Physics and Chemistry of the Earth xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Physics and Chemistry of the Earth

journal homepage: www.elsevier .com/locate /pce

Productivity of irrigation technologies in the White Volta basin

E.A. Ofosu a,b,c,*, P. van der Zaag a,b, N.C. van de Giesen b, S.N. Odai c

a Department of Management and Institutions, UNESCO-IHE, Delft, The Netherlandsb Civil Engineering & Geosciences, TU Delft, The Netherlandsc Department of Civil Engineering, KNUST, Kumasi, Ghana

a r t i c l e i n f o

Article history:Available online xxxx

Keywords:Smallholder irrigationTomato cultivationWater productivityWatershed managementEndogenous development

1474-7065/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.pce.2010.07.005

* Corresponding author at: Department of ManageSCO-IHE, Delft, The Netherlands. Tel.: +31 644168708

E-mail address: ericad200@yahoo.com (E.A. Ofosu

Please cite this article in press as: Ofosu, E.Adoi:10.1016/j.pce.2010.07.005

a b s t r a c t

Parts of the White Volta basin in northern Ghana and southern Burkina Faso have witnessed a spectacularrise of irrigated agriculture since about 2000, largely without government support, and seems to havebeen triggered by a strong and growing demand for vegetables, notably tomatoes in the urban centresof southern Ghana. It is interesting to note the variety of different irrigation technologies that individualand groups of smallholder farmers adopted, adapted and implemented. Some technologies are well-known, such as those associated with conventional sources of water like small and large reservoirs; oth-ers have been rarely described in literature, such as temporal shallow wells and alluvial dugouts.

This paper describes and characterises these different irrigation technologies and conducts a compar-ative analysis of their productivities, in terms of crop yield, water use and financial returns. The study wasconducted in three neighbouring and transboundary watersheds (Anayari, Atankwidi and Yarigatanga)located in the Upper East Region of Ghana and southern Burkina Faso. For the study, 90 tomato farmerswith different irrigation technologies were surveyed during one crop season (2007/2008).

The results show that adequate fertilizer application is the major contributor to irrigation productivity.Technologies characterised by relatively small farm sizes are better managed by the surveyed farmersbecause they are able to provide adequate water and crop nutrients thus resulting in higher productivity,and high profit margins.

Apart from technologies that depend on reservoirs, all other technologies surveyed in the paper arefarmer driven and required no government support. This ongoing type of endogenous irrigation develop-ment provides a strong backing that the way forward in sub-Saharan Africa is for governments to createpolicies that facilitate poor farmers becoming irrigation entrepreneurs. Such policies should aim toenhance the reliability of markets (both input and output) as the driving force, and facilitate people’saccess to land and water.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Irrigation has ensured significant increase in global food supplyand raised millions out of poverty (Faurès et al., 2007). For manydeveloping countries investment in irrigation will continue to rep-resent a substantial share of investment in agriculture, but the pat-tern of investment will focus more on enhancing productivity ofexisting systems through upgrading of infrastructure and reform-ing management (Faurès et al., 2007). Studies have shown thatan increase in irrigation productivity which results in improvedfarm income creates an increase in demand for local non-tradablegoods and services, which offer labour opportunities to the poorestsegments of the rural population, promotes local agro-enterprisesand stimulates the agricultural sector as a whole (Lipton et al.,

ll rights reserved.

ment and Institutions, UNE-; fax: +31 152122921.

).

., et al. Productivity of irrigati

2003; Smith, 2004; Hussain and Hanijra, 2004). In most parts ofsub-Saharan Africa, increasing agricultural productivity is oftenthe only way out of poverty, and new irrigation development canbe a springboard for economic development (Faurès et al., 2007).In the face of many disappointing experiences with irrigationdevelopment in sub-Saharan Africa, investing in irrigation requiresempirical evidence of the potential productivities of the prevailingirrigation technologies, and a comparative analysis to inform irri-gation policy.

At this stage it is important to define the term irrigation tech-nology as employed in this study. Generally, an irrigation technol-ogy diverts water from a source, conveys it to cropped areas of thefarm and distributes it over the area being irrigated (James, 1993).Following the above definition, in this study we define irrigationtechnology as an ‘‘ensemble” (specific combination) of differentmethods and techniques for diverting and/or pumping, storing,transporting and distributing (ground, surface and rain) water toagricultural crops. Since the technology is a combination of

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

2 E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx

techniques, in order to simplify their naming, we denote the differ-ent irrigation technologies in this study by the source of water theyabstract from.

So far little is known on the productivity of irrigation technolo-gies adopted by smallholder farmers and the factors influencingproductivity in the Volta basin. Few research activities have beenconducted on the productivity of small and medium reservoirs inthe White Volta sub-basin. Faulkner (2008) looked at the perfor-mance of two small reservoirs (Tanga and Weega) in the UpperEast Region of Ghana and found that because of the high wateravailability in the Tanga system water management was relaxed,which resulted in inefficient irrigation water-use methods. By vir-tue of this the Tanga system was less profitable in water and land.Faulkner et al. (2008) argues that the increased profit per unit ofcultivated land was not due to water stress or irrigation techniqueor management but rather due to differences in fertilizer, pesticideand seed inputs. Mdemu (2008) found that the contribution of irri-gation to tomato production in the Upper East Region of Ghana washigh, that irrigation plots were over-irrigated by 11–70%, and thatwater use efficiencies were higher under small reservoir irrigationthan in systems with large reservoirs. Both studies only consideredirrigation technologies that sourced water from reservoirs. Mean-while there are other prevailing irrigation technologies with differ-ent water sources in the White Volta sub-basin which also need tobe investigated. Also the different aspects of the technologies suchas water source, water consumption, farm size and fertilizer inputon the productivity of technologies, essential information forimproving irrigation productivity, were not analysed. Finally, ageneral comparison of all prevailing irrigation technologies, usingproductivity factors is necessary to assess the strength and weak-nesses of the various irrigation technologies to help guide irriga-tion development policies.

A general overview of the trend of irrigation development inthe White Volta sub-basin shows that the construction of smallreservoirs started in the 1950s and 1960s followed by the con-struction of large reservoirs and associated irrigation schemesin the 1970s and 1980s. Such irrigation developments have beenpopular with many development agencies, and have been de-scribed in the scientific literature (Faulkner et al., 2008). How-ever, the prospects of vegetable production has triggered the

Fig. 1. The location of study area

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

initiation and upscaling of several alternative irrigation technolo-gies since the early 1990s. These technologies have expandedover the past two decades significantly due to increasing marketfor the products in the urban centres in southern Ghana, but theprecise rate and pattern of expansion is not yet precisely knownand subject to detailed study.

The focus of the study therefore is to assess the productivities ofthe prevailing irrigation technologies and the contributing factorsto the difference in productivities using common parameters. Thisdemands the selection of a common irrigated crop practised underall the prevailing irrigation technologies which happens to betomatoes.

2. Study area

The White Volta has a total catchment area of 105,000 km2. Thecatchment area in Ghana is 49,200 km2 with the rest of areas inTogo and Burkina Faso. The climate is semi-arid with annual rain-fall on the basin ranging from 1010 mm/a in the north to1140 mm/a in the south; pan evaporation is in the order of2540 mm/a; and runoff from within the basin averages about96 mm/a (Wiafe, 1997). The predominant soils are Plinthic ferral-sols (groundwater laterites), Eutric nitosols (savannah ochrosols)with their intergrades (Brammer, 1962; Adu, 1995). Degradedsavannah occurs in uncultivated areas with patches of reservedsavannah woodlands (Agyepong, 1997). The predominant landuse is arable agriculture and widespread grazing of large numbersof cattle and other livestock (up to 100 cattle/km2; FAO, 1991).Extensive valley bottoms in many parts of the basin, particularlyin the guinea savannah areas, have in recent years been cultivatedfor rice under rain-fed conditions (Agyepong, 1997).

Irrigation farming in the study area is mainly practised duringthe dry season (October–April). In addition to the governmentdeveloped irrigation systems which are mainly the small reservoirand large reservoir irrigation there are other irrigation technolo-gies developed by the farmers and groups of farmers scatteredacross the basin. These are located in areas of rich alluvial depositsusually found along streams or rivers and in flood plains. Crops irri-gated include rice, tomatoes, pepper, onions, cabbage and lettuce.However tomato irrigation is the most extensive and it is practised

10 km

in Ghana and Burkina Faso.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx 3

under all types of irrigation technologies. Irrigation of tomatoeshas been the main contributor to the upscaling of irrigation devel-opment in the basin within the past two decades.

The study was conducted in three neighbouring watersheds,Anayari (464 km2), Atankwidi (275 km2) and Yarigatanga(352 km2) all located in the White Volta sub-basin (Fig. 1). All pre-vailing irrigation technologies are identified in these three water-sheds. It should however be noted that not a single of thewatersheds has all the prevailing irrigation technologies practisedin it. These three watersheds are transboundary with the majordownstream parts located in the Upper East Region of Ghana andthe remaining upstream parts in southern Burkina Faso (Fig. 1).

Apart from the fact that the watersheds are located in both Gha-na and Burkina Faso, it is interesting to note that the developmentof these irrigation technologies is very similar in both countriesand influences each other since they all produce for the same mar-ket. This makes it relevant to compare the situation in both coun-tries also for further policy analysis.

3. Methodology and data collection

The study involved a ground survey that identified all irrigationtechnologies practised in the watersheds. On the basis of the sur-vey a target number of farmers were allotted for various irrigationtechnologies within each watershed for data collection. Withineach watershed a number of irrigation sites were selected for eachirrigation technology. Knowing the target number for each irriga-tion site a random sampling was employed in selecting farmers.However, the leaders (i.e. executive members of WUA at small res-ervoirs and senior farmers at shallow wells) at the irrigation siteswere instrumental in convincing fellow farmers to assist in the re-search. This resulted in site leaders influencing the selection offarmers.

Not all selected farmers could read and write, yet both literateand illiterate farmers were selected. In some instances only onefarmer was educated amongst a group of farmers within the samelocation. Literate farmers and educated children of illiterate farm-ers were trained to enter the records of the daily activities andwere given notebooks. In other situations, literate farmers wereasked to assist their fellow illiterate farmers in recording their dai-ly activities. In view of these difficulties not all data were requiredfrom illiterate farmers such as pumping duration, water level mea-surements and other field measurements. In such instances the re-cords provided by the literate farmers were used to represent allfarmers. The records made by farmers were inspected every fort-night. Field assistants also visited farmers on a weekly basis to as-sist them in measurements and the recordings. Some of themeasurements such as irrigation water use were taken at thebeginning, the middle and the end stages of the irrigation season.

Table 1Number tomato farmers surveyed and analysed for the 2007/2008 season.

Area Technology (water source) C

A

Upper East Region, Ghana Small reservoirs 1Permanent shallow wellsLarge reservoirsRiverine waterTemporal shallow wellsRiverine alluvial dugouts

Southern Burkina Faso Small reservoirsPermanent shallow wells

Total 2

Associations (WUAs).

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

A preliminary survey was conducted in the 2006/2007 irriga-tion season on a total of 85 tomato growing farmers selected fromthe Ghana section of the study area but only 51 of the recordscould be used as a result of missing records and unreliable infor-mation. With the experience obtained from the 2006/2007 season,the number of farmers was increased to 135 in the 2007/2008 sea-son maintaining most of the farmers used in the previous seasonand also farmers were selected from the Burkina Faso section ofthe study area. At the end of the season 114 of the farmers note-books could be retrieved. The reason is that some travelled withtheir notebooks at the end of the season, some also lost their note-books due to rain and fire while others left their farms in the mid-dle of the season due to several reasons thus stopped recording andtook the books away. Out of the 114 farm notebooks, 90 containedcomplete data sets and were found useful, while the remaining hadinformation gaps. The records of the 90 farmers of the 2007/2008irrigation season were finally used for the analysis (Table 1).

The data collected were diverse but common for all irrigationtechnologies. It included primary and secondary data/informationfrom the farmers, farmer organisations and related institutionssuch as Ministry of Food and Agriculture (MOFA) and IrrigationDevelopment Authority (IDA). The methodology employed includefield measurements, daily recording of activities (water consump-tion, time input, farm inputs, operating costs and sales), field inter-views (farmers, farmer organisations and related institutions), fieldcounts, irrigation water-use measurements, yield measurementsand market information.

4. Field results

This section analyses the data of tomato farmers listed in Table 1for the 2007/2008 irrigation season. The investments in water andwater infrastructure for various technologies were analysedincluding the seasonal expenditure on water per each technology.Also seasonal application of water, fertilizer and pesticides wereestimated for the various irrigation technologies. The productivi-ties (land and water) of the irrigation technologies were compared.A regression analysis was used to generate a relationship betweencrop yield and fertilizer input, water source and water applicationmethod. The final stage of the analysis looked at the socio-eco-nomic impact of the irrigation technologies in terms of farmerprofit, labour creation and gender involvement. This section startswith describing the various irrigation technologies, distinguishingtheir various sources of water.

The data analysis was aggregated for irrigation technologies andcountry, meaning that farmers under one technology on one side ofthe border were assessed together irrespective of the watershedthey were located in. All data are analysed based on the farm sizeof the farmer and results have been normalised for a 1 ha farm size.

atchment area

nayari Atankwidi Yarigatanga Total

5 0 20 357 0 0 70 0 6 60 0 8 80 8 0 80 16 0 16

0 4 0 40 4 2 6

2 32 36 90

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

4 E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx

4.1. Description of irrigation technologies

Small reservoirs are characterised by embankments connected atone end with an earth/concrete channel spillway. In addition, smallreservoir irrigation schemes have developed irrigable fields down-stream of the impoundment, furnished with water abstracting andtransporting facilities. Water abstraction from the reservoir andtransportation to the developed irrigation field is by means of grav-ity through tunnels, pipes or open channels fitted with controlvalves.

Permanent shallow wells irrigation technology is one of the old-est irrigation technologies practised in the sub-basin. The perma-nent shallow wells technology is made up of a permanentlyconstructed well. The well is either lined with concrete, sandcreteor stones to stabilise the walls. Some of the wells may be unlinedbut the owners have plans to line them eventually. Wells havelined walls which are raised above the ground level to prevent run-off and silt from entering the wells. This technology is practised bylandowners who live in areas with a high groundwater table butnot necessarily near a stream.

Large reservoirs in the local context are surface water storagesystems developed for irrigation purposes by the government forthe local people to be managed and operated by public institutions.The International Commission on Large Dams (ICOLD) defines largedams as impoundments of at least 15 m high or storing more than3 million m3 of water (ICID, 2000). There are two large reservoirs inthe study area, namely Vea and Tono, developed in 1980 and 1985respectively.

One of the most reliable water sources which farmers find rel-atively easy and cheap to access is flowing water in perennial riv-ers and also upstream water releases or return flows from largeirrigation schemes (e.g. Vea and Tono). This irrigation technologyis called riverine water irrigation. This technology has been prac-tised in the sub-basin since 1992 around Pwalugu in the UpperEast Region and has expanded over the years. Also due to the reg-ulating effect of dams, small rivers like the Nakambe have becomeperennial streams in the lower reaches in Burkina Faso, evenimproving supplies to towns in northern Ghana (ICOLD mem-bers-Burkina Faso, 2001). This has contributed to the upscaling ofthis irrigation technology. Meanwhile, since the eradication of riverblindness, the use of land along the river banks of the White Voltabasin for irrigated farming has increased considerably (Birner et al.,2005).

The use of temporal shallow wells is one of the new irrigationtechnologies that have emerged in the sub-basin within the pasttwo decades. It is common in the Upper East Region, specificallyin the Atankwidi and Anayari sub-catchments where groundwaterlevels range between 3 and 8 m throughout the dry season. Tem-

Table 2Cost of water-use for irrigation technologies in Upper East Region and Burkina Faso.

Irrigation technology Investmentcost ($/ha)

Lifespan(years)

Depreciationcost – [X] ($/ha/y)

Maintecost – [

Upper east region

Small reservoir 33,000 25 4650 100Permanent shallow well 4800 30 672 400Large reservoir 15,000 50 2029 150Riverine water 1700 a 8 360 50Temporal shallow well 1200 0.5 1200 0Riverine alluvial dugout 700 a 5 200 50

525 b 0.5 525

Burkina FasoSmall reservoir 38,000 25 2500 100Permanent shallow well 6500 30 385 600

a Petrol pump, pipes.b Constructing the dugout.

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

poral shallow wells tend to be used in cases where the irrigatorrents the land from the landowner for just the dry season. As a con-dition for obtaining the land for irrigation the next season, the irri-gator agrees with the landowner to refill all the wells at the end ofthe season with sand for reasons of preventing farm animals andpeople from falling into them and also to save their farming areafrom being reduced. This explains why the wells are temporal.

Riverine alluvial dugouts are developed mainly in non-peren-nial streams in the sub-basin. Riverine alluvial dugouts are locatedin the river channel and fed with underground water. From inter-views conducted, this irrigation technology has been practisedsince 1995 in the Atankwidi and Anayari sub-catchments of theWhite Volta. Riverine alluvial dugout irrigation schemes havenon-permanent irregularly shaped depressions of various sizesand depths (3–5 m) located in the river channel. These alluvialdugouts are spaced at least 5 m apart and dotted along the riverchannel. All these dugouts are temporal structures which last onlyone dry season. They are filled up with silt during the rainy seasonsuch that by the onset of the next dry season all dugouts arecovered.

4.2. Investments in water and seasonal expenditure on water

Investment in infrastructure and equipment for irrigation iseither by government, development agency or private individualor group of individuals. In the White Volta sub-basin governmentsand development agencies invest in the development of large res-ervoir irrigation schemes and small reservoirs, while other tech-nologies in the sub-basin are developed by private individuals orfarmer groups.

Farmers irrigating from the large reservoir scheme are chargedUS$56/ha for irrigating a tomato crop, US$42/ha for rice andUS$28/ha for vegetables. This fee has to be paid before the irriga-tion commences. Under the small reservoir irrigation technologieswater levies are charged differently. While at some small reser-voirs water is charged per bed (US$0.2/bed size of 1 m � 50 m),at others irrigators are charged per plot (US$1/plot area of0.05 ha), and the rest are charged a general fee irrespective of farmsize cultivated (US$2/farmer/season). These water levies are usedfor the maintenance of canals and organisational activities andare reviewed upwards almost every year by the Water User Asso-ciations (WUAs).

In the remaining technologies, which are riverine water, river-ine alluvial dugouts, temporal shallow wells and permanent shal-low wells, a farmer individually decides when and how toirrigate. Most of these farmers are either landowners or are relatedto the landlord. In some instances the farmers rent the land fromthe landlord for just the dry season. All these farmers do not pay

nanceY] ($/ha/y)

Water-usecost – [Z] ($/ha/y)

Total cost[X + Y + Z] ($/ha/y)

Seasonal expenditure($/ha/y)

47 4797 470 1072 400

57 2236 57550 960 600

0 1200 1200755 1530 1330

80 2680 800 985 600

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx 5

a water levy but rather invest in water abstraction equipments,infrastructure (permanent and temporal) and energy to be ableto irrigate the crops. Table 2 shows the investment cost and sea-sonal expenditure on water (abstraction and distribution) incurredon the irrigation technologies for tomato cultivation.

The infrastructure and equipment used in the irrigation are af-fected by wear and tear or obsolescence over an accounting periodcalled depreciation which reduces their economic value and pro-ductivity annually (Lund, 1989). The annual capital costs of the irri-gation facilities are estimated using a discount rate based on theNational Central Bank discount rate of December 2007 (Ghana –13.5%/yr and Burkina Faso – 4.25%/yr, CIA – The World Factbook,2009). The economic lifetime of small reservoirs is chosen as25 years based on the fact that many of the reservoirs that wereconstructed in the 1950s, 1960s and 1970s in the Upper East Re-gion of Ghana were non-functional by the 1980s and had to berehabilitated in the 1990s by the World Bank and FAO. The eco-nomic lifetime of the large reservoir was chosen as 50 years basedon the design life of the reservoir and that of the remaining tech-nologies were obtained from field interviews.

The annuity factor, a, was determined from the followingequation:

a ¼ i� ð1þ iÞn

ð1þ iÞn � 1ð1Þ

where i is the discount rate and n is the economic life (years).The annual capital cost or depreciation is then calculated by

multiplying the investment cost with the annuity factor a. Thesum of the annual maintenance cost, annual depreciation costand the operating and maintenance cost of the facility gives the to-tal annual expenditure on irrigation water by the technology(Grant et al., 1990).

Table 2 shows that the small and large reservoir are the mostexpensive technologies in terms of investment, operation andmaintenance of irrigation water infrastructure and technologywith the riverine water irrigation being the least (about 20% ofthe small reservoirs). Interestingly, the less expensive technologiesin terms of investment are developed privately while small reser-voirs and large are mostly developed by the government. Thisshows that a development policy aimed at promoting the shallowwell technology is more economical than small and largereservoirs.

Also the seasonal expenditure on water-use by farmers, whichincludes the amount of money spent on water abstraction andmaintenance by the farmers either as water levy, on fuel, pump re-pairs or digging every season, reveals that the riverine alluvial dug-out technology and temporal shallow well are the most expensivetechnologies with the cheapest being the small and large reservoirirrigation (25% of the riverine alluvial dugout) which are mostlydeveloped by government and development agencies. The majorcontribution to the high seasonal expenditure on water for theseirrigation technologies is energy and labour.

Fig. 2a. Seasonal irrigation water applied.

4.3. Plot size and fertilizer use

Within the study area, the measured farm sizes for small reser-voirs range from 0.005 to 0.58 ha, namely 0.005–0.09 ha for per-manent shallow wells, 0.125–1.0 ha for large reservoirs, 0.2–1.0 ha for riverine water, 0.006–0.08 ha for the temporal shallowwells and 0.05–0.7 ha for riverine alluvial dugouts.

The application of chemical fertilizers and agro-chemicals iscommon to all irrigation technologies in the sub-basin. The surveyin the three watersheds showed that 100% of farmers use NPK, 32%use Urea and 59% use Sulphate of Ammonia for their crops. In bothcountries less than 10% of farmers use manure to support their

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

chemical fertilizer because most farmers gather animal droppingsduring the dry season for their rainy season crops. Fertilizer appli-cation is an expensive component of irrigation farming. The cost offertilizer was fairly stable during the 2007/2008 irrigation season;the cost of NPK (15-15-15) was US$0.50/kg, US$1.12/kg for Ureaand US$0.64/kg for Sulphate of Ammonia.

After the fruiting of the tomatoes the farmers spray the toma-toes with other yield inducing chemicals such as harvestmore,superforce, 19-19-19 and growfull. The quantities applied per farmarea however differed from farmer to farmer due to differencesin individual financial capacity and the perception of adequacy.For this reason the nutrient intake by the crops differs from farmerto farmer which translates to yield differences.

The farmers control diseases and pests with agrochemicals,which are applied at different quantities by the farmers and thisis also based on differences in farm sizes financial capacity andexperience.

4.4. Amount of water applied by irrigation technologies

Watering of crops is the main activity of irrigation farmingthroughout the season. Water applied differs for each technologyand depends on the water application method and water availabil-ity. The watering schedules for small and large reservoirs are deter-mined by the WUA and the management of the large irrigationscheme respectively. There are no measures to ensure or regulatethe quantity of water per plot, rather farmers decide on the quan-tity of water they apply on their farms without regulation or tech-nical advice. The WUAs plan their own water scheduling for theirrigation season based on the water level in the reservoir at vari-ous stages of the season. As a result the water schedule differsamong the small reservoirs and can change during the irrigationseason.

Farmers using shallow wells and abstracting water with ropeand bucket water their crops twice a day (morning and evening),every day or thrice a week. The bucket sizes range from 8 to 14 li-tres. Farmers use their own discretion to decide how many bucketsof water is needed per bed, and depends, among others, on the croptype and the dimensions of the bed.

In general farmers try to be efficient in water-use due to thescarcity of water in the dry season. Farmers who abstract waterwith motorized pumps try to be extra efficient in energy consump-tion due to the cost of fuel. Similarly, farmers applying water man-ually are also efficient because they apply water directly the rootsof the crops.

Figs. 2a and 2b show the amount of water applied per irrigationtechnology for tomato production and the corresponding expendi-ture on water, respectively. The cost of water use is not related tothe water consumption but rather to the technology.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

Fig. 2b. Seasonal expenditure on irrigation water.

6 E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx

4.5. Crop productivity

The productivity of the irrigation technologies is measured interms of crop yields. The yields of the crops were measured inthe field by weighing harvested products on the farm. Tomatoesare harvested and sold in boxes/crates, basins or buckets. (Thereare two different sizes of boxes/crates (105 kg and 156 kg of toma-toes), while a basin carries 34 kg and the bucket carries 21 kg oftomatoes.) Farmers were provided with field-books in which theyrecorded harvesting dates; the amount of boxes/bucket/basin har-vested; the type of box; number of labourers employed, labour cost

Fig. 3a. Average tomatoes yield for 2008.

Fig. 3b. Fertilizer input for tomatoes.

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

and selling price. Farmers gave the labourers who helped in theharvesting about 2 kg each of tomatoes at the end of each harvest-ing as part of their wages. These were also factored into the com-putation of the total yield (Fig. 3a). Also the amount (by weight) offertilizer applied was estimated (Fig. 3b), showing that fertilizerapplications tend to be very high: on average more than 1000 kg/ha. Comparing Figs. 3a and 3b suggests a positive correlation be-tween crop yield and the amount of fertilizer applied.

5. Data analysis

This section analyses the factors affecting the yield of all thetechnologies using a simple regression analysis model. It also anal-ysis the water productivity of all the technologies as well as the va-lue of water obtained. The final analysis is on the socio-economicimpact of the various irrigation technologies.

5.1. Analysis of crop yield of irrigation technologies

A regression analysis was performed of all 90 tomato farmersincluded in the analysis to establish the relationship between cropyield and other input parameters, such as fertilizer, water con-sumption, water source, and water application method, character-istic of the different irrigation technologies in the sample. Thefollowing mathematical equation was established:

Y ¼ �2:2þ 201:4� A f þ 17:7� D grwþ 17:6� D man

r2 ¼ 0:76ð2Þ

where Y is the crop yield (kg/m2), A_f is amount of fertilizer applied(kg/m2), D_grw is dummy variable for if the source of water isgroundwater (D = 1) or surface water (D = 0), D_man is the dummyvariable for if water is applied manually (D = 1) or mechanically(D = 0).

Further statistical analysis on the results showed that the esti-mates of the parameters are significant at 5% level. Fig. 4 showsthe relationship between observed and simulated crop yields ofall 90 farmers included in the analysis.

Eq. (2) shows that the amount of fertilizer applied to the crophas a significant impact on crop yield: 1 kg of chemical fertilizerapplied results in a yield increase of 200 kg tomatoes. This figureis higher compared to the O/N (output/nutrient) ratio in West Afri-ca for other crops such as maize (range of 0–54) and groundnuts(range of 4–21) (Yanggen et al., 1998), implying a high yield re-sponse to fertilizer for tomatoes in the study area. The value/costratio (VCR), which is a rudimentary indicator of potential profit-ability, is also measured based on the yield response to fertilizer,the cost of fertilizer (see Section 4.3) and the sales value of toma-toes (see Fig. 6b). The VCR for applying fertilizer to the tomato cropranges from 5 to 120 in the study area. According to Kelly (2005)the rule-of-thumb for VRCs is that they must be at least two beforea farmer will consider fertilizer use, while in high-risk productionenvironments the minimum VCR for adoption may be 3 or 4. Thisimplies that fertilizer use in tomato production has a high profit-ability in the study area.

Eq. (2) indicates that irrigation technologies that rely ongroundwater achieve significantly higher crop yields than thoserelying on surface water. We have no satisfactory explanation ofthis finding, but it may be associated with factors that are associ-ated with irrigation technologies that abstract groundwater ratherthan the groundwater itself. Eq. (2) also indicates that manualabstraction technologies are more productive than motorisedabstraction (pumps). This is perhaps because manual applicationof water by bucket is more precise and closer to the roots, whichenhances crop production.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

y = 0.89xR2 = 0.72

0

20

40

60

80

100

120

0 20 40 60 80 100 120

observed crop yield (kg/m2)

sim

ulat

ed c

rop

yiel

d (k

g/m

2 )

Fig. 4. Performance of a regression analysis of crop yield of irrigation technologies.

E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx 7

5.2. Water productivity

Water productivity is a measure of performance generally de-fined as the physical quantity or economic value generated fromthe use of a given quantity of water (Molden et al., 2003). Increas-ing water productivity to obtain higher output or value for eachdrop of water use is key to the efficient use of water in the sub-ba-sin and therefore a very important factor in the comparative anal-ysis of irrigation technologies. The productivity of the variousirrigation technologies in the sub-basin is an important variablefor up-scaling irrigation technologies: the higher the productivityof a technology, the more efficient the scarce water will be used.

Water productivity equals the crop yield divided by the volumeof water applied. The average water productivity can then be com-puted for each technology (Figs. 5a and 5b).

Fig. 5a. Water productivity for tomatoes in 2008.

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

Fig. 5a shows that the temporal and permanent shallow wells(producing 60 kg of tomatoes per m3 of water applied), and river-ine alluvial dugouts (45 kg/m3) are the most water productivetechnologies in the Upper East Region of Ghana. However the samecannot be said of permanent shallow wells in Southern BurkinaFaso. Water productivity of permanent shallow wells and smallreservoirs in Southern Burkina Faso are about 50% and 30% of thatin the Upper East Region. The large reservoir irrigation scheme isthe technology with the lowest water productivity (9 kg/m3). Thisconfirms the results of Mdemu (2008) who found that the waterproductivity of a small reservoir (Dorongo dam) was higher thanthat of a large reservoir scheme (Tono) in the Upper East Regionof Ghana. The observed water productivity for the shallow wellsare higher than most published figures and are due to the high fer-tilizer application on these fields. For example in Molden et al.(2007) the fertilizer application ranges from 9 kg/ha in sub-Saha-ran Africa to 135 kg/ha in South-East Asia resulting in a water pro-ductivity for tomatoes ranging from 5 to 20 kg/m3.

Fig. 5b. Value of water for tomato production in 2008.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

Fig. 6a. Farm-gate prices for tomatoes in 2007/2008.

Fig. 6b. Trend of farm-gate prices for 2007/2008 tomatoes in 2007/2008.

Fig. 7b. Farmer profit for tomato production.

8 E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx

The value of water obtained under tomato cultivation by theseirrigation technologies is estimated using the average market priceof tomatoes per each technology and their corresponding waterproductivities. Figs. 6a and 6b give the recorded on-farm tomatoprices for various irrigation technologies and the average tomatomarket prices for 2007 and 2008. These figures show that the mar-ket prices are not the same for all irrigation technologies and thatthey fluctuate throughout the harvesting season.

5.3. Economic impact of irrigation technologies

Due to the semi-arid nature of the sub-basin which is character-ised by short rainy seasons coupled with the fact that over 85% of

Fig. 7a. Gross profit margin for tomato production.

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

the inhabitants are subsistence farmers, there are little employ-ment opportunities during the dry season. This accounts for thehigh rural–urban migration during the dry season to do menialjobs in the urban centres. Creating job opportunities for the inhab-itants during the dry season improves local livelihoods and thus re-duces poverty. The additional income made by the farmer andlabourer through irrigation farming goes to support their families(in education, feeding, shelter and health), thereby reducing pov-erty in society in general.

The profit margin is computed as a ratio between the farmers’profit and total cash revenue from sales and converted into a per-centage. The profit is the difference between the income madefrom the sales and expenditure incurred in the crop production,which excludes depreciation costs of irrigation assets and infra-structure, as well as the labour costs of the farmers themselves.The profit made in the season represents the direct income of thefarmer at the end of the season.

Fig. 8a. Cost of hired labour in production.

Fig. 8b. Participation of women farmers.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

Table 3Comparison of irrigation technologies using productivity factors.

Parameter Small reservoir Permanent shallow wells Large reservoir scheme Riverine water Temporal shallow wells Riverine alluvial dugouts

Developmentandmanagement

Very expensive;developed bygovernment and NGOs,managed by thecommunity

Manageable cost, developedandmanaged by individuals andfarmer groups

Very expensive; developed bygovernmentand managed by public institution

Manageable cost, developed andmanagedby individuals and farmer groups

Manageable but expensive dueto shortlife-span of wells; developedandmanaged by individuals

Expensive due to short life-span ofdugouts,labour and energy cost; developedandmanaged by individuals

Plot size Has relatively small farmsizeswhich are manageableby thepoor farmer

Has relatively small farmsizes whichare manageable by the poorfarmer

Has relativelylarge farm sizes whichrequires more financial resource for theuser

Has relatively large farm sizeswhichrequires more financial resourceforthe user

Has relatively small farm sizeswhich are manageable by thepoor farmer

Has relatively small farm sizeswhichare manageable by the poor farmer

Fertilizer andchemicalinput

Adequate amounts areapplieddue to relatively smallfarmplots

Adequate amounts areapplieddue to relatively small farmplots

Poor farmers are not able to provideadequate amount as a result of largefarm plots

Poor farmers are not able toprovideadequate amount as a result oflargefarm plots

Adequate amounts are applieddue to relatively small farmplots

Adequate amounts are applied duetorelatively small farm plots

Water use The cost of water-use isthecheapest. Water – use iscontrolled and notflexible

The cost of infrastructure ismanageable. Water-use isuncontrolled and flexible

The cost of water-use is the cheapest.Water – use is controlled and notflexible

The cost of infrastructure ismanageable.Water-use is uncontrolled andflexible

The cost of infrastructure ismanageable. Water-use isuncontrolled and flexible

Water abstraction cost isexpensive.Water-use is uncontrolled andflexible

Crop yield Has good crop yield Has very high crop yield Has the least crop yield Has low crop yield Has very high crop yield Has high crop yieldWater

productivityHas good waterproductivityand good value of water

Has the highest waterproductivity,but value for water is reducedbypoor market for products

Has the least water productivity as wellas low value for water

Has low water productivitybut good value for water dueto good market for products

Has very high waterproductivity andgood value for water

Has good water productivity andgoodvalue of water

Povertyreduction

Creates moreemployment perunit area for farmersespeciallywomen; profit of aboutUS$400/farmer/season

Creates more employmentperunit area and has profit ofaboutUS$220/farmer/season

Low profits are recorded especially bypoor farmers. Provides income to societythroughlabour generation; profit of US$450/farmer/season

Low profits are recorded especiallyby poorfarmers. Provides income tosocietythrough labour; profit of aboutUS$1050/farmer/season

Creates more employment perunit area forfarmers (both gender); profit ofabout US$420/farmer/season

Good profits, provides a lot ofemployment for farmers andlabour forthe youth. Low womenparticipation;profit of about US$600/farmer/season

E.A.O

fosuet

al./Physicsand

Chemistry

ofthe

Earthxxx

(2010)xxx–

xxx9

Pleasecite

thisarticle

inpress

as:O

fosu,E.A

.,et

al.Productivity

ofirrigation

technologiesin

theW

hiteV

oltabasin.

J.Phys.

Chem.

Earth(2010),

doi:10.1016/j.pce.2010.07.005

10 E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx

Figs. 7a and 7b show that the large reservoir irrigation farmerhas the lowest profit margin amongst all the irrigation technolo-gies. The temporal and permanent shallow wells have the highestprofit margins making the two technologies the most economicallyviable irrigation technologies. Interestingly, despite the relativelylarge farm plots of the large reservoir irrigation scheme, the profitmade per farmer is almost equal to that of the temporal shallowwell farmers. The average profits for the season were for small res-ervoirs US$420/farmer, permanent shallow wells US$225, largereservoir irrigation US$470, riverine water US$1050, temporalshallow wells US$420/farmer and riverine alluvial dugoutsUS$620/farmer. Thus unless the productivity of the large irrigationscheme is improved the users are not better off than those usingmore expensive technologies on comparatively smaller plots.

5.4. Social impact-employment/labour and women participation

The operational costs spent on labour hired from the inhabit-ants of the sub-basin during the irrigation season helps reducepoverty in the sub-basin. The irrigators employ basically the youthand women for their labour work. Typically, the youth are hired asfarm assistants by farmers using riverine water, large irrigationschemes and the riverine alluvial dugouts. Also the youth are hiredfor land preparation, transplanting, weeding, digging and harvest-ing. Women are mostly hired for transplanting and harvesting oftomatoes, rice and pepper by farmers.

Fig. 8a shows that if 1 ha of tomatoes is cultivated using tempo-ral shallow wells, an amount of about $2000 goes to the commu-nity in the form of labour mainly for digging the wells. Thelowest contributor of income in the form of hired labour to thecommunity is the riverine water technology.

The trend of marginalising women in society has contributed tothe high poverty levels amongst women. Improving the participa-tion in the ongoing irrigation development in the sub-basin isessential for poverty alleviation. In Fig. 8b it is observed that sometechnologies are more favourable towards women and may be dueto factors such as existing land tenure arrangements, cost of thetechnology and the motivation they may be receiving from themen. The promotion of irrigation technologies should considerassociated factors that will improve women participation.

6. Conclusion

A summary of the strengths and weaknesses of the irrigationtechnologies in the sub-basin in terms of productivity is presentedin Table 3. Both the farm records and the regression analysis haveshown that apart from water being a pre-requisite for successfulirrigation adequate fertilizer application is the major factor toachieving high irrigation productivities. Farmers are willing to ap-ply fertilizer because of high returns to the investment, and this isfacilitated by a relatively strong and predictable market fortomatoes.

The impact that an irrigation technology has on the irrigationproductivity has got to do with the control over the water re-sources by the farmer and the size of the farm irrigated by thetechnology. Technologies characterised by relatively small farmsizes are better managed by the surveyed farmers because theyare able to provide adequate water and crop nutrients thus result-ing in higher productivity. The large irrigation scheme where farmplots are relatively large has a low productivity (in terms of land,water and profit margin).

The most productive irrigation technologies are the temporaland permanent shallow wells, followed by the riverine alluvialdugout. Moreover, these highly productive irrigation technologiesalso achieve good profit margins and provide income opportunities

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

to the wider society in terms of labour and also have higher womenparticipation. This confirms the claim by Chambers (1988) thatsmall farms can be more efficient than large farms, and the claimof Obeng-Asiedu (2004) that small-scale technologies can be prof-itable, financially sustainable and able to do better than large scaleirrigation schemes in the Volta basin.

The findings imply that in order to achieve high impact, irriga-tion development in sub-Saharan Africa should consider the eco-nomic status of the users and their ability to make the best outof the technology in terms of productivity. Also, the technologyshould give control over the water to the farmer. The resulting eco-nomic activities have a large positive spinoff in terms of job oppor-tunities, especially for the youth. Also, participation of femalefarmers was also significant in most technologies surveyed.

Finally apart from technologies that depend on reservoirs, allother technologies are farmer driven and required no governmentsupport. This ongoing type of endogenous irrigation developmentin the study area provides a strong backing that the way forwardin sub-Saharan Africa is for governments to create policies thatfacilitate poor farmers becoming irrigation entrepreneurs. Suchpolicies should aim to enhance the reliability of markets (both in-put and output) as the driving force, and facilitate people’s accessto land and water. The role of access to credit and agriculturalextension services in irrigation development remains unclear.

References

Adu, S.V., 1995. Soils of the Nasia Basin, Northern Region, Ghana. Memoir No. 11.Soil Research Institute, Kwadaso, Kumasi.

Agyepong, G.T., 1997. Land Use/Land Cover; Volta Basin System. InformationBuilding Block, Ghana Water Resources Management Study, vol. I. Accra.

Birner, R., Schiffer, E., Asante, F., Osman, G., McCarthy, N., 2005. GovernanceStructures for Water Resources Management in the White Volta Basin, Ghana.Integrating Governance and Modelling. Challenge Program on Water and Food.CGIAR, Accra, Ghana.

Brammer, H., 1962. Soils. In: Wills, J.B. (Eds.), Agriculture and Land use in Ghana.OUP, London.

Chambers, R., 1988. Managing Canal Irrigation: Practical Analysis from South Asia.Cambridge University Press, New York.

CIA – The World Factbook, 2009. <https://www.cia.gov/library/publications/the-world-factbook/print/gh.html>.

FAO, 1991. Land Resource Management Study. Report No. 103/91 CP-GHA 28. FAO,Rome.

Faulkner, J.W., Steenhuis, T., van de Giesen, N., Andreini, M., Liebe, J.R., 2008. Wateruse and productivity of two small reservoir irrigation schemes in Ghana’s uppereast region. Irrigation and Drainage 57, 151–163.

Faurès, J., Svendsen, M., Turral, H., 2007. Reinventing irrigation. In: Molden, D. (Ed.),Water for Food, Water for Life: A Comprehensive Assessment of WaterManagement in Agriculture. Earthscan/International Water ManagementInstitute, London/Colombo, pp. 353–394 (Chapter 14).

Grant, E.L., Ireson, W.G., Leavenworth, R.S., 1990. Principles of EngineeringEconomy, eighth ed. John Wiley, New York.

Hussain, I., Hanijra, M.A., 2004. Irrigation and poverty alleviation: review of theempirical evidence. Irrigation and Drainage 53 (1), 1–15.

ICID, 2000. Role of Dams for Irrigation, Drainage and Flood Control, InternationalCommission on Irrigation and Drainage. Position Paper. InternationalCommission on Irrigation and Drainage ICID, New Delhi.

ICOLD – Members Burkina Faso, 2001. Position on the report from The WorldCommission on Dams. Dams and Development a new Framework for Decision-making, 2000. Earthscan, London and Sterling, VA. <http://www.dams.org/report/reaction/icold_burkina_faso.htm>.

James, L.G., 1993. Principles of Farm Irrigation System Design. Kreiger PublishingCompany, Malabar, Florida.

Kelly, V.A., 2005. Farmers’ Demand for Fertilizer in Sub-Saharan Africa. Agricultureand Rural Development Discussion Paper. The World Bank, Washington, DC.

Lipton, M., Litchfield, J., Faurès, Jean.-Marc., 2003. The effects of irrigation onpoverty: a framework for analysis. Water Policy 5 (5–6), 413–427.

Lund, Jay R., 1989. The value and depreciation of existing facilities: the case ofreservoirs. The Engineering Economist 35 (1), 1–24.

Mdemu, M., 2008. Water Productivity in Medium and Small Reservoirs in the UpperEast Region of Ghana. PhD dissertation. Rheinischen Friedrich, Wilhelms,Universität zu Bonn, Bonn.

Molden, D., Murray-Rust, H., Sakthivadivel, R., Makin, I., 2003. A water-productivityframework for understanding and action. In: Kijne, W., Barker, R., Molden, D.(Eds.), Water Productivity in Agriculture: Limits and Opportunities forImprovements. CAB International, Wallingford, UK.

Molden, D., Oweis, T.Y., Steduto, P., Kijne, W., Hanjra, M.A., Bindraban, P.S., 2007.Pathways for agricultural water productivity. In: Molden, D. (Ed.), Water for

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),

E.A. Ofosu et al. / Physics and Chemistry of the Earth xxx (2010) xxx–xxx 11

Food, Water for Life: A Comprehensive Assessment of Water Management inAgriculture. Earthscan/International Water Management Institute, London/Colombo, pp. 279–310 (Chapter 7).

Obeng-Asiedu, P., 2004. Allocating Water Resources for Agricultural and EconomicDevelopment in the Volta River Basin. Doctoral thesis. University of Bonn, Bonn.

Smith, L., 2004. Assessment of the contribution of irrigation to poverty reductionand sustainable livelihoods. Water Resources Development 20 (2), 243–257.

Please cite this article in press as: Ofosu, E.A., et al. Productivity of irrigatidoi:10.1016/j.pce.2010.07.005

Wiafe, Kwabena, 1997. Irrigation. Information Building Block. Ghana WaterResources Management Study. WRRI, Accra.

Yanggen, D., Kelly, V., Reardon, T., Naseem, A., 1998. Incentives for Fertilizer Use inSub-Saharan Africa: A Review of Empirical Evidence on Fertilizer Yield Responseand Profitability. MSU International Development Working Paper No. 70.Michigan State University, East Lansing.

on technologies in the White Volta basin. J. Phys. Chem. Earth (2010),