Agricultural irrigation is the world’s largest consumer of freshwater across all economic sectors, accounting for 90% of freshwater utilized globally [1
]. However, irrigation efficiencies are low. In large open-canal irrigation systems, some 70% of irrigation water does not reach the intended crop [2
]. Nonetheless, to achieve water and food security goals in a sustainable manner, water use has to be better optimized. Moreover, to advance sustainable agriculture, approaches are needed that integrate environmental, economic, and social justice concerns [4
]; see also http://asi.ucdavis.edu
. One such approach is integrated water resources management (IWRM). IWRM is a process by which water is allocated between different uses and users in a coordinated, balanced manner [5
According to the World Bank [6
], Yemen’s freshwater resources are among the lowest per capita in the world. In addition to limited freshwater supplies, the country has a growing population. Water scarcity has been further exacerbated by government policies aimed at boosting the agricultural sector, such as low-interest loans and cheap, subsidized diesel fuel. In addition, the government’s institutional capacity to implement water laws has remained weak [7
]. Due to this combination of factors, Yemeni farmers have become increasingly dependent on groundwater for irrigation.
Since 1970, Yemeni lands irrigated from wells expanded 10-fold, from 40,000 ha to 400,000 ha in 2008, and the number of wells rose from just a few thousand to 50,000 over the same period [3
]. In addition, the number of harvesting and diversion structures has increased in upstream areas, which has led to reduced surface flows, dried-up wells, and water being lost from dam reservoirs through evaporation [8
]. As a result, groundwater levels have rapidly dropped. In some cities, such as Sana’a, wells more than 1000 m in depth are sometimes needed to access the water required [9
Agriculture is Yemen’s third largest economic sector, after services and industry (including oil). It contributes some 20% of the country’s gross domestic product (GDP), although the sector’s importance is greater than its sheer size, as agriculture provides food and income for a large segment of Yemenis, especially rural populations [6
]. Agriculture is also the country’s largest user of freshwater, accounting for 93% of freshwater consumption on an annual basis. This substantial utilization of water for agriculture raises the question of whether water is being allocated to its highest value uses and if farmers are stimulated to use water efficiently. In other words, is water being utilized in such a way as to accrue the greatest benefits to society?
Conceptually, a better understanding of the economic value of water is important to support policymaking for the development, conservation, and allocation of water across places, uses, users, and time periods [10
]. Indeed, water allocation decisions are particularly key in areas where water is scarce and there is growing demand and competition for access to water [11
]. More efficient allocation of water to those crops that can generate the greatest value for society, and the economy can help optimize basin management as well [12
]. Estimating the value of irrigation water serves not only to guide allocations of water between different crops, it can also guide the allocation of water to different production and/or irrigation techniques within a single crop type. Calculations of the economic value of irrigation water are useful both in regions with and without abundant water resources, as such calculations enable comparisons of farming profitability between regions with scarce water resources or where only rain-fed agriculture is possible versus where irrigation water is abundant. This provides an indication of the economic effect of increased irrigation [13
The current study sought to estimate the economic value of water for irrigation of a set of dominant crops. The aim of the study is to support water resource management and improve water allocation efficiency [14
]. A measure of the value of water in alternative uses can substantiate policy decisions related to the development, allocation, and use of water resources [10
]. In a market system, water’s value is represented by its price, with the price steering allocations to those uses that offer the greatest economic returns [10
This paper argues that changing groundwater levels and the freely available spate water for irrigation in the Wadi Zabid region of Yemen are propelling changes in the production value of water and hence changes in cropping patterns. Specifically, the sensitivity of the value of water to the increased depth of groundwater pumping is investigated in the midstream, downstream, and coastal areas of the study region. To that end, we posed three research questions: (1) What is the cost of pumping one unit of groundwater from different depths? (2) What is the economic value of water in terms of the production of major agricultural crops? (3) What impacts have declining groundwater levels had on the value of water for particular crops and hence on cropping patterns?
This is very novel research, as it clarifies the impact of spate water availability and groundwater extraction from different (declining aquifer) depths on economic returns to land and water for different crops, both now and in the future. Part of this study’s novelty and importance lies in the location of the study area, as Wadi Zabid, Tihama Plain, Yemen, is not a data-rich environment. Combining the various elements to demonstrate that returns to land and water are highly sensitive to changes in groundwater depths over time, and the availability of spate water for irrigation is an important advancement of current work in this field.
The current study focused on the Wadi Zabid region of Yemen. The wadi is one of the catchments of the Tihama coastal plain. The plain is considered one of Yemen’s most fertile areas. Agriculture here provides for a large part of Yemen’s food needs from cereals, vegetables, and fruits. The area is also characterized by good groundwater aquifers, which are recharged during the rainy season. The wadi originates in the western highlands of Ibb and Dhamar governorates, passes through the highlands of the Jabal Ras directorate, and continues through Al-Jarrahi, Zabid, and Al-Tuhita directorates, discharging into the Red Sea in heavy rainfall years. This study covered the plains area of the wadi, which is divided into a midstream, downstream, and coastal area (about 46 km by 20 km altogether). The center of the study region has the coordinates 317,122UTM-E and 1,564,732UTM-N (Figure 1
In the study region, agricultural land use increased from 10,509 ha in 1972 to 27,786 ha, in 2014, according to remote sensing image analysis [16
]. The population was 312,408 at the time of the latest census, which was 2004 [17
]. Using a yearly population growth rate of 2% to 3%, the population can be estimated as 430,514 in 2016 (see https://www.worldometers.info
Residents of the region depend mainly on agriculture for their livelihoods, in addition to raising livestock and fishing. Rainfall on the Tihama plain is very scarce, ranging from 100 mm per year along the coast to about 350 mm midstream in the wadi (near the foothills). Groundwater is the primary source of irrigation water, although spate flows are available in some parts of the wadi. Rules established under Sheikh Al-Jabarti more than 600 years ago gave the upper riparian area first priority for spate water use. Spate waters are divided among three groups in the midstream area of the wadi, with no water rights reserved for the downstream and coastal areas.
Wadi Zabid is known for its variety of crops, including vegetables (e.g., onions, tomatoes, okra, legumes, zucchini, hot pepper, and mulukhiyah), fruits (e.g., banana, mango, date palm, watermelon, cantaloupe, guava, papaya, and citrus), and cereals (sorghum, millet, maize, and sesame), in addition to fodder crops and crops such as cotton, Jasminum sambac, and henna.
The number of wells in the study region has increased significantly over time, from 859 in 1975 with abstraction rates of 81.7 Mm3
] to 7802 in 2008, with abstraction rates of 444.2 Mm3
2. Materials and Methods
The residual value method was applied to estimate the economic value of water, based on crop budgets at the farm gate. This method, a deductive approach, is the most commonly used technique for water valuation [20
]. Young and Loomis [21
] presented extensive information on methods for calculating the economic value of water.
Nevertheless, the current study derived the value of water based on crop budgets (as in the Appendix A
), which were represented by the net profit from crop production divided by the quantity of irrigation water applied. The net profit is the total revenue earned from a crop minus the total (variable) costs incurred for its production, including the cost of irrigation [22
The economic value of groundwater in irrigated agriculture was estimated separately for the midstream, downstream, and coastal areas of the study region. This enabled us to understand and compare the impact of declining groundwater levels on the profitability of crops in the various areas, due to the increased cost of pumping from greater depths.
2.2. Data Collection
A field survey was carried out in mid-2016 by two teams of four multidisciplinary researchers. The first team conducted questionnaire-guided interviews and discussions with farmers. The second team carried out an investigation of well water levels (measured by electronic gauge) and well yields (estimated by a hydrogeologist in the region). Changes in well depths and spate water availability were observed across the wadi. Seventy-nine farms in the midstream, downstream, and coastal areas were randomly selected to conduct the farmer interviews (Figure 1
). We expected to find differences between farms in the quantities of agricultural inputs used and outputs produced. So, the plan for the field survey was to examine in detail changes in crop budgets on each farm—that is, the sum total of the inputs used subtracted from the earnings from the crops produced on that single farm. These results would enable us to estimate the sensitivity of returns to land and water to changes in the groundwater pumping depth.
Unfortunately, during the survey, we faced various obstacles in collecting the required data. The hard daily life circumstances experienced by the farmers at the time of the field survey in mid-2016 affected their responses during the interviews. Indeed, farmers were confronted with numerous hardships associated with the unstable political situation in the country at the time. Moreover, post-harvest losses in the study area are high; this was especially due to the high temperatures, which can reach 40 °C, and the farmers’ lack of cold storage facilities.
Illiteracy was another obstacle. Half of our farmer respondents were illiterate (Figure 2
a). Therefore, many were not accustomed to precisely calculating inputs and outputs. Furthermore, the majority of the region’s farmers were smallholders. Some 56% of farms occupied 3.6 ha or less (Figure 2
b). This raised another hurdle, as some of these smallholders were discreet about sharing privileged information about the specifics of their farming operation, considering this information a trade secret. Although farmers were reluctant to talk, we were able to convince them to talk by explaining that the results can later be used to improve agricultural practices in the region. However, not all farmers gave complete answers to all the questions. They elaborated more in regard to the costs of production and costs of pumping, but gave incomplete answers regarding production and revenues.
Some of these obstacles were alluded to previously by Hellegers and Perry [24
] (p. 33):
It is important to note that such returns are difficult to compute precisely in the absence of a major modelling exercise. First, the precise technical coefficients (yield/ha, water use, etc.) will vary across farms and by year. Second, some inputs are difficult to capture accurately because they are not monetized (like family labor), or may be subject to distortions due to taxes or subsidies.
To overcome the limitations of the collected data, crops were chosen in the study region for which complete data were available from representative farms, omitting farms for which exaggerated values were collected (very high and very low figures). Out of the 79 farms from which data were collected, five representative farms were selected for calculation of the crop’s budgets and the return to land and water. Data from 50 farms of the 79 farms were used to calculate the cost of diesel to pump one cubic meter of groundwater from various depths. Nonetheless, many of the farmers interviewed were unable to specify crop production values in terms of quantity (kg) and price (US$ per kg), as they had sold the product in a different way, such as per tree, basket, carton, or bundle. Information related to weight (kg) could be gathered from a few farms (six farms out of 79 farms) and from two key informants. Thus, we calculated estimates based on the data that were provided. For example, banana production was estimated by multiplying the number of cartons produced per hectare according to the farmers by the average weight per carton. Other crop production figures were similarly obtained.
2.3. The Cost of Pumping One Unit of Groundwater
We calculated the cost of diesel to pump one cubic meter of groundwater from various depths based on data collected in mid-2016. From the results of the questionnaires and discussions with farmers, from 50 of the 79 farms, we first calculated the cost of groundwater pumping per hour (US$/hr) from a certain depth at each farm. This was done using the farmers’ responses in regarding (i) pumping depths (the number of pipes installed in the well and the length of each pipe), (ii) the number of operating hours obtained using 20 liters of diesel (this is one drum, which represents a unit of volume for diesel in Yemen), and (iii) the price of per drum of diesel. After that, the cost of pumping one cubic meter of groundwater (US$/m3) from a certain depth on each farm was calculated based on the cost of groundwater pumping per hour (US$/hr) at that depth and well yield (m3/hr).
2.4. To Calculate the Returns to Land and Water
As the study region spans a relatively small, 46 km by 20 km area, crop budgets were assumed to be similar; that is, farm-to-farm differences in inputs for and earnings from particular crops were taken to be minimal. A more significant difference was the depths from which groundwater had to be pumped for irrigation and the availability of spate water, according to farms’ location in the midstream, downstream, or coastal area. The average pumping depths for midstream, downstream, and coastal areas were obtained from the contour map of pumping depths. The map was drawn using the kriging interpolation method based on well information from the field visit, groundwater-level measurements, and the drilling depth of 248 wells, alongside well information obtained from farmers. Spate water availability was also known, which was determined by a field visit and descriptions of water distribution rules and spate water rights from the literature, as found in Tipton and Kalmbach [25
] and IIP [26
]. Spate water availability was determined as follows: (i) traditional rules dictate that only the three groups in the midstream (around the constructed weirs) had spate water rights, so the farmers downstream and in the coastal area had no spate water rights; (ii) the number of days that spate water was available annually (days/yr) was known; (iii) discussions with farmers indicated that they did not use groundwater for irrigation during the periods in which spate water was available.
2.5. The Effect of Increased Groundwater Pumping Depth on the Economic Value of Water
An Excel spreadsheet was used to build an economic model for investigating the impact of changes in groundwater pumping depths (mbss) on net returns to land (US $/ha) and the value of water (US $/m3). As noted, to calculate the value of water, we subtracted the cost of all production inputs (including the cost of irrigation) from the total income from production and divided the result by the total volume of water applied. This represents the net profit gained by farmers from each unit of water applied.
The cost of pumping one unit of groundwater differed across the study region due to differences in the groundwater level and thus in groundwater pumping depths. With increasing groundwater level depths, pumping costs will dramatically rise. Indeed, the cost of pumping represents a primary production cost for most crops in the study region. Our analysis found this cost to be of foremost significance in diminishing returns to land and water in the context of a falling groundwater table. The cost of pumping certainly plays a large role in determining cropping patterns in Wadi Zabid. Any incentive that reduces the cost of pumping—for example, an energy subsidy (e.g., for diesel fuel or solar panels)—would dramatically increase returns to land and water. Furthermore, it could be expected to initiate a change in cropping patterns toward crops with higher water requirements. Hellegers, Perry, and Al-Aulaqi [33
] observed that direct incentives to farmers in the form of high diesel subsidies and support for more efficient irrigation techniques encourage groundwater abstraction rather than reducing irrigation demand. In contrast, raising the cost of inputs, such as energy, is an effective way to reduce demand for irrigation [3
]. This is confirmed by the results of the current study, as the highest returns to land and water for the major crops were found in the midstream wadi area. The reason why is that farmers in the midstream area did not rely exclusively on groundwater for irrigation, but also had access to freely available spate water, unlike farmers in the downstream and coastal areas.
Midstream farmers would continue to make a profit from cultivating all crops, despite the expected increase in groundwater level depths. In contrast, downstream farmers were found to have the lowest economic returns to land and water, due to their total dependence on groundwater for irrigation. Moreover, in the downstream area, groundwater had to be pumped up from much greater depths: twice the depths in the midstream and coastal areas. Furthermore, no spate water reaches the downstream area.
In the downstream area, returns to land and water showed the highest sensitivity to and greatest impact of further drops in groundwater levels. In the coastal area, most crops would still continue to return some profit in the scenario of further drops in the groundwater level, except for date palm, which would become unprofitable if the current groundwater level dropped by more than 40 m. However, a further drop in groundwater levels in the coastal area would also lead to seawater intrusion. This would likely degrade water quality and risks rendering entire groundwater aquifers unfit for agriculture. Field surveys show that seawater upconing has already occurred on one date palm farm in Al-Fayza village, leading to the dying out of all the trees on the affected farm [35
Although sorghum was found to deliver the highest return to water, farmers in the midstream area of the wadi preferred banana, which also delivered a high return to water, although less than sorghum, and also delivered the highest return to land. While banana had the highest return to land, it also required application of the largest amounts of irrigation water of all the crops examined. Discussing the larger issue of the value of cash crops versus food security crops for developing countries, Achterbosch et al. [37
] indicated that as long as a balance is maintained, cash crops do have an important role in ensuring food security at both the micro and macro levels. That is because cash crops provide the income that households need to purchase other essentials required for their well-being and food security. However, these authors did note that the economic and environmental risks associated with cash crops should be guarded against and mitigated.
In regard to government incentives, Young and Loomis [21
] noted that some developing country governments have sought to keep the prices of agriculture products low in order to ensure low food prices for consumers. However, intentionally keeping prices of agriculture products low could disrupt the working of the market, diminishing the economic value of water. In fact, the current study found the opposite. Government incentives, particularly subsides on fuel, contributed to lower production costs (represented by the cost of the irrigation applied) below their costs at the global level (in a free market context). Thus, farmers earned more profit (i.e., returns to land and water were greater), and therefore, the economic value of water was higher in such cases than it would have been without such incentives. In fact, these policies encourage the expansion of agricultural lands and irrigation demand, rather than their reduction. In developed countries, production inputs such as fuel and labor are more expensive, and the prices of outputs are higher as a result.
The pumping depths were calculated based on the field measurements and information on the wells collected during the field survey and farmer interviews. The well yield data were verified by comparing them with figures from the nearest wells for which data were available from other studies in the region, the data of the well inventories of NWRA [19
] and DHV [32
]. The average was found to be within the same range, between 6 l/s and 11 l/s. The availability of the spate water for the entire midstream region was assumed in this study to be equivalent to that of the group with the highest water rights, although other groups had less water rights. Moreover, even within the different water rights groups, there were differences between farms as stipulated by traditional spate water distribution rules. In fact, there is no recent, accurate map of spate water distribution in the midstream region. Thus, there is also a diminution in the returns to land and water for the farms in the midstream. The extent of that diminishment depends on the percentage of irrigation provided by spate water on the various farms. Among midstream farms with less spate water rights, the returns to land and water would be approximately equal to the returns of the farms in the coastal areas. This is because there were no substantial differences between the two areas in depths of groundwater pumping.
The current study found that due to differences in groundwater depths, the cost of pumping one unit of groundwater was different in the midstream, downstream, and coastal areas of the study region of Wadi Zabid. Groundwater levels were found to be especially deep in the downstream area of the wadi compared to the midstream and coastal areas. Pumping costs in the downstream area were double those in the midstream and coastal areas. The continuing fall of the groundwater level here will result in a rapidly increasing cost of pumping one unit of groundwater. The highest returns to land and water were found in the midstream area of the wadi, followed by the coastal area. Returns were lowest in the downstream area because of the greater groundwater pumping depths and lack of freely available spate water for supplementary irrigation.
Throughout the study region, crops ranked as follows, from highest to lowest returns to water: sorghum (food), mango, banana, sorghum (feed), and date palm. Regarding returns to land, the ranking was banana, mango, sorghum (food), date palm, and sorghum (feed), except in the downstream area, where sorghum (feed) came before date palm. The dominant crops in the midstream area were banana, mango, and sorghum. In the downstream area, they were mango, sorghum, and to a lesser extent, date palm. In the coastal area, the dominant crops were date palm and sorghum. Banana and mango were rarely cultivated in the coastal area because of the high salinity of some wells. Sorghum, which is a drought-resistant crop that requires the lowest quantities of irrigation water, was cultivated over the entire region.
A future scenario that assumes a continuing drop in groundwater levels would have significant impact on economic returns to land and water, particularly in the downstream area of the wadi. Here, all the cultivated crops would become economically unprofitable for farmers. A falling groundwater table would have the least effect on returns to land and water in the midstream area because of the more moderate pumping depths here, as well as the free availability of spate water. In the coastal area, although the immediate impact on economic returns would also be low, further falling groundwater levels would threaten aquifer quality due to the risk of seawater intrusion.
Regarding water reallocation, this study found sorghum (food) to provide the highest return to water but only a moderate return to land. Nonetheless, this crop has social benefit (food security), and requires less irrigation water application. Sorghum varieties for both food and feed are known to be drought-resistant crops [38
]. To encourage the reallocation of water to crops with low water requirements, such as sorghum, government incentives would need to be oriented toward supporting the marketing of drought-resistant crops to assure profitable sale prices for farmers within the wadi. In addition, support for other less thirsty crops, such as peanuts and sesame, which can be cultivated with sorghum in mixed cropping, intercropping, and crop rotation systems, could enhance returns to land. Banana provides a moderate return to water and the highest return to land, but it also has a high annual requirement for irrigation water. Therefore, a groundwater balance study is recommended to further investigate the effect of banana farming on groundwater aquifers. Neighboring countries such as Saudi Arabia, according to Ouda [41
], have adopted agricultural policies oriented toward food self-sufficiency (e.g., stimulating production of wheat, vegetables, and fruit). Their encouragement and support of farmers has enabled them to achieve wheat self-sufficiency, with surpluses for export. However, those policies have also resulted in the depletion of scarce groundwater resources. Irrigation water demand increased almost threefold, from some 8 km³ in 1980 to some 22.3 km3
in 1994. The effect of excessive extraction of groundwater for irrigation is reflected in the decline in groundwater levels. In some aquifers, groundwater has declined by more than 200 meters in the past two decades [42
]. It is worth mentioning here that the reduction of post-harvest losses is a promising strategy for increasing marketable output. Post-harvest losses of fruits and vegetables reach some 50% [43
], and for cereal grains reach up to 60% [44
]. Agricultural extension offering farmers training and best practices for reducing these losses could be particularly important in regions such as the study area, where illiteracy is high and environmental conditions are harsh. Appropriate harvesting, handling, packaging, storage, and transportation can make important inroads in reducing produce losses. Indeed, preserving an existing crop constitutes a more economically and environmentally effective option than seeking to produce more agricultural produce in an area with such scarce water resources. Another policy that could be considered is e.g., the encouragement of fishing to reduce the stress on scarce water resources, especially in view of the study region’s coastal proximity. Moreover, water and food production-related policies could be reoriented toward support for the marketing of agricultural products at profitable prices for farmers rather than economic incentives that do not considerably reduce water demand. As suggested by the region’s farmers, any reallocation of water should consider the whole catchment of Wadi Zabid, including the upstream area, where cash crops are cultivated [35