Saving the Colorado River Delta: How Much is It Worth?

1. Introduction

The Colorado River Delta is not only North America’s largest wetland but also one of the greatest desert estuaries in the world (Glenn et al. 1996; Vanderpool 2018). It supports diverse kinds of ecosystems like the riparian and the brackish systems (Glenn et al. 1996; MacDougal 1904). Extensive damming of the Colorado River for over a century (1892–1992) has substantially affected its flora and fauna (Carlson and Muth 1989; Hanemann2002). Agricultural, electrical, urban, and industrial sectors in nine states across the United States (US) and Mexico utilize water from the river, preventing a substantial amount of it from reaching the Delta. Acute deficiencies due to droughts in the past two decades have endangered species including the desert pupfish and Yuma clapper rail (Pitt et al.2000). These issues have affected the routes of migratory birds that are part of the Pacific Flyway and also assisted in the growth of invasive species (Medellín-Azuara et al.2007). Additionally, low water levels have affected the livelihood of local indigenous peoples, known as the Cucapa, who rely on the river system for their daily survival (Luecke et al.1999). Solutions to this problem require conservation and strategic management of water in the river.

The principal purpose of this paper is to assess cost-effective ways of regaining water for the Delta from the river system’s users. To this end, water agencies in California have been chosen as the area of study as they are the oldest and have the most extensive water rights to the river. More specifically, the agency based in the Imperial Irrigation District (IID) has Colorado River water rights dating back to the 19th century. Furthermore, it is one of the nation’s largest irrigation districts. The IID uses its current entitlement of 3.1 million MAF (million acre-feet¹) per year to irrigate roughly 475,000 ac of farmland. Additionally, its proximity to the Delta ensures there is minimal systemic water loss. This makes it cost-effective to divert a small quantity of water from this economy and to be chosen as the area of study. Results indicate that fallowing 20,000 ac of alfalfa land in the IID will yield a minimum cost of $13 million annually to conserve water for the Delta.

1.1. Background

Historically, the river’s management has been a contentious issue and only grows more complex due to the increasing annual demand for water from an increasing population (Robbins 2019a). In the 1920s, the first treaty was established to divide water amongst all the states located in the basin of the Colorado River. A convoluted set of agreements among the administrators known as The Law of the River was signed into action in 1922 (Pitt et al.2000). This compact divided the Colorado River into two parts: the Upper Basin and the Lower Basin (see Figure 1). Both regions receive 7.5MAF per year for consumptive purposes, plus 1.5MAF for Mexico, whereas the annual average water flow in the river is less than 16.4MAF (Huckleberry and Potts2019). Utah, New Mexico, Colorado, and Wyoming constitute the Upper Basin, and California, Arizona, and Nevada in the US and Sonora and Baja California in Mexico constitute the Lower Basin.

The Upper Basin delivers 7.5MAF and 750, 000 AF (acre-feet), which is half of its delivery obligation to Mexico, to Lee Ferry (where the Lower Basin starts) (Gelt1997). The Paria River, which flows one mile above Lee Ferry, has an average flow of 20,000AF (Colorado River Basin Natural Flow and Salt Data 2019; Kuhn2005). This water is counted as part of the Upper Basin and is, therefore, deducted from the total release from the Glen Canyon Dam flowing into Lee Ferry (Upper Colorado River Basin Compact 1948). The Upper Basin delivers a total of 8.23MAF of water, fulfilling its obligation.

There are several tributaries between Lee Ferry and Lake Mead, and they contribute roughly 770,000AF of water to the Lower Basin. Lake Mead, situated behind Hoover Dam, is the distribution center of water to the states in the Lower Basin (Wang and Schmidt2020). The total amount of water flowing into Lake Mead is 9MAF. The problem lies in negligence of two factors that account for water loss: reservoir evaporation from Lake Mead, and the systemic losses from the Hoover Dam to the US–Mexico border; and the Lower Basin’s water delivery obligation to Mexico. This creates a situation where the Lower Basin draws more water from Lake Mead for both its own and Mexico’s consumptive use than the amount being delivered to the Hoover Dam (Fleck2010).

In the early years after policymakers approved the Law of the River, hydrological supply was able to meet the demands from the growing cities and towns nearby. Water provisions from the river began to fall after 1998 due to drought combined with the increase in demand. The problem persists today as shown in Figure 2 (USBR 2012). States in the Lower Basin draw more water from Lake Mead than what flows into it, leading to declining water levels. Thus, the challenge with this agreement is that water from the river was over-allocated based on erroneously high estimates of the river’s water flow (Pitt et al.2000; Kuhn and Fleck2019). In reality, the river’s hydrological volume is not sufficient to provide water to all the states and Mexico.

The states have initiated several water management programs to address this issue and most recently, in January 2020, the Drought Contingency Plan (DCP) was set into motion (Federal Register2019). This plan requires states in the Lower Basin to reduce their water intake because of shortages caused by the prevailing drought. The amount of water that should be divided amongst the states depends on the elevation of Lake Mead. As a result, the river’s water flow can be managed (the Agreement Concerning Colorado River Drought Contingency Management and Operations [“Companion Agreement”], Attachment B). However, the DCP expires in 2026, meaning the authorities have time to create a new water management plan.

Several conservation projects have been initiated to protect the Delta’s ecosystems due to concerns about its water flow decline. The first successful project, named Minute 319, delivered approximately 100,000AF of water to the Delta in 2014, representing roughly 1% of the river’s annual flow. This amount was engineered to mirror the historical natural spring snowmelt that created the vast riparian forest in the Delta (Kendy et al.2017). Biologists and environmentalists studied the effects this project had on the native habitat. The river once again connected to the ocean after several decades of being dry. Local flora species increased, as did the numbers of waterbirds and riparian birds. Figure 3 shows the before and after aerial photos of the Delta region where the pulse flow was made. A remarkable difference can be seen. Subsequently, the local people, many of whom had never seen foliage in this area before, were surrounded by greenery and water (Sonoran Institute2016).

Minute 319 was a one-time pact between Mexico and the US, and political strategies were employed to convince people on both sides to conserve water for restoration purposes. It took the authorities 10 years to come up with this agreement. Decentralized methods of regulating and budgeting restoration and monitoring processes further convoluted the collaboration efforts (Kendy et al.2017). Based on the record of the partnership between the two countries, it can be assumed that a substantial effort will have to be made on both sides to come up with a long-lasting yearly accord to revive the Delta. This study focuses on analyzing the trade-offs and costs of diverting 100,000AF (derived from Minute 319) of water to the Delta for ecological restoration from the IID in California in the US.

1.2. Past literature on the Colorado River Delta

Studying water management in the arid American West has gained momentum in the past few decades as a result of a persistent drought attributable to anthropogenic causes (Park Williams et al. 2020). Jenkins et al. (2001, 2004) devised an economic–engineering model, called CALVIN, to study improvements in California’s water supply structure. The results imply that water supply efficiency might reduce scarcity and that environmental flows from different sectors have a high opportunity cost. This analysis is based in California, but the model can be used to study other regions. Medellín-Azuara et al. (2007) did so by expanding the CALVIN model to Mexico to model economic values of water in Mexicali Valley for agricultural and urban uses if a certain amount of water is contributed to the Delta’s restoration. The results indicate it would cost users of the Colorado River $50/AF (/acre-foot) to direct 32,428.6AF of water to the Colorado River Delta annually. This cost does not include the revenue lost from forgoing agricultural production, but it includes water scarcity and operating costs like pumping and treatment. A major drawback of the study is that the authors assumed trading water from the US would cost them $30/AF. No analysis was done on how this cost was derived, rendering this value unreliable.

Bark et al. (2014) focused on the economic point of ecological restoration in the Delta by studying opportunity costs, marginal analysis, and Pareto-improving compensation in building the Delta repair options. Results show that opportunity costs of irrigating water are low for small changes in water availability but there is a caveat. The short-run opportunity costs are lower if the water is transferred from one agricultural field to another, but the cost is high if the water is transferred from fields to cities. The authors noted that collecting water for a one-time event like pulse flow is affordable, but it would prove to be expensive if transferring water to the Delta were to be a regular event. This study is an amalgamation of previous researches done on reviving the Delta.

Building on previous works, this paper analyzes the trade-offs and associated costs of implementing policies that would help divert water from Southern California to the Colorado River Delta. The area of focus is the IID because it is the largest consumer of water from the river. Moreover, its proximity to the Delta means water conserved here has a higher probability of actually making it to the region without significant systemic losses. Several different crop patterns have been evaluated to understand which pattern would yield the best trade-off with minimal costs.

2. Model Development

Re-allocating water from the users of the Colorado River to the river Delta is quantified using simulations. In this study, the source of water is assumed to be the normal distribution of historic water flow in the Colorado River, devoid of any political influence or regulations. The mean and the standard deviation of the normal distribution are calculated from the United States Bureau of Reclamation’s (USBR) 24-month studies’ data. The USBR data on water released from Lake Powell and river tributaries such as the Virgin River and the Little Colorado River, that lie between Lake Powell and Lake Mead, cover the years from 1991 to 2019 (Annual Operating Plan for Colorado River Reservoirs 2019). The USBR study observes reservoir levels of the river and projects the water flow two years into the future, in an annual report called the Operation Plan for Colorado River System Reservoir (Stern and Sheikh2021). Tributaries beyond Lake Mead have not been included as their contributions are negligible (Operation Plan for Colorado River System Reservoirs 2019). Based on the data, the mean flow is calculated to be 10,123 kAF (kilo acre-feet²) and the standard deviation is 1,717kAF.

Many hydrology-based studies have predicted the volume of water on the Colorado River through the year 2050 (Milly and Dunne2020; James2020). Milly and Dunne (2020) estimated there will be a reduction in the river by 14–26% in that time. Following their results, in this study, the upper estimate of 26% has been distributed evenly throughout the time period from the year 2020 to 2050.

Using all the above values, Monte Carlo simulations are run for the water flow in the Colorado River for a period of 31 years, beginning at Lake Mead. Subsequently, this amount of water is distributed to states in the Lower Colorado River Basin, namely California, Arizona, and Nevada, and two states in Mexico (Sonora and Baja California), and the river’s Delta. This apportionment of water from Lake Mead has been made according to the DCP convention, which was enacted in January 2020. Figure 4 shows the schematic diagram for the model.

Further, an agricultural model is built inside California’s sub-model based on the IID’s specifications. This is where numerical simulation plays the role of minimizing the cost of conserving water for the Delta. The IID in California mostly uses water from the river, which totals 3.1MAF out of 4.4MAF allocated to the state. Approximately 97% of the entitlement is used for agricultural purposes, which makes it one of the top 10 agricultural regions in the country. Four crops grown in the IID (alfalfa, wheat, broccoli, and lettuce) have been chosen for this model based on land and water used by their production and profitability for the agricultural district. These four crops feature prominently in top agricultural produce in the IID. Alfalfa occupies 30.6% of total irrigated land in the IID, lettuce totals 6.5%, wheat 4.9%, and broccoli 2.7% (Water Department, Imperial Irrigation District2018). Additionally, it is of not much consequence whether the crops are annual or perennial in nature; the amount of water utilized in their production each year matters more.

For the agricultural model, the lost revenue has been calculated by multiplying the price of the crop by the quantity lost from fallowing. Evapotranspiration functions of each crop are used to calculate how much can be produced given the amount of water and acreage. Crop evapotranspiration functions give the relationship between yield and the water required for it (Allen et al. n.d.). The evapotranspiration functions for all the crops are given as follows :

Function (1) denotes the evapotranspiration function of alfalfa taken from Grismer (2001), and function (2) is for wheat, which is modeled after Roth et al. (1981). Functions (3) and (4) are the evapotranspiration functions for broccoli and lettuce, respectively, taken from Sanchez et al. (1996). The evapotranspiration function for lettuce is similar to that for broccoli because an appropriate function needed for the study of this crop specifically was not available. Also, both are cool season vegetables with similar evapotranspiration rates, making the assumption more reasonable (Johnson and Cody2015).

The functions are calibrated in the model to the same units to maintain consistency throughout the study. The amounts of water required for the production of alfalfa, wheat, broccoli, and lettuce in an acre of land one inch deep are $x_a$, $x_w$, $x_b$, and $x_l$, respectively. The amount of alfalfa produced in an acre of land is ${\mathrm{\text{yield}}}_a$, ${\mathrm{\text{yield}}}_w$ is the amount of wheat produced in an acre of land, ${\mathrm{\text{yield}}}_b$ is the amount of broccoli produced, and ${\mathrm{\text{yield}}}_l$ is the lettuce produced in an acre of land. The nitrogen required to produce wheat is $n_w$, and $n_b$ and $n_l$ refer to the nitrogen amounts required to produce broccoli and lettuce, respectively. The value for $n_w$ is taken from Roth et al. (1981) and the values for $n_b$ and $n_l$ are taken from Sanchez et al. (1996).

A production function approach of estimating water consumption in crop production is used because this research is policy-driven (Tidwell et al. 2006). The policymaker’s objective is to minimize the lost revenue from fallowing, derived from Tidwell et al. (2006), discounted at 3% for a 31-year time period :

A discount factor, denoted by $\delta$, is included in the model to show how individuals value environmental resources in the present versus in the future. The discount factor taken in this model is 3%, as suggested by NOAA (n.d.). This ensures the losses are in net-present value. Time is denoted by $t$.

The parameter ${\alpha }_i$ is the fraction that shows how much money a crop earns when 1 AF (1 acre-foot) of water is used in its production and $i=\text{alfalfa}$ ($a$), wheat ($w$), broccoli ($b$), and lettuce ($l$). This fraction helps in analyzing the objective function such that the profit of the crops is distributed equitably. This ensures the crops with a higher return per AF (per acre-foot) of water are not fallowed by a significant amount. If all the values of ${\alpha }_i$ are summed, it equals 1, i.e., ${\sum }_1^4{\alpha }_i=1$. This helps in setting up a boundary to the function to keep the losses at a reasonable value. Total acres of land covered by an individual crop are denoted by $a_i$, which aids in calculating the total loss and is the control variable for the model. Concisely, the function demonstrates the discounted value of losses from fallowing the four crops. All prices, $p_i$, have been taken from the April 2019 USDA report (USDA2019). The prices have been changed to dollars per pound for consistency and are in terms of the 2019 dollars. The choice variables are $a_a$, $a_w$, $a_b$, and $a_l$, as in, the water saved through agricultural practices relies on $a_i$.

The constraint to the objective function is the amount of water available for the Delta after it has been allocated to the four crops considered in the model :

The amount of water available to the crops is denoted by $z_t$. In this case, it is the amount of water available in the river to produce alfalfa, wheat, broccoli, and lettuce in the IID. The term after $z_t$ shows how much water is used in agricultural production in the model’s two scenarios. The amount of water the crop $i$ uses per acre is $x_i$.

Table 1 gives the values of the parameters used in the model. Using data from the detailed Crop Report of 2018 (Crop Report, IID 2018), the economic value of each crop per AF of water used in their production was calculated. Then, the ratios (${\alpha }_a$, ${\alpha }_w$, ${\alpha }_b$, and ${\alpha }_l$) for each crop were evaluated to easily compare the revenue each crop earned and set a boundary to the cost minimization expression, which was discussed earlier. As mentioned before, the values for $n_w$, $n_b$, and $n_l$ have been taken from previous literature. Values for $x_i$ have been taken from Johnson and Cody (2015).

The crops were chosen based on the revenue they earn per AF consumption of water. Broccoli and lettuce bring in substantial returns to the farmers with broccoli earning $4,903/AF and lettuce $5,250/AF. Alfalfa has a low return value per AF of water consumption ($282/AF) but has an enormous market outside the US (Culp and Glennon2012). Wheat has a substantial consumer base as well with a return of $386/AF of water consumption.

In what follows, two distinct scenarios are investigated to provide different insights into the agricultural production–water trade-off in the IID. The first scenario is fallowing, which is a common occurrence in the region. The last year of fallowing IID farms was 2017 (Fallowing Programs, IID n.d.). Different permutations of crop patterns are fallowed in the model to see which one yields the minimum cost of conserving 100,000AF of water per year. The cost is lost revenue that arises from not growing crops. The second scenario is sub-surface drip irrigation, which can be used in alfalfa production. Its high fixed cost of infrastructure largely dissuades farmers from using this type of irrigation. However, this is one of the best technologies to increase water use efficiency in agricultural practices, especially in times of droughts (Putnam2012; Hearden2017). The annual fixed cost of implementing drip irrigation is $232 per acre and it saves 1.8AF/acre/year (IID Water Newsletter2006). These data were used to find the minimum cost of saving 100,000AF of water through implementing sub-surface drip irrigation in alfalfa production using the same objective function.

3. Simulations

Numerical simulations are performed in GoldSim Academic 12.1.3 using the IID’s parametrized model over a 31-year planning horizon. In each period, the policymaker optimally selects the amount of land to fallow for each crop, $a_i$, of the total land of crop $i$, to minimize the revenue lost from doing so. This would divert 100,000AF of water annually to the Delta, which is a rounded approximation of the minimum water requirement in the Delta for a sustainable ecosystem needed yearly (USGS2016). In the initial period, the choice variables can be characterized as follows: $a_{0a}$ (the number of acres used for alfalfa production) is 155,171 ac; $a_{0w}$ (acres used for wheat production) is 24,932 ac; $a_{0b}$ (acres used for broccoli production) is 13,726 ac; and $a_{0l}$ (land used for lettuce production) is 30,194 ac. In the model, the acres are constrained to the current values for each crop. The state variable $z_t$ has an initial value of 876,640AF. Other variables can be characterized as follows: $x_a$ (the amount of water used for alfalfa production per acre) is 5AF; $x_w$ (the amount of water needed for wheat production per acre) is 1.4AF; $x_b$ (the amount of water required for broccoli production per acre) is 1.5AF; and so is $x_l$, which is the water required per acre of lettuce production (Johnson and Cody2015). The water needs are being met in each year of simulation and 100,000AF/year are diverted by fallowing pieces of land or by applying sub-surface drip irrigation.

Two of the crop patterns for fallowing are presented here to compare costs in assessing which one results in minimizing the loss of revenue:

(1)

All four crops lose acres of land. The model is restricted by holding the current cropping pattern constant to reflect the distribution of each crop in the IID in real life. The amount lost has been calculated using how much land the crops take up in the IID. This is done as opposed to dividing the water conserved equally amongst the crops to make the division proportional to the land engaged in each crop’s production. This is a more accurate representation of reality.

(2)

Alfalfa is the only one to lose acres of land. The unrestricted model yielded alfalfa as a reasonably inexpensive alternative to conserving water from the IID. Alfalfa is the least profitable among the four crops and substantial quantities of water are consumed in its production. Alfalfa was the single largest user of water in California with 5.2MAF per year in 2010 and its economic value is $175 per AF of water utilized in its production.

4. Results and Discussion

4.1. Results

Figure 5 shows the Colorado River’s simulated flow throughout the time horizon and Table 2 contains the corresponding decadal values. Water flow is simulated using the average flow to Lake Mead between the years 1991 and 2019. There is a decrease in the simulated flow as it edges toward the end of the time period. This depicts the hydrological loss owing to climate change. The average simulated flow in the time horizon is 8,201,156.25AF and the average flow from 1991 to 2019 is 10,123,000AF as calculated from USBR’s 24-month studies.

Figure 6 shows the simulations of water distributed to California, Arizona, Nevada, Mexico, and the Delta. Table 3 shows the corresponding decadal values of water distribution. It is clear from the figure and the table that there is no water left in the river for the Delta after the states get their share. Furthermore, water supply to the states decreases as years pass by. It portrays the effects of drought and reduced water level in the river due to climate change.

Figure 7 depicts the simulations of water distributed to California, Arizona, Nevada, Mexico, and the Delta, when 100,000AF of water are rerouted from the IID to the Delta. Table 4 shows the corresponding simulated decadal values of water distribution. There is a slight decrease in the water delivered to California as 100,000AF of it are rerouted to the Delta. Further, if we look at Figs. 6 and 7, we see that for certain years, the distribution of water to the states is higher than the volume of water in the river during that year. This is not a problem with the model; rather, the extra water being delivered is from what has been stored in Lake Mead from previous years.

Let us consider the first combination of crops. The graphs (Figs. 8 and 9) show lost revenues from fallowing all different types of crops and the cost of implementing sub-surface drip irrigation in lands used for alfalfa production enough to conserve 100,000AF of water. Clearly, the graphs show that it is more expensive to fallow lands in this scenario (around $70 million) than implementing sub-surface drip irrigation ($30 million). The cost of proportionate fallowing of the crops increases from $66 million in the first time period to $72 million in the last year of simulation. As for sub-surface drip irrigation, the cost increases from $12 million in the first time period to $32 million in the last time period. This is because lettuce and broccoli are highly valued in the market, and they bring substantial returns to the farmers. In 2018, broccoli had a gross value of $100 million and lettuce $200 million, albeit they occupy fewer acres of land compared to alfalfa (Crop Reports, IID 2018). Implementing sub-surface drip irrigation in alfalfa farms is more economical than fallowing expensive crops. Additionally, costs increase every year due to hydrological scarcity.

Figures 10 and 11 show how much it will cost to divert 1AF of water to the Delta in the two scenarios. As before, it will be more expensive to fallow lands — which starts from $660/AF and ends at $720/AF in the last year of simulation — than to implement sub-surface drip irrigation, which progresses from $120/AF to $310/AF (1 acre-foot) as years pass by.

Fallowing results in the loss of around 35,000 ac of crops, whereas there is no crop loss under the sub-surface drip irrigation scenario.

Now, let us consider the next combination of crops for fallowing: only alfalfa is fallowed, and the costs are compared between the two different scenarios of fallowing alfalfa lands and implementing sub-surface drip irrigation in alfalfa lands.

Figure 12 shows the cost to divert 100,000AF of water to the Delta. Compared to the cost of applying sub-surface drip irrigation as seen in Figure 9, the lost revenue from fallowing is low, starting at $5.5 million and ending with $14 million in the last time period. Drip irrigation has high installation, maintenance, and energy costs and severe pest management issues. All these external costs are not economically valuable for low-value crops like alfalfa (Blake2015). Again, as seen in Figure 12, the cost continues to rise as water in the river becomes scarcer; thus, the cost curve does not flatten out.

Figure 13 shows the marginal cost per AF of water rerouted to the Delta of fallowing alfalfa lands. It is calculated by dividing total lost revenue by the quantity of water saved. The cost of delivering under the scenario of fallowing is lower ($55–136.67) than with sub-surface drip irrigation ($120–310), as shown in Figure 10.

It is found that 20,000 ac of land must be fallowed if water conservation is entirely dependent on alfalfa production and no sub-surface drip irrigation is installed.

4.2. Discussion

Broccoli and lettuce earn more revenue per AF of irrigation water followed by wheat and alfalfa. Alfalfa has the lowest return per AF of water used in its cultivation. It is also the most water-consuming product in the IID amongst the four crops. Naturally, costs are higher when all four crops are fallowed to conserve water. The costs represent lost revenue that could have been earned if water had not been conserved. The costs of fallowing only lettuce or broccoli are even higher, and they do not yield enough water for the Delta, i.e., the water saved is below 100,000AF. They are high-priced vegetables that give large revenues to the farmers. Broccoli production earns $4,903/AF of water used in its growth and lettuce production earns $5,250/AF. So, cases where only broccoli and lettuce farmlands are fallowed, have been excluded. Figures A.1 and A.2 in Appendix A depict the costs of completely fallowing broccoli and lettuce lands, respectively. Because foregoing production of these two vegetables is not enough to conserve 100,000AF of water, few lands from alfalfa production have also been fallowed to make up for the required amount.

Next, I consider wheat and alfalfa. Even though the price of wheat is a bit lower than alfalfa, the return value of consuming 1AF of water in irrigation of wheat is higher than that of alfalfa. That is, the marginal revenue from using 1AF of water for irrigation purposes is higher for wheat with $386 and $282 for alfalfa. Thus, the case where wheat farmlands are fallowed has been dropped. In addition to this, wheat does not consume much water and its production does not occupy enough land in the IID to release 100,000AF. Fallowing wheat lands will yield little water at a much higher cost. Figures A.3 and A.4 in Appendix A depict the costs entailed in fallowing wheat lands. Figure A.3 shows the lost revenue from fallowing when alfalfa and wheat are foregone proportionate to how much land they consume in the IID. Figure A.4 shows the lost revenue from fallowing entire wheat lands and a few from alfalfa. As before, fallowing all lands of wheat does not yield 100,000AF of water. Therefore, few acres of land from alfalfa have been included to reach the desired quantity of water.

This leaves alfalfa as a suitable candidate for fallowing compared to the different crop patterns discussed in the previous paragraphs and sub-surface drip irrigation. On top of having the most acres of land dedicated to its production in the IID, alfalfa also consumes a significant amount of water. Even though alfalfa hay has a high demand due to a surge in ranching in California and China, its revenue is low at approximately $1,500 per acre (Fox2015; Culp and Glennon2012). Moreover, the marginal revenue from using 1AF of water for irrigation is low compared to other crops. Hence, it makes sense to fallow alfalfa lands only, as the cost of doing so ranges from $55/AF to $136.67/AF. This value closely matches what is observed in real life. In 2017, when the IID lands were last fallowed, the farmers were paid $175 per acre on average (IID n.d.).

It is worthwhile to mention one of the other market costs of fallowing alfalfa lands: loss of labor. According to Medellín-Azuara et al. (2015), alfalfa farms in California create a total of 5,347 jobs or 0.77 jobs per thousand AF (or 0.00077/AF) of water used in the production of alfalfa. Using this broad estimation, diverting 100,000kAF of water from alfalfa lands in the IID will yield a loss of 770 jobs annually. Unemployment due to fallowing alfalfa lands is smaller compared to fallowing other crops like grains and vegetables.

As for the second scenario, utilizing sub-surface drip irrigation in alfalfa lands for conserving water for the Delta cuts both ways. This method of irrigation not only conserves water for ecological restoration but helps increase the productivity of alfalfa. Despite its prohibitive cost, this irrigation process is gaining traction amongst Western farmers in the US due to its capacity to improve both yield and water-use efficiency (Putnam et al. 2016; Blake2015). Thus, it is not necessary to fallow lands in this situation.

Now that the costs have been analyzed, it is imperative to compare them to the benefits of water flow to the region. A broad study on the benefits of ecological restoration has not yet been undertaken. However, a rough approximation that served as the guideline for reviving the local ecosystem in the Delta was estimated early in the 2000s (Flessa2004). This estimate was derived from a previous study (Costanza et al.1997) that provided the dollar value of ecosystem services across 11 different biomes. Using the previously established number, the author gave his own estimation on the benefits derived from the Colorado River Delta’s ecosystem. Figure 14 shows the economic benefit that can result from an increased inflow of water using the aforementioned estimation. The actual value was in terms of the dollar’s value in 2004 ($208/AF). It has been adjusted for inflation to 2019 and discounted to find the present value. The benefit starts from $300/AF in 2020 to $700/AF in 2050. It is evident from these numbers that the benefits surpass the costs of rerouting water to the Delta. However, it should be noted that other costs like unemployment and food insecurity were not factored and would have otherwise raised the cost.

5. Conclusions and Policy Implications

In this study, a hydro-agricultural economic model was built to investigate water resource management in the IID in California. This agricultural turf has the senior-most rights to water from the Colorado River and thus is its biggest consumer (Culp and Glennon2012). More water has been apportioned than the river can provide, and climate change is adding to the woes of the bleeding river (Stern and Sheikh2021). As a result, the Delta of this river has dried up, leading to the endangerment of several of its natural inhabitants.

A binational agreement must be made to protect the Delta’s environment. Several papers have been published regarding how Mexico can contribute to the cause. This paper deals with how farmers in the IID, being the senior-most recipient of water from the river, can contribute to the issue. The results indicate that the most cost-effective option would be to fallow 20,000 ac of alfalfa to save 100,000AF of water for the Delta. This would cost the government authorities between $5.5 million and $13 million per year or between $55/AF and $136.67/AF per year over a 31-year planning period.

These results have important implications for policymakers. Past environmental flows to the Delta have been a conglomerated effort on behalf of environmentalists, scientists, and government officials from both sides of the international border. Water saved for this purpose has been a long-term endeavor of around 10 years, with Mexico paying most of the costs (IBWC2012; U.S. Department of State2009). This study provides new insight into ways in which the US and Mexico can secure the future of ecosystems like the Colorado River Delta. The cost of the trade-off of conserving water for the Delta shall be dealt with at the government level, similar to what has been done previously.

However, there are several limitations to this analysis. First, this model can be improved further to get inter-temporal solutions to this problem of ecological restoration using dynamic optimization. For this, more micro-level data will be required, which may prove difficult to procure. Second, more crops from the IID can be included, along with other hydrological options like a hydroelectricity model, the improved efficiency of canal infrastructure, and the urban use of water. Including other sources of water in the IID will enable policymakers to expand more resources for the Delta. Third, there has not been a comprehensive study done on the economic benefits of ecological restoration in the Delta. Policymakers must compare the costs and benefits of spending money for rehabilitation purposes to strike a deal between the nations to revive the Delta. In addition, a study must be done to see how the agricultural trade-offs and non-market costs will affect its market, i.e., the repercussions and consequences of fallowing croplands in the market (e.g., issues with food security).

Furthermore, institutional costs must be examined to understand water governance and transboundary water management between the two nations. The institutional framework of this river system is resilient to reforms and market-based water allocation (Garrick et al.2013). Such an administrative structure will influence the Delta’s ecosystem and relationship between the US and Mexico. It is difficult to put a number on such a cost, but it is imperative that it be studied extensively.

In conclusion, it is important to remember how heavily contested the river water is. The voluntary nature of transferring water takes the form of a Coasian bargain. The political side of this dilemma should be addressed in a broad agreement between the nations to obtain permanent water rights for the Delta, along with establishing environmental and economic terms. The authorities can look toward several water transfer programs in California between the agricultural and urban sectors [for instance, the QSA Agreement between the IID and San Diego County Water Authority in 2003 (San Diego County Water Authority2003)] to gauge how a deal can be struck in the future. Furthermore, the marginal opportunity cost increases as more water is transferred from the agrarian to the municipal sector, rendering a future permanent water transfer deal for the Delta an expensive one (Bark et al.2014). However, given that Minute 319 has been successful in bringing back the Delta’s ecosystem, it proves that the two countries are willing to devise new methods of water management for restoration purposes.

Acknowledgment

The author would like to acknowledge the writing assistance from Dr. Benjamin A. Jones, Dr. Janie Chermak, and John Fleck of the Department of Economics at the University of New Mexico. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author. All mistakes remain the author’s responsibility.

Notes

¹ An acre-foot equals 325,851 gal of water. It is enough to fill an acre of land, roughly the size of a football field, one foot deep.

² 1kAF $=1,000$AF.