Quantitative Analysis of Correlation between US Army Installation Characteristics and Water Price

Hur, Andy Y.; Garfinkle, Noah W.; Ploschke, Christine M.; Guest, Jeremy S.; Chini, Christopher M.

doi:10.1061/JWRMD5.WRENG-6189

Open access

Case Studies

Oct 18, 2023

Quantitative Analysis of Correlation between US Army Installation Characteristics and Water Price

Authors: Andy Y. Hur, Ph.D. https://orcid.org/0000-0003-4572-905X [email protected], Noah W. Garfinkle [email protected], Christine M. Ploschke [email protected], Jeremy S. Guest, Ph.D. [email protected], and Christopher M. Chini, Ph.D., A.M.ASCE [email protected]Author Affiliations

Publication: Journal of Water Resources Planning and Management

Volume 150, Issue 1

https://doi.org/10.1061/JWRMD5.WRENG-6189

PDF

Abstract

This study applied statistical approaches (e.g., Spearman’s rank correlation coefficient, regression tree analyses) to characterize Army site-specific factors (e.g., water demand, installation size, climate zone, Army mission type, utility privatization, water source, population, installation status on mission capacity) correlated with Army installation water unit price (

USD \cdot {kgal}^{- 1}

). The results from Spearman’s rank correlation coefficient for individual factors showed that annual water consumption, size of installation, and population were the major influencing factors (positive correlation) to annual water billed. For Army installation water unit prices, however, negative correlation with annual water consumption and less significant correlation with climate zone were observed. This could lead to a failure in promoting water conservation in water-stressed areas with limited water supplies. From the results of regression tree analyses with combination of characterized variables, installation mission type, type of primary water source, and assured access to water were statistically significant factors to Army installation water unit prices. The regression trees provided coarser but actionable insights while clustering water unit prices by the influencing factors. The results of this study support site-specific reconsideration of water pricing and further development of installation water security and resiliency through deeper understanding of factors correlated with installation water unit price. Also, this research adds to existing studies on water infrastructure system characteristics, specialized use cases and water price across scales and locations at the United States.

Practical Applications

Implementations of emerging water conservation technologies are often challenging due to the low market price of water, which could lead to a limited cost savings relative to the investment. While disruptions to water supply can affect many communities and businesses, understanding of factors correlated with water price may provide implications to promote water conservation efforts especially in water stress regions. This study explores significance of site-specific factors correlated with US Army installation water price to support prescriptive modifications to water pricing and further development of water security and resiliency.

Introduction

In conventional cost-benefit evaluation methods for water, return on investment is poor in many regions due to low market price of water (Olmstead and Stavins 2009; Rasoulkhani et al. 2018). This issue may limit adoption rates for other capabilities (i.e., high-cost water conservation technologies) required to support emergency operations of mission critical facilities (e.g., server cooling systems) where the willingness-to-pay is much higher than the existing costs (Department of the Army 2014). Potable water prices within US Army installation are typically lower than that of residential, commercial, and industrial uses in the civilian sector. In 2016, for example, the average potable water price among Army installation was

2.75 $ \cdot {kgal}^{- 1}

(

0.73 $ \cdot m^{- 3}

), 18% lower than the US average price for water (Department of the Army 2016a; DOE 2017). For evaluation of water project investments, total costs and benefits of implementing the water conservation capabilities should be captured and compared with the alternative water supply options. The types of water markets supplying Army water users are mainly categorized into self-supplied, government owned contract operated, purchased water distributed through government owned infrastructure, and purchased water distributed through privatized infrastructure. In general, prices charged to Army water end users are based on the market costs of collecting, treating, and distributing water (Department of the Army 2018). However, the price variabilities across Army installations cannot be fully explained by these factors. Therefore, the price of water may not fully capture the true value or uncertainties of water, especially when considering non-market valued factors (Grafton et al. 2020). For example, mission critical buildings have additional value factors to consider for resiliency and risk. Costs associated with mission security and resiliency (e.g., ability to maintain mission execution capabilities during reduced water availability or emergency operation) are crucial to evaluating the true cost of water in order to justify and prioritize resiliency investments for emergency off-grid water capabilities. Besides the cost-based pricing, one of the previous Army studies (Department of the Army 2014) reviewed other approaches for assessing resource value across water use applications, broadly categorized into revealed preference and stated preference (Table 1). However, both approaches still remain underutilized due to the complexity and time- and resource-intensive processes.

Table 1. Summary of water pricing methods along with description thereof

Method	Description
Cost-based pricing	This approach uses cost-related data and expected costs for provision of water.
Revealed preference (i.e., hedonic pricing)	This approach provides water price as a function of various variables (i.e., quantity demand).
Stated preference (i.e., survey)	This approach utilizes responses from individuals through questionnaires and considers water user preferences (i.e., willingness-to-pay).

Source: Data from Department of the Army (2014).

In the context of current Army Directive 2020-03 Installation Energy and Water Resilience Policy, installations must create strategies for maintaining critical operations (e.g., information system processing centers, hospitals) for a two-week period under emergency situations (resource-limited settings) (Secretary of the Army 2020). Army cost benefit analysis (CBA) and life cycle cost analysis (LCCA) of water conservation technologies (i.e., water reuse systems) are widely used to assess the financial feasibility of alternative strategies to meet these needs (Department of the Army 2016b; Noshadravan et al. 2017). However, current assessment tools often ignore unmonetized resources, biasing results to favor low-cost alternatives. As a result, these analyses yield recommendations that may leave installations more vulnerable to environmental risks or disasters and potentially jeopardize the ability to perform critical missions. By expanding these analyses to better understand the impacts to mission-critical functions under the context-specific constraints in military settings, decision makers can better justify investments for water security in a prudent and responsible manner that results in long-term cost savings through resiliency (Jenicek et al. 2018; Quitana et al. 2020).

Statistical methods such as regression analysis have been widely used to measure the relationships between price and its associated influencing factors, with focus on price prediction and price determination using limited sets of variables (Bjornlund and Rossini 2005, 2007; Brookshire et al. 2004; Brown 2006; Payne et al. 2014; Xu et al. 2019). Zhang et al. (2022) examined influencing factors such as utility privatization for water price and water affordability for large community water systems in the US. Also, a number of studies show the need for methods (e.g., nonmarket valuation) of assessing the value of unpriced environmental resources (Baker and Ruting 2014; Gibson et al. 2016; Rogers et al. 2015). Whitehead and Rose (2009) included non-market values in decision-making process by incorporating benefit transfer methods to a CBA to mitigate natural hazard (e.g., earthquake, wind, flood event). However, these valuation methods still remain underutilized in some areas where they are relatively difficult to quantify, particularly in financial-equivalent terms (F. Zhang and J. Fogarty, “Nonmarket Valuation of Water Sensitive Cities: Current Knowledge and Issues,” Working Paper 1513, The University of Western Australia, Crawley, Australia). For the unique characteristics of military settings, past studies have reviewed nonmarket valuation methods (quantitative, qualitative, and monetary) and developed a framework to evaluate Army infrastructure projects that have significant internal and external nonmarket costs and benefits (Aligned Incentives, LLC 2016; Department of the Army 2016b). However, the framework is too broad for large-scale application, complex, and does not directly address the costs and benefits of the water resiliency. Therefore, there is a need to measure Army context-specific variables (e.g., installation mission type, water source, utility privatization, installation status on mission capacity) by utilizing statistical approaches that provide a graphical visualization for water security and resiliency planning. For this study, regression tree analyses were selected due to their ability to efficiently illustrate clustering of prices by explanatory variables.

The objectives of this study were (1) to characterize Army specific factors influencing installation water price, (2) to measure site-specific factors that are correlated with water price, and (3) to formulate recommendations for water pricing to support installation water security and resiliency planning. The scope of the study selected 63 installations [22 of Army Materiel Command (AMC), 15 of Installation Management Command (IMCOM)—Sustainment, 12 of IMCOM—Readiness, 14 of IMCOM—Training] for detailed data collection and analysis. Water measures categories from Installation Status Report—Mission Capacity were also studied. The water unit price at Army installations was examined by mission type and function, as well as the critical site-specific factors to water price for Army operations and mission executions. This study is primarily intended to support policymakers in understanding emergent patterns and correlations in the price variations of water serving Army installations, as well as to support prescriptive modifications to pricing and valuation (for instance, cost-benefit analysis) methodologies incorporating factors such as infrastructure condition and the important missions enabled by potable water access. In addition, this study adds to the field of water price analysis and its unity through a nonparametric analysis for the existing studies on the demand response and system characteristics related to water price at a national scale.

Materials and Methods

Data

Variations in the price of water across 63 United States Army installations were explored using the combination and analysis of multiple data sets, each providing important information about the quantity and prices of water demanded, the missions served, and the status and context of the installation under study.

Information about water price and demand was collected for each installation over the period between 2015 and 2019 from the Army Energy and Water Reporting System (AEWRS) (Department of the Army 2018). Each data point consisted of the reported price of water in dollars per thousand gallons, and total water demand in thousands of gallons. Potable water intensity was also reported in various metrics from AEWRS, including: gallons per square foot, thousands of gallons per person, and gallons per person per square foot. Gross annual population and facility area measures used in these metrics were also retained for comparisons of system size.

Building upon review of existing literature and study priorities, additional data sets prioritized characteristics of the water systems, climate of the installation, and available measures enabling comparison of Army water systems. Key characteristics about each water system included the predominant water source- ground versus surface- and whether the system was publicly or privately owned and operated. Climate data included heating and cooling degree days and the predominant ASHRAE climate zone of each installation.

The US Army maintains an enterprise-scale reporting system for installations and their infrastructure. For this study, the installation status report—mission capacity (ISR-MC) (HQDA 2019; Smith et al. 2022) was utilized. This annual status report asks each installation to rank their infrastructure capabilities with ordinal scores between one and four over a series of questions, divided into composite score areas of: assured access, infrastructure condition, system operation, and critical mission sustainment. For this study, responses to both individual questions and composite scores were assessed.

Data Preparation

All data elements described in the previous section were encoded in tabular format and ingested using Python 3.6 for data cleaning, processing, and analysis. Regression and regression tree analyses were completed utilizing the Pandas (version 1.1.3), SciKitLearn (version 0.23.2), and StatsModels (version 0.12.0) packages.

Data processing began by removing any installations which report a unit price of water of zero dollars. In these analyses, 13 of the original 63 installations lack annual water price data, resulting in 50 installations down selected for further analysis. The second step in data processing was to convert each observation into an appropriate format for regression analysis. For instance, all ISR-MC data was reported in color scores (green, amber, red, black), with each translated into ordinal numbers from one to four for rating composite scores in a scale of 0–17 (black), 17–50 (red), 50–84 (amber), and 84–100 (green). Nonordinal categorical variables- for instance climate zones- were converted into Booleans through one-hot encoding where appropriate.

In addition to annual values, statistics such as average, minimum, and maximum were also calculated for all time-series variables. Initial linear regression models were developed in order to support prediction of water prices given numeric data. These models were complemented with regression trees, allowing for more nuanced policy-level interpretations of factors which may influence the price of water on Army installations (Fig. 1).

Fig. 1. Clustering Army installation water prices by influencing factors using regression tree models. Each node represents a decision rule that splits data into two or more homogeneous sets. A root node represents the entire data set. Splitting internal nodes are repeated based on the evaluation criteria until leaf nodes meet the terminal conditions such as minimum number of samples and maximum depth of the tree.

Descriptive analyses began by exploring how the price of water varies across 50 Army installations spread across the United States. These study installations served varied purposes, broadly described as IMCOM-Readiness (a home installation for troops), IMCOM-Training (an installation at which troops temporarily reside for basic or specialized training), AMC (Army Materiel Command, an installation which is primarily focused on industry activity), and IMCOM-Sustainment (installations which support troops without serving primarily as a home installation). Each installation also varied in water source (predominantly ground or surface), as well as whether the water system was owned and operated by the US Federal Government or a private contractor (Table 2).

Table 2. Categorized context-specific variables along with descriptions thereof

Category	Variable	# of installations	Description
Installation mission type	AMC	18	Support industry production and maintenance
	IMCOM-readiness	11	Support troops serving as a home installation
	IMCOM-sustainment	10	Support troops without serving as a home installation
	IMCOM-training	11	Support transient troops for basic or specialized training
Water source	Surface water	30	Primary water source from lake or river
Water source	Groundwater	20	Primary water source from well
Utility privatization	Public	30	Utility(ies) publicly owned and operated
Utility privatization	Private	20	Utility(ies) privately owned and operated
Installation status (mission capacity)	Assured access	50	Access to contingency water supplies during unplanned water outage
	Infrastructure condition	50	Overall condition of the installation’s water treatment, storage, and distribution systems
	System operation	50	Perform all annual required OM&T and ability to operate and prioritize water distribution during periods of scarcity
	Critical mission sustainment	50	Ability to meet water requirements for critical missions and supporting facilities and infrastructure

Additional context about the infrastructure systems and the missions they serve were generated from the Army’s installation status report (ISR) mission capacity (MC) questionnaire. This set of questions asked of each Army installation seeks numeric scores for a variety of questions designed to provide four composite scores to Army leadership: assured access, infrastructure condition, system operation, and critical mission sustainment.

Correlation Analyses with Individual Variables

Spearman’s rank correlation coefficient for each independent variable was measured to identify the effects on installation water price by the individual factors [Eq. (1)]. To quantify the correlation, Army data (water price, water consumption, ISR, population, installation size) and ASHRAE climate zone for fiscal year 2019 were used. The correlation was measured as the Spearman’s rank correlation coefficient (

ρ

) and the critical values (two-tailed) at significance level of 0.05 were

\pm 0.285

from the total 50 observations. The coefficient represented strong positive correlation (toward

ρ = 1

), strong negative correlation (toward

ρ = - 1

), and no correlation (

ρ = 0

) between the variables

ρ = 1 - \frac{6 \sum_{i = 1}^{n} {[R (x_{i}) - R (y_{i})]}^{2}}{n (n^{2} - 1)}

(1)

where

ρ

= Spearman’s rank correlation coefficient;

n

= number of observations;

x_{i}

= value of an independent variable for observation

i

;

y_{i}

= installation water price for observation

i

; and

R (\cdot)

= relative rank of

x_{i}

and

y_{i}

across the

n

observations.

Regression Tree Analyses with Combined Variables

Regression trees is useful to efficiently illustrate clustering of prices by explanatory variables. In addition, this statistical approach provided a graphical visualization for understanding and communicating complex relationships between dependent (e.g., water price) and independent (e.g. utility privatization, infrastructure condition) variables. Regression trees operated as a continuous analogy to classification trees, in which observations were continuously subdivided based upon their features.

All regression trees implemented in this study utilized the SciKit Learn decision tree library, which operates utilizing the common and well accepted C4.5 algorithm (Witten et al. 2016). Two key parameters specified during the course of this study were the minimum number of samples per leaf node and the maximum depth of the tree. These parameters helped users to balance precision of results with creating coarser but potentially more concise and actionable insights.

Results and Discussion

Influence of Individual Installation Specific Factors on Water Price

Building upon analysis of the impact of installation specific factors upon the price of water supply, a natural progression was to examine whether the individual factors impact the total water expense and unit water price paid by an installation. Several previous studies in water resource management showed multiple factors (e.g., water demand and consumption, water quality, regional characteristics, population, condition of water distribution system) influencing price of water from suppliers (Bjornlund and Rossini 2005; Brookshire et al. 2004; Brown 2006; Payne et al. 2014; Xu et al. 2019). Spearman’s rank correlation coefficient for factors studied (climate zone, population, installation size, annual water consumption, ISR-critical mission sustainment, ISR-assured access, ISR-infrastructure condition, ISR-system operation) were evaluated to elucidate the correlation between each factor and installation water prices. Results revealed the different trends among the factors for total water expense installations paid annually and water unit price (Fig. 2). As expected, population, installation size, and annual water consumption were highly correlated (

ρ > 0.7

) with total water expense. This baseline result is consistent with the expectation that installation water bill will be dependent on water consumption, number of stationed and transient population, and gross square footage of installation. Conversely, climate zone as a function of the number of heating degree days was negatively correlated with total water expense, which indicated that installations underlying higher climate zone were billed less total water expense than installation located in lower climate zone. In accordance with ASHRAE standards, higher climate zone number corresponds to more heating degree days and sites located at higher latitudes. Although it may be reasonable to expect that sites with higher climate zone numbers would have lower water demand, a recent study has shown a lack of correlation with climate patterns (Chini and Stillwell 2018).

Fig. 2. Water price sensitivity across the influencing factors as measured by the Spearman’s rank correlation coefficient.

In contrast to total water expense, installation water unit price was negatively correlated among the factors. With Spearman’s rank correlation coefficient of

\pm 0.285

at significance level of 0.05, water consumption was one of the major sensitivity factors (negative correlation) to water unit price. This implies that many Army installations may utilize the declining block rate for the water rate structure, often observed in large farming or heavy industry areas (Boyer et al. 2012). The implication using inappropriate utility rate structure may lead to discouraging water conservation efforts in water stressed regions. In addition, the result showed that infrastructure condition was the next statistically significant factor (negative correlation) to the installation water unit price. This implies that the water unit price may reflect required or realized infrastructure investments. Deteriorating water infrastructure is receiving a great deal of interest across the United States (Lee et al. 2019). Limited replacement and rehabilitation of aging water systems has resulted in creating a high risk of water system failure and extreme replacement expenditures when failure occurs (Selvakumar and Tafuri 2012). Alternatively, the result indicated that climate zone was not statistically correlated with water unit price. This finding implies that geographic region factor has a minimal effect on the installation water unit price, which could lead to a failure in promoting water conservation in areas with limited water supplies.

Water Price Driven by Installation Water Systems Characteristics

Installation water systems characteristics were used to identify factors related to water unit price variation across study installations. The combination of variables for installation water systems characteristics consisted of installation mission type, type of primary water source, system privatization, and climate zone. As a result from regression analyses, installation mission type, specifically whether the installation belongs to IMCOM-Readiness, was the most statistically significant factor to installation water unit price among the combination of variables (Fig. 3). IMCOM-Readiness installations paid 55% less for a unit of water on average (

2.59 $ \cdot {kgal}^{- 1}

or

0.68 $ \cdot m^{- 3}

) than non-IMCOM-Readiness installations. Also, AMC installations paid 23% less for on average unit water price (

4.95 $ \cdot {kgal}^{- 1}

or

1.31 $ \cdot m^{- 3}

) than IMCOM-Training and IMCOM-Sustainment installations. Among IMCOM-Training and IMCOM-Sustainment installations, installation mission type was statistically less significant for further classification and the primary water source became a dominant influencing factor to installation water price. In this case, installations with surface water as a primary water source paid 36% less for an average unit water price (

5.14 $ \cdot {kgal}^{- 1}

or

1.36 $ \cdot m^{- 3}

) than water system supplied primarily by groundwater. Likewise, 31% less water unit price (

2.14 $ \cdot {kgal}^{- 1}

or

0.57 $ \cdot m^{- 3}

) was observed within IMCOM-Readiness installations where the primary water source was surface water. Conversely, AMC installations had significantly less water unit price (

1.69 $ \cdot {kgal}^{- 1}

or

0.45 $ \cdot m^{- 3}

) if their primary water source was groundwater. Some of the factors had relatively high

P

values (paired T-test) and the null hypothesis was not rejected (i.e., no significant mean difference between two groups) at given significance level of 0.05. However, it is important to understand that this does not prove the null hypothesis and observed effects still exist between groups (Amrhein et al. 2019). Hence, identified site-specific factors with effects on installation water price should be considered for installation water security and resiliency planning. As more data become available, the traditional approach (i.e., ordinary least squares) can be applied to support findings through an action oriented statistical approach.

Fig. 3. Descriptive analysis of water unit price variations by installation features. Each node includes decision rule, average value, standard deviation, and number of samples.

IMCOM-Readiness installations provide mission readiness for combat and support workforces and their families. The installations under this functional command tend to have a large number of stationed personnel, and therefore, on average annual water consumption was 779,000 kgal (828%), 659,000 kgal (391%), 472,000 kgal (214%) greater than installations under AMC, IMCOM-Sustainment, IMCOM-Training, respectively. Also, AMC installations provide mission support for industry production and maintenance. This functional command tend to have relatively smaller number of stationed personnel. Thus, potable water consumption (

gal \cdot {person}^{- 1} \cdot {year}^{- 1}

) was relatively higher than other functional command. These findings could suggest diminishing unit prices of water being observed due to economies of scale. In addition, the result indicated that while the primary water source, whether it was surface water or groundwater, was correlated to installation water unit price, other more local factors (e.g., water source and treatment within base, location, number of source water) likely influenced much of the individual installation in each functional command.

Water Price Driven by Installation Mission Capacity

It is also important to understand the relationship between water unit price and installation mission capacity in order to realize the significance of factors (water infrastructure condition, system operation, critical mission sustainment, and assured access) to water unit price across the study installations. By expanding a set of variables to include installation mission type, primary water source, system privatization, and ISR-MC water measures, assured access to water was the most statistically significant factor for installation water unit price. Installations paid 70% more for an average water unit price (

7.07 $ \cdot {kgal}^{- 1}

or

1.87 $ \cdot m^{- 3}

) if their access to water was amber or worse (Fig. 4). This could be due to the lack of contingency water supplies during unplanned water service outages from the primary water supplier. Within amber or worse in assured access to water, installation mission type, specifically whether the installation belongs to IMCOM-Readiness, was the next classification of the statistically significant factor to installation water unit price. IMCOM-Readiness installations’ unit water price was 81% lower (

1.58 $ \cdot {kgal}^{- 1}

or

0.42 $ \cdot m^{- 3}

) than non-IMCOM-Readiness installations (AMC, IMCOM-Sustainment, IMCOM-Training). On the contrary, primary water source was the next influencing factor to the water unit price if installations maintained green or better in assured access to water. Within the green or better in assured access, an average installation water unit price was 44% higher (

4.68 $ \cdot {kgal}^{- 1}

or

1.24 $ \cdot m^{- 3}

) if installation’s primary water source was surface water as oppose to groundwater.

Fig. 4. Statistical analysis of water unit price by installation features and water systems. Each node includes decision rule, average value, standard deviation, and number of samples.

Based on ISR-MC water measures, the amber or worse in assured access included single primary water supply source used only by the installation or shared with other water users, lack of access to contingency water supplies for all end uses, and over 65% installation water consumption from the primary water supply capacity. The water unit price was correlated with conditions of assured access to water (i.e., whether installations are capable of water supplies to meet the water demand during unplanned water service outages from the primary water supplier). Therefore, higher unit price of water was observed when installations had amber or worse in assured access to water.

In addition, water utility privatization was considered as another important influencing factor to water rates. Generally, water privatization can improve infrastructure and provide competitive rates and safe water to the end users (Lieberherr and Truffer 2015). In some cases, privately owned water systems lead to higher annual water bill (Zhang et al. 2022). In contrast, utilities owned by installations determine their water rates based on the total cost to the Government, including transmission losses, operation and maintenance costs, capital charges, and administrative overhead (Department of the Army 2018). In cases where installations had green or better in assured access and utilized groundwater as a primary water source, water unit price from utilities that were not privatized was 66% lower than privatized utilities. This may be due to lack of capabilities (e.g., regulation) to appropriately quantify the water rates billed to customers.

Common water utility rate structures include flat rates (a fixed charge regardless of consumption); uniform rates (charges based on consumption but at same rates for all end uses); tiered rates (charges based on consumption but different rates in steps); and seasonal rates (charges vary seasonally). To support Army installation water security and resiliency planning, it is essential to take into consideration such options for future rates structures. Annual installation water consumption footprints (

gal \cdot {sf}^{- 1}

or

kgal \cdot {person}^{- 1}

) revealed different characteristics based on Army missions (Fig. 5). Because AMC installations are primarily focused on industry production, water consumption footprint is highly correlated with building size. Alternatively, IMCOM installations directly support troops and, subsequently, water consumption footprint is highly correlated with the number of personnel. This implies that further disaggregation of installations into missions and water uses may provide different pricing structures depending on the criticality of various missions.

Fig. 5. Installation potable water consumption footprints by mission type: (a) footprints measured as gallons per square foot per year; and (b) footprints measured as kilo-gallons per person per year. Data is sorted by the median value in descending order. Boxes extend from 25th to 75th percentiles. Within the boxes, 50th percentiles are marked by a horizontal line. Whiskers extend to the minimum and maximum value, excluding outliers.

Conclusions

This study characterized Army specific factors to examine the determinants of installation water prices using statistical approaches including Spearman’s rank correlation coefficient and regression tree. The key conclusions are as follows:

•

Water prices across Army installations varied depending on the characteristics of installations and their water systems. Among individual factors studied, annual water consumption, size of installation, and population were the major influencing factors (positive correlation) to annual water billed.

•

For Army installation water unit prices (

$ \cdot {kgal}^{- 1}

), however, negative correlation with annual water consumption and less significant correlation with climate zone were observed. Since declining block rate for the rate structure tends to discourage water conservation practices, an appropriate water rate structure is needed to promote water conservation efforts especially in water stress regions.

•

With combination of characterized variables, installation mission type, type of primary water source, and assured access to water were statistically significant factors to Army installation water unit prices. Regression tree analyses results provided coarser but actionable insights while clustering water unit prices by the influencing factors.

Based on these results, the factors correlated with installation water unit price need to be considered in the strategic planning for water disruption events that may interrupt executions on Army missions. For military settings, data-driven approaches may provide comprehensive information (quantitative and qualitative) to achieve installation water security and resiliency. Future studies should focus on addressing causation problems to fully understand the statistically significant factors to installation water unit price identified in this study. Although this approach is developed to examine various context-specific factors correlated with Army installation water price, future studies may apply this approach in different contexts to support sustainable design and planning.

Data Availability Statement

The raw data are not publicly available due to information security. Derived data supporting the findings of this study are available on request from the corresponding author.

Acknowledgments

This research was funded by the Office of the Assistant Secretary of the Army for Installations, Energy, and Environment. The views expressed are those of the authors and do not necessarily reflect the official policy or position of the Department of the Air Force, the Department of Defense, or the US government. Distribution statement A: Approved for public release, distribution is unlimited.

Author contributions: Conceptualization, Andy Y. Hur and Noah W. Garfinkle; methodology, Andy Y. Hur and Noah W. Garfinkle; software, Noah W. Garfinkle; validation, Andy Y. Hur, Noah W. Garfinkle, and Christine M. Ploschke; formal analysis, Andy Y. Hur; investigation, Noah W. Garfinkle; resources, Christine M. Ploschke; data curation, Noah W. Garfinkle; writing—original draft preparation, Andy Y. Hur; writing—review and editing, Noah W. Garfinkle, Christine M. Ploschke, Jeremy S. Guest, and Christopher M. Chini; visualization, Noah W. Garfinkle; supervision, Christine M. Ploschke, and Jeremy S. Guest; project administration, Andy Y. Hur; and funding acquisition, Andy Y. Hur. All authors have read and agreed to the published version of the manuscript.

References

Aligned Incentives, LLC. 2016. Deriving water valuation factors for Ft. Riley. Washington, DC: Office of the Assistant Secretary of the Army.

Abstract

Practical Applications

Introduction

Materials and Methods

Data

Data Preparation

Correlation Analyses with Individual Variables

Regression Tree Analyses with Combined Variables

Results and Discussion

Influence of Individual Installation Specific Factors on Water Price

Water Price Driven by Installation Water Systems Characteristics

Water Price Driven by Installation Mission Capacity

Conclusions

Data Availability Statement

Acknowledgments

References

Information

Published In

Copyright

History

ASCE Technical Topics:

Authors

Affiliations

Metrics

Citations

Download citation

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!