See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/319995242

Water allocation and management along the

Santa Cruz border region

Article · September 2017

DOI: 10.1016/j.wre.2017.09.004

CITATIONS READS

0 47

4 authors, including:

Sanchari Ghosh Maria Eugenia Ibarraran Viniegra

Northeastern State University Universidad Iberoamericana Puebla
10 PUBLICATIONS 17 CITATIONS 38 PUBLICATIONS 235 CITATIONS

SEE PROFILE SEE PROFILE

Keith Willett
Oklahoma State University - Stillwater
39 PUBLICATIONS 155 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Energy Policy View project

Western water markets and water trading View project

All content following this page was uploaded by Sanchari Ghosh on 25 October 2017.

The user has requested enhancement of the downloaded file.

 Water Resources and Economics 19 (2017) 1–17

Contents lists available at ScienceDirect

Water Resources and Economics

journal homepage: www.elsevier.com/locate/wre

Water allocation and management along the Santa Cruz

border region
Sanchari Ghosh a, *, Maria E. Ibarraran b, Keith D. Willett c, Gerardo Sanchez Torres
Esqueda d
a
Department of Accounting and Finance, Northeastern State University, 700 N Grand Avenue, Tahlequah, OK 74464, USA
b
Department of Social Science, Universidad Iberoamericana Puebla, 72810 San Andres Cholula, Puebla, Mexico
c
Department of Economics and Legal Studies in Business, Oklahoma State University, 339 Business Building, Spears School of Business, Stillwater, OK
74078, USA
d
Department of Civil Engineering, Universidad Autonoma de Tamaulipas, 89339 Tampico, Tamaulipas, Mexico

A R T I C L E I N F O A B S T R A C T

Keywords: Water management along the transboundary Santa Cruz River basin overlapping Arizona and
Transboundary Mexico, faces a host of physical, economic and institutional challenges. The situation is worsened
Integrated approach by a failure of bilateral strategies to achieve an administratively feasible solution to these prob-
Net benefits lems. The current study, utilizing data from the Santa Cruz Active Management Area in Arizona
Sectoral water use
and from Nogales in Sonora, Mexico, builds an integrated framework of intersectoral water
US-Mexico
allocation within each region, incorporating relevant economic, physical and institutional con-
straints. Results from the baseline model and from three scenarios representing population growth
and water scarcity, suggest higher optimal water use and net benefits in the urban residential
sector for both regions. The outcomes conform to previous studies emphasizing the need for better
water supply infrastructure and reuse of treated wastewater to counter demand from a growing
population and climate induced water shortages in the region.

1. Introduction and background

Transboundary water management along the US-Mexico border has gained a prominent place in recent years due to burgeoning
population in the border cities, rapid growth in economic activities and unfavorable climatic patterns. Some of the worst ramifications of
climate variability include low precipitation, streamflow variability, abrupt changes in the timing of snowpack melt and severe droughts
in several parts of California, Arizona, Colorado and Texas as well as in the Mexican border states of Baja California, Sonora, Chihuahua,
Coahuila, Nuevo Le on, and Tamaulipas. These outcomes coupled with the high demand for water, has inevitably made it a scarce
resource in terms of quantity and often in terms of quality, as proper infrastructure for treating effluents is often ineffective and non-
existent. A fundamental problem that pervades water management along the border has been the failure of binational strategies to
arrive at a consensus on many water related issues, in some of the most affected regions of the border.
The Santa Cruz River Basin in Arizona is such a case. The basin covers 716 square miles of the Santa Cruz River Valley along the
border. The river flow originates in Arizona, flows south into Mexico, and then turns north towards the Mexican border through Nogales,

* Corresponding author. 700 N Grand Avenue, Tahlequah, OK 74464, USA.

E-mail addresses: 11.sanchari@gmail.com, ghoshs@nsuok.edu (S. Ghosh), mariaeugenia.ibarraran@iberopuebla.mx (M.E. Ibarraran), keith.willett@okstate.edu
(K.D. Willett), gsanchezt@docentes.uat.edu.mx (G. Sanchez Torres Esqueda).

https://doi.org/10.1016/j.wre.2017.09.004
Received 21 August 2016; Received in revised form 1 September 2017; Accepted 5 September 2017

2212-4284/© 2017 Elsevier B.V. All rights reserved.

S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

Sonora, and crosses back to the U.S., flowing north until it meets the Gila River. Interconnected with the basin is the Santa Cruz aquifer,
which is the main source of ground water for communities in the Santa Cruz County in Arizona (the most important being the City of
Nogales) and also for Nogales, Sonora in Mexico.
Both the cities of Nogales, Arizona and Nogales, Sonora, have witnessed rapidly growing populations. While the population of
Nogales, Sonora, increased by 50% during the 1990s and is expected to increase by 86% between 2000 and 2020, the population of
Nogales, Arizona, is expected to grow by 67% during the same time period [15,25]. Rapid population growth and almost complete
dependence upon the aquifer by the two cities have left wells severely depleted in many cases [1,10]. The regions have inadequate water
storage capacities and population growth puts more stress on water resources and infrastructure. This has led to vulnerability and water
shortages during droughts particularly for Nogales, Sonora [25]. Furthermore, low and highly variable surface water flows have affected
the amount of recharge to the aquifer as well as the maintenance of the riparian ecosystem in the upper Santa Cruz Basin in Arizona.
The flow of wastewater from Nogales, Sonora to the predominantly US sponsored Nogales International Wastewater Treatment Plant
(NIWTP) and the subsequent use of this treated water, constitutes another source of contention for this region. The establishment of a
second treatment plant in the Los Alisos Basin south of Sonora, Mexico in response, means that Mexico can now retain a portion of its
effluent and reduce its share of the cost for effluent treatment in the United States [33]. A portion of the wastewater from Nogales,
Sonora, now flows into the Los Alisos Water Treatment Plant (LAWTP) instead of being diverted for treatment entirely to NIWTP.
Treated wastewater from the LAWTP is used to recharge the Los Alisos aquifer from where Nogales, Sonora draws a portion (26%) of its
water supply. The LAWTP thus, has the potential to reduce the amount of wastewater being directed to the NIWTP, where release of
treated wastewater contributes to instream flow that supports and restores riparian vegetation in the Santa Cruz River in Arizona. It
deserves mention here that, though a portion of the treated wastewater from NIWTP helps to maintain riparian flows in Arizona, the
legal right to the water belongs to Mexico since it originates from Nogales, Sonora [42].
As is evident from the discussions above, water management along the Santa Cruz River Basin is fraught with a milieu of physical,
economic and institutional challenges, not very different from the complexities of border water management seen in the Lower Rio
Grande Basin or the Baja California region. What makes the Santa Cruz case unique, however, is the evolution of the problems over time
that has affected some of the fastest growing cities along the border, with water from the Santa Cruz aquifer serving not only a large
municipal population but also some booming industrial activities, such as the maquiladoras on the Mexican side. Researchers have
identified the interconnected problems of increasing population, rising water demand and the depleting Santa Cruz aquifer and have
advocated better coordination in bilateral water management [14,15,44]. For instance, Frisvold and Caswell [15] approach the US-
Mexico transboundary water problem as a typical Nash bargaining game, suggesting that technical support, assistance, water rights
and regulations all affect the outcomes of the negotiations. Employing three game theoretic scenarios for reduction of polluted
wastewater in the US-Mexico border, Fernandez [14] demonstrates that a Stackelberg solution leads to cheaper costs of pollution
abatement in Mexico than domestic measures.
In addition, a long standing debate (spearheaded primarily by members of the International Boundary and Water Commission
(IBWC)) has continued, on how to achieve an administratively feasible solution to the above problems, given different layers of gov-
ernance involved in regulating water use in each country. As explained in Milman and Scott [29] and Megdal and Scott [28], a missing
element here is the coordination or the absence of an institutional consensus between the two countries, given disparities in problems
faced and information asymmetries.
Over the last few decades, economists have recognized climate variability manifested through water scarcity and droughts, as one
major source of transboundary water conflicts [32]. Several studies have identified the costs and benefits of water allocation at a basin
scale, where water may be considered a scarce commodity affected by physical, economic and institutional factors [4–6,22,27,45,46].
Most of these studies conclude that an integrated water management approach is ideal in such situations, since it can specify both spatial
and temporal policy goals in terms of such factors [45]. Though the applicability of this approach has shown success in the United States
and elsewhere where climate variability is an issue, for the Santa Cruz Basin, it is yet to be explored. Some studies in this region have
conducted simulation based water demand and supply analysis considering projected growth rates in population along this border
[1,31,39,40]. Yet, optimal or efficient allocation of water and resultant benefits and costs of water allocation across sectors and regions
are not given sufficient importance. This is relevant, mainly because of the negotiations and project developments that have taken place
in recent years in order to solve water quantity and quality problems.
Our main contribution in this paper is to highlight and assess some of the economic consequences of the challenges facing water
allocation across the two border cities, under existing constraints. The paper intends to provide further insights on how the problems
facing this region may be interconnected across space, using an integrated approach that takes into account the physical and economic
constraints in water allocation and management in the region.
To this end, we first develop an integrated (hydroeconomic) optimization model of water use across the three major sectors —
agriculture, residential and non-residential (industrial and commercial) — for the Santa Cruz border region. This “baseline” model
provides a holistic assessment of the economic costs and benefits of water allocation, instrumental in driving future water management
in this transboundary peri-urban region. Second, we attempt to determine how three different scenarios namely, a shortage in water
availability, an increase in population and a combined growth in population and water shortages, influence water use, prices, and the
economic benefits and costs of water allocation for both regions. These “scenarios” are designed to compare the relative outcomes in
terms of water use, net benefits and opportunity costs of water use in each sector, under the baseline conditions and under alternative
scenarios.
The following section describes the modeling framework for integrated water management where water is shown to have competing
demands across sectors. In Section 3, we highlight specific characteristics of the study region and outline the data sources. Section 4
discusses the results from the baseline model and the scenarios and then conducts some sensitivity analysis of the main model

2
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

parameters. Section 5 briefly touches upon certain equity implications. We conclude with Section 6, delineating some of the salient
contributions of the present study and its limitations and scope for further extensions.

2. Model and methodology

An integrated water resource management framework may be interpreted as a joint maximization of net benefits, within a spatially
differentiated river basin model [4,5,43,46], so as to achieve an economically efficient (optimal) allocation of water across all uses. An
integrated water management approach is utilized in this research for two key reasons. First, as previously mentioned, studies by
Frisvold and Caswell [15], Varady et al. [44], Fernandez [14], and Bark et al. [3] have provided economic solutions through Coasian
bargaining and strategic game theoretic models, but there remains a lack of an integrated approach that incorporates institutional,
economic and hydrological constraints faced in water management. Second, an assessment of the net benefits across the major water
users in the region, along with the allocative and distributional impacts of water shortages driven by population and growth in economic
activities, has two major implications. First, it illustrates the optimal allocation of water use within sectors based upon the population
served in the border region, including the twin cities of Nogales, where sectoral water demand has shown diverse growth in recent years.
In addition, by including both physical and economic constraints, it provides an estimate of the opportunity cost of water allocation with
different levels of water availability within and across sectors in the region. In what follows, we explain the main elements of our model
—the net benefits from water use in each sector, the relevant constraints and the relationships depicting the flow of wastewater between
the two treatment plants.

2.1. Sectoral net benefits from water use

The determination of optimal net benefits from competing demands for water follows the approach taken in Ward et al. [45], Maneta
et al. [27], Howitt et al. [24], Booker and Young [4] and many others who have applied basin scale spatial optimization techniques to
assess the relative benefits and costs of water allocation.1These methods can be applied at a smaller regional level, as is done in the
current paper. Maximization of the net benefits (or economic surplus) obtained from water use across all sectors under relevant con-
straints, provides optimal levels of water use in each sector, the economic returns from water use and the opportunity costs of water
allocation within and across sectors in the region.

2.1.1. Net benefits or returns from water use in agricultural production

A deductive method of determining net returns to irrigation water, through varying input constraints or water prices in a linear
programming framework, is often adopted to evaluate agricultural water demand in the absence of actual data for estimation of an
agricultural demand function. As noted by Booker et al. [6], a major limitation of this method is the ex-ante assignment of costs of pre-
owned inputs into the model, for which Howitt [23] offered an improvement through the introduction of the positive mathematical
programming (PMP) procedure. The PMP methodology essentially allows greater flexibility than the linear constraints by inserting a
nonlinear yield or cost function for the production activities. It employs observed values for crop acreage and input and output
parameters from the base year to generate self-calibrating models of agricultural production and resource use, consistent with micro-
economic theory and accommodating variations in land quality, risk behavior and rotational constraints [23]. For determining net
benefits in the agricultural sector, we apply a simplified version of the PMP method as used in studies by Gohar and Ward [17], Dagnino
and Ward [11] and Gohar et al. [18]. (For details, refer to the Appendix).
We assume that agricultural producers maximize net economic returns to irrigation water subject to regional level constraints on
total agricultural water withdrawal. Let the subscript r represent the “region” for our study area. (If r is 1, the region is Sonora, Mexico
and if r equals 2 the region is SCAMA (Arizona)). Producers' profits π rA , or net returns from water use as a function of the amount of land
devoted to agriculture takes the following form:

X
K
π rA ¼ ½Prk yrk ðLrk Þ  Crk  Crw Lrk
(1)
k¼1
ðr ¼ 1; 2Þ

where Prk denotes the region specific crop price, yrk ¼ frk ðLrk Þ represents region specific crop yield, Crk and Crw represent region specific
non-irrigation costs of crop production and pumping costs respectively, and Lrk is the amount of irrigated land devoted to crop k in
region r: The model makes the simplified assumption that agricultural activity takes place at the regional level r; something that will be
relaxed for water use benefits in the non-agricultural sectors.
A balance constraint for irrigation water used for agriculture in region r is represented as follows:

X
K
wrAk Lrk ¼ WAr
(2)
k¼1
ðr ¼ 1; 2Þ

1
In Mexico, an example of this type of study is Soria et al. [41].

3
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

where wrAk is the amount of water applied for production of crop k per unit of land area and the variable WAr represents the total amount
of irrigation water used in region r: By the assumptions inherent in our PMP methodology, wrAk is fixed per unit of land.
Agricultural water demand in each region is constrained by the level of water available to be withdrawn annually as shown by the
following constraint:

X
K
wrAk Lrk  SA
(3)
k¼1
ðr ¼ 1; 2Þ

where, SA represents the total amount of agricultural water available for withdrawal in region “r”.2

2.1.2. Net benefits from water use in the residential and non-residential sectors
The municipal demand for water is divided into residential and non-residential demand, the latter comprising commercial and
industrial water demand. The determination of net benefit functions for residential and non-residential water use follows the point
expansion method of water use valuation [20,26]. The point expansion methodology with a mathematical programming model has been
widely used in studies on urban water demand [6,45]. It relies on a given elasticity estimate for urban residential/non-residential use
and a point estimate on the marginal benefit function (observed value of a quantity of water and its corresponding marginal value or
price) for a baseline year. In the absence of consistent historical data to estimate urban water consumption for residential and non-
residential purposes, this method serves as a suitable technique for our analysis.
Suppose, water demand in each sector “i” is represented by a linear marginal benefit function having the following form:

b ai  b
pi ¼ b b i wi

where, b a i and b
p i and wi denote the sectoral level price of water and water use respectively and b b i represent the intercept and the slope
parameters (refer to the Appendix for derivation of b b
a i and. b i Þ:
A linear marginal benefit schedule implies a nonlinear quadratic total benefit function for water use for each of the residential and
non-residential sectors as follows:

w*  
TBI ðwi Þ ¼ ∫ 0 i bai þ b
b i wi dw

The net benefit from water use in each sector is then expressed as the joint maximization of the consumer and producer surplus (total
benefits) after subtracting the total costs of water delivery for each sector in each region.
For the residential sector ðRÞ; the net benefit takes the following form:

X
C
 R R   R 
NBRr ¼ Br wcr  Ccr
R
wcr
(4)
c¼1
ðr ¼ 1; 2Þ

where BRr ðwRcr Þ refers to the total benefit from residential ðRÞ water use for each city “c” in region “r” and Ccr
R
ðwRcr Þ denotes the total cost of
residential water delivery for the same in region r: Similarly, the net benefit for the nonresidential sector ðNRÞ takes the following form:

X
C
  NR   NR 
r ¼
NBNR BNR wcr  Ccr
NR
wcr
r (5)
c¼1
ðr ¼ 1; 2Þ

where the terms BNRr ðwcr Þ and Ccr ðwcr Þ have the same interpretations as above, except that they denote nonresidential water use in
NR NR NR

each city of region. r:

As in the agricultural sector, a balance equation aggregates all of the nonagricultural water use (WrRNR Þ in the regions specified in our
model.

X
C X
C
wRcr þ cr ¼ WrRNR
wNR
(6)
c¼1 c¼1
ðr ¼ 1; 2Þ
Finally, constraints are set on the total amount of water available ðSRNR Þ for withdrawals for both residential and nonresidential
purposes in each region r is as follows:

2
This constraint and the one below (equation (7)) are incorporated based on email conversations with ADWR officials and our coauthors from Mexico. According to
email conversations with Pam Muse from ADWR, these are flexible in many cases, for example, farmers often go over their pumping limit and charge the excess use on
their “flexibility account The constraints for Nogales, Sonora are obtained from CONAGUA [57] and INEGI [56].

4
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

X
C X
C
wRcr þ cr  SRNR
wNR
(7)
c¼1 c¼1
ðr ¼ 1; 2Þ
Note that these constraints are usually determined by the water regulatory bodies in each region, based on historical water use and
water rights. So, a total water demand and availability constraint for all three sectors may be expressed by the following equation:

X
K X
C X
C
wrAk Lrk þ wRcr þ cr  S
wNR (8)
k¼1 c¼1 c¼1

where S denotes the total water availability at the sectoral level in region r.

2.2. Waste water relationships

The next step in our model is to relate the flow of wastewater from the non-agricultural sectors to the treatment plants located in both
regions. First, it is assumed that a constant proportion of non-agricultural water use gets converted into wastewater every year. Sec-
ondly, of the total amount of wastewater generated from each region, a certain proportion flows to each wastewater treatment plant. In
order to preserve the tractability of the model, the proportions are considered predetermined or exogenous to the model, but may be
endogenized by changing the cost of treatment and the optimal choice of treatment facilities/plants.
The wastewater flow relationship may be expressed as

WrW ¼ δWrRNR
(9)
ðr ¼ 1; 2Þ
where WrW represents the amount of wastewater flow generated in region “r” and “δ” denotes the proportion of residential and non-
residential water that is converted to wastewater flow in each region. In addition, for Nogales, Sonora, a proportion of the wastewater
flow is sent to either the NIWTP or the LAWTP or to both. Let θ ð0  θ  1Þ denote the proportion of the wastewater flow from Nogales,
Sonora being transferred to LAWTP and WLAW denote the total amount of wastewater flow to LAWTP. Then we have:

WLAW ¼ θW1RNR (10)

The wastewater flows to the NIWTP originate from both Nogales, Arizona and Nogales, Sonora. Let parameter α denote the pro-
portion of the wastewater flow from Nogales, Arizona that is transferred to the NIWTP. Since all wastewater from Nogales, Arizona flows
to NIWTP, α ¼ 1.
Thus, the total wastewater flow to NIWTP (WNW Þ takes the following form:

WNW ¼ ð1  θÞW1RNR þ W2RNR (11)

It may be noted that the wastewater treatment costs are included while maximizing the net benefits from water use and are not
explicitly shown to affect the annual level of flow to each treatment plant.
Although wastewater flows are part of the model constraints, the benefits of treated wastewater for the Santa Cruz River in Arizona,
are not included since under current regulations, Mexico has the legal right to the treated water flowing from Nogales, Sonora. Also, lack
of data precluded the option of incorporating the amount of treated wastewater affecting the non-agricultural water use in Nogales,
Sonora.
The model objective may then be expressed as:

X
2
 
B¼ π rA þ NBRr þ NBNR
r (12)
r¼1

where B represents the net monetary benefits from the three sectors and is maximized subject to equations (2) and (3) and (6)–(11)
above.

2.3. Baseline model

The baseline model maximizes the net monetary benefits (B) subject to the constraints noted above. The baseline model is repre-
sented as:

X
2
 
Max B ¼ π rA þ NBRr þ NBNR
r (13)
r¼1

subject to equations (2) and (3) and (6)–(11) above. The baseline model provides outcomes in terms of water use, prices or marginal net
benefits from water use, net benefits or returns from each sector and the amount of wastewater flow from each region. Each of the

5
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

scenarios described below provides variations in the base outcomes as water availability changes and/or demand for water changes.

2.4. Scenario 1: Drought related water shortages

For Scenario 1, we hypothesize drought related water shortages leading to reduction in water availability. This is represented by
reducing the total amount of water available to be withdrawn in each region as depicted by equation (14) below:

X
K X
C X
C
wrAk Lrk þ wRcr þ cr  ρS
wNR (14)
k¼1 c¼1 c¼1

where “ρ” refers to the proportion of water available under varying levels of water shortages. For instance, ρ ¼ 0:7 for a water supply
reduction of 30% during a drought.
So for Scenario 1, the optimization model is modified as follows:

X
2
 
Max B ¼ π rA þ NBRr þ NBNR
r (15)
r¼1

subject to equations (2), (3), (6), (7), (9), (10), (11), and (14).

2.5. Scenario 2: Population growth

One of the major drivers of increased water demand in the Santa Cruz border region has been the high growth rates of population
experienced particularly in Nogales, Mexico. As has been reported in Scott et al. [39], the population of Nogales, Sonora, has been
experiencing an average annual growth rate of 1.6% over the last decade, while the Santa Cruz County itself showed about 1.3% average
growth rate during the same time period. The above figures are used to build a population growth rate scenario to assess the impact upon
water demand and net benefits from water use in each sector over 2025. The population for 2025 (Pop) is projected from the baseline
population levels for 2010, for each of the twin cities of Nogales as follows:

Pop ¼ popbase ð1 þ gÞt

where, popbase refers to the population in 2010 for each city and g denotes the respective projected growth rates over the time period t.
Growth in urban population will potentially influence water demand in the residential sector and the parameters of the net benefit
function in this sector, which modifies the net benefit function for each region as follows:

X
C
 0R  R   
NB0R
r ¼ Br wcr  CrR wRcr
(16)
c¼1
ðr ¼ 1; 2Þ

r ðwcr Þ represents the new residential benefit function under population growth in Nogales, Sonora and in Nogales, Arizona.
where, B'R R

So, the optimization problem for Scenario 2 takes the following form:

X
2
 
MaxB ¼ π rA þ NB0R
r þ NBr
NR
(17)
r¼1

subject to equations (2) and (3) and (6)–(11).

2.6. Scenario 3: Population growth and drought related water shortages

The final scenario assumes that both Nogales, Arizona and Nogales, Sonora, experience increase in population similar to Scenario 2
and at the same time, face drought related water shortages (Scenario1). This affects the water availability constraints in the model as well
as the net benefit functions in the residential sector in each region. For scenario 3, the optimization problem may be expressed as:

X
2
 
MaxB ¼ π rA þ NB0R
r þ NBr
NR
(18)
r¼1

subject to equations (2), (3), (6), (7), (9), (10), (11), and (14).

3. Study region and data

For the purposes of our analysis, the study region consists of the Santa Cruz Active Management Area (SCAMA) that covers a portion

6
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

of the border region in Arizona overlaid by the Santa Cruz (SC) aquifer, and the City of Nogales in Mexico. On the Arizona side, we
include three major cities within SCAMA —Nogales, Tubac and Rio Rico —which are served by the Santa Cruz aquifer, while Nogales,
Sonora, is served by the SC aquifer, the Los Alisos (LA) Aquifer and to a certain extent by the Nogales aquifer,3 on the Mexican side of the
border (Fig. 1).

Fig. 1. Study region for Arizona.

For SCAMA, agricultural activities demand the largest percentage of water (52%) followed by municipal (40%) and industrial
activities (8%). As noted previously, the municipal and industrial sectors are disaggregated into residential and non-residential sectors.
For Nogales, Sonora, urban water use for residential and non-residential sectors constitutes 78% of total water consumption, with
agriculture and livestock accounting for the rest.
According to current SCAMA regulations, groundwater and surface water are conjunctively managed, hence no distinction is made in
terms source of water. Agricultural water is drawn from irrigation grandfathered rights, while non-agricultural water (water for resi-
dential and non-residential purposes) is drawn from Type I and II rights, domestic wells (exempt) and service area rights.4 An irrigation
grandfathered right confers the right to irrigate specific plots of land that had been irrigated with groundwater between 1975 and 1980.
A Type I right is associated with land permanently retired from farming and converted to a non-irrigation use, e.g., building a new
industrial plant or a subdivision. A Type II right is based on historical pumping of groundwater for a non-irrigation use and equals the
maximum amount pumped in any one year between 1975 and 1980. Unlike grandfathered and Type I rights, groundwater may be sold
under Type II rights without the associated land (Arizona Department of Water Resources (ADWR)).5
We introduce two modifications in our dataset. Due to paucity of data on the actual amount of water drawn and costs of pumping
from domestic wells and Type I and Type II rights, it is assumed that a substantial portion of non-agricultural water is served by private
water companies which have access to service area rights. Service area rights authorize cities, towns, private water companies, and
irrigation districts to withdraw groundwater to serve their customers (ADWR). The four largest water providers— the City of Nogales,
Valle Verde Water Company, Rio Rico Utilities, and Tubac Water Company— provide 60% percent of water in the four municipalities of
SCAMA, with the City of Nogales providing 52% of municipal water need. Second, though a large percentage (86%) of industrial (non-
residential) water is used by turf related facilities (owned by golf courses like the Rio Rico and Tubac golf courses) utilizing Type II
rights, benefits from such water use cannot be readily determined. Thus, we include their estimated demand for water [1], leaving out
details regarding their costs and demand functions.6
For Nogales, Sonora, water services are more centralized in nature with water use regulated by the National Water Commission of
Mexico (CONAGUA), which is the source of our data for sectoral level demand for water. However, a large portion of water is drawn
from privately owned wells [7,16], which usually go unreported (since they are out of service connections) and so it was not possible to
impute values to such water use.
Conjunctive management of water in SCAMA adds a layer of complexity to the governing rules for allocating water rights, since some

3
The SC aquifer serves 72% of the water needs for Nogales, Sonora.
4
All of these rights and permits have a water allocation associated with them [1].
5
Apart from irrigation grandfathered rights, Type I and Type II rights and the service rights mentioned below, withdrawal permits may allow new withdrawals of
groundwater for non-irrigation purposes.
6
Determination of turf related benefits from water use is not straightforward, due to these users exercising their own pumping rights (from self-supplied wells) which
do not require reporting under current groundwater management rules.

7
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

rights used for the same purpose may have claims to both groundwater and surface water. So, ADWR has specified some annual
allotment of water for all grandfathered, Type I and Type II rights, including groundwater rights used for irrigation of golf courses.
Unless exempted from reporting their water use (small domestic wells or when water is drawn from self-owned pumps), these allotments
are strictly enforced under the 100 year Assured Water Supply rules of SCAMA. Data for the maximum annual allocations for each sector
are obtained from ADWR [1] and from individual surface water provider reports. For Nogales, Sonora, the constraints are derived from
the total amount of water licensed for use by each sector based upon reports published by CONAGUA [57] and INEGI [56].
The parameters measuring the proportion of wastewater going from each region to the treatment facilities are drawn from published
literature [7,33] and the annual municipal water use report for 2010 submitted to the Arizona Department of Water Resources (ADWR).
For instance, it is estimated that 70% of the municipal water (residential and nonresidential water consumption) reaches the sewer
system for the cities of Nogales, Sonora and Nogales, Arizona [7]. It is further assumed that while the total amount of wastewater
generated from Nogales, Arizona reaches the NIWTP, for Nogales, Sonora, 60% of the wastewater reaches NIWTP and the rest is treated
at the newly constructed LAWTP [32].

3.1. Economic parameters

Though most of the economic parameters pertain to the baseline year of 2010, some prices and costs are adjusted to reflect the base
year values.7 In the agricultural sector, parameters like crop level prices, land in production and input costs as obtained from the Food
and Agricultural Policy Research Institute (FAPRI), United States Department of Agriculture (USDA-NASS) and the National Institute of
Statistics and Geography (INEGI), are used for calibrating the functional forms in Section 2 to observed data in each region.8 While the
Mexican data is reported for 2010, for SCAMA, data from a three-year average yield and three-year average of land acreage devoted to
irrigated crops (USDA-NASS) are used for the initial calibration. The baseline calibration produces values for the slope coefficient and
the intercept term for the quadratic crop yield or response function defined in Section (2.1.1).
Sectoral level average consumption of water for the baseline year is available from ADWR for the SCAMA region and from INEGI for
Nogales, Sonora. Individual utility company data is used to derive the marginal water prices (or rates) for the baseline year. Since in
most cases, the rate structure varies with the type of water use (residential or commercial), we utilize previous data on volumetric
consumption to compute water rates charged in each sector.
To obtain a definite point elasticity of demand for water, we explore recently published estimates for residential and non-residential
water demand for each region. These estimates are chosen based upon similarities in socio economic and demographic characteristics
between the sampled populations in each study and our study region. For Nogales, Arizona, we select the elasticity of demand for water
from Yoo et al. [48], while studies by Grafton et al. [19] and Garcia-Salazar et al. [16] provide pricing and elasticity estimates for
residential water consumption in Mexico. The elasticity estimates for non-residential water use for Nogales, Arizona and Nogales,
Sonora, are obtained from studies by Renzetti [36] and Garcia Rojas [21]. respectively.
The selection of the residential demand elasticity from a study conducted in Phoenix, is motivated by two factors. One, is the lack of
published data on elasticity of residential water demand for specific cities in Arizona. Two, though the study by Yoo et al. [48] is based
on Phoenix, the elasticity of demand for residential water is measured for households in different neighborhoods as classified by
demographics and income. Hence, these estimates provide a rough approximation of the water demand elasticity for the study region in
Arizona.
As shown in Table 1 below, the Santa Cruz aquifer supplies water to SCAMA, while Nogales, Sonora, draws water from three aquifers
— Santa Cruz, Los Alisos and Nogales — with the largest percentage of water (72%) drawn from the Santa Cruz aquifer. Institutional
regulations dictate that a certain percentage of the water available from these aquifers in each region may be extracted for agricultural
and non-agricultural activities. For SCAMA, values for total water permitted to be drawn for the above activities are taken from ADWR
[1], conforming to Assured Water Supply (AWS) rules for Active Management Areas. For Nogales, Sonora, the values for maximum
amount of water allowed to be extracted annually (water licenses) are obtained from CONAGUA [57]. Note that these values pertain to
annual water availability for a normal year and are not outcomes of any treaty obligation between the two regions.

Table 1
Sources of water supply in the study region.

Estimated stock of water Sonora (cu m)

SCAMA (cu m)

Santa Cruz 48,845,808 30,288,447

Los Alisos 0 10,713,181
Nogales 0 1,129,732

Source: Arizona Department of Water Resources (ADWR) and CONAGUA

7
For instance, some sectoral level water prices and costs are available for 2009 or 2008 and the values are adjusted to 2010 dollars.
8
While yield data is obtained from USDA Agricultural Census, data for crop acres comes from the USDA Cropland Data Layer and is validated through the five-year
Agricultural Census data for Arizona.

8
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

4. Results and discussions

The results for the baseline (or benchmark) optimization model and the three scenarios are summarized in Tables 2–4. Tables 2 and 3
describe the optimal water use and the prices in the three sectors, while Table 4 illustrates the net monetary benefits for each sector. The
baseline model assumes water availability and population levels in both regions are at their 2010 values and replicates empirically the
baseline model of Section 2.3. As shown in the first two columns of Table 2, optimal water use in the residential sector in SCAMA seems
to fall short of the non-residential sector use by 18%. In Nogales, Sonora, water use in the residential sector is found to exceed that of the
non-residential sector by 23%, with residential water use accounting for 54% of the total sectoral water use. Also, as evident from
Table 4, the net benefits from water use in the residential sector in Nogales, Sonora, is found to exceed that of the non-residential sector
by 60%. This indicates that in the baseline year, there is high demand for residential water use and higher marginal valuation of water in
that sector. Total water used and net returns from agriculture is found to be low in both regions. The optimal levels of wastewater flow
from both Nogales to the NIWTP (Table 5 below) conforms to estimated levels of wastewater flowing to each treatment plant from the
two border regions in 2010.
Two things may be noted about the agricultural water use and net returns from this sector. First, agriculture seems to have the lowest
net returns as well as the lowest optimal level of water use in both regions.9 Second, with changes in water availability and demand,
water seems to have been reallocated from the two non-agricultural sectors, with little variation in agricultural water use. Thus, in the
following subsections, results for the residential and non-residential sectors are analyzed, while a discussion of the factors behind this
noticeable lack of response of irrigation water use is reserved for a later overview of the results.

4.1. Scenarios

4.1.1. Scenario 1: Climate variability and drought

The Santa Cruz Basin on both sides of the border experiences an arid climate with an average annual precipitation of around 14.3
inches over Nogales and Rio Rico in Arizona (National Climate Data Center) and wide variation in streamflow (ADWR). In Nogales,
Sonora, severe water shortages stemming from expanding population is compounded by prolonged heat waves, low precipitation (less
than 5 inches on an annual average—Prichard et al. [33]) and variability in surface water flows. The latter often results in low levels of
groundwater recharge especially during dry seasons, even when over pumping from wells reduces the average depth of water [10]. Since
surface water and groundwater are hydraulically connected, variability in streamflow translates into variability in groundwater levels
through reduction in recharge and tributary underflows [1,9,40].
In this scenario, we consider drought related water shortages as manifestations of climate variability in the study region. High
average temperature and low precipitation are both negatively correlated with streamflow in the Santa Cruz River Basin and low surface
water flows affect the total stock of groundwater.10 As streamflow recharge varies during drought, so does the amount of water available
in the aquifer.
To estimate the variability in streamflow recharge during drought, we draw data from stochastic recharge levels during dry periods
[1]. These stochastic recharge values are determined through simulating natural streamflow recharge over the period 2011–2025,
utilizing historical observations on recharge values over 1985–2010. An upper and lower level of confidence around the mean recharge
value are constructed for Nogales, Arizona, indicating upper and lower bounds of variability in recharge under dry conditions. This
variability translates into a reduction of around 30% of water availability for the SCAMA region. We call this a moderate drought
situation. A severe drought situation is indicated by a 50% cutback in water supply. For Nogales, Sonora, however, without access to
stochastic recharge data, groundwater availability from the Santa Cruz (SC) aquifer is varied through changes in the recharge levels of
the Santa Cruz aquifer on the Mexican border by 30% and 50%, respectively.11 Our assumptions are not implausible under conditions of
severe drought when groundwater depth falls below normal levels on both sides of the border [1,10,33].
Thus, Scenario 1 assumes maximization of sectoral net benefits subject to water availability constraints encapsulated by equation
(14). Scenario 1 in Tables 2–4 illustrate the changes in outcomes on sectoral water use and net benefits in each region relative to the
baseline values. For instance, when water availability is lowered by 30% (moderate drought), optimal levels of water use in the residential
sector falls by 17% and 14% in SCAMA and in Nogales, Sonora, respectively, while water use in the non-residential sector reduces by
51% in Nogales, Sonora. These results suggest that when water supply is scarce, the intersectoral reallocation of water is determined by
the use with higher marginal value in both regions. These marginal values are reflected by the higher prices of water in each sector, with
optimal prices in the residential and non-residential sectors in SCAMA increasing by almost 26% from their baseline values. Under
conditions of severe drought, optimal water use in the non-residential sector in Nogales, Sonora, falls by 84% while net benefits are
reduced by 79% relative to their baseline values. Severe water shortages result in an optimal price increase of 69% and 24% in SCAMA
and in Nogales, Sonora, respectively.

9
One limitation of utilizing fixed water per unit of land is underestimating the intensive margin of agricultural water use, which may have an impact upon the results.
10
In SCAMA there is no provision of mitigating low water storage with access to renewable surface water like in the Tucson Active Management Area (TAMA) or other
active management areas having access to CAP water supplies.
11
This approach conforms to Scott et al. [39] who state “The generally shallow Mexican portion of the aquifer means that recharge variability translates into water
supply variability on an annual basis” (pg.167).

9
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

Table 2
Water use in the various sectors for the base scenario and the three alternate scenarios.

Baseline Scenario 1 Scenario 2 Scenario 3

Moderate drought Severe drought

SCAMA MX SCAMA MX SCAMA MX SCAMA MX SCAMA MX

(cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m)

Agriculture
Water use 2,433,804 1,373,614 2,433,804 1,373,614 2,433,804 1,373,614 2,433,804 1,373,614 2,433,804 1,373,614
Perc. Change from base
Residential
Water use 6,481,908 16,105,764 5,391,598 13,811,637 3,569,643 12,282,218 7,370,270 17,874,457 5,796,031 15,957,172
Perc. Change from base 17 14 45 24 14 11 11 1
Non-residential
Water use 7,654,850 12,334,251 6,410,832 6,096,374 4,332,030 1,937,790 7,687,591 10,565,558 6,006,400 3,950,839
Perc. Change from base 16 51 43 84 0.43 14 22 68

Note: SCAMA refers to the Santa Cruz Active Management Area while MX refers to Nogales, Sonora in Mexico. Scenario 1 represents drought related water shortages in
both regions, with a moderate drought associated with a 30% reduction in water availability and a severe drought denoting a 50% reduction in water availability.
Scenario 2 represents projected population growth in Nogales, Arizona and Nogales, Sonora. Finally, Scenario 3 refers to a moderate drought situation with population
projections similar to Scenario 2.

Table 3
Prices in the various sectors for the base scenario and the three alternate scenarios.

Baseline Scenario 1 Scenario 2 Scenario 3

Moderate drought Severe drought

SCAMA MX SCAMA MX SCAMA MX SCAMA MX SCAMA MX

($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m) ($/cu m)

Residential
Prices 0.62 0.82 0.78 0.94 1.05 1.02 0.62 0.93 0.82 1.01
Perc. Change from base 25.91 14.56 69.21 24.27 0.00 12.81 32.24 22.40
Non-residential
Prices 0.62 0.82 0.78 0.94 1.05 1.02 0.62 0.93 0.82 1.01
Perc. Change from base 25.91 14.56 69.21 24.27 0.00 12.81 32.24 22.40

Note: SCAMA refers to the Santa Cruz Active Management Area while MX refers to Nogales, Sonora in Mexico. Scenario 1 represents drought related water shortages in
both regions, with a moderate drought associated with a 30% reduction in water availability and a severe drought denoting a 50% reduction in water availability.
Scenario 2 represents projected population growth in Nogales, Arizona and Nogales, Sonora. Finally, Scenario 3 refers to a moderate drought situation with population
projections similar to Scenario 2.

4.1.2. Scenario 2: Population growth

The second scenario considers increases in population in Nogales, SCAMA and Nogales, Sonora, as described in Section 2.6 above. As
expected, growth in population in both Nogales, drives up the demand for water in the residential sector in the regions by around 14%
and 11% respectively. However, as demand increases in the residential sector in Nogales, Sonora, optimal water use falls by 14% in the
non-residential sector. This may be explained by an intersectoral reallocation of the average annual amount of water available for use in
the two sectors. With population growth, a high demand from one sector (residential) raises the opportunity cost of water in the other
(non-residential) sector.12 Thus, as seen from Tables 3 and 4, optimal price (marginal value) of water goes up by 12.81% and net benefits
in the residential sector by 20%, relative to their baseline values in Nogales, Sonora. By contrast, in SCAMA, prices are similar to their
baseline values suggesting that higher residential demand may be satisfied by the average annual availability of water. We find an
increase in optimal water use in the non-residential sector to the extent of 0.43%, which is a spillover from high urban demand for water
in general, as a fallout of the growth in population.

4.1.3. Scenario 3: Population growth and water shortages due to climate variability
The final scenario considers changes in both population and in water availability, as captured by a moderate drought (a 30%
reduction in water availability—Scenario 1) and by similar growth rates of population as seen in Scenario 2. Interestingly, non-residential
water use in SCAMA reduces by a large extent (22%) from the baseline value, with optimal prices under this scenario increasing by 32%
in this region. This may be explained by water supply shortfalls affecting the non-residential sector, with higher population having
already created an upsurge in water demand in the urban residential sector. The optimal net benefits in the residential sector in SCAMA
shows an increase of 6% from the baseline values, as demand and high prices raises the valuation of water in this sector, though total
water use goes down. For Nogales, Sonora, optimal levels of water use and net benefits in the non-residential sector are reduced by 68%

12
The opportunity cost when nonresidential water supply becomes a limiting constraint with an increase in population in Nogales, Sonora is found to be $381.14.
Similar shadow prices also emerge from the model implementation for Scenarios 1 and 3, but they are not reported in the paper due to brevity. These are available from
the corresponding author upon request.

10
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

Table 4
Net benefits/returns in the various sectors for the base scenario and the three alternate scenarios.

Base Scenario 1 Scenario 2 Scenario 3

Moderate drought Severe drought

SCAMA ($) MX ($) SCAMA ($) MX ($) SCAMA ($) MX ($) SCAMA ($) MX ($) SCAMA ($) MX ($)

Agriculture
Net benefits $242,649 $37,562 $242,649 $37,562 $242,649 $37,562 $242,649 $37,562 $242,649 $37,562
Perc. Change from base
Residential
Net benefits $3,380,066 $10,052,970 $3,292,575 $9,448,508 $2,755,864 $8,892,814 $3,741,839 $12,099,200 $3,584,679 $11,431,100
Perc. Change from base 3 6 18 12 11 20 6 14
Non-residential
Net benefits $3,956,617 $3,972,068 $3,856,792 $2,328,502 $3,244,419 $817,536 $3,588,459 $3,930,704 $3,420,622 $1,625,741
Perc. Change from base 3 41 18 79 9 1 14 59

Note: SCAMA refers to the Santa Cruz Active Management Area while MX refers to Nogales, Sonora in Mexico. Scenario 1 represents drought related water shortages in
both regions, with a moderate drought associated with a 30% reduction in water availability and a severe drought denoting a 50% reduction in water availability.
Scenario 2 represents projected population growth in Nogales, Arizona and Nogales, Sonora. Finally, Scenario 3 refers to a moderate drought situation with population
projections similar to Scenario 2.

and 59% respectively from their baseline values. This is not unexpected. As Scott et al. [39] opines, higher demand for water as
population expands in this area, may be increasingly served by non-aquifer sources like treated wastewater, a provision that is not
considered in our present framework. Thus, as water demand increases in the residential sector along with moderate shortfalls in supply,
scarce water is reallocated from non-residential use towards residential urban use. At the same time, higher residential water demand in
Nogales, Sonora, drives up prices by 22% and results in residential net benefits increasing by 14% relative to their baseline values.
Table 5 illustrates the optimal amounts of wastewater generated from each region and flowing to the two treatment plants and it is
evident that water supply shortfalls have reduced the amount of wastewater flow relative to the baseline. For example, with a severe
drought (Scenario 1) there is a 50% reduction in the amount of wastewater flows to NIWTP and LAWTP from Nogales, Sonora. For
Nogales, Arizona, lower water use in the residential sector reduces the amount of wastewater flowing to NIWTP by 44%. Again, a
reduction in water supply along with a growing population (Scenario 3), leads to a 24% decrease in the total amount of wastewater
flowing to NIWTP from Nogales, Sonora, and from Nogales, Arizona. For the two regions separately, Scenario 3 leads to a 17% decrease
in wastewater flows from Nogales, Arizona to NIWTP and a 30% reduction in flows from Nogales, Sonora to each of the two treatment
plants, NIWTP and LAWTP.

Table 5
Annual wastewater flows in the base scenario and the three alternate scenarios.

Base Scenario 1 Scenario 2 Scenario 3

Moderate drought Severe drought

NIWTP LAWTP NIWTP LAWTP NIWTP LAWTP NIWTP LAWTP NIWTP LAWTP
(cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m) (cu m)

SCAMA 9,895,729 8,261,701 5,531,171 10,540,506 8,261,701

MX 11,944,811 7,963,199 8,361,366 5,574,244 5,972,399 3,981,599 11,944,811 7,963,199 8,361,366 5,574,244
Total 21,840,540 7,963,199 16,623,067 5,574,244 11,503,570 3,981,599 22,485,304 7,963,199 16,623,067 5,574,244

Note: SCAMA refers to the Santa Cruz Active Management Area while MX refers to Nogales, Sonora in Mexico. Scenario 1 represents drought related water shortages in
both regions, with a moderate drought associated with a 30% reduction in water availability and a severe drought denoting a 50% reduction in water availability.
Scenario 2 represents projected population growth in Nogales, Arizona and Nogales, Sonora. Finally, Scenario 3 refers to a moderate drought situation with population
projections similar to Scenario 2.

Several interesting discussions emerge from the above results. The outcomes seem to closely follow the pattern observed in some
recent studies on this region. For instance, Scott et al. [39] conclude that a possibility of a growing population with no additional sources
of water supply on the Arizona side of the border, may be reflected by reallocation of water from agriculture and reuse of treated
wastewater, often interpreted as a demand side management of water [1,39]. While, for Nogales, Sonora, the impacts of a larger
population and water supply variability will be felt through higher residential demand, expansion of economic activities and the need
for more water supply infrastructure (The LAWTP being one such infrastructure catering to the needs of a growing population in
Nogales, Sonora). The results also point towards population growth in Nogales, Sonora being responsible for high water prices within
the non-agricultural sectors. This is an economic as well as a social concern and we take up discussions pertaining to such social costs in
the next section.
Yet, there is the lack of a robust impact upon agricultural water use for the baseline and each scenario, which may be explained by
three factors. In both regions, agricultural water use has been stable over the years [1,57] and it has been shown that with future growth
in water demand in SCAMA, consumption of agricultural water has not been affected to a large extent (ADWR, 2012). Secondly, the
irrigation grandfathered water rights that govern agricultural water use in SCAMA, dictate that only land certified to use such rights for
historic agricultural acres can be irrigated, with no new land put to production [1]. The same applies to Nogales, Sonora, where

11
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

agricultural activity is relatively marginal, with water being scarce and low value crops being grown in the region. In that case, land in
agriculture is bound by historical average acres in both regions, an artifact of the existing water rights as well as the empirical model.
Finally, our economic modeling for agriculture has assumed that water use is fixed on the intensive margin, crop prices are constant and
only land use varies with changes in total water availability. (This assumption is plausible on a regional scale). Thus, with initial low
levels of water use, agriculture has witnessed very low opportunity costs of water shortages due to moderate or severe droughts or due to
population growth in the urban sector. The selection of land and crop variety has not also responded to changes in water availability to
any noticeable extent (these are not presented in our main results).
A related concern is the reallocation of water among the non-agricultural sectors, even with severe water shortages. Though water
markets for agricultural producers exist in Nogales, Sonora, lower prices due to low-valued crops grown, have failed to encourage water
leasing or transfers especially with climate variability and droughts [57]. On the other hand, Active Management Area rules in SCAMA
have tied water rights with land such that irrigation water rights may not be sold without the associated land [1].

4.2. Sensitivity tests

In the baseline model, price elasticities of water demand were selected in order to reflect the demographic composition of the
population. So for instance, with a population with low per capita income in Nogales, Arizona and in Nogales, Sonora, demand elas-
ticities higher than unity were taken. However, in order to determine how the results are sensitive to lower price elasticities, the
elasticity coefficient for Nogales, Sonora was lowered to 0.44 [13]. In the baseline model, we find slightly higher benefits for the
residential sector in the region which is not unexpected, as lower elasticity implies people are less likely to be sensitive to price changes.
While results for the three scenarios are similar to what are obtained before, for Scenario 3, we find higher water use relative to the
baseline for the residential sector in Nogales, Sonora. It may be inferred that water shortages fail to prevent higher water demand in this
sector as consumers become less responsive to higher prices (lower elasticity of demand).
We also modified the number of non-residential units in Nogales, Sonora based upon estimates drawn from INEGI [56]. For the
baseline scenario as well as for the other scenarios, we find a reduction in net benefits for the non-residential sector in Nogales, Sonora.
Interestingly, a larger number of non-residential units lowers the total amount of water used in the non-residential sector as prices go up.
The only exception is Scenario 2, where an increasing population keeps demand for water in the non-residential sector higher than the
baseline. In short, this implies that growth in the commercial and/or the industrial sectors and net returns from these sectors are
negatively affected by water scarcity and population growth.

5. Equity issues

Because of data constraints equity issues have not been considered in the models, but discussion of equity implications hold priority
in such transboundary water management. As stated above, Nogales, Sonora, is an urban area with a population close to 300 thousand
—almost ten times that of Nogales, Arizona—and with a projected population growth higher than that of the latter. Yet, per capita water
consumption in Nogales, Arizona is around four times than that of Nogales Sonora. This in itself, illustrates economic and social dis-
crepancies across these two cities [38]. A reduction of 14% in optimal residential water use in Nogales, Sonora, due to a moderate
drought, is likely to hit the lower income groups the most because of their lower ability (and capability) to pay. It may force consumers to
purchase water from private providers at a high cost.13 Thus, the impoverished section of the population could face adverse circum-
stances as the water scarcity situation gets worsened by population growth and increasing water demand.
Another significant issue that has direct impacts upon efficiency and equity on both sides of the border, but one that is highly
debated, is the status of wastewater flows. Currently, Mexico sends two thirds of its wastewater to the NIWTP in Arizona (costing Mexico
around $200,000 per year), with treated effluent from the plant supporting the riparian corridor north of the border. While Sprouse [42]
summarizes some possible management options for this wastewater that could potentially benefit both countries, access to clean water
still eludes most communities of Nogales, Sonora [33] perpetuating inequality. There have been discussions around the possibility of
Mexico treating its own wastewater at the LAWTP, which may save the country significant treatment costs and allow it to keep its water
for residential or other uses. This could also help Mexico reap some benefits since the value of water to Mexico is larger than its
treatment costs [37]. However, as documented by Prichard and Scott [34], interbasin water transfers from the Los Alisos Basin to
Nogales, Sonora, have affected irrigators as well as residents in the adjoining regions, raising concerns about its long term feasibility. In
short, the protracted issues surrounding the treatment and use of wastewater in the region, have impacts upon both equity and efficiency
that would need to be studied further.

6. Conclusions

Transboundary water resource management is a complex problem and as has been demonstrated in several river basins spanning
the US- Mexico border, it is usually exacerbated by a host of physical, economic and institutional factors. The Santa Cruz River Basin
overlapping the twin cities of Nogales in Arizona and in Sonora, Mexico, epitomizes such a situation, where most policy level studies
have emphasized upon better coordination and bargaining mechanisms for border water management. The current study, utilizing
data from the Santa Cruz Active Management Area (SCAMA) in Arizona and from Nogales, Sonora, builds an integrated optimization

13
Water delivery trucks are commonly seen in less better-off neighborhoods in this region as noted by our Mexican coauthors.

12
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

model of water use across sectors to determine some of the economic outcomes of water allocation in this region, under existing
constraints. Results from the baseline model and from three scenarios representing population growth and water scarcity, suggest
higher optimal water use and net benefits in the urban residential sector for both regions. For Nogales, Sonora, in particular, severe
droughts are shown to potentially lead to water scarcity and large reduction in benefits in the non-residential sector. These outcomes
conform to conclusions reached by some studies, that high population growth amidst climate variability in Nogales, Sonora, have
culminated into water shortages for most communities. The outcomes also seem to support the rationale for the existing interbasin
water transfers from the Los Alisos Basin to Nogales, Sonora. In response to population growth and resulting water demand in Nogales,
Sonora, treated wastewater has been transferred from LAWTP since the last decade (though at the expense of communities residing
adjacent to the Los Alisos Basin).
With moderate and severe droughts and with population growth, water use is generally observed to be reallocated from the non-
residential to the residential sectors, indicating higher marginal valuation of water in the residential sector. This has important
implications for policy makers who should plan demand management according to where water is valued most, especially when there
are shortages. Even if there is no inter country trading of water, intersectoral reallocation of water is possible to counter the water
scarcity in one sector. Otherwise, residential water users may end up paying a higher price for water when demand is high due to
drought related shortages and with population growth and drought. A relevant example is the change in block rate structures that are
implemented by utility companies to counter water scarcity or to promote water conservation during droughts.
There are several limitations of the current study. The study makes a very simplifying assumption that commercial and industrial
users have similar elasticities of demand for water. Also, due to lack of data on self-owned well pumping, the study confines itself to
water supplied by largest providers in each region. Some of the pumping, treatment and water delivery costs for Mexico are adjusted
based upon data from Arizona. Last, but not the least, the study falls short of modeling benefits from treated wastewater for both regions,
a topic that is beyond the scope of the present analysis but a direction for future research. A dynamic modeling of the stock of water and
assessing the impact of the scenarios considered, is also an avenue for further research.
Wilchens [47] and Nakao et al. [30] have noted that water management and adaptation is one of the most important mechanisms to
make border regions viable and sustainable in the medium and long run. The current research fills an important gap in the work done on
the Santa Cruz border region, by underscoring the need for integrating economics with simulation based studies for addressing water
management policies for the future. This is one of the first studies to attempt an integrated approach for water resource management in
the Santa Cruz River Basin and results from this study may be extended by including other parts of the basin, subject to data availability.
Nevertheless, it adds to economic policy debates centered around transboundary water allocation, where the interlinkages amongst
institutional regulations, socio-economic and demographic factors and physical scarcity often play a major role in determining the
efficiency and efficacy of water management.

Acknowledgements

Funding: This work was supported by Oklahoma NSF (IIA-1301789) EPSCoR, ROA þ grant.
The authors would like to thank graduate students — Cristina Pastrana Lopez and Edgar Luis and undergraduate student, Charles
Coffman— for providing valuable research assistance during the course of this study. The authors also extend their gratitude to the
reviewers for their insightful comments and suggestions which have helped in vastly improving the manuscript.

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.wre.2017.09.004.

Appendix

Derivation of the slope and intercept parameters for the agricultural profit function

In accordance with the Ricardian theory of rent where crop yield is subject to diminishing returns to heterogeneous land quality, we
assume the following crop response function [11].

yield ¼ B0 þ B1 *land (A1)

where, crop yield (yieldÞ declines with expansion of land acreage ðlandÞ . This specification assumes a fixed ratio of irrigation water
being applied per unit of land, such that the total water applied per crop equals the proportion of water applied per acre (Bw Þ times the
total land acreage. Thus, land takes on the role of the dependent variable as given by

Water
Land ¼ (A2)
Bw
Substituting the above expression into the yield equation (1), we obtain

13
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

 
Water
yield ¼ B0 þ B1 * (A3)
Bw
The net profit level of the agricultural water user, may be expressed as a function of the land in production in the following manner

π ¼ ðPk *yield  CÞ*Land  Pw *water (A4)

where, Pk is the crop price, C denotes the non-irrigation costs of production and Pw is the price of irrigation water.
Substituting the expression for crop yield from (3), we may rewrite the profit level in the agricultural sector as:
  
Water
π ¼ Pk * B0 þ B1 *  C ðWater=Bw Þ  Pw *water (A5)
Bw
The only unknown parameters in the above function are the intercept ðB0 Þ and slopeðB1 Þ.
This last expression in (A5) when differentiated with respect to water applied, synopsizes the classical microeconomic theory of
water use being expanded until the value of the marginal product of water equals the price of water. It forms the basis of our baseline
PMP calibration, which recovers the estimates of B0 andB1 , based on observed data on the known parameters:

B1 ¼ ½ðPk *yieldÞ =ð Pk *landÞ (A6)

B0 ¼ ½yield  B1 *land  (A7)

Derivation of the slope and intercept parameters for residential and non-residential benefit functions

Let i ði ¼ 1; 2Þ denote the category or sector of water use (residential or non residential), wi the quantity of water used in sector i, and
MBi ðwi Þ the marginal benefit function for water use, in sector i. We assume that the marginal benefit function or the inverse demand for
water is linear, so that

MBi ðwi Þ ¼ ai þ bi wi (A8)

This marginal benefit function in the residential sector is affected by population and household size while in the non-residential
sector, it may be influenced by growth in industrial and commercial activities.
We also note that

dMBi ðwi Þ
¼ bi (A9)
dwi
Let the price elasticity of demand for water coefficient be denoted as εi . This elasticity coefficient is defined as

dwi MBi ðwi Þ

εi ¼ (A10)
dMBi ðwi Þ wi
Using equation (A9) to rewrite equation (A10), we obtain the following:

1 MBi ðwi Þ
εi ¼ (A11)
bi wi
The parameters ai and bi may now be determined by the point expansion process. First, we can solve equation (A11) for bi , which is
the slope of the sectoral water demand function

1 MBi ðwi Þ
bi ¼ (A12)
εi wi
Let the exogenously determined value for the elasticity coefficient be denoted as εi , the known value for wi be denoted as wi and the
corresponding marginal benefit of water be MBi ðwi Þ. We can substitute these values into equation (A12) to compute the slope parameter,
b
b i : The value of the intercept (ai Þ can then be obtained by substituting the value of bi into equation (A8) as follows:

a i ¼ MBI ðwi Þ  b
b b i wi (A13)
Once these slope and intercept parameters are calibrated using actual data for water demand, prices or rates per unit of water
delivered and costs of water distribution/delivery, they are inserted back into equation (A8). Finally, the total benefit function is
determined by the area under the marginal benefit function corresponding to the quantity of water consumed in sector i. Let TBi ðwi Þ
represent total benefits of water consumption for sector i: Integrating the marginal benefit function (A8), generates a quadratic total
benefit function for water use in sector i as follows:

14
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

*
wi
 
TBI ðwi Þ ¼ ∫ b ai þ b
b i wi dw (A14)
0

Fig. A1. Streamflow data for the Nogales region of SCAMA for 1980–2014.

Fig. A2. Average precipitation in inches in SCAMA over 1980–2012.

Fig. A3. Mean temperature in SCAMA over 1980–2012.

15
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

Fig. A4. Simulated recharge data for the Santa Cruz aquifer in SCAMA over 2012–2025.

References

[1] Arizona Department of Water Resources, Draft Demand and Supply Assessment 1985-2025 Santa Cruz Active Management Area, 2012, pp. 1–130.
[3] R.H. Bark, G. Frisvold, K.W. Flessa, The role of economics in transboundary restoration water management in the Colorado River Delta, Water Resour. Econ. 8
(2014) 43–56.
[4] J.F. Booker, R.A. Young, Modeling interstate and intra state markets for Colorado river water resources, J. Environ. Econ. Manag. 26 (1994) 66–87.
[5] J.F. Booker, Hydrologic and economic responses of drought under alternative policy responses, Water Resour. Bull. 31 (5) (1995) 889–906.
[6] J.F. Booker, R.E. Howitt, A.M. Michelsen, R.A. Young, Economics and the modeling of water resources and policies, Nat. Resour. Model. 25 (2012) 168–218.
[7] Camp Dresser and McKee, Wastewater characterization and flow projections, Prog. Rep. Int. Bound. Water Comm. (1997) 1–50.
[9] K.M. Cobourn, Externalities and simultaneity in surface Water-Groundwater systems: challenges for Water rights institutions, Am. J. Agric. Econ. 97 (3) (2015)
786–808.
[10] F. Corkhill, L. Dubas, Analysis of Historic Water Level Data Related to Proposed Assured Water Supply Physical Availability Criteria for the Santa Cruz Active
Management Area: Santa Cruz and Pima Counties, Arizona, 2007. Modeling Report 18.
[11] M. Dagnino, F.A. Ward, Economics of agricultural water conservation: empirical analysis and policy implications, Int. J. Water Resour. Dev. 28 (2012) 577–600.
[13] M. Espey, J. Espey, W.D. Shaw, Price elasticity of residential demand for water: a meta-analysis, Water Resour. Res. 33 (6) (1997) 1369–1374.
[14] L. Fernandez, Transboundary water institutions in action, Water Resour. Econ. 1 (2013) 20–35.
[15] G.B. Frisvold, M.F. Caswell, Transboundary water management: game-theoretic lessons from projects in the US-Mexico border, Agric. Econ. 24 (2000) 101–111.
[16] J.A. García-Salazar, E.G. Soria, M.F. Hernandez, Water demand and distribution in the comarca lagunera, mexico, Agrociencia 40 (2006) 269–276.
[17] A.A. Gohar, F.A. Ward, Gains from expanded irrigation water trading in Egypt: an Integrated basin approach, Ecol. Econ. 69 (2010) 2535–2548.
[18] A.A. Gohar, F.A. Ward, A.A. Amer, Economic performance of water storage capacity expansion for food security, J. Hydrol. 484 (2013) 16–25.
[19] R.Q.T. Grafton, H.T. Kompas, M. Ward, Residential Water Consumption: a Cross-country Analysis, 2009. Environmental Economics Research Hub Research
Reports, Research Report No. 23.
[20] R.C. Griffin, Water Resource Economics: the Analysis of Scarcity, Policies, and Projects, The MIT Press, Cambridge, MA, 2006.
[21] Guerrero Garcia Rojas, H, Industrial Water Demand in Mexico: Econometric Analysis and Implications for Water Management Policy, Universite de Toulouse 1-
Sciences Sociales, 2005. Pour le Doctorat en Sciences Economiques.
[22] J.J. Harou, J. Medellín-Azuara, T. Zhu, S.K. Tanaka, J.R. Lund, S. Stine, M.A. Olivares, M.W. Jenkins, Economic consequences of optimized water management for
a prolonged, severe drought in California, Water Resour. Res. 46 (2010) 1–12.
[23] R.E. Howitt, Positive mathematical programming, Am. J. Agric. Econ. 77 (1995) 329–342.
[24] R.E. Howitt, J. Medellín-Azuara, D. McEwan, J.R. Lund, Calibrating disaggregate economic models of agricultural production and water management, Environ.
Model. Softw. 38 (2012) 244–258.
[25] H. Ingram, D. White, International boundary and water commission: an institutional mismatch for resolving transboundary water problems, Nat. Resour. J. 33
(1993) 153–176.
[26] D.L. James, R.R. Lee, Economics of Water Resource Planning, McGraw-Hill, New York, 1971.
[27] M.P. Maneta, M.O. Torres, W.W. Wallender, S. Vosti, R. Howitt, L. Rodrigues, L.H. Bassoi, S. Panday, A spatially distributed hydroeconomic model to assess the
effects of drought on land use, farm profits, and agricultural employment, Water Resour. Res. 45 (2009) 1–19.
[28] S.B. Megdal, C.A. Scott, The importance of institutional asymmetries to the development of binational aquifer assessment programs: the Arizona-Sonora expe-
rience, Water 3 (2011) 949–963.
[29] A. Millman, C.A. Scott, Beneath the surface: intra national institutions and management of the United States—Mexico transboundary Santa Cruz aquifer, Environ.
Plan. Gov. Policy 28 (2010) 528–551.
[30] M. Nakao, D. Wichelns, J. Montgomery, Game theory analysis of competition for groundwater involving el paso, in: Paper Presented at Moving with the Speed of
Change,” the 2002 Annual Meeting of the American Agricultural Economics Association in Long Beach, CA, July, Texas and Ciudad Juarez, Mexico, 2002,
pp. 28–31.
[31] L.M. Norman, M.L. Villarreal, R. Niraula, T. Meixner, G. Frisvold, W. Labiosa, Framing scenarios of binational water policy with a tool to visualize, quantify and
valuate changes in ecosystem services, Water 5 (2013) 852–874.
[32] Sheila M. Olmstead, Climate change adaptation and water resource management: a review of the empirical literature, Energy Econ. 46 (2014) 500–509.
[33] A. Prichard, C. Scott, P. Vandervoet, S. Megdal, Drought and urbanization: water supply challenges of Nogales, Sonora, in: Arizona Hydrological Society Sym-
posium, September, 2010, pp. 1–4.
[34] A. Prichard, C. Scott, Interbasin water transfers at the US–Mexico border city of Nogales, Sonora: implications for aquifers and water security, Int. J. Water Resour.
Dev. 0 (2013) 1–17.
[36] S. Renzetti, The Economics of Water Demands, Kluwer Academic Publishers, Boston, 2002.
[37] R. Sanchez, Water conflicts between Mexico and the United States: towards a transboundary regional water market, in: E. Brans, E. de Haan, A. Nollkaemper,
J. Rizema (Eds.), The Scarcity of Water: Emerging Legal and Policy Responses, Kluwer Law International, London, 1997.
[38] R. Sanchez, F. Lara, Manejo Transfronterizo del Agua. Estrategias y opciones para Nogales Sonora, El Colegio de la Frontera Norte, Nogales, 1992.
[39] C.A. Scott, S. Megdal, L.A. Oroz, J. Callegary, P. Vandervoet, Effects of climate change and population growth on the transboundary Santa Cruz aquifer, Clim. Res.
51 (2012) 151–170.
[40] E. Shamir, S.B. Megdal, C. Carrillo, C.L. Castro, H.I. Chang, K. Chief, F.E. Cokhill, S. Eden, K.P. Georgakakos, K.M. Nelson, J. Prietto, Climate change and water
resources management in the upper Santa Cruz River, Arizona, J. Hydrol. 521 (2015) 18–33.
[41] E.G. Soria, J.A.G. Salazar, M.F. Hernandez, Demanda y distribuci on del agua en la Comarca Lagunera, Mexico, Agrociencia 40 (2) (2006) 269–276.

16
S. Ghosh et al. Water Resources and Economics 19 (2017) 1–17

[42] T.W. Sprouse, Equitable management of mexican effluent in ambos Nogales”, J. Southwest 45 (4) (2003) 595–610 (Winter).
[43] M. Torres, R. Howitt, L.N. Rodrigues, Modeling the economic benefits and distributional impacts of supplemental irrigation, Water Resour. Econ. 14 (2016) 1–12.
[44] R.G. Varady, C.A. Scott, M. Wilder, B. Morehouse, N. Pineda- Pablos, G.M. Garfin, Transboundary adaptive management to reduce climate change vulnerability in
the western U.S.–Mexico border region, Environ. Sci. Policy 26 (2013) 102–113.
[45] F.A. Ward, J.F. Booker, A.M. Michelsen, Integrated economic, hydrologic and institutional analysis of policy responses to mitigate drought impacts in Rio Grande
Basin, J. Water Resour. Plan. Manag. 132 (6) (2006) 488–506.
[46] F.A. Ward, M. Pulido-Velazquez, Economic costs of sustaining water supplies: findings from the Rio Grande, Water Resour. Manag. 26 (10) (2012) 2883–2909.
[47] D. Wilchens, The role of “virtual water” in efforts to achieve food security and other national goals, with an example from Egypt, Agric. Water Manag. 49 (2001)
131–151.
[48] J. Yoo, S. Simonit, A.P. Kinzig, C. Perrings, Estimating the price elasticity of water demand: the case of Phoenix, Arizona, Appl. Econ. Perspect. Policy 36 (2014)
333–350.

Web references

[49] American Community Survey Available at: http://www.census.gov/acs/www/

[50] Arizona Department of Water Resources. Available at:http://www.azwater.gov/azdwr/
[51] Food and Agricultural Policy Research Institute. Available at:http://www.fapri.org/
[52] International Boundary and Water Commission. Available at:http://www.ibwc.state.gov/home.html
[53] National Agricultural Statistics Service Cropland Data Layer. Available at:https://www.nass.usda.gov/Research_and_Science/Cropland/SARS1a.php
[54] National Agricultural Statistics Service. Available at:http://www.nass.usda.gov/
[55] National Climate Data Center. Available at:http://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets/climate-normals/1981-2010-
normals-data
[56] National Institute of Statistics and Geography. Available at:http://www.inegi.org.mx/
[57] National Water Commission of Mexico. Available at: http://www.conagua.gob.mx/home.aspx
[58] National Population Council of Mexico. Available at:http://www.conapo.gob.mx/
[59] Santa Cruz Active Management Area. Available at:http://www.azwater.gov/AzDWR/WaterManagement/AMAs/SantaCruzAMA/default.htm

17

View publication stats