Irrigated Agriculture in Djibouti: An Economic and Physical Analysis of Irrigation Systems
1 Irrigated Agriculture in Djibouti: An Economic and Physical Analysis of Irrigation Systems Based on the Aden Atteyeh S...
Irrigated Agriculture in Djibouti: An Economic and Physical Analysis of Irrigation Systems Based on the Aden Atteyeh Sougal Family Farm
Aden Atteyeh Sougal – Agronomist/ Team Lead Bryan Pon – Economist Matthew Bates – Engineer/ Economist Anna Petersons – Irrigation Specialist
D-Lab, Spring 2009, Professor Kurt Kornbluth
Contents 1. Introduction 2. Demand Estimation 3. Technology Selection 4. Pump-System Design a. Pump Calculations b. Pump Selection c. Solar Panel Selection d. Additional Pump-System Components
5. Irrigation a. Irrigation Background b. Irrigation Suppliers and System Design
6. General System Recommendations 7. Economic Analysis a. b. c. d. e. f.
Land-Use System (LUS) Analysis LUS Results Total Cost of Ownership (TCO) Analysis TCO Results Incremental Approach Costs Under Two Production Alternatives
8. Conclusions and Next Steps
1. Introduction In Djibouti, the agricultural sector contributes just 3% of GDP, and only a few people work in farming. Due to the Djibouti climate (arid to semi-arid) and the scarcity of fresh water resources (~150 mm rainfall/year), only irrigated and seasonal agriculture is possible. Djibouti farmers use diesel engine water pumps, which have significant costs to purchase (~$2,000) and operate (~$1,700/hectare). These high costs result in elevated prices for locally produced agriculture products compared to imported fruits and vegetables. Djibouti therefore imports most of its fresh vegetables and fruits from neighboring countries, including Ethiopia, Yemen, Kenya, and Europe/France. Many other factors, such as global climate change, the decreasing availability of fresh water, the increasing price of gasoline and diesel, and the increase in the Euro exchange rate, also affect the cost of doing agriculture in Djibouti. These forces combine to create a significant food-security situation in Djibouti.
The project team explored solutions and innovations which could alleviate constraints and allow local farmers sufficient financial incentive to stay on their land, while providing increased food security and economical opportunities. In developing this analysis we used as a model a real Djibouti farm, that belonging to Aden Atteyeh Sougal, which provides extensive, if complex, data to build our project analyses on. While the team evaluated technologies and system requirements based on the Aden family farm, it is our hope that this research forms the basis for ongoing exploration of economically viable irrigation systems for Djibouti smallholders in general. The research question the team sought to answer was “How does the economic performance of current diesel pumping technology and basin irrigation compare to solar energy pumping and drip irrigation for small scale (0.5-2 ha) animal husbandry and high-value crops in Djibouti?”
2. Demand Estimation The purpose of our agricultural pumping in Djibouti is to irrigate fodder crops and high-value fruit and vegetable crops. While in reality, Aden’s family and other Djiboutian farmers grow a variety of crops throughout the year, we focused on just three archetypical crops that his family grows: napier grass (Pennisetum purpureum), sweet peppers, and melons. This allowed us to estimate crop water requirements and thus total system water requirements. For details, please consult the attached spreadsheet in Appendix 1, Agricultural Pumping in Djibouti, Irrigation and Greenhouse Gas LCA. Daily water needs were calculated at one-week intervals. Assuming adoption of drip irrigation for 1 hectare of production, as well as unchanged water use for livestock, in June, a peak of 80 m3 of water was required per day; in September, water needs were as low as 35 m3. The estimated daily water demand hydrograph is shown for each week of the year in Figure 1, below.
Figure 1. Estimated daily water demand to meet current crop requirements throughout the year.
3. Technology Selection Our initial problem statement did not specify a particular technology to achieve our water supply and delivery goals. One of our first tasks, after determining the water needs, was to evaluate potential technologies for suitability to Djibouti and Aden’s family’s farm. In this process we looked at three general categories of technologies: developed wind technology, developed solar technology, and “appropriate technologies” in general. While wind technology can be used to pump water either mechanically or through electrical generation, solar pumping is strictly electrical, and the appropriate technologies are strictly mechanical. After evaluating several of the appropriate technologies like treadle pumps, rope-and-washer windmill pumps, India Mark II hand pumps, etc., we decided that pumps in this category would be unable to deliver the water volume necessary or deliver water to the required head. Though perhaps most appropriate for the region, pumps in this category were ruled out for our specific application. In deciding between developed wind and solar technologies, availability of local information has been the deciding criteria. While global data on solar radiation has little special variability within any given region and is readily available at fine scales, data on wind velocities is generally unknown at find scales and is intensely variable with local terrain. Given that costs are similar for solar- and wind-powered pumping, and lacking any quantitative wind data other than Aden’s description that “it is often windy,” we decided that the uncertainty was too great to justify any sizable investment in wind power. After evaluating parameters at Aden’s family’s farm and finding no significant barriers to photovoltaics, we decided to focus our problem statement and proceed with a system design centered on solar-powered pumping.
4. Pump-System Design Pump Calculations The maximum water demand predicted by our agricultural model, peaking at 80 m3/day, occurs during the month of June. Based on this delivery volume and NASA surface meteorological data indicating a June monthly average of 6.27 equivalent peak-sun-hours per day 1 (e.g. 6.27 kWh/ m2/day ), we used our custom Flow and Head Loss Engineering Calculations spreadsheet in Appendix 2 to determine the head and flow requirements for the pump. These calculations indicated that the necessary pump must be capable of delivering at least 56 gpm to a head of 85 feet (37psi), including calculated frictional head losses of 5.4 feet, to meet peak demand. Designed the system around the mean, instead of peak, flow, however, may make more efficient use of our solar energy investment. Assuming an annual mean of 51.5 m3/day (which is only exceeded in 16 weeks per year, due to the asymmetrical nature of the demand hydrograph), and the average annual solar insolation of 6.19 peak-sun-hours, our pump will need to supply at least 37 gpm to a head of 81.3 ft,
NASA Surface meteorology and Solar Energy - Available Tables, http://eosweb.larc.nasa.gov/cgibin/sse/grid.cgi?email=mebates%40ucdavis.edu&step=2&lat=11&lon=42&num=223102&p=grid_id&p=swv_dwn &p=exp_dif&p=avg_dnr&p=srf_dwn0&p=avg_kt&p=avg_nkt&p=clr_sky&p=clr_kt&p=clr_nkt&p=lwv_dwn&p= daylight&veg=17&hgt=+100&submit=Submit.
including frictional losses of 2.5 ft. Having decided to size our pump around the mean delivery volume, peak delivery can still be achieved with additional solar power. Pump Selection After researching various pump specifications and operating ranges we selected two makes of pumps that seem suitable for our high-head, high-flow, direct-solar application. Our initial choice, the Grundfos SQFlex pump series, is specifically designed for wind and solar pumping, and offers an incredibly wide range of direct DC operating voltages, functioning from 30v to 300vDC. Though appealing for a variety of reasons, the SQFlex pumps are not quite capable of delivering the necessary flow to the height of Aden’s family’s tank. In general, however, we still recommend the Grundfos SQFlex pumps as the first choice for direct-DC solar-pumping applications in Djibouti. The alternative pump series, which we recommend for Aden’s family’s particular setup, is the Lorentz PS pump series, also designed with solar and wind energy in mind. The operating voltage for these pumps peaks at 200vDC, and is designed for a nominal 72-96vDC. The particular pump model that best fits Aden’s family’s flow and head requirements is the Lorentz PS1800 C-SJ8-5 pump 2. The pump-system costs of $1650 include the pump ($1000), an optional-but-beneficial pump controller ($600), and a waterlevel sensor ($50) for the tank. A modified pump performance curve showing the relationship between water delivered and solar investment is shown in Figure 2.
Figure 2. Pump performance curve for the Lorentz PS1800, showing flow delivered in three PV scenarios.
In the long term, it is our hope that appropriate direct-solar DC pumps like the Lorentz and Grundfos designs can be locally available in Djibouti. Lorentz pumps are currently locally available in twenty 2
LORENTZ PS1800 C Submersible Water Pump for Solar Operation, http://www.lorentz.de/offgrid/en/products/ps/1800.
countries in Africa and the Middle East 3, the nearest being Yemen, Kenya, Uganda, and Sudan. The prices quoted above for Aden’s family are based on US pump purchase 4. Shipping costs for all pump and solar components, from San Francisco to Djibouti, is estimated at $8005. Solar Panel Selection Once the pump operating parameters were known, a corresponding array of solar panels was selected. For the Lorentz pump, the nominal operating voltage ranges from about 72v to 96vDC, with a maximum open voltage of 200vDC. Usable power, as identified in the pump performance curve, ranges from about 200 to 1800Watts. The integrated Maximum Power Point Tracking component of the pump controller continuously maximizes pump performance through a wide range of meteorological conditions and varying photovoltaic energy output. One goal of the solar-panel selection process has been to choose a set of panels that can provide usable power even when installed incrementally. The Sharp ND-224U1F Solar Panels recommended here meet this goal, yet do not do so exclusively. We recognize that there are many panels and configurations that meet this criterion and that ultimate panel selection will likely depend on local availability 6. The Sharp panels we recommend here each produce 224 Watts of power at 36.6 Volts and 8.33 Amps, and cost $899 apiece 7. Thus, incremental configurations of two, three, four, and six panels are possible, supplying 448, 672, 896, and 1344 Watts of power at 73.2 or 109.8 Volts at each acquisition stage. Should more uniform voltage be desired at each incremental step, a larger array of smaller panels (e.g. standard 12 Volt) can be used instead of the suggested smaller array of larger panels. Additional Pump-System Components In addition to the pump and solar panels, there are several other improvements that can be made to the general system to improve efficiency and ease of use. PVC joints and valves have been extraordinarily expensive in Djibouti, reducing their overall importance to system design (the current system has no valves aside from one at the outlet of the main storage tank). However, if Aden is going to be traveling back to Djibouti with luggage, or if any system components are going to be shipped from America, we have an opportunity to supply a number of valves and junctions at US rather than Djiboutian costs.
LORENTZ Dealers, http://www.lorentz.de/offgrid/en/contact/worldwide#.
Lorentz PS1800 pump and controller prices ($1600) have been quoted by Innovative Solar Solutions, Inc., of Charlotte, NC, and come with a two-year manufacturer’s warranty. Buy Lorentz Pump System, http://www.innovativesolar.com/solar-water-pumps-27/submersible-pumps-59/lorentz-104/lorentz-ps1800-solarpump-system-695.html. 5
Online Calculator for OCDEAN CONSOLIDATION RATES, http://www.freight-calculator.com/worldocr_cal.asp.
According to Aden, solar panels are locally available in Djibouti, though price and performance details are not currently known; importing panels may remain a viable option.
Sharp ND-224U1F solar panel prices ($899 each) have been quoted by Wholesale Solar of Mt. Shasta, CA, and come with a twenty-five-year manufacturer’s warranty. Sharp ND-224U1F Sharp ND-224U1F 224 watt Solar Panel, http://www.wholesalesolar.com/products.folder/module-folder/sharp/sharpND224U1F.html.
The three distribution tanks in the current system are not interconnected. Whenever the diesel pump needs to supply water to a different tank, two pipes, one wrapped with rubber, are wedged together to form a connection. We suggest adding a four-way 3” PVC junction to permanently link the pump with each of the three tanks. Valves can be added to each branch to allow for future flow customization and maintenance. A short length of PVC and a fifth valve can be used to alternate between solar and diesel pumping sources. To keep water from spilling through the lower-elevation livestock tanks, we recommend installing two float valves to halt flow when the tanks are full. As 3” float valves tend to be rather expensive, we suggest reducing the pipe diameter from 3” to 1” at the inlet to each livestock tank and then using a readily available 1” float valve to regulate flow. The minor losses associated with this 3” to 1” reduction are not expected to significantly impact the fill rate of the livestock tanks. Diagrams of these recommended improvements are included in Appendix 3. Approximate costs of system improvement, including solar panel wiring and frame are given in Table 1, below. Table 1. US costs of system improvement components.
3" to 1" PVC reducers
unit price $5
standard 1" float valves
3" 4-way PVC junction
3" 3-way PVC junction
3" PVC valve
Miscellaneous Equipment and Parts
Homemade 2x4 frame to hold solar panels
5. Irrigation Irrigation Background Aden’s family currently irrigates a series of crop beds with raised edges (basins). These are irrigated by flooding the basin, by placing a 3-inch hosepipe at the high end of the basin and using a slight slope to move water across the basin. Water efficiency is defined as: water used by crop (transpired)/ total water applied. The efficiency of basin-flood irrigation depends on slope, run length (how long the basin is), soil cultivation (how smooth the surface is) and, most importantly, soil type. Soil type—sandy, loamy, or clay—influences infiltration and percolation (how quickly water moves into and down through soil). With flood irrigation, inefficiencies arise because more water infiltrates at the head or the bottom of the bed, depending on slope, and because the entire cropping area is wetted and exposed to evaporation. Also, run length is limited by the ability of the soil to move water horizontally as opposed to vertically. This means Aden’s family must use short beds. Drip irrigation, on the other hand, uses plastic laterals to move water horizontally across the field. This eliminates most of the uneven application and deep percolation associated with surface irrigation. In addition, smaller amounts of water can be applied over a longer time period, so that water moves via unsaturated flow, rather than saturated flow. This results in more even water application, less stress to respiring plant roots, and less evaporation from the soil surface. Drip irrigation’s water use efficiency
usually ranges from 90-95% 8; we assumed a low figure of 90% efficiency when calculating water needs. Efficiency for surface irrigation can range from 15-75% 9. We chose the assumption that Aden’s family currently irrigates at 45% efficiency; this is the middle of the surface irrigation efficiency range, and also half that of our drip irrigation efficiency assumption. In a drip irrigated field, the entire area can be planted, as pathways are not necessary for moving irrigation hoses or constructing basin boundaries. Thus, the percentage of cropped area per hectare increases greatly. We are assuming that with drip, Aden’s family could plant two 0.5 hectare blocks, with access paths around the exterior of each block only. In our model, one block is devoted to napier grass yearround, whereas the other is used to grow sweet peppers in the winter and melons in the summer. For specific cropping dates, see attached spreadsheet in Appendix 1, Agricultural Pumping in Djibouti, Irrigation and Greenhouse Gas LCA. Irrigation Suppliers and System Design In searching for an appropriate drip irrigation system, we used two primary constraints: system pressure and system cost. Our tank is located 5 meters above the highest field; thus, for that field, the maximum pressure available is 7 psi. Most irrigation systems operate at high pressures; drip systems in general operate at comparatively lower pressures, but most conventional drip irrigation emitters are still designed for higher pressure than Aden’s tank can provide; these values range from 10-20 psi. However, a few manufacturers are producing low-pressure drip, for one of two markets: home gardeners in developed countries, and small-scale growers in the developing world. The former tend to be relatively expensive and difficult to scale up; the latter are ideal for our purposes, but it can be difficult to find specifications and pricing information. For example, Netafim has partnered with development organizations in West and East Africa to provide high-quality, low-pressure drip irrigation; however, these systems are not commercially available, and components costs are reportedly comparable to Netafim’s other high-quality, high-price irrigation products. IDE is perhaps the largest supplier of irrigation designed for small-scale growers; however, their for-profit organization in charge of manufacturing and selling irrigation, Global Easy Water Products (GEWP), does not publish price lists, due to fluctuating materials costs. Likewise, Jain Irrigation’s low-pressure system, Chapin drip, is reportedly sold in the US and India, but pricing information has proved difficult to acquire. One U.S. manufacturer named T-Tape, however, produces low-pressure drip in appropriate quantities for our project, and Dripworks, a retailer targeting home, small-scale, and medium-scale commercial growers, lists prices for laterals and system components. From Anna’s work with driptech inc., a relatively new irrigation manufacture, we have access to GEWP prices, as of last December. We used these two sources to construct and price system designs; the components and pricing lists are located in the Agricultural Pumping in Djibouti, Irrigation and Greenhouse Gas LCA spreadsheet in Appendix 1, in the sheet labelled “Irrigation system design.” Both systems use a filter and laterals spaced at 1 meter, with stopcock takeoffs: these allow each lateral to be turned on or off independent of the others. This is useful for managing pressure: fields may be irrigated in segments to maintain adequate water pressure for even distribution. Considering that Aden’s family cropping system is much more diverse than our 8
The California Irrigator’s Pocket Guide, NRCS/NCAT, 2005, p. 27
simplified version, it is also useful for managing mixed fields: several smaller sections of different crops can be irrigated specific to their needs. Each system also contains 3” PVC submains and fittings; these may be available in Djibouti city or may need to be imported from Dubai. While T-Tape is rated for run lengths up to 600 ft (~200 meters), IDE does not recommend that GEWP lateral run lengths exceed 40 meters. Thus, the system designs are slightly different, as shown below: the GEWP scenario has two submains, and thus requires twice as much PVC and twice as many takeoffs. Note: the diagrams below are for each 0.5 hectare field; the total scenario has two such fields, sharing one filter, tank, etc.
IDE/GEWP Submain Laterals 50 meters
Submain Laterals 100 meters
Figure 3. Field layouts appropriate for using Dripworks’ and IDE’s drip irrigation systems. Due to the lower cost of laterals, takeoffs, and filters, the GEWP system is still less expensive, with an initial cost of $911 and lateral replacements every 4 years at $500, compared to a T-Tape initial cost of $1865, with lateral replacement costs of $1365 every 4 years. Taking into account price and adaptability to mixed cropping systems, we decided to use IDE/GEWP products in our scenario. In addition, shipping costs from India should be lower than those from the US (Dripworks ships T-Tape orders from northern California), and there is a possibility that materials and/or replacement parts could be obtained from IDE’s Ethiopia office.
6. General System Recommendations In the process of evaluating the proposed solar pumping and drip irrigation technologies for Aden’s family’s farm, we have also identified several operating practices that can be followed to potentially reduce costs or improve yields in the baseline or improved systems. Our recommendations are as follows: 1) Protect the electric pump. We recommend adding a coarse filter at the well side of the pump intakes. The Lorentz submersible well pump is designed to last for many years of successful operation, but is only warranted by the manufacturer for the first two years. We suspect the local diesel pumps effectively come with no warranty. Though both pumps are designed to be able to pass smaller suspended particles, limiting the amount of course material that passes through the pumps impellers may increase pump lifetime. 2) Safeguard against diesel contamination. We recommend filtering all fuel before use in the diesel pump. Fine suspended particles in the fuel, often picked up during transportation and storage, can decrease engine life upon combustion. By filtering the fuel, we can insure that peak engine performance is maintained for as long as possible. Periodic draining and cleaning of the pump’s fuel tank should help remove any particles that have accumulated in the tank. Even particles less than the width of a human hair can be damaging to an engine’s finely-tuned fuel injection system. 3) Normalize water use. Though output of solar energy does vary seasonally, it varies much less than current-crop water estimates. If a full-scale solar-pumping system is implemented, much of the potential energy will be wasted if water demands fluctuate greatly throughout the year. Modifications to the base-case cropping patterns to either reduce peak demand or increase nonpeak demand may reduce system costs or increase profits.
7. Economic Analysis and Recommended Investment Approach The team took multiple approaches to evaluating the economic viability of the alternate irrigation and pump technologies. The core objective was to verify whether the proposed alternate systems could be economically viable not only for Aden’s family farm, but also for other smallholders in Djibouti facing similar constraints as Aden’s farm. The economic analysis took two forms: a land-use system (LUS) analysis, and a total cost of ownership (TCO) comparison. Land-Use System (LUS) Analysis An LUS analysis is an economic tool for estimating the net present value of a specific farm system. It takes into account the cost of all inputs—e.g., fertilizer, seeds, hired and family labor, irrigation, gasoline to bring products to market, etc.—and revenue from all outputs to determine economic performance for that specific use of the land. For the LUS to be a practical tool, it must take into account the real-world variables and constraints about the particular land system being studied, such as weather, topography, access to markets, subsidies, etc. The result of an LUS analysis is an economic appraisal of the net present value of that particular system in that particular region of that particular country; in other words, it hopes to answer the question, “Is this a good business to get into?” The LUS as a tool also offers the ability to
easily substitute different prices for inputs/outputs to see how they affect economic performance. For example, you can double the price of fertilizer within the spreadsheet and see how that affects profitability over the years. Compiling a complete LUS in the time allotted for the project was challenging, especially given the complexity of the system. Aden’s farm currently plants napier grass year round, but only uses the output as fodder for their livestock (does not sell it). They also alternate planting melons, sweet peppers, and other high-value vegetables, switching from summer and winter, which they do sell at the market. The difficulty in separating out input costs per each sub-system was very high. For example, it was difficult for Aden to know how many person-days his hired laborers actually spend on melons vs. napier grass since both are being worked on simultaneously; similarly it was hard to determine a clear way of assigning diesel fuel costs for the current irrigation pump to one crop vs. another. That said, now that Aden understands how the tool works, he can track and categorize his input costs in a manner more aligned with the model, so that the model should become more accurate over time. We are also creating a separate LUS analysis that will isolate just one high-value vegetable, so that he can better evaluate the economic performance of a smaller system. Isolating this sub-system might also provide better guidance to the central question of replicability, or in other words, whether such a farm system makes good business sense for other smallholders to invest in. LUS Results The LUS showed an annual profit ranging from $5,890 - $8,806 for the first few years, with clear variance in accordance with periodic operating expenses for repairing and replacing the diesel pump. By the end of the system lifetime of 20 years, the discounted annual benefits drop to $1,282 (with a discount rate of 10%). Unfortunately, we could not validate the estimated profit numbers against real figures, as Aden does not currently track discrete revenues (they are immediately invested back into the system). The baseline LUS resulted in a highly positive net present value of $42,541, which would indicate that the production system was a good business to get into. However, due to aforementioned challenges with entering precise data about costs, we are not using the LUS results as a basis for recommended action at this time. With additional refinement by Aden with help from Bryan, the level of confidence in the LUS results should rise. See Appendix 5, the Djibouti LUS Baseline spreadsheet for all figures, specifically the worksheet “Economic Performance.” Total Cost of Ownership (TCO) Analysis The team also performed multiple TCO analyses comparing baseline vs. alternate technologies. By isolating estimated costs of the irrigation system and the pump system and comparing them to theoretical costs of proposed alternatives, the team sought to develop a simplified model that controlled for all of the production variable present in the LUS analysis. The TCO analysis assumed that the context for evaluation was a Djibouti smallholder using diesel-pump, furrow irrigation for small-scale agriculture (i.e., Aden’s family farm). The TCO analysis evaluates alternative options potentially available to this smallholder. Note that because diesel pumps need to be replaced every ~3 years, there are frequent entry points for the options presented by the TCO analysis.
TCO Results As expected, the solar-powered electric pump system has high up-front capital costs, but very low operating costs, compared with the baseline system of diesel pumping, which has low capital costs but high operating costs. This is the typical cost/investment structure facing many renewable energy vs. fossil fuel energy technology choices. Specific Findings: 1. Drip irrigation increases water use efficiency from ~50% to ~90%, which may offer the most gains through increased production land area and yields for same amount of water 2. Initial upfront costs of solar-powered pump system were $6,400 - $8,000 3. Payback period of investing in solar-powered pump system was ~4 years (calculated as cumulative costs of solar vs. cumulative costs of baseline, independent of product output) 4. An incremental investment approach to solar pumping might be viable if drip irrigation is implemented first, thereby reducing water demand and diesel pump operating costs, freeing up capital to be invested in solar pumping (see “Incremental Approach” below) Assuming that neither Aden’s family farm nor other smallholders in Djibouti have available resources to afford an one-time $8,000 investment for the solar alternative, the team explored different scenarios for how such a system might become viable. The obvious solution is financing, but because Djibouti has relatively few credit options for a typical smallholder, we decided against a simple reliance on credit schemes. Instead, the team developed multiple models by which a smallholder could incrementally invest in the solar pumping technology while still maintaining production through use of diesel. In other words, the farmer continues to use the diesel system to produce crops for sale, but gradually transfers more and more investment away from diesel to solar, until the point when the solar is completely online and the farmer can stop investing in diesel. What makes these scenarios possible is the introduction of drip irrigation technology. With drip irrigation, the smallholder can effectively double the crop area under production for an equal amount of water—or maintain the same crop area for about half the water. Under the latter scenario, halving the water requirement almost halves annual operating costs for the diesel system (gas and oil). By freeing up this capital, the smallholder can start to invest in the solar pumping system over time. Incremental Approach We believe the best strategy is a hybrid approach in which a series of smaller investments in increased solar capacity gradually replace the need for diesel. Under this strategy only the drip irrigation system would be purchased in the first year or two, with the savings from this investment helping finance the well pump and first two solar panels (448 Watts) by the fourth year. Further savings accrued with the partial solar pumping will allow for subsequent incremental solar panel purchases, until the full system has been acquired. Depending on the number of years in which diesel power is still employed, this incremental diesel/solar approach reduces optimal lifetime system savings by $5,000 to $10,000 as compared to the immediate-solar approach. However, $40,000 to $50,000 can still be saved above the
total base-case expenses over the expected system lifetime of twenty to twenty-five years. The most important implication of the incremental approach is that it totally removes the barrier to entry. While we believe a combination of drip irrigation and solar-powered pumping is the best solution, either of these technologies could be installed without the other. Installing solar pumping would reduce capital and operating costs but not improve irrigation efficiencies to allow for increased yields. Installing drip irrigation only would improve water and land use efficiencies and allow for increased production, but not address the continuous diesel consumption. Costs Under Two Production Alternatives There are two scenarios for which we can model an alternative pumping system with drip irrigation. The efficiencies of drip irrigation allow for either a one-half reduction in water applied and land cultivated, while maintaining yield, or a doubling of yield while maintaining the amount of water applied and land cultivated, as compared with the base-case. A cost comparison, several relevant graphs, and a recommended timeline for component purchases are presented in Appendix 4, Purchase Strategy Timelines and Graphs, for the base-case system, a diesel and drip irrigation system, and an incremental solar and drip irrigation system, for both of the land-use scenarios. In Figure 4, below, total system capital and operating costs are compared for the base-case, diesel with drip irrigation, and solar with drip irrigation strategies under the fixed-yield scenario. For this scenario, expected annual costs with incremental system purchases are uniformly less than same-year costs for the base-case system, removing any barrier to entry. Barring any unforeseen technological complications, any farmer already operating with similar annual costs can reduce his costs and begin accruing savings before the end of the first year.
Figure 4. Comparison of capital and operational expensed to produce the same crop yield. The double-yield scenario assumes that that once a farmer switches to drip irrigation and is able to halve the land usage for his customary crops, he will want to plant the remainder of his land and also double his yield. In Figure 5, below, total system capital and operational costs are compared for the base-case, diesel with drip irrigation, and solar with drip irrigation strategies under the double-yield scenario. For
this scenario, the expected annual costs of the diesel pumping and drip irrigation strategy remains similar to the annual base-case system costs. Almost negligible savings of approximately $3,500 in capital and operating costs are expected over the course of twenty-five years, but yields will doubled from the second year onwards. With the incremental solar pumping and drip irrigation strategy, savings of $35,000 to $40,000 over twenty-five years, also with doubled yields, can be expected over the fixed-yield base case estimate. We believe this is the optimal scenario, featuring both considerable savings and substantial new profits.
Figure 5. Comparison of capital and operational expensed to double the base-case yield.
8. Life Cycle Analysis of Greenhouse Gas Emissions Though our primary focus was economic sustainability, we investigated alternative energy sources because we wanted an environmentally sustainable solution. In order to test our assumption that solar powered pumping would have a lower environmental impact than diesel pumping, we used CarnegieMellon’s Green Institute Design LCA tool, the Economic Input-Output Life Cycle Assessment (EIOLCA) 10. This gave us an estimate of greenhouse gas emissions produced by the manufacturing of various system components, based on economic sector and component price. We tested the baseline scenario of diesel pumping with flood irrigation, compared to the proposed system of solar pumping with drip irrigation, measuring greenhouse gases in estimated CO2 equivalents and estimated CO2. In the solar scenario, all emissions are embodied; none are produced in use. In the diesel scenario, some emissions are embodied, but the majority are in-use, from burning diesel and oil. Figure 6, below, shows the emissions results of our calculations following estimates for a total life cycle of 20 years.
Carnegie Mellon University Green Design Institute. (2009) Economic Input-Output Life Cycle Assessment (EIOLCA) US Dept of Commerce 1997 Industry Benchmark (491) model [Internet], Available from: [Accessed 5 Jun, 2009]
96 metric tons CO2 and CO2 eq. in 20 years
6 metric tons CO2 eq. in 20 years
Pump Parts Diesel Oil
Figure 6. Twenty-year CO2 and CO2 equivalent emissions for the base-case and proposed solar systems. The values in Figure 6 show an order-of-magnitude difference between the scenarios; from the perspective of greenhouse gas emissions, the proposed solar pumping and drip system is more environmentally sustainable. Though its initial impact is high, this initial impact is still smaller than even one year in the bas-case system, as shown in Figure 7, below.
Emissions (CO2 and CO2 Equivalents) for Existing (red) and Proposed (blue) Systems 6
Submains, fittings Solar Panels
Metric Tons CO2 & 3 CO2 Equivalents
2 3 4 5 6 7 8
9 10 11 12 13 14 15 16 17 18 19 20
Figure 7. Calculated annual emissions for the existing diesel and proposed solar pumping systems. For more detailed calculations, please see the attached spreadsheet Agricultural Pumping in Djibouti, Irrigation and Greenhouse Gas LCA, sheet “LCA,” in Appendix 1. These LCA estimates are somewhat misrepresentative in that they are a gross generalization by economic sector, and are not specific to the goods in questions. In addition, the LCA results do not take into account the transportation of materials to Djibouti. Disposal energy costs are also not included, thought we believe it to be unlikely that waste materials will enter a formal disposal/recycling chain. In summary,
though this is a rough calculation, the order-of-magnitude difference between the systems is unlikely to be reversed should more detailed calculations be performed.
9. Conclusions and Next Steps The project team performed extensive analyses of the physical system requirements, lifecycle environmental costs, production system profitability, total cost of ownership, and potential investment strategies. Given the research question of whether alternative irrigation system technologies are economically feasible for the Aden family farm, the results seem to indicate significant potential for alternative systems. Summary of findings: 1. The implementation of a drip irrigation system seems to be a clear win, with significant improvements over the baseline scenario of basin/furrow irrigation. In comparison, drip irrigation uses 50% less water to irrigate the same area of crops, or, conversely, can water twice the amount of crop area for the same amount of water. Given the resource scarcity of fresh water in Djibouti, this is an extremely important benefit. While current retail distribution of drip systems in Djibouti is difficult to assess, product availability in other East Africa countries seems strong, and the lowtech nature of the system should pose no problems in regards to specialized parts or training for maintenance. 2. Environmental cost in terms of CO2/CO2e is order of magnitude less with solar system compared to baseline diesel system. Rough estimates from an industry-based calculator show the CO2 and CO2 equivalent gases released during manufacture and 20 years of operation for the baseline diesel and solar system to be 96 metric tons and 6 metric tons, respectively. 3. The current baseline system is expensive for many reasons, not just capital outlay and operating expenses. For example, the Aden family farm completely replaces the diesel pump every 3 years—sooner than the expected lifetime— not because the whole pump is bad, but simply because there are no parts available for repairs. This greatly increases system costs beyond what the theoretical cost of a diesel set-up should be. Therefore, the real-world context of the market and availability of goods and services in Djibouti should not be underestimated in any recommendations for alternative systems. 4. Solar systems have a greatly lower TCO, with payback within approximately 4 years compared to the baseline diesel system. The virtual elimination of operating expenses (diesel fuel and oil) and expensive diesel pump replacements means that a full solar-powered pump system could save $40,000-$50,000 over the 25-year lifetime of the system. 5. High upfront costs of solar system may make incremental investment more feasible. With an estimated upfront cost of $6,400 to $8,000, the full solar system is likely out of reach for the vast majority of smallholder farmers in Djibouti, and sub-Saharan Africa in general. While it is tempting to assume that financing can be obtained, Aden confirms that loans are hard to come by, and are especially difficult for agriculture. The team therefore identified options for the incremental purchase of solar system components over time, building up to an operational system that gradually phases out the diesel until it is completely solar-powered by year 7. The capital
required for the gradual investment comes from costs savings derived from the drip irrigation— the increased efficiency of drip requires less water to be pumped, essentially halving operating expenses for the diesel pump and freeing up operating funds to invest in solar. Next Steps While the project team is excited with the preliminary findings above, there remain significant questions in terms of actual implementation of the proposed solar and drip irrigation system. Given that the Aden family farm, and by extension other smallholders in Djibouti, have significant resource constraints, caution is prudent with any new production system. The team has identified some clear next steps to validating the analysis and preparing for implementation. 1. Determine logistics of importing solar and drip irrigation equipment to Djibouti. As noted above, there exist significant barriers and hidden costs to current diesel pumping technologies in Djibouti, and it will be important to learn how and which of these also affect solar and drip technologies. Possible issues include: import duties, VAT, transportation costs, breakage, insufficient demand/unwillingness to import, time lag of importing, etc. It is possible that the only way to truly identify these issues is to try to import products. 2. Work toward producing enough funds to invest in new technologies. Although the team established alternative investment scenarios which should avoid additional upfront costs beyond the baseline, these scenarios assume readily accessible products. For a pilot project, the Aden family farm may have to purchase and import multiple products at once to avoid additional costs associated with multiple transactions. This may require budget planning. 3. Create formal experiments of crop production using drip irrigation and/or solar pumping system. Part of the ongoing research will include more formal experimentation of specific crop production, including quantification of inputs, to enable more accurate LUS data and results. By isolating and formalizing specific production systems (e.g., melons only), the team can offer better recommendations regarding the economic viability for going into agriculture production in Djibouti, and how that specific system may respond to the proposed alternative irrigation systems.