Introduction
Hydroponics is a system of growing plants without soil that can offer improved growth and environmental control through a diverse array of methods (Bradley and Marulanda, Reference Bradley and Marulanda2000; Maboko et al., Reference Maboko, Du Plooy and Bertling2011), including both passive and active systems (Harms and Combrink, Reference Harms and Combrink2000; Kratky, Reference Kratky2003). Efficient hydroponic production systems can reduce water use by up to 90% and land requirements by 75% (Bradley and Marulanda, Reference Bradley and Marulanda2000; Trejo-Téllez and Gómez-Merino, Reference Trejo-Téllez, Gómez-Merino and Asao2012). Hydroponic systems most commonly focus on high-value crops in areas with limited access to soil and are becoming more prevalent around the world (Trejo-Téllez and Gómez-Merino, Reference Trejo-Téllez, Gómez-Merino and Asao2012). However, the potential for hydroponics is only beginning to be explored in developing countries (Baumgartner and Belevi, Reference Baumgartner and Belevi2001; Du Plooy et al., Reference Du Plooy, Maboko, Van den Heever and Chiloane2012).
Simplified systems adapted to local materials and resources have been shown to be feasible and profitable in multiple developing countries (Bradley and Marulanda, Reference Bradley and Marulanda2000), and hydroponically cultivated crops have been shown to contain similar, and sometimes higher, levels of vitamins and minerals than soil-based cultivation systems (Gruda, Reference Gruda2009). This could be particularly important in sub-Saharan Africa, which has a high rate of both malnutrition and urbanization (FAO, 2012). Hydroponic production systems offer an alternative growing method with the potential to increase yields and bring agriculture to areas lacking access to traditional soil-based growing (Gruda, Reference Gruda2009).
Sub-Saharan Africa is home to the greatest concentration of food insecure people, where one in four people is chronically hungry (FAO, 2014). More common than undernutrition is malnutrition, where people lack the essential vitamins and minerals they need to grow and lead healthy lives. This lack of critical micronutrients is a ‘hidden hunger’ that impacts an estimated two billion people worldwide (WHO and FAO, 2006).
African Leafy Vegetables (ALVs) such as vegetable amaranth (Amaranthus cruentus) offer a locally appropriate way to address these issues. ALVs are a diverse and widespread set of vegetables that have traditionally been consumed across Kenya and sub-Saharan Africa (Muhanji et al., Reference Muhanji, Roothaert, Webo and Stanley2011) but in many cases have been ignored at the expense of introduced vegetables such as kale and cabbage (Adeka et al., Reference Adeka, Maundu and Imbumi2008; Okeno et al., Reference Okeno, Chebet and Mathenge2003). ALVs like vegetable amaranth are the cheapest source of many nutrients in Kenya (Orech et al., Reference Orech, Christensen, Larsen, Friis, Aagaard-Hansen and Estambale2007; Uusiku et al., Reference Uusiku, Oelofse, Duodu, Bester and Faber2010) and in some parts of western Kenya, ALVs contribute as much as 33% of total vitamin A to the diet (Oiye et al., Reference Oiye, Shiundu and Oniang'o2009). This is extremely important in an area of the world where the daily intake of fruits and vegetables is well below dietary recommendations, and affordability of vegetables remains a challenge for the poor (FAO, 2012).
Vitamin A, iron and zinc are some of the most common and critical micronutrient deficiencies (WHO and FAO, 2006), all of which are present in ALVs (Orech et al., Reference Orech, Christensen, Larsen, Friis, Aagaard-Hansen and Estambale2007; Uusiku et al., Reference Uusiku, Oelofse, Duodu, Bester and Faber2010). Carotenoids, zinc and iron are all already found in ALVs, and it may be possible to increase their concentration through nutrient solution manipulations, which would help these vegetables address micronutrient deficiencies.
For Kenya's growing urban population, the availability and accessibility of vegetables is limited by seasonal fluctuations in climate (Nekesa and Meso, Reference Nekesa, Meso and Guarino1997) and developing supply chains (Ayieko et al., Reference Ayieko, Tschirley and Mathenge2005). ALVs are grown primarily as a rain-fed crop, which restricts their production season to the two rainy seasons. Urbanization is also occurring at a higher rate in Africa than anywhere else in the world (FAO, 2012), and demand for ALVs in urban Kenya has already outstripped supply (Mwangi and Kimathi, Reference Mwangi and Kimathi2006). If hydroponic production in urban and peri-urban areas were able to provide reliable access to nutritious ALVs, this could improve food security outcomes for Kenya's urban residents.
If hydroponic cultivation of ALVs is shown to be practical, profitable and acceptable to consumers, this technology could provide an opportunity for small businesses and entrepreneurs to capitalize on the urban ALV market and these innovative production practices. If these vegetables can provide improved nutrient concentrations, they may demand a higher price or be used to target at-risk populations.
Nutrient solution impacts on nutritional density
Of the 17 essential nutrients required by plants, 14 of them must be supplied by the nutrient solution (Mattson and Peters, Reference Mattson, Peters and White2014). Altering the concentration of nutrients can affect bioaccumulation in many species (Palaniswamy et al., Reference Palaniswamy, Bible and McAvoy2004; Gruda, Reference Gruda2009; Sonneveld and Voogt, Reference Sonneveld and Voogt2011), though most research has focused on hydroponically-grown tomatoes rather than leafy vegetables (Trudel and Ozbun, Reference Trudel and Ozbun1970; Paiva et al., Reference Paiva, Sampaio and Martinez1998; Fanasca et al., Reference Fanasca, Colla, Maiani, Venneria, Rouphael, Azzini and Saccardo2006; Ramírez et al., Reference Ramírez, Díaz Serramno and Muro Erreguerena2011). Higher potassium concentrations have been shown to increase total tomato carotenoid concentration (Trudel and Ozbun, Reference Trudel and Ozbun1970; Ramírez et al., Reference Ramírez, Díaz Serramno and Muro Erreguerena2011), although β-carotene concentrations decreased (Trudel and Ozbun, Reference Trudel and Ozbun1970; Dumas et al., Reference Dumas, Dadomo, Di Lucca and Grolier2003). Altering the ratios of calcium, magnesium and potassium revealed that total carotene and β-carotene concentrations decreased with high calcium concentrations (Paiva et al., Reference Paiva, Sampaio and Martinez1998; Fanasca et al., Reference Fanasca, Colla, Maiani, Venneria, Rouphael, Azzini and Saccardo2006). There may also be an interaction effect between nitrogen and potassium in β-carotene synthesis (Wang et al., Reference Wang, Wu, Ding, Zhou and Lin2013). Iron, zinc and manganese concentrations have been shown to correlate with leaf concentrations, but this was only true in cucumber and not tomato leaves (Sonneveld and Voogt, Reference Sonneveld and Voogt2011). Competition between these metal ions is also strongly impacted by pH, especially zinc and iron (Berry and Knight, Reference Berry and Knight1997; Bugbee, Reference Bugbee2003). Thus, the nutrient supply to the plant must be closely monitored and regulated to create an optimal growth environment for the plant and to ensure high nutrient concentration in edible biomass. However, the focus on tomato production has left many gaps in the literature regarding leafy vegetables, such as amaranth.
Economic considerations
In many developing countries, initial investment costs may be the biggest barrier to the adoption of hydroponic systems. Minimizing these costs and showing that this investment will be profitable in the long run are critical to success. While many have proposed systems that are purportedly adapted to developing countries (Bradley and Marulanda, Reference Bradley and Marulanda2000; Harms and Combrink, Reference Harms and Combrink2000; Stajano, Reference Stajano2004), little has been done to evaluate their benefit-cost ratio. In order to understand the full implications of hydroponic agriculture for ALVs, the importance of economic sustainability cannot be overlooked.
Bradley and Marulanda (Reference Bradley and Marulanda2000) performed a profitability analysis but excluded the effects of time and ignored alternative options, including traditional soil-based agriculture. Others have evaluated greywater reuse systems (Godfrey et al., Reference Godfrey, Labhasetwar and Wate2009) and ornamental plant production via hydroponics (Grafiadellis et al., Reference Grafiadellis, Mattas, Maloupa, Tzouramani and Galanopoulos2000; Papadopoulos et al., Reference Papadopoulos, Chatzitheodoridis, Papadopoulos, Tarelidis and Gianneli2008) but there is still a substantial lack of literature regarding the profitability of hydroponic production of food crops for local consumption in developing countries. The adaptability of hydroponic production could be well suited to the diverse urban and peri-urban environments found in developing countries and may offer solutions not found in traditional soil-based agriculture.
Objectives
The aim of this study was to evaluate the potential for hydroponic production of vegetable amaranth in Kenya. The first part of this research evaluated the nutrient density of vegetable amaranth produced in different nutrient solutions for its potential to address micronutrient deficiencies and nutritional security. The second part of this research evaluated the economic feasibility of hydroponic production using a benefit-cost analysis of three different hydroponic systems that were constructed primarily from locally available materials in Eldoret, Kenya.
Methods
Nutrient concentration
Plant materials and growing conditions
Vegetable amaranth was used as a model crop. Seeds from variety AM38 from the AVRDC World Vegetable Center in Tanzania were sown into germinating trays and transplanted at 5 weeks into the Kratky non-circulating hydroponic systems (Kratky, Reference Kratky2003). Kratky hydroponic systems use a pot suspended over a reservoir of nutrient solution, which is depleted as the plant transpires; no pump is required for these passive systems. For our experiment, 2 L tubs were filled with 1.5 L of nutrient solution and covered with a lid. The net pot suspended from the lid reached the top 2 cm of solution and was filled with an inert clay substrate to hold plant roots (Turface® MVP clay, Profile Products LLC., Buffalo Grove, IL). Each plant was grown in a separate tub. The nutrient solution used as a base was the ‘Modified Sonneveld Solution’ for lettuce, herbs and leafy greens as described by Mattson and Peters (Reference Mattson, Peters and White2014). Plants were maintained in a greenhouse (24°C daytime/18°C night, photoperiod maintained at 12 h with supplemental lighting), and pH of 5.5–7.0. Plants were harvested 3 weeks later and biomass data collected. Plant material for zinc and iron measurements was dried in an oven at 65°C for a minimum of 3 days until weights stabilized. Fresh plant material was frozen at −20°C for carotenoid analysis. Plant material included both leaves and stems for all nutrient analyses.
Nutrient concentration evaluation
Zinc and iron analysis
Zinc and iron analysis followed AOAC official protocols 985.01 (zinc) and 999.10 (iron) (Latimer, Reference Latimer2012). Homogenized samples of dried plant material were used to create three replicates of 0.5 g each. Ten mL of concentrated nitric acid was added to each sample, which was open digested for 15 min prior to sealing the vessels for microwave digestion. Digested contents were rinsed clean with 20 mL of ultrapure water for a total of 30 mL for each sample. Elemental concentrations were determined via inductively coupled plasma (ICP) spectroscopy and quantified against standard solutions of known concentrations.
Carotenoid analysis
Carotenoid content was used as a proxy for vitamin A, as many carotenoids are precursors to vitamin A (Deman, Reference Deman1999). Carotenoid analysis followed AOAC official protocol 941.15 (Latimer, Reference Latimer2012). Frozen plant material was homogenized and 10 mL of chilled acetone was added to a representative sample of approximately 2 g (Biswas et al., Reference Biswas, Sahoo and Chatli2011). Each sample was protected from light and chilled at 0°C for 15 min and then homogenized for 2 min. Samples were then centrifuged at 1370 × g for 5 min and supernatant collected into a separate test tube. The original sample was then re-extracted as described above. Supernatants were pooled and filtered through Whatman No. 42 filter paper. Carotenoid and chlorophyll absorbances of the extracts were determined using a multi-wavelength analysis at 470, 661.6 and 644.8 nm with a UV-Vis spectrophotometer against standards of known concentrations. Carotenoid concentration was then calculated based on the equations of Lichtenthaler and Buschmann (Reference Lichtenthaler and Buschmann2001). Each sample was run in triplicate.
Experimental design and data analysis
A Definitive Screening Design (DSD) with four replicates was used for this experiment (Jones and Nachtsheim, Reference Jones and Nachtsheim2011). This design is built around six factors, each of which have been shown to be associated with plant bioaccumulation of carotenoids, iron, or zinc. These six factors are: iron, zinc, manganese (Sonneveld and Voogt, Reference Sonneveld and Voogt2011), calcium, magnesium (Paiva et al., Reference Paiva, Sampaio and Martinez1998; Fanasca et al., Reference Fanasca, Colla, Maiani, Venneria, Rouphael, Azzini and Saccardo2006) and potassium (Trudel and Ozbun, Reference Trudel and Ozbun1970; Paiva et al., Reference Paiva, Sampaio and Martinez1998; Dumas et al., Reference Dumas, Dadomo, Di Lucca and Grolier2003; Ramírez et al., Reference Ramírez, Díaz Serramno and Muro Erreguerena2011; Wang et al., Reference Wang, Wu, Ding, Zhou and Lin2013). Each factor was held at a low, medium and high level in different treatments, as shown in Table 1. The DSD allows for second-order interaction effects to be included, which is appropriate given the numerous interaction effects found between many of these nutrients (Gruda, Reference Gruda2009; Wang et al., Reference Wang, Wu, Ding, Zhou and Lin2013). It requires 13 treatments to detect and identify any significant effects among these six factors, without confounding any two-factor interactions with each other (Jones and Nachtsheim, Reference Jones and Nachtsheim2011). For each of the three response variables (zinc, iron and carotenoid concentrations) the following model is used to evaluate all main effects, two-way interactions and quadratic effects:
$$Y = b_o + \sum b_ix_i + \sum b_{ij}x_ix_j + \;\sum b_{ii}x_i^2 + {\rm \varepsilon} $$
where b i , bij and b ii are regression coefficients and ε is an error term. Stepwise forward regression was used for model selection in stages, as recommended by Jones and Nachtsheim (Reference Jones and Nachtsheim2011), first with main effects, then two-way interactions, and finally quadratic effects. The experiment was conducted as a randomized complete block design with four replicates and was repeated twice. All analyses were carried out in R 3.1.2 (Wickham, Reference Wickham2011; R Core Team, 2015).
Table 1. The six-factor, three-level Definitive Screening Design matrix for the evaluation of relative importance of selected nutrients for carotenoid, zinc, and iron accumulation in amaranth (Amaranthus cruentus).
Benefit-cost analysis
The hydroponic systems
This study analyzed the benefits and costs of three different hydroponic systems. The first was a Kratky non-circulating system (Kratky, Reference Kratky2003) similar to the ones described above, but much larger. This system is especially low-input; fixed costs include the plastic tank used as a suspension system while variable costs included the seedlings and fertilizer. The second system was a Nutrient Film Technique (NFT) system (Gruda, Reference Gruda2009), which circulated a shallow stream of nutrient solution over the exposed roots of plants. This system was slightly more input-intensive and incurred fixed costs for the materials used to build the circulating system, including a solar-powered pump. Variable costs included seedlings, fertilizer and plant growth media. The third system was a modified ebb-and-flood table for seedling production, in which a bed of course sand was periodically flooded with nutrient solution and drained (Harms and Combrink, Reference Harms and Combrink2000). This system was designed to produce seedlings for the other two systems, not to grow vegetables to a harvestable stage. Fixed costs included the tank used to hold the nutrient solution and a solar-powered pump. Variable costs included seeds, fertilizer, cloth and sand that was used as a growing medium.
The three hydroponic systems were compared with soil-based production and to purchasing vegetables at the local market. Typically, soil-based ALV cultivation does not involve any chemical fertilizer, herbicide, or pesticide application (Abukutsa-Onyango, Reference Abukutsa-Onyango2007). Fixed costs were the traditional farming implements of a hoe and rake. Variable costs were the seeds used. Purchasing vegetables from local markets accrued no fixed costs, but it was assumed that an average time of 2 h would be necessary to reach the market and return. Net benefits for all systems were calculated from yield of saleable biomass.
Estimation of benefits and costs
Both net present value (NPV) and benefit cost ratio (BCR) were used to evaluate each system and its alternatives. A time horizon of 5 years was used, given the short production cycle of these crops and the short amount of time both entrepreneurs and urban growers would need to see a return on their investment. Due to the time value of money, future values were discounted at rates of 10% and 1% to be able to compare with present values. The rate of 10% is consistent with other values in the literature (Nkang et al., Reference Nkang, Ajah, Abang and Edet2007; Obiri et al., Reference Obiri, Bright, McDonald, Anglaaere and Cobbina2007), while a rate of 1% was included for comparison. NPV was calculated on a per meter basis, given the small amount of area available to urban agriculture and backyard farming with the following formula:
$${\rm NPV} = \; \mathop \sum \limits_{t = 1}^{t = N} \displaystyle{{B_t - C_t} \over {{(1 + r)}^t}}$$
Where B t and C t are the benefits and costs during the period t, respectively. The discount rate is represented by r, t is the time period (1, 2, 3, … n) and N is the number of years. The BCR was calculated as:
$${\rm BCR} = \displaystyle{{\mathop \sum \nolimits_{t = 1}^{t = N} \displaystyle{{B_t} \over {{(1 + r)}^t}}} \over {\mathop \sum \nolimits_{t = 1}^{t = N} \displaystyle{{C_t} \over {{(1 + r)}^t}}}}$$
Data on material costs were evaluated based on field research in the town of Eldoret, Kenya. All materials were purchased locally with the exception of the solar pumps. Benefits were based on market data of the price of vegetables in the local market from samples taken regularly over the course of 5 months (Table 2). The price for seedlings and labor was based on a similar survey of local nurseries and businesses. The costs and benefits in terms of time, harvest and inputs were based on field research over two harvest cycles (Table 3). It was assumed that hydroponics would enable year-round production (12 months) while soil-based production would be limited to rainy seasons and only feasible 6 months out of the year.
Table 2. Constants and their assumed values in Kenyan Shillings (KSH) held through each comparison scenario (Note: 100 KSH = 0.99 USD).
Table 3. Time, harvest and input costs in Kenyan Shillings (KSH) for each of the three systems and their soil-based comparison (Note: 100 KSH = 0.99 USD).
Certain costs were left out of this study; the price of land was not included, since it is assumed that the land would be available to the grower for either soil-based or hydroponic production. For the purposes of this study, water was assumed to be free, though this may accrue large time costs to acquire depending on the situation. The initial startup costs were not treated as a loan with interest to be repaid, and the health benefits of consuming vegetables were not included in the scope of this paper. The benefits of having a stable food source were also not quantified, though hydroponic production would allow year-round production unlike traditional rain-dependent soil-based agriculture.
Sensitivity analyses were conducted to find breakeven points for NPV for time cost, vegetable price, vegetable harvest and input costs for each scenario. Three hypothetical scenarios were evaluated as well, the first assuming that the hydroponic systems double their yield and the price for vegetables triples. These conditions may be reflective of remote areas with poor soil-based agricultural potential (which would also increase the price for produce) and with improvements in hydroponic production. The second hypothetical scenario assumes that hydroponic production doubles, vegetable price doubles and the cost of inputs falls by 60%. This may be the case with improvements in hydroponic production in remote areas where significant amounts of the fixed costs associated with hydroponic systems can be salvaged from recycled or other free materials. The third scenario assumes that reaching the market takes six hours instead of two and that the price of vegetables doubles. This could be possible in remote areas with poor access to vegetables and other produce.
Results
Nutrient concentration
Zinc, iron and carotenoid concentrations were modeled with the stepwise forward model shown in Equation 1 (Table 4). Yield was not significantly impacted by treatment for either leaf area or fresh weight (Fig. 1). Zinc tissue concentration was predicted by high calcium, magnesium and zinc concentrations in the nutrient solution (Fig. 2a). Increasing zinc nutrient solution concentrations from 5 to 25 µmol L−1 increased mean tissue concentration by 29.5%.
Figure 1. Leaf area (cm2) and fresh weight (g) for plants grown in the 13 different nutrient solution treatments (Table 1). There were no significant differences between treatments (p > 0.05).
Figure 2. Mean concentration of (a) zinc, (b) iron, and (c) carotenoids in plants grown with low (−1), medium (0), and high (+1) concentrations of nutrient levels (see Table 1) and across experimental runs. Error bars represent ± 1 standard error from the mean and different letters above bars indicate significant differences based on Tukey HSD at p < 0.05.
Table 4. Model coefficients for the main and quadratic effects for the zinc, iron and carotenoid models with their respective P values in parentheses, as compared with the -1 level.
No interaction terms were significant. **P < 0.01, *P < 0.05.
Iron tissue concentrations were significantly affected by soluble iron and magnesium concentrations in the nutrient solution, though the effect of magnesium was quadratic. This indicates that at both low and high levels of magnesium, iron concentrations increased, which was significant when analyzed by analysis of variance (ANOVA) (Fig. 2b). Higher iron concentrations in the nutrient solution also increased plant tissue concentrations by as much as 20.4% (Fig. 2b).
Carotenoid concentrations were only significantly increased by iron concentrations in the nutrient solution, though manganese and magnesium were included in the model as well. The effects of iron were not significant when analyzed with an ANOVA (Fig. 2c) though mean carotenoid concentration increased 11.3%.
Benefit-cost analysis
Benefit-cost outcomes were consistent across the two alternatives: soil-based production and purchasing. With a 5-year time horizon, none of the hydroponic systems were profitable by either NPV or BCR measure (Table 5), but soil-based production was also unprofitable. Breakeven values for time cost indicated that if the opportunity cost was close to 0, soil-based production would become profitable. Of the three systems, the NFT system was the closest to profitability but the time cost would have to increase by 122% or the cost of inputs would have to fall close to 0 for this system to break even as compared with soil-based production with a 10% discount rate. The discount rate of 1% consistently had lower NPV and BCR values for all comparisons. When the alternative is purchasing vegetables, none of the hydroponic systems show BCR exceeding 0.60. The price of vegetables would have to increase by a minimum of 1027% or the cost of inputs would have to fall below 0 for any of the hydroponic systems to become profitable. Across all scenarios, the Kratky system consistently had the lowest BCR and NPV values.
Table 5. Net present value (NPV) and benefit cost ratio (BCR) in Kenyan Shillings (KSH) for each of the three hydroponic systems compared to soil-based production and purchasing vegetables with a 5 year horizon and 1% and 10% discount rates.
Breakeven values for time cost, price of vegetables or seedlings, harvest and inputs are also given (Note: 100 KSH = 0.99 USD).
Sensitivity analysis allowed us to explore other scenarios where hydroponic systems may be more favorable. Under a hypothetical scenario where soil-based production labor requirements doubled, the yield of hydroponic systems doubled and the price of vegetables tripled (Scenario 1), both the NFT and Ebb and Flood system become profitable when compared with soil-based production (Table 6). When compared with purchasing vegetables, only the Ebb and Flood system has a positive NPV. By both NPV and BCR measures, the Ebb and Flood system is more profitable than the NFT system, though the Kratky still is not cost-effective. Under an alternative hypothetical scenario where production doubles, vegetable price doubles, time cost doubles and the cost of inputs falls by 60% (Scenario 2), both NFT and Ebb and Flood systems remain profitable when compared with either soil-based production or purchasing vegetables (Table 6). Soil-based production and Kratky systems remain unprofitable in both of these scenarios. Under the final scenario where vegetable prices double and the time to reach the market triples, NFT and Ebb and Flood systems are profitable again when compared with purchasing vegetables. Hydroponic systems compared with soil-based production remain unchanged, but both NFT and Ebb and Flood show potential to be profitable in this scenario.
Table 6. Hypothetical net present value (NPV) in Kenyan Shillings and benefit cost ratios (BCR) for the three hydroponic systems under three scenarios for a 5 year horizon and 10% discount rate.
Scenario 1: assumes the labor for soil cultivation doubles, the yield of the hydroponic systems double, and the price of vegetables triples. Scenario 2: assumes the yield of hydroponic production doubles, the vegetable price doubles, the cost of time doubles and the cost of inputs falls by 60%. Scenario 3: assumes the price of vegetables doubles and time cost to reach the market triples.
Discussion
Our results have shown that hydroponic nutrient solutions can be manipulated to increase amaranth tissue concentrations of zinc, iron, or carotenoids, but not all three simultaneously. The conditions that increase zinc, iron and carotenoid concentrations are different (Table 4), indicating that targeting just one of these critical nutrients may be most effective in producing high-value, nutritionally-dense amaranth. Zinc and iron concentrations were most effectively improved by increasing their respective elemental concentrations in the nutrient solution, though coefficients from their models indicate that they may have an antagonistic effect on the other, as shown by Sonneveld and Voogt (Reference Sonneveld and Voogt2011). There were no significant effects on yield, which suggests that producing nutritionally dense amaranths does not require a trade-off with biomass production at the nutrient levels tested here.
Hydroponic production systems may not be financially viable in the context of western Kenya. Unlike other attempts at evaluating the economic value of hydroponic systems (Papadopoulos et al., Reference Papadopoulos, Chatzitheodoridis, Papadopoulos, Tarelidis and Gianneli2008), we used a benefit-cost analysis approach. Other benefit-cost analyses of hydroponic systems focused on high value crops such as flowers (Grafiadellis et al., Reference Grafiadellis, Mattas, Maloupa, Tzouramani and Galanopoulos2000) or melons (Shaw et al., Reference Shaw, Cantliffe, Rodriguez and Karchi2007) and found positive results when compared with other traditional production systems. By taking into account the time value of money, we were able to model opportunity cost and the time horizon necessary to be profitable, which has been excluded from other economic evaluations (Bradley and Marulanda, Reference Bradley and Marulanda2000). ALVs are not generally high-value crops, which may explain some of the more negative results shown in our models. The hydroponic yields were also much lower than yields from soil-based systems (Table 3), even with the season extension that the hydroponic systems can offer. Under conditions where their yield doubles, their value increases, or the alternative measures taken to obtain them become more costly, the hydroponic production systems become more profitable (Table 6). Under the average conditions found in western Kenya, however, none of the hydroponic systems were found to be profitable in comparison with soil-based production (Table 5). When purchasing vegetables is possible within the local area, this may be the most cost-effective option.
One interesting outcome from these results is the negative NPV and BCR values shown for soil-based production of ALVs. This changes when the opportunity cost for time is set to 0 and it becomes profitable. Farmers may not consider their own time to be worth money and given the difficulty in finding unskilled labor employment opportunities, ALV production via traditional methods will likely continue. Greater opportunity costs will improve the profitability of both NFT and Ebb and Flood systems relative to soil-based production and purchasing vegetables. Under more extreme conditions where the value of produce is higher and soil-based production is more time-consuming, the Ebb and Flood system for seedling production may also become profitable but this would have to be adapted to increase yield as well. Overall, the Kratky system was never shown to be profitable and urban residents would be better off purchasing their vegetables. Only when markets are difficult to access and vegetable prices are high does hydroponic production become the most profitable option.
Summary and Conclusions
Hydroponic production of amaranth is possible using local materials in western Kenya and has the potential to produce high-value, nutritionally-dense products via nutrient solution manipulation. These techniques could also be applied to other ALV species, though this warrants further investigation. The optimal nutrient solutions for high zinc, iron and carotenoid plant tissue concentrations will differ but it may be possible to develop specific nutrient solutions for producing vegetables to target malnourished populations.
Hydroponic production of amaranth was shown to be profitable only under alternative scenarios that do not reflect current conditions in Eldoret, Kenya. Purchasing vegetables was shown to be more cost-effective than growing them under most conditions. It would not be wise to promote hydroponic production of vegetables for all urban residents in developing countries, but only when a very specific set of conditions is met. Hydroponic production may be better suited to higher value crops, areas where input costs are lower, or where vegetable markets are distant or not available.
Acknowledgements
This work was funded by the Borlaug Fellowship for Global Food Security, the David L. Boren Fellowship, and the Clifford B. Kinley Trust.
