choice between the various systems. This requires knowledge of system parameters that is not available at present. It is therefore recommended to: measure the net cultivated area among vegetable growers in Bamako;
13

See: http://www.geocities.com/anpipniger/CompC.htm#L'Exhaure

Spray Irrigation

11

measure the uniformity and depth of application by setting small cans in a square grid in the vegetable beds; compare the depth of application with calculated evapotranspiration; measure the pressure-discharge relationship in spraying heads; if the typical pressure loss in spraying heads is about 4 m, ask farmers to try more efficient spraying heads with pressure losses of 1 to 2 m (see above); measure engine speed-fuel consumption relationships for typical user conditions; describe the well hydrology and measure well yields in different seasons; describe the cropping system over a full year in relation to dropping and rising groundwater levels in order to better understand how vegetable growers cope with water shortage at the end of the dry season; scout in countries, such as China, for affordable, low-lift pumpsets that are capable of a discharge of 1,5 to 2 l/sec at a total dynamic head of 5-8 m while maintaining reasonable engine and pump efficiency; examine the scope for using electric pumps, e.g. 220 V, 300-400 W submersible dewatering pumps; and check whether any progress has been made with improving the fuel efficiency of the Shandong SPP50-14 (PPIP/ANPIP, 2000).

ACKNOWLEDGEMENTS
Special thanks are due to Laurent Maurice and Andr Lgier of Sahel Consult, Bamako (Mali) and to Frdric Louvet of the AFVP in Niamey (Niger).

REFERENCES
ANPIP/PPIP (2000) Banc dessais motopompes. Auteurs: Stphan Abric, Gert Jan Bom, Jean Guillaume Lacas, and Halirou Oumarou. ASA (1987) Irrigation of agricultural lands ASAE (1983) Design and operation of farm irrigation systems

Spray Irrigation

12

ICRA (2000). Intensification durable de l'agriculture urbaine et priurbaine Lom - Togo: cas du marachage. Srie Documents de Travail no. 91. Auteurs: Komlan Dossa, Moussa Guira, Bernadin Loko, Boureima Traor et Jacintha Vigelandzoon. Kay, M. (1983) Sprinkler irrigation: equipment and practice Maurice, L. (2002) Private communication Pair, C.H. (1968) Water distribution under sprinkler irrigation. In: Transactions of the ASAE 11(5):648-651. Mott, R.L. (2000) Applied fluid mechanics Tindall, H.D. (1975) Commercial vegetable growing Van t Hof, S. (2000) Pompage basse pression pour le dveloppement de la petite irrigation : Slection de matriels adapts Van t Hof, S. (2002) Explanatory note on PUMPSELECT.exe (underway) Van t Hof, S. (2001a) Roving course on pump selection: Burkina Faso and Niger Van t Hof, S. (2001b) Roving course on pump selection: Mali.

Spray Irrigation

13

APPENDIX: Spray Irrigation in Bamako

In the Bamako-type of spray irrigation, the length of the lay-flat hose is 2550 m and its diameter is 40 or 50 mm. The spraying heads are locally made. So far, no measurements have been made of their dimension, number and size of holes, or discharge/pressure relationship. In a 0.5 ha vegetable garden, there may be 5 wells. A portable, self-aspiring pumpset is carried from one well to the other to empty them one by one. With a discharge of 1-2 l/sec, it takes 15 to 30 minutes to empty each well, a bit less during the dry season when the groundwater level reaches a seasonal low with a static head of 2.5 m below the surface. The static water level in the well is between 0.5 and 2 m, depending on the season. Typically, the well diameter is 2 m at the surface and 1.5 m at the bottom. Fuel consumption is in the order of 0.5 l/hour.

Figs 1, 2: Considering the small storage in the well, water efficiency is of paramount importance

Spray Irrigation

14

Figs 3,4: Priming & spraying