Previous Page Table of Contents Next Page

Chapter 4 - Comparison of energy alternatives for small-scale irrigation*

* by James P. Cunningham, Economist

1. Technical calculations
2. Economic analysis


The objective of this chapter is to illustrate the overall logical framework and the specific calculating methods employed in choosing among alternative sources for rural energy supply, sizing of systems and predicting their viability under a range of prospective economic and field conditions. In showing how quite different technologies can be compared, the farm and market data have been constructed as a composite of cases. Consistency of assumptions on basic economic and environmental parameters has been checked and where necessary imposed. But it is to be stressed that conditions in any new real situation may vary widely. No general conclusions are implied by these illustrative calculations, as to the relative feasibility of the technologies examined.

The problem is to determine whether any of several energy supply technologies can be financially and/or economically viable, and to choose among them the best single- or combined-technology system for local conditions. The general approach to solution is first to examine the viability of a base case constructed under a set of conditions judged initially to be most probable. Then tests are made to see how this predicted outcome may change if natural and market conditions, or the management of the farm enterprise itself, are not as foreseen: so-called "sensitivity analysis".

The base case is a hypothetical village of 500 persons, situated in a semi-arid area of 600 mm annual rainfall, 500 mm of which occurs during the primary crop season. The primary output to be provided is irrigation water, and windmills or small diesel pumps on individual farms would perform only this function. Village-scale central plants, either diesel or wood-fired steam, could also provide off-peak surplus energy for food processing as well as a community service center, but still no allowance is made for electrification of individual households.

For the base case, it is assumed that there is not an active hire-labor market, so that the farm size is determined by the area which can be tended under irrigation by the members of a single household. Technical, financial and economic calculations must be performed:

Technical calculations required for a new project include -

- Estimation of irrigation water requirements on the basis of local soil and climatic conditions, and the water-use physiology of crops which are marketable and/or acceptable for on-farm household consumption. Both the season total and its periodicity are needed.

- Estimation of irrigation power requirements on the basis of local surface or ground water conditions, and the water requirements.

- Plant sizing estimation on the basis of the power requirements, and the intensity of available solar, wind or other noncommercial energy supplies, for technologies using them.

Financial calculations correspond to the farm and project budget sheets, testing viability on the basis of purchased input costs and marketed product incomes. Economic calculations are distinguished by an attempt to include also the values of non-market costs and benefits.

1. Technical calculations

1.1 Farm size and crops
1.2 Water requirements
1.3 Energy requirements
1.4 specific fuel consumption
1.5 Plant sizing to meet peak loads

1.1 Farm size and crops

The basic purpose of small-holder irrigation in semi-arid areas is to set a "safety net" against catastrophic drops in food production coming with cyclical failures of rainfall, in areas which can at least marginally support subsistence grain production on their average precipitation.

To extend cultivation into previously uncropped areas not only pushes the limits of both natural and financial resources, but may also ask too much in the way of change in work habits. Similarly, while marketing of some produce may be necessary to generate money income sufficient at least to pay the project's financial operating costs, the area which will support subsistence food consumption can be taken as the minimum for a base case.

A conventional planning concept is that smaller plots require greater concentration on cash crops to be viable, but in some thorough case studies it has been found that the smaller the farm, the more likely it is to produce only basic foods. Since choice of crops was voluntary in these cases, this finding supports the concept that smallholder irrigated farming mostly to supply food for the household may sometimes be more sustainable economically as well as socially, than an abrupt venture into dependence on external marketing of produce.

In any event only site-specific study can make clear whether increased dependence on cash crops implies an increase or decrease in farm size.

With active nearby markets for both subsistence foodstuffs and cash crops, coupled with wide relative price margins, it is even conceivable that a strategy of growing only cash crops could finance both the project expenses and the household purchases of basic foods, from a cultivated area smaller than needed for the project farmers to grow subsistence quantities of their own basic food crops.

On the other hand, evaluations of other smallholder Irrigation projects have reported instances of farmers neglecting the tending of irrigated land, leaving it idle to devote their labor to adjoining dryland plots, when not allowed to grow food grains on the project. They were reluctant to rely wholly on a combination of several favorable outcomes in both natural and market factors.

In project planning it therefore seems prudent to allow at least enough land to grow staple crops for household consumption, if only to assure retaining the commitment of participants, though ample instances of project failure by underestimating land requirements suggest that the farmers' assessment of the required margin for risk is incorrect. After considering this question in depth, an unpublished 1987 wood-energy feasibility study by this author arrived at seven persons per hectare as a planning figure or 71.4 hectares for the village of 500, 75 hectares allowing 3.6 hectares or 5 percent of the area to be used for experimental and extension purposes.

1.2 Water requirements

In general, project water supply requirements in cubic meters per period are calculated as follows:

summed over the i crops to be grown, in which:

A = Area

V = Volume of irrigation water needed in total per period

Ep = Overall project water use efficiency, itself the product of

Ep = (EC x Eb x Ea)

in which:

Ec = Conveyance efficiency (from a distance, for large projects)

Eb =Field canal efficiency

Ea =Field application efficiency

ETcrop = Evapotranspiration rate specific to each crop

Pe = Effective precipitation

Ge = Groundwater contribution (directly taken up by plants assumed insignificant in this case)

Wb = Stored soil water (also assumed insignificant on arid lands)

LR = Leaching requirement, additional water above direct crop requirements, needed to avoid salt buildup in the root layer

The variability of precipitation typically becomes more pronounced as the average declines, and the probability distribution becomes skewed as well: much of the recorded rainfall for a year or even several years may in fact come from a few very heavy rains of short duration, with almost all other showers being too light to have any effect on rain-fed crops or to generate any runoff. Surface streams of local origin may then be even more variable than the rainfall, a major obstacle to gravity-flow irrigation.

As would be reasonable in a semi-arid area, it is assumed that of the 600 mm total annual rainfall, 500 mm occurs in a four-month rainy season. With only slight loss of realism, it is further assumed that all the individual precipitation events occurring outside the rainy season are too light to be effective, that is to result in storage of water in the root zone long enough for it to be used by plants.

The effective fraction of rainfall is jointly determined by climatic, topographic and soil factors in a complex relation, and is furthermore not independent of the choice of crops, in that the below-ground growth habits of the plants will determine the depth of the root zone. The rest of the rainfall which is not effective, is accounted for in three ways: direct evaporation from the surfaces first struck, surface run-off, and deep percolation below the root zone of saturated soil.

Surface evaporation is a clear loss and ways are sought to minimize it. But the latter two phenomena can be important to arid-zone agriculture because run-off feeds surface or underground streams, and percolated water works its way down to the groundwater, so recharging the two main resources from which water for irrigation and other purposes must be drawn. Nevertheless they represent losses from direct utilization by plants.

In some countries locally-applicable formulae have been derived for determining the effective precipitation, and should be resorted to in the project-siting phase. In general the relation between effective and recorded rainfall is nonlinear, but according to "default" tables the effective fraction might vary in a range from about one-half to two-thirds of recorded monthly rainfall over a range of precipitation values which would bound the monthly totals likely in the hypothetical project area during the rainy season: say from 75 to 200 mm/month.

Taking account of the skewed distribution of precipitation events and the resulting heavier weighting of the higher effective precipitation fractions, an overall value of 0.6 is assumed for the preliminary plant-sizing calculations. This would yield 300 mm of effective rainfall on average for the rainy season.

Maize is chosen as a reference crop for plant-sizing purposes, because it is an efficient user of water although it has a high total seasonal water requirement, and so affords high yields: up to 9 tons per hectare. Under the semi-arid conditions we are studying, the specific evapotranspiration for maize would be about 800 mm over a five-month growing season. A variety of alternative crops could be grown within the same season-total and peak-month water requirements.

Crops offering high yields per unit area are especially desirable under irrigation, not only to minimize water losses in distribution and application, but also because more compact plantings allow for less expenditure of effort in maintaining canals and ditches, and more efficient use of field inputs in general.

As for the leaching requirement, the record of irrigation project performance in semi-arid provides ample warning not to underestimate this factor. Accumulation of salts in field soils is one of the most common longer-term causes of project failure, which can occur with surprising rapidity, and one far advanced is frequently prohibitively costly to reverse. The leaching requirement is determined as:

in which:

ECw = Salt concentration in the irrigation water, expressed as its electrical conductivity measured in mmhos/cm¹

ECe = Crop salt tolerance, electrical conductivity of the soil saturation extract in mmhos/cm

Le = Leaching efficiency of the soil, i.e. its internal drainage

¹ micro mho/centimeter or 10-9 1/ohm/cm

The salt tolerance of maize is 2.5 mmhos/cm to realize 90 percent of its maximum potential yield under otherwise optimum conditions, and it must be recognized that this sets a definite upper bound to the salt concentration in irrigation water utilizable for the particular crop. In practice, over a broad range of crops it is recommended that the irrigation water salinity should not be more than about two-thirds that of the maximum tolerable in the saturation soil extract. Maize will tolerate salinity of 5.9 mmhos/cm at a sacrifice of 50 percent of its potential yield.

Higher levels of water salinity would most likely dictate changing staple crops: for example sorghum which is often grown even under tainted conditions in sub-humid areas, will tolerate salinity up to 5.1 mmhos/cm for a 90 percent yield, and can still produce half of its maximum potential with the soil leachate measuring 11.0 mmhos/cm. Provided that input water within a tolerable range of quality can be found, however, the higher water-use efficiency of maize would largely compensate for its greater leaching requirement.

Finally, the soil leaching efficiency varies with soil type, from as little as 30 percent for very finely-divided clays of massive structure, up to 130 percent for coarse, loose sands. Though shallow impermeable layers are not uncommon in the areas we have in mind, the soils are otherwise generally rather light, and it seems reasonable initially to count on an efficiency of leaching towards the upper end of this range, say about 83 percent.

Supposing that the actual irrigation water analysis at the project site shows an electrical conductivity of 2.0 mmhos/cm - the one relatively favorable assumption concerning the state of nature which is made in the base case - the leaching requirement defined in formula 2 becomes:

Unlike the other variables above, the overall project water use efficiency is only partly determined by natural factors. To a significant degree it will be an outcome of the effectiveness of initial project design, implementation and management.

Assuming unlined canals dug and maintained by the farmers themselves, distribution efficiencies anywhere from 0.4 to 0.9 can be anticipated, but efficiencies of field application by the furrow system should fall between 0.55 and 0.75: once in the field, the likelihood of water bypassing altogether the root zones of plants is significantly reduced. With mid-range assumptions on both of these two components, overall water-use efficiency would be 0.424.

Substituting all of these values into the main irrigation water formula (1), we obtain the total seasonal water requirement:

This will be used in turn in estimating the minimum total annual power to be generated by the village central plants, either diesel or steam. Ten percent greater water-use efficiency is assumed for individual on-farm wind or diesel pumps, due to shorter conveyance channels, and the corresponding total seasonal water requirement is 1.044 million m³.

1.3 Energy requirements

Requirements for diesel oil, fuelwood or wind energy depend on the power needed to lift the seasonal water volume estimated above, and on the generating technology utilized. The net hydraulic energy needed to lift the seasonal water volume is obtained by:

in which H is the pumping head and the denominator is a physical constant.

The pumping head is not exactly synonymous with the height the water is to be lifted: because of pipe friction the "effective head" will be at least slightly greater than the height, and may be appreciably so if a horizontal distance is also traversed. For present purposes a simple vertical lift is assumed, from a streambank pumpset or a borehole of sufficient diameter to minimize the frictional component.

Experts in irrigation say that most viable projects are confined to pumping heads of not more than five meters or so, except when geared to market production of vegetables. In order to extend the potential siting as broadly as possible in semi-arid areas, to places where adequate flows of groundwater are found only at somewhat deeper levels, in the base case analyzed here the pumping head is doubled to ten meters.

If the project proves oat under these somewhat more adverse conditions, it will also increase the confidence of success when sited in more favorable circumstances. Substituting in formula 3 we have:

for village-central plants, 28.4 mWh for individual farm installations. A 40-percent "wire to water" efficiency in converting electric to hydraulic energy is achievable by the more powerful, high-head pump units which will be needed Co deliver water at the required peak rate in central plants. So the required 31.6 mWh of pumping power per season corresponds to an electrical energy requirement (Elec) of:

Elec = (31.6 mWh/.40) = 79.0 megaWatt-hours

Assuming 85% efficiency for the electric generation step, the engine shaft power to be provided is

(79.0/.85) = 92.9 mWh

1.4 specific fuel consumption

Diesel fuel

These requirements are taken from standard engineering tables and only one significant remark is necessary: large diesels in the 25-50 mW range can be more than four times as efficient as the 1-3 mW engines under consideration for individual farm pumps, accounting for their much higher fuel costs recorded in Table 1.


With steam (or gasifier) technology 2 to 3 kilograms of wood are sufficient to deliver 1 kiloWatt-hour of engine shaft power. At 3 kg/kWh and 750 kg/cubic meter of fuelwood, this means that about .005 cubic meters of fuelwood are sufficient to produce 1 kWh of electricity. So the total seasonal electric power requirement for pumping alone implies a wood fuel consumption of

(79,000 kWh) x.005 = 395 cubic meters

To this must be added household fuel needs, customarily estimated to be on the order of 2 kilograms per capita per day for cooking only, and siting is anticipated in an area where fuel is not used for heating. The cooking use for our village of five hundred works out to

(500 x 2 x 365)/725 = 487 cubic meters

so the combined demands for cooking and power generation can be supplied by harvesting the growth from 882 hectares, at the very modest arid-zone annual increment of one cubic meter per hectare, and this means that wood for the village and the project can be gathered within a radius of 1675 meters - little more than a thousand paces. This in turn implies that the wood harvesting and transport arrangements are not very demanding; both might even be done by hand Wind power

By substituting basic physical constants in the complex estimating formula cited by Vel and van Veldhuizen (28), a simplified calculation of the monthly water output, Vm of a multiblade windmill is obtained by:

in which

A = swept rotor area in square meters
H = pumping head in meters, as before
Sd = design windspeed of the windmill, meters/sec
Sm = monthly average wind speed, meters/sec

(Note that the final factor in this equation is a correction factor to be applied only in the case of monthly windspeeds greater than the design speed).

Thus once a windmill design is chosen appropriate to the windspeeds experienced in the project area, for a given pumping head the water output estimate varies directly with rotor area alone. Windmill cost estimates also vary directly with the rotor area

The windmill in this example Is capable of pumping 3.64 cubic meters of water per day per square meter of rotor area at a head of 10 meters, operating at an average windspeed of 4.5 meters per second.

1.5 Plant sizing to meet peak loads

To make the plant sizing decision, in addition to seasonal water needs it is also necessary to know the peak load which will be placed on the plant. Owing to the possibility of some soil water storage in between irrigation applications, it is sensible to define the peak on a monthly basis. For grains and many other crops the heaviest water use falls in the critical mid-season stage between the attainment of fall ground cover an 1 the onset of maturation.

For maize during one month when the ears are developing, 260 of the total 800 millimeters of water must be available for full yield to be realized. Such a marked peak is more the rule than the exception in field crops, and it has significant implications for plant costs and configuration; on the other hand it also creates opportunities for utilizing substantial surplus generating capacity for other uses.

About 100 of the 300 millimeteres of effective rainfall should also be available during this month on average, but owing to the variability discussed earlier, the peak-load sizing problem here takes account of the possibility of total rain failure during the month of greatest water need. To insure that the growing crop receives 90 percent of its total water requirement under this worst case, the hydraulic energy requirement during this month becomes:

(.9 x 260/800) x 31.6 mWh = 9,243 kWh

Under conditions in project areas it is not realistic to expect that more than 13 hours a day of continuous engine operation could be reliably sustained even for a month. Delivery of the required power in 540 running hours implies a single steam or internal combustion engine of at least 50 kW capacity, or individual-farm engines of 2.9 kW capacity. The corresponding individual windmills would need a rotor radius of 3.2 meters, rather large for farm handling, giving added force to the support for dual-energy installations emerging from the financial analysis of the following section.

2. Economic analysis

2.1 Project costing assumptions and estimates
2.2 Base case technology choice alternatives
2.3 Base case cost and revenue comparisons
2.4 Economic pricing
2.5 Sensitivity analyses

2.1 Project costing assumptions and estimates

For the following financial analyses most of the basic technical parameters are adapted from recent FAO studies (e.g.12,20) of wood-fueled gasifiers and steam engines for electric power generation in developing countries, and another sourcebook by the organization dealing with windmills among other pumping power sources (15). Rescaling and adjustment of installation cost items has been done on the basis of standard cost factors, with adjustment for load factor and operating time.

Planning costs for irrigation works such as distribution channels, wells, and storage tanks in the case of windmills, have been adapted from several case studies. These have been normalized to the conditions of the example case, but again the costs of ditching, grading etc. may well vary considerably in different field conditions.

It has been assumed that field application ditches will be dug and maintained by the farmers themselves in their own interest. Indeed some reports of other smallholder irrigation projects indicate that requiring early involvement in such works can be a significant factor in ensuring that farmers will later feel they have a "stake" in the success of the operation.

The importance of such motivating measures can be gauged from the fact that failures in land preparation, ditch maintenance and management of water distribution are some of the most often cited causes of project failure.

Another, more minor question concerns provision for household water requirements of the village. When set in comparison with the volume of irrigation water needed, this element is comparatively so small that it can be taken care of in "rounding up" the irrigation power requirement for presentation: the difference between 31.6 and 32 mWh hydraulic energy corresponds Co 9,000 cubic meters of water, equivalent to more than 80 liters/capita per day, an ample ration by arid-zone standards.

In central plants a modest additional power load of 10 mWh could allow for some community services at a few locations: communications, a dispensary, a lighted hall for evening use, perhaps central communal refrigeration and food preparation facilities for household use. While no explicit account has been taken of secondary uses for the individual on-farm installations, the same off-peak surplus power is available with the diesels for farmers who might choose to purchase a small generator or harness the engines to other direct-drive applications, and windmills can also be alternatively used at least for grinding.

For both diesel and steam engines an annual operating time of 5,000 hours is assumed, which would correspond to an average of 12 operating hours per day with one month of complete shutdown, an occurrence which might be desirable but perhaps not possible to avoid.

2.2 Base case technology choice alternatives

In the base-case comparisons presented in Table 1, for central plants wood-fired steam has the desirable characteristics of a long-known and thoroughly proven technology, and a very long lifetime with proper maintenance. Operation itself is relatively simple but the temperatures and pressures maintained mandate adequate training for safety as well as efficiency. The boiler as well as the external-combustion engine are relatively costly, however, and any but the highest-quality feed water must be treated to avoid corrosion and scale buildup.

On individual farm installations the windmill is also well known and of tested effectiveness, and maintenance is simple though it cannot be neglected. With periodic replacement of bearings it is long-lived. Again capital costs are relatively high.

Diesel is today the best known of the three technologies compared, and the simplest both to operate and maintain, and the engine is also by far the least expensive in initial investment though its useful life is relatively short, partially offsetting the reduction in annual capital charges. A more important drawback is the cost of diesel oil, which most developing countries have to import, and price fluctuations are so great as to render the lifetime cost of diesel generation the least certain of the three alternatives.

There is a possibility of partially mitigating both the high capital costs of wind or steam power, and the high diesel operating costs using dual power sources, so that the base load is supplied by capital-intensive equipment which is almost always fully utilized; while peak capacity is chosen to be relatively inexpensive in investment, compensating for higher operating cost during its briefer running time. Table 1 includes this option for both village-central and individual installations.

2.3 Base case cost and revenue comparisons

With respect to total and unit water costs, the most striking contrast in Table 1 is between the village central plants as a group, and the individual farm installations. The difference between these two groups is far too large to be removed by any plausible change in the base-case assumptions, and clearly calls for re-examining the common notion that decentralized energy technologies are automatically less capital intensive.

On the other hand there is an element of coincidence in the close similarity of total annual costs within the central-plant group, between steam, diesel and combined plants. Largely it is an artifact of the relatively low level of capacity utilization, .36 on annual basis, so that the higher fuel costs for diesel very nearly balance the higher capital costs for steam, with a further levelling due to the substantial fraction of waterworks construction costs in total investment for all plants.

The similarity would disappear with an alternative set of starting assumptions, specifically to find paying uses for the off-peak surplus power, such as in running local industry like a sawmill. Then the higher fuel costs of diesel begin to dominate even at the base-case unit fuel price of $.20 per liter. These alternative assumptions were not made part of the base case in this illustration because it would not then be possible to compare the central plants on equal terms with the individual household installations for which off-farm use of surplus power is not feasible.

Table 1: Cost Structures of Alternative Rural Energy Technologies

Base assumptions:

Diesel fuel price

$0.25/kg (=1.251)

Wood preparation cost


Interest rate

8 percent

Crop sale revenue


Lifetimes: Large Diesel 10 yr, Small Diesel 7 yr, Steam & Wind 20 yr

50-kiloWatt central plants

Equivalent it household plants



Dual S/D



Dual W/D















(Water pump)

Switches & Wiring






















Freight & Insurance














(Tech. Contingency)







Bldgs & Foundations






(Site Preparation)




Well drilling







Intake pipe







Storage tank



Conv. channels




Distn channels














(Cons. Contingency)


















85700 312660


Ann. capital chgs



















Water treatment



(Maint. & Repairs)




Fuel costs




















Oils & Chemicals %







However it is apparent even in the base case that individual farm installations can be made significantly less costly by combining a windmill for base-load water supply with a diesel for peak-period operation, each half the capacity of the corresponding stand-alone units. Besides there are the values of redundancy, in the simplest case just Co insure against mechanical failure of either the windmill or the diesel pump.

The dual-plant strategy can also provide insurance against adverse events in another very important way. As Figures 1 and 2 based on the table illustrate, in addition Co reducing total annual costs significantly for the individual-farm installations, a cost structure much better balanced between capital and operating costs is achieved with the dual wind/diesel option, and there is similar though less-dramatic balancing of cost structure in the dual steam/diesel option for village central plants.

The insurance value of these changes in cost structure would be realized in the event that there were substantial future increases in either oil prices to which the diesel-only plants are especially vulnerable, or in the maintenance and replacement costs which weigh most heavily for the capital-intensive installations, the steam and especially the windmill options.

Besides capital and fuel costs, the major remaining cost items are attributable to local labor, either directly in wage payments or through purchase of fuelwood. Of courses it might well prove feasible to permit payment of irrigation fees in kind with fuelwood, reducing further the cash income requirement to cover the project costs.

2.4 Economic pricing

As it is the break-even annual income per farm varies from $ 259 for a village central steam plant, to $ 671 for Individual windmills operated alone. Nevertheless the choice between these two options is far from obvious, even if they were taken Co be the only ones under consideration in order to avoid dependence on oil fuels entirely:

It is to be stressed that there are nonfinancial costs of social organization and management which must be incurred in order Co collectively operate the central plant, and which in particular cases may more than compensate for the financial cost differential.

When an attempt is made to take account of the "hidden" costs of this kind, financial analysis becomes economic analysis. Hidden costs (or benefits) are of two kinds: externalities and opportunity costs. Externalities are costs borne by parties other than the project participants, while opportunity costs arise from nonproject earning activities which the participants themselves may have to give up in order to join the project.

Figure 1. Comparative Cost Structures Village Central Park

Figure 2. Comparative Cost Structures Individual Farm pumps

In this case an example of an external cost would be loss of income by nearby nonparticipating farmers practicing rainfed agriculture, if competition from better-quality produce of irrigated plots were to reduce the market value of their crops. External benefits might accrue by improved nutrition of project children eliminating the need for their participation in supplementary feeding programmes.

Opportunity costs would be incurred if project farmers had to give up off-farm wage labor in order to devote the necessary time to tending irrigation machinery and waterworks, and to the more-intensive crop cultivation needed to make the project pay its way.

If there were opportunities for wage labor, the project wage bill which here includes only the financial costs of wages actually paid to maintenance workers, should be increased by the amounts of the incomes farmers could otherwise have earned elsewhere.

Similarly, if prices of either fuels or farm produce are artificially elevated or suppressed from market levels as is done in many countries, it is the so-called "shadow" prices, prices which would prevail in unregulated markets, further adusted for any non-market costs or benefits, which should be used in evaluating the economic viability of the project.

As an illustration of such adjustments, suppose that the prevailing world-market price of oil, plus refining and distribution costs, would dictate a current price of $.35 per liter for diesel oil, and it is maintained at $.20 by government subsidy. From the data on which Figures 3a and 3b are based, the differences in project net incomes for the four affected technology options would be as follows:

Differences in Financial and Economic Incomes Due to Diesel Oil Subsidy:


Village Central

Individual Farm


Dual S/D


Dual W/D

Financial analysis:

$ 45,457

$ 44,809

$ 16,713

$ 23,621

Economic analysis:

$ 39,765

$ 43,243

$ -7.761

$ 16.391

Added economic benefits can be accounted in the same way. Suppose that prior to implementation of the project, 140 children in the village were receiving nutrition supplements at a cost of $.25 per child per day, and increased on-farm food availability allows this program to be discontinued. The differences between financial and economic income measures become:

Differences in Financial and Economic Incomes Due to Nutrition Benefits:


analysis: $ 45,457

$ 43,243

$ 16,713



analysis: $ 58,232

$ 57,534

$ 29,433


2.5 Sensitivity analyses

Figures 3a and 3b themselves are constructed by recalculating the project costs for each technology option over a range of possible fuel prices which might prevail during the project lifetime, to show the sensitivity of net incomes to such changes. With respect to fuel price, the sensitivity analysis can be interpreted as showing the year-to-year variation in incomes which might actually occur as oil prices fluctuate, since this is a part of operating costs.

On the other hand the sensitivity to different interest rates analyzed in Figures 4a and 4b is useful mainly for planning purposes, since the lending rate will be fixed at the time the project financing is negotiated.

All the village central plants show positive net incomes over a wide range of fuel prices: for the diesel-only installation the breakeven income corresponds to a fuel price of $.25 per kilo, and pairing diesel and steam engines in a dual-fuel plant substantially reduces vulnerability to oil price fluctuations.

In contrast individual-farm diesel pumps evidently would run a loss at any fuel price above $.30 per kilo, and the improvement through coordinated operation of a small windmill and a diesel pump is again dramatic.

The village-scale plants display distinctly more similarity in their characteristics of sensitivity to variation in the interest rate, and none of the three options are much disadvantaged over a broad range of lending rates.

The more highly capitalized single-farm pumping installations are more sensitive to the interest rate, net income of windmills falling to zero at an interest rate of 15 percent, but again there is an appreciable advantage in dual-source operation.

In examining these cases it must be stressed that due to their hypothetical nature, the absolute levels of net incomes corresponding to different fuel price or interest-rate levels have little general significance. There is greater validity in the relative positions and slopes of the functions for the different technology options.

Finally, Figure 5 portrays sensitivity Co fuel price variation in terms of the internal rate of return, a business budgeting measure useful for comparing enterprise earnings with the interest incomes of alternative investments. The IRR is the interest rate which equates the discounted values of the expected annual series of project costs and benefits. It is found by setting the present value of an annuity continuing for the project lifetime, equal to the result of:

in which

Io = initial investment B - C
B = annualized benefits
C = annualized costs

Note that due to the discounting factor, the internal rate of return goes to zero more rapidly with increasing fuel prices, than does the simple net income measure of the preceding sensitivity analyses.

Figure 3a. Sensitivity to Fuel Price Village Central Plants

Figure 3b. Sensitivity to Fuel Price Individual Farm Pumps

Figure 4a. Sensitivity to Lending Rate Village Central Plants

Figure 4b. Sensitivity to Lending Rate Individual Farm Pumps

Figure 5. IRR Sensitivity to Fuel Price Individual Farm Pumps


1. Blackie, M.J. ed. African regional symposium on small holder irrigation. University of Zimbabwe, Harare 1984 (papers and proceedings)

2. Brouwer, C. & M. Heibloem. Irrigation water needs. FAO Irrigation Water Management Training Manual no. 3, 1986

3. Carruthers, I. & C. Clarke. The economics of irrigation. Liverpool U. Press, 1981

4. Doneen, L.D. and D.W. Westcot Irrigation practice and water management. I&D 1/R.1, 1984

5. Doorenbos,. & A.H. Kassam. Yield response to water. I&D 33,1979

6. Doorenbos, J. and W.O. Pruitt. Guidelines for predicting crop water requirements. I&D 24, 1984

7. Durst, Patrick B. Wood-fired power plants in the Philippines: financial and economic assessment of wood supply strategies. Biomass 11 (1986) p. 115-133

8. FAO. Land evaluation criteria for irrigation. World Soil Resources Report No. 50, 1979

9. FAO. Report of the 1985 technical consultation of the European cooperative network on rural energy, biomass conversion for energy. Freising, Federal Republic of Germany, October 1985. Conference on New and Renewable Sources of Energy report no. 10

10. FAO Land and Water Development Division. Guidelines: Land evaluation for irrigated agriculture. SB 55, 1985

11. FAO Land and Water Development Division. Soil survey investigations for irrigation. SB 42, 1986

12. FAO Mechanical Wood Products Branch. Wood gas as engine fuel. Forestry paper 72, 1986

13. Finkel, H.J. CRC handbook of irrigation technology (2 vol) CRC, Boca Raton, Florida 1982

14. Florin, Reto Personal communication, June 1987

15. Fraenkel, P.L. Water lifting devices. I&D 43, 1986

16. Kraatz, D.B., J. Stoutjesdijk & P. Maele. Report on a survey of existing and potential farmers' self-help irrigation schemes in Liwonde Agricultural Development Division, Malawi. FAO/UNIDO/MLW/80/013, 1983

17. National Academy of Sciences, United States. Energy for rural development: Renewable resources and alternative technologies for developing countries. Washington 1976

18. Philippines National Irrigation Administration. Improvement of irrigation facilities through groundwater development, Philippines. Manila 1975

19. Project Preparation and Monitoring Bureau, Tanzania Ministry of Agriculture. Mwamapule village irrigation project. Dar es Salaam 1980

20. Roveda, E.B. Electricity from wood through steam engines. FAO Forestry Paper, forthcoming

21. Rufiji Basin Development Authority. A tentative evaluation of the potential for agricultural development on the lower Rufiji flood plain. Tanzania Ministry of Agriculture, Dar es Salaam 1979

22. Rukuni, M. An analysis of economic and institutional factors affecting irrigation development in communal lands of Zimbabwe. University of Zimbabwe doctoral dissertation, 1984

23. Togo Government/FAO. Consultation on irrigation in Africa. I&D 42, 1987

24. Trossero, M.A. Personal communications, 1987

25. UNDP/FAO. Kaombe smallholder irrigation scheme, Malawi: Project findings and recommendations. AG:DP/MLW/80/038, 1981

26. United States Forest Service. Potential for fuelwood and charcoal in the energy systems of developing countries. Submitted to FAO Forestry Department and U.N. Conference on New and Renewable Sources of Energy, 1980

27. Van Doorne, J.H. A review of small-scale irrigation schemes in Kenya. FAO 1985.

28. Vel, J.A.C. and L.R. van Veldhuizen. A model for the economics of small-scale irrigation with windmills in Sri Lanka. Netherlands Ministry of Development Cooperation, Amersfoort 1981

Previous Page Top of Page Next Page