Water–Energy–Carbon Nexus and the Impact of Real Water Losses in Urban Water Supply: A Case Study of the Metropolitan Waterworks Authority, Thailand

Abstract

Urban water supply systems require considerable electrical energy inputs across all operational processes: raw water abstraction, treatment, transmission, and distribution. Consequently, water loss within these processes represents not merely a loss of water volume, but also additional energy consumption and an increase in carbon emissions, given that electricity generation relies predominantly on fossil fuels. This study applied two methodological approaches to analyze the role of water loss within the Water–Energy–Carbon (WEC) Nexus of the Metropolitan Waterworks Authority (MWA), Thailand, over the period 2017–2024. The first method utilized a detailed WEC linkage analysis to balance water inputs and outputs in each process to quantify specific losses: raw water, in-plant, transmission, and distribution losses. The second method applied the International Water Association’s Leakage Emissions Initiative framework, focusing specifically on potable real water loss in distribution process, which constituted the largest volume (64.85% of total losses) and embodied the highest specific energy consumption. Based on the first method, the average annual potable real water loss was 534.71 MCM/yr (23.58% of water supplied to distribution), corresponding to embedded energy and carbon emissions of 103.76 GWh/yr (24.89% of total energy consumption) and 49,562 tCO₂e/yr (24.89% of total carbon emission), respectively. Although the second method was considerably simplified, the estimated energy and carbon emission values were only slightly higher than those derived from the detailed method, demonstrating the second method’s effectiveness as a streamlined assessment tool. These findings underscored that water loss reduction initiatives are essential for minimizing energy consumption and carbon emissions, thereby supporting Thailand’s pathway toward Net Zero emissions by 2050.

1. Introduction

Climate change has emerged as one of the most critical global challenges of the twenty-first century, driven by global warming resulting from increasing greenhouse gas (GHG) emissions. Among these gases, carbon dioxide (CO₂) is the primary contributor. Consequently, the Intergovernmental Panel on Climate Change (IPCC) emphasizes that limiting the rise in global temperature to 1.5 °C requires achieving global net-zero CO₂ emissions by 2050 [1].

Thailand has committed to the Paris Agreement through the national targets of achieving carbon neutrality by 2050 and net-zero GHG emissions by 2065, as announced at the 26th Conference of the Parties. These targets are now integrated into national strategies and development plans to support energy, economic, and social transitions toward a low-carbon future [2]. According to Thailand’s Fourth National Communication in 2018, the energy sector remains the largest contributor to national GHG emissions, accounting for approximately 69% of the total, highlighting the critical importance of improving energy efficiency in achieving the country’s net-zero pathway [3].

Among energy-intensive public utilities, urban water supply systems are especially important. They require substantial energy inputs across all operational processes—ranging from raw water abstraction, treatment, and long-distance transmission under high pressure, to distribution across extensive service areas—thereby resulting in indirect CO₂ emissions from electricity consumption [4,5,6,7,8,9]. Lee et al. [10] have highlighted that urban water systems are more complex than rural systems due to higher demand, multi-stage pumping requirements, and topographical variations. Therefore, water supply management must be regarded as a key component within the broader Water–Energy–Carbon (WEC) nexus to optimize energy efficiency and support the attainment of the IPCC’s global net-zero CO₂ emissions goal.

The urgency of this decarbonization effort is underscored by recent projections from Vecchia and Franzke (2025) [11], who identified that under high-emission scenarios (e.g., SSP3-7.0), the onset of unprecedented ‘Day Zero Drought’ events is expected to accelerate, potentially emerging as early as the 2020s and 2030s. Consequently, minimizing water losses serves a dual critical purpose: it not only preserves essential freshwater resources but also substantially reduces energy-related carbon emissions. This aligns directly with global mitigation strategies aimed at decelerating the anthropogenic warming trends that drive these imminent water scarcity crises.

The WEC nexus is widely applied in urban water systems, including water supply and wastewater services [4,6,12]. However, carbon emissions associated with water loss are still rarely addressed explicitly. In parallel, the energy–water literature has developed extensive energy audit approaches and efficiency metrics for water networks and urban wastewater systems [5,13,14,15,16,17,18,19], supported by operational audits and mass-balance assessments in drinking water networks [20,21]. These studies indicate that water loss—particularly in distribution systems—can be a major source of energy wasted through water loss, and that loss reduction can significantly improve energy efficiency [22,23]. Accordingly, substantial energy wasted through water loss is likely to result in considerable electricity-related carbon emissions, highlighting the need to explicitly incorporate water loss into WEC-based carbon accounting. Although some research [10,24,25] has begun to adopt WEC-based perspectives in urban water systems, water loss is still largely treated as a service performance issue, rather than being explicitly recognized as a source of avoidable energy use and associated carbon emissions embedded in water abstraction, treatment, transmission, and distribution processes. To bridge this gap, the International Water Association (IWA) established the Leakage Emissions Initiative (LEI) [26]. This initiative explicitly identifies water loss as a substantial source of avoidable carbon emissions and advocates for leakage reduction projects to be recognized as eligible mechanisms for generating carbon credits, thereby incentivizing water utilities to contribute directly to climate change mitigation.

In Thailand, existing research has primarily focused on the energy consumption of water supply systems or has examined water loss only at the distribution stage, without incorporating system-wide interactions across the full water supply chain [21,23,27,28,29]. As a result, comprehensive assessments of the WEC nexus remain limited, and the embedded carbon associated with water loss has not yet been systematically quantified. Emerging evidence suggests that water loss is a major source of avoidable energy use and unnecessary carbon emissions within urban water supply systems; however, the absence of an integrated analytical framework constrains the ability to evaluate system-level impacts or identify effective carbon-reduction pathways. This research gap highlights the need for holistic approaches that capture the full implications of water loss within national and global decarbonization efforts.

Beyond energy considerations, water loss also has substantial consequences for water quality management, particularly regarding disinfectant residuals. Some studies have extended the scope of network auditing by establishing a link between water loss and chlorine loss, conceptualizing a “Water Quality Audit” or “Chlorine Mass Balance” analogous to traditional water and energy balances. For example, Lipiwattanakarn et al. (2020). [20] pioneered the Free Residual Chlorine mass audit framework, demonstrating that disinfectant mass is diminished not only through chemical decay but also through physical leakage, which often exceeds the decay component in magnitude. Subsequently, this theoretical framework was subsequently refined by Wongpeerak et al. (2023) [21] to quantify specific disinfectant mass balance components and later integrated into computational tools enabling simultaneous audits of water, energy, and chlorine [23]. Collectively, these findings indicate that water loss reduction strategies generate dual benefits: improving energy efficiency and preserving essential disinfectant levels for public health protection.

Furthermore, the COVID-19 pandemic induced unprecedented disruptions to urban water consumption patterns across residential and non-residential sectors. In many regions, the pronounced decline in non-residential demand, contrasted with a substantial rise in residential water usage—particularly during lockdowns, reflecting the broader socioeconomic disturbances triggered by the crisis [22,30,31,32,33]. In addition, shifting water demands can lead to changes in pressure distribution in pipe networks and impact energy consumption across water supply processes [34]. However, there is a lack of rigorous quantitative assessment within the WEC nexus framework that provides empirical evidence on holistic resource efficiency across the pre-pandemic, pandemic, and post-pandemic phases, especially regarding the impact of water loss.

Therefore, the current study aimed to assess the interrelationship among water, energy, and carbon within the Metropolitan Waterworks Authority (MWA), Thailand. To achieve this, the evaluation applied two methodological approaches: (1) a comprehensive WEC nexus analysis covering all four core operational processes, and (2) a targeted assessment based on the IWA–LEI framework, focusing specifically on treated water processes. This specific focus was adopted because, typically, potable water loss constitutes the largest volume and represents the most energy-intensive component of the system. Additionally, the study compared WEC outcomes before and after the COVID-19 pandemic. The findings should support strategic planning for resource efficiency and low-carbon water utility management in metropolitan contexts. To summarize the core contributions and stimulate reader interest, the main highlights of this study are:

• Comprehensive evaluation of the WEC nexus in Thailand’s metropolitan water supply, incorporating a comparative analysis across the pre-COVID-19, lockdown, and post-COVID-19 periods.
• Separation of real and apparent water losses in energy and carbon accounting, showing that only real physical losses represent actual environmental waste.
• Highlighting active leakage control as a key driver for substantial WEC benefits and carbon credits.

2. Study Area and Methodology

2.1. Study Area

The case study (Figure 1) focused on the MWA, a state-owned enterprise responsible for the provision of potable water supply services in Thailand. In the Central Region of Thailand, the MWA serves three major provinces—Bangkok, Nonthaburi, and Samut Prakan—which collectively constitute the core metropolitan area of the country and are located within the lower Chao Phraya River basin.

MWA’s primary mandate is the production and distribution of treated water for domestic, commercial, industrial, and public uses. As of 2024, water supply services were provided to 2,644,177 customers (service accounts) across a total service area of 3195 km², administratively organized into 18 operational branches. Approximately 50% of the supplied water was allocated to household consumption [35], highlighting the importance of reliable service provision and effective demand management in the residential sector.

In terms of operational scale, the MWA produces 2038.13 million cubic meters (MCM) of treated water annually [35], with authorized consumption of 1514.82 MCM. Therefore, the potable water loss is 523.31 MCM (25.67%). This annual supply is delivered through an extensive pipeline network with a total length exceeding 40,000 km, reflecting the large scale and complexity of the metropolitan water supply process.

Figure 2 illustrates that MWA’s raw water relies on two major rivers, namely the Chao Phraya River and the Mae Klong River. Raw water abstraction and conveyance are carried out through two principal raw water canals: the east raw water canal and the west raw water canal. The east raw water canal conveys raw water pumped from the Chao Phraya River, whereas the west raw water canal conveys raw water from the Mae Klong River completely through gravity-driven flow.

Raw water conveyed through the east raw water canal is transported to three major water treatment plants: the Bangkhen Water Treatment Plant (WTP), the Samsen WTP, and the Thonburi WTP. In contrast, raw water conveyed through the west raw water canal is transported to the Mahasawat WTP.

Following treatment, a large portion of the treated water from the Bangkhen and Mahasawat WTPs (the two largest) is conveyed initially by 4 pumping stations (PSs) through the 191 km of transmission tunnels and conduits prior to entering the distribution network far from WTPs mostly in the south and the east of the whole service area. The balance of the supply is served directly to customers in the vicinity. In contrast, all potable water produced at the Samsen and Thonburi WTPs is supplied directly to the distribution process to service customers in the vicinity due to their smaller capacities.

The distribution process comprises 17 PSs and 13 direct supply points, which collectively regulate system pressure and ensure adequate water delivery throughout the service area. The overall distribution process consists of primary trunk mains with a total length of approximately 1841 km and distribution pipes extending 39,518 km, enabling the delivery of potable water to end users across the metropolitan region.

At this scale, the water system requires continuous and substantial energy input. The average annual electricity consumption of the MWA process is approximately 433,456.88 MWh [36], representing a considerable share of operational expenditure and a major source of indirect greenhouse gas emissions. Therefore, effective energy management is essential to enhance the operational sustainability of the water supply process and to minimize its associated environmental impacts, providing a strong foundation for integrated analysis of water supply operations, energy consumption, and environmental performance.

2.2. Operational Processes and Water Loss Classification

The diagram in Figure 3 illustrates the structural configuration of the MWA urban water supply system, comprising 4 sequential operational processes: Raw Water Process, Treatment Process, Transmission Process, and Distribution Process. Collectively, these processes deliver water for authorized consumption. Each process is associated with distinct forms of water loss that correspond to its operational characteristics and hydraulic conditions along the supply chain.

In the Raw Water Process, water is abstracted and conveyed from natural sources to treatment facilities. Losses arising at this process are classified as Raw Water Loss and may result from leakage, overflow, evaporation, infiltration, or operational discharges associated with intake and conveyance activities.

During the Treatment Process, raw water undergoes physical and chemical procedures to meet potable water standards. The operational requirements inherent to these processes generate In-plant Water Loss, which primarily consists of the essential process water needed for filter backwashing and clarifier desludging. To optimize resource efficiency, large-scale facilities typically utilize recovery systems to recycle a significant portion of this washwater back into the treatment cycle. Therefore, within the context of our macroscopic water balance, this term collectively encompasses both the necessary functional process water and any unintended physical leaks or overflows within the plant boundaries.

After treatment, potable water enters the Transmission Process driven by transmission pumping stations. This process transports high volumes of water via large-diameter pipelines from treatment plants to reservoirs located at distribution pumping stations near service areas.

Subsequently, the water enters the Distribution Process, which consists of a complex network of smaller-diameter pipes designed to deliver water directly to customers as Authorized Consumption. Notably, service areas located in the vicinity of water treatment plants are supplied directly via the Distribution Process, thereby bypassing the Transmission Process.

Water losses occurring within these interconnected stages are collectively defined as Potable Water Loss. While transmission losses are generally limited due to the structural robustness and minimal branching of the infrastructure, distribution losses tend to be higher due to the extensive network length and numerous service connections.

Within the distribution process, water losses occur in both physical and non-physical forms and can be classified into real losses and apparent losses. Real losses refer to physical water losses resulting from leakages and pipe bursts within the distribution network, including losses from transmission mains, distribution pipelines, service connections, valves, and other hydraulic components. Apparent losses, in contrast, do not involve physical water loss but arise from customer metering inaccuracies, unauthorized consumption, and system data handling error.

While real water losses adversely affect the operational efficiency of the distribution process by causing unnecessary energy consumption and extra carbon emissions into the environment, apparent losses (often termed commercial losses) do not have the same environmental or operational impact. Apparent losses represent water that is consumed by users; therefore, reducing them primarily increases billed revenue without altering the total system inflow or its associated energy and carbon footprints. Consequently, only real water losses yield no service value to end-users; nonetheless, considerable energy is consumed in these processes. Furthermore, this unnecessary energy consumption translates directly into increased indirect greenhouse gas emissions. To accurately reflect this distinction, while apparent losses are accounted for to complete the macroscopic water, energy, and carbon balances, they are clearly separated from real losses and explicitly excluded from all energy and environmental impact calculations in this study.

2.3. Methodology

This study applied a systematic framework to quantify the impacts of real water losses on water quantity, energy consumption, and carbon emissions within the MWA system. The methodology allows for a comprehensive evaluation across the WEC nexus dimensions.

2.3.1. Step 1: Study Area Identification and Data Collection

The study area was defined and all the required baseline datasets were assembled for subsequent operational process-level analysis. The MWA was selected due to its large-scale operations and its importance as one of the largest urban water suppliers in Thailand. The MWA water supply system was disaggregated into four operational processes, as shown in Figure 2 and Figure 3.

Operational datasets covering the period 2017–2024 were compiled, including

• Water volume data, used to establish the water balance and quantify water loss for each process.
• Electricity consumption data, enabling calculation of Specific Energy Consumption (SEC) at the process level.
• Emission factor (EF) data, used to convert electricity consumption into carbon dioxide equivalent (CO₂e) emissions.

2.3.2. Step 2: Calculate Water Losses in Study Area

A disaggregated framework was used to quantify water loss at each operational process. Each process (Raw Water, Treatment, Transmission, Distribution) was assessed individually to ensure precise quantification of water volumes and losses. Notably, the IWA/American Water Works Association (AWWA) water balance framework [26] forms a subset of this procedure, focusing specifically on determining losses of treated water during the transmission and distribution stages.

• Water Loss Calculation

Let the subscript i denote the sequential operational processes: Raw Water (raw), Treatment (tre), Transmission (tra), and Distribution (dis). The water loss for each process i was calculated based on the mass balance principle, as expressed in Equation (1):

$$V_{\mathrm{loss},i}=V_{\mathrm{in},i}-V_{\mathrm{out},i}$$

(1)

where $V_{\mathrm{loss},i}$, $V_{\mathrm{in},i}$, and $V_{\mathrm{out},i}$ represent the water loss, inflow volume, and outflow volume for process i, respectively. All volumes are expressed in millions of cubic meters per year (MCM/yr).

• Water Loss Component Calculation

For the Raw Water, Treatment, and Transmission processes, all water losses were classified as real losses. In contrast, water loss within the Distribution process was divided into apparent losses ($V_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$) and real losses ($V_{\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$) using Equation (2) and Equation (3), respectively:

$$V_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}=V_{\mathrm{cmi},\mathrm{d}\mathrm{i}\mathrm{s}}+V_{\mathrm{uc},\mathrm{d}\mathrm{i}\mathrm{s}}+V_{\mathrm{sdhe},\mathrm{d}\mathrm{i}\mathrm{s}}$$

(2)

$$V_{\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}=V_{\mathrm{loss},\mathrm{d}\mathrm{i}\mathrm{s}}-V_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$$

(3)

where $V_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$ comprises customer metering inaccuracies ($V_{\mathrm{cmi},\mathrm{d}\mathrm{i}\mathrm{s}}$), unauthorized consumption ($V_{\mathrm{uc},\mathrm{d}\mathrm{i}\mathrm{s}}$), and systematic data handling errors ($V_{\mathrm{sdhe},\mathrm{d}\mathrm{i}\mathrm{s}}$).

Based on MWA operational criteria and the findings from a study on apparent losses [37], $V_{\mathrm{cmi},\mathrm{d}\mathrm{i}\mathrm{s}}$ was estimated at 3.04% of the annual billed authorized consumption. According to [38], the auditor may choose to use the worldwide default value of 0.25% of the water supplied volume for $V_{\mathrm{uc},\mathrm{d}\mathrm{i}\mathrm{s}}$ during the top-down water audit. However, MWA conservatively assumes $V_{\mathrm{uc},\mathrm{d}\mathrm{i}\mathrm{s}}$ to be 1% (four times the default value) [37]. $V_{\mathrm{sdhe},\mathrm{d}\mathrm{i}\mathrm{s}}$ was considered negligible (and set to zero) due to the high reliability and precision of the MWA billing database system. These audited parameters were kept constant throughout the 2017–2024 study period due to the consistency in MWA’s metering technology and apparent loss control operations. While minor variations in these apparent loss assumptions could inversely affect the calculated real loss volumes in a top-down water balance, they do not significantly alter the core results and conclusions of this study. All components were reported in MCM/yr.

• Authorized Consumption per Capita

Total authorized consumption per capita per day ($V_{\mathrm{cap},\mathrm{tot}}$) and the residential authorized consumption per capita per day ($V_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{r}\mathrm{e}\mathrm{s}}$) were calculated using Equation (4) and Equation (5), respectively:

$$V_{\mathrm{cap},\mathrm{tot}}={\displaystyle \frac{V_{\mathrm{auth},\mathrm{tot}}\times {10}^9}{P\times D}}$$

(4)

$$V_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{r}\mathrm{e}\mathrm{s}}={\displaystyle \frac{V_{\mathrm{auth},\mathrm{res}}\times {10}^9}{P\times D}}$$

(5)

where $V_{\mathrm{auth},\mathrm{tot}}$ and $V_{\mathrm{auth},\mathrm{res}}$ represent the annual total and residential authorized consumption volumes (MCM/yr), respectively. $P$ denotes the served population, and $D$ represents the number of days in the year (365 or 366 if a leap year). The factor ${10}^9$ converts the volume from millions of cubic meters to liters. The resulting indicators were expressed in liters per capita per day (L/cap/d).

2.3.3. Step 3: Energy Assessment

The energy implications of water supply operations were quantified across all processes. Electricity consumption values used in this study are expressed in gigawatt-hours per year (GWh/yr). The assessment encompassed the evaluation of point and accumulated energy, the estimation of energy loss associated with water loss, the calculation of SEC, and the evaluation of electricity consumption per capita.

• Point Energy and Accumulated Energy

Let $E_i^{*}$ denote the point energy, defined as the electricity consumed explicitly within process i. The accumulated energy at process i ($E_i$) represents the cumulative embedded electricity in the water flow from the raw water stage up to process i. It was calculated as the summation of the point energy of the current process and all preceding processes, as shown in Equation (6):

$$E_i=\sum _{\mathrm{r}\mathrm{a}\mathrm{w}}^i{E}_i^{*}$$

(6)

• Energy Loss Associated with Water Loss

The electrical energy corresponding to water loss in process i (${EL}_i$) quantifies the embedded energy wasted due to water loss. It was estimated based on the ratio of water loss ($V_{\mathrm{loss},i}$) to the inflow volume ($V_{\mathrm{in},i}$), multiplied by the accumulated energy ($E_i$) at that process, as shown in Equation (7):

$${EL}_i={\displaystyle \frac{V_{\mathrm{loss},i}}{V_{\mathrm{in},i}}}\times E_i$$

(7)

It should be noted that for the distribution process (i = dis), $V_{\mathrm{loss},i}$ will be divided into $V_{\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$ and $V_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$ to represent the energy losses associated with real and apparent water losses in the distribution process (${EL}_{\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$ and ${EL}_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$), respectively. However, consistent with the nature of these losses, only ${EL}_{\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$ represents actual physical energy waste into the environment, whereas ${EL}_{\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$ is evaluated strictly for the utility’s accounting purposes to represent the embedded energy of lost revenue potential.

The total energy loss across all processes (${EL}_{\mathrm{tot}}$) was calculated using Equation (8):

$${EL}_{\mathrm{tot}}=\sum {EL}_i$$

(8)

• Specific Energy Consumption (SEC)

The specific energy consumption for process i ($SE{C}_i$) in kWh/m³, representing the accumulated energy intensity relative to the final stage of system input ($V_{\mathrm{i}\mathrm{n},\mathrm{d}\mathrm{i}\mathrm{s}}$), as shown in Equation (9):

$$SE{C}_i={\displaystyle \frac{E_i}{V_{\mathrm{i}\mathrm{n},\mathrm{d}\mathrm{i}\mathrm{s}}}}$$

(9)

It should be noted that $SE{C}_i$ was calculated on an annual basis, providing a dynamic yearly value throughout the study period.

• Electricity Consumption per Capita

The overall energy performance was evaluated relative to the served population (P) based on the total electricity consumption per capita ($E_{\mathrm{cap},\mathrm{tot}}$) and the residential electricity consumption per capita ($E_{\mathrm{cap},\mathrm{res}}$) in kWh/cap/d using Equation (10) and Equation (11), respectively:

$$E_{\mathrm{cap},\mathrm{tot}}={\displaystyle \frac{E_{\mathrm{dis}}\times {10}^6}{P\times D}}$$

(10)

$$E_{\mathrm{cap},\mathrm{res}}={\displaystyle \frac{V_{\mathrm{auth},\mathrm{res}}}{V_{\mathrm{a}\mathrm{u}\mathrm{t}\mathrm{h},\mathrm{tot}}}}{E}_{\mathrm{cap},\mathrm{tot}}$$

(11)

where the ratio between $V_{\mathrm{auth},\mathrm{res}}$ and $V_{\mathrm{a}\mathrm{u}\mathrm{t}\mathrm{h},\mathrm{tot}}$ represents the proportion of authorized consumption attributed to the residential sector.

2.3.4. Step 4: Carbon Emission Assessment

The carbon emission implications of water supply operations were quantified directly from the energy assessment results by converting electricity consumption into CO₂e emissions using the national grid emission factor. This encompassed the evaluation of point and accumulated carbon emissions, the estimation of carbon emissions associated with water loss, the calculation of GHG intensity per unit volume of supplied water, and the carbon emission per capita.

• Point Carbon Emissions and Accumulated Carbon Emissions

Let $C_i^{*}$ denote the point carbon emissions and $C_i$ denote the accumulated carbon emissions for process i in tonnes (t) of CO₂e (tCO₂e). These values were calculated by multiplying the corresponding annual electricity consumptions (point and accumulated) by the national grid emission factor, as shown in Equations (12) and (13):

$$C_i^{*}=E_i^{*}\times EF\times 1000$$

(12)

$$C_i=E_i\times EF\times 1000$$

(13)

where the national grid emission factor ($EF$) adopted in this study was 0.4788 kgCO₂e/kWh, representing Thailand’s average electricity carbon intensity over the 2018–2022 period, which was assumed to remain constant throughout the study timeframe (2017–2024) [39].

• Carbon Emissions Associated with Water Loss

Similarly, the carbon emissions associated with water loss in process i ($C_{\mathrm{loss},i}$), expressed in tCO₂e, were computed using Equation (14):

$$C_{\mathrm{loss},i}={EL}_i\times EF\times 1000$$

(14)

Consequently, following the same principle for energy losses, the carbon emission associated with water losses in the distribution process ($C_{\mathrm{loss},\mathrm{d}\mathrm{i}\mathrm{s}}$) is separated into $C_{\mathrm{loss},\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$ and $C_{\mathrm{loss},\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$. Only $C_{\mathrm{loss},\mathrm{real},\mathrm{d}\mathrm{i}\mathrm{s}}$ represents actual unnecessary carbon emissions released into the environment due to physical water waste. In contrast, $C_{\mathrm{loss},\mathrm{app},\mathrm{d}\mathrm{i}\mathrm{s}}$ merely represents the embodied carbon of lost revenue potential, which is evaluated for the utility’s comprehensive accounting purposes.

The total annual CO₂e emission linked to water loss across all processes ($C_{\mathrm{loss},\mathrm{tot}}$) in tCO₂e was computed using Equation (15):

$$C_{\mathrm{loss},\mathrm{tot}}=\sum _i{C}_{\mathrm{loss},i}$$

(15)

• Greenhouse Gas (GHG) Potential

To evaluate the environmental impact per unit of water supplied, the GHG potential at process i ($GH{G}_i$) was calculated using Equation (16):

$$GH{G}_i=SE{C}_i\times EF\times 1000$$

(16)

where $GH{G}_i$ is expressed in gCO₂e/m³, and the factor 1000 converts kgCO₂e to gCO₂e.

• Carbon Emission per Capita

The total ($C_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{t}\mathrm{o}\mathrm{t}}$) and residential ($C_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{r}\mathrm{e}\mathrm{s}}$) carbon emissions per capita, expressed in gCO₂e/cap/d, were calculated to assess the daily individual carbon emission associated with water use, based on Equation (17) and Equation (18), respectively:

$$C_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{t}\mathrm{o}\mathrm{t}}=E_{\mathrm{cap},\mathrm{tot}}\times EF\times 1000$$

(17)

$$C_{\mathrm{c}\mathrm{a}\mathrm{p},\mathrm{r}\mathrm{e}\mathrm{s}}=E_{\mathrm{cap},\mathrm{res}}\times EF\times 1000$$

(18)

2.3.5. Step 5: WEC Balances According to IWA-LEI

While the disaggregated analysis in Steps 1–4 provided a comprehensive evaluation of the WEC nexus across the entire supply chain—from raw water abstraction to distribution—it is critical to recognize that not all water losses carry the same environmental weight. Water losses occurring after the treatment process, specifically within the potable water transmission and distribution systems, have the greatest impact on carbon emissions because treated water embeds the highest accumulated SEC, having already undergone energy-intensive abstraction, treatment, and pumping processes. Consequently, leakage in the distribution network, typically accounting for the largest proportion of total system losses, equates to a substantial waste of energy and a direct increase in avoidable greenhouse gas emissions.

Recognizing this critical link, the IWA established the LEI to standardize the quantification of the carbon footprint associated with potable water losses [26]. The practical application of this framework has been demonstrated in Musoma, Tanzania, where targeted leakage reduction strategies were shown to substantially mitigate carbon emissions [40]. Furthermore, the AWWA reinforced this urgency through its committee report on the LEI, highlighting the pivotal role of leakage management in climate action strategies for water utilities [41].

Aligning with these global standards, this step specifically focuses on the Distribution Process, which had the highest volume of water loss within the MWA system. By applying the IWA-LEI standard WEC balance framework, the current study aimed to isolate and quantify the environmental burden of distribution losses, providing a clear benchmark for decarbonization through leakage management as follows.

• IWA Standard Water Balance

First, the MWA water audit data was structured according to the standard IWA Water Balance framework [26]. The Water Supplied was classified into Authorized Consumption and Water Losses, which were further subdivided into Apparent Losses and Real Losses. This volumetric breakdown is consistent with the distribution process evaluation in Step 2.

• Energy and Carbon Balances

The IWA-LEI framework extends the IWA water balance into Energy and Carbon Balances by assigning energy intensity and carbon emission values to each component. In the current study, the energy intensity applied to water losses was derived from the accumulated Specific Energy Consumption (${SEC}_{\mathrm{d}\mathrm{i}\mathrm{s}}$) at the distribution process input. This approach relies on the assumption that all water entering the distribution network carries the same uniform embedded energy.

Accordingly, the energy embedded in water losses ($E_{\mathrm{loss},\mathrm{LEI}}$) expressed in GWh/yr and the associated carbon emissions ($C_{\mathrm{loss},\mathrm{LEI}}$) expressed in tCO₂e/yr were calculated based on Equation (19) and Equation (20), respectively:

$$E_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{L}\mathrm{E}\mathrm{I}}=V_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{d}\mathrm{i}\mathrm{s}} \times SE{C}_{\mathrm{d}\mathrm{i}\mathrm{s}}$$

(19)

$$C_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{L}\mathrm{E}\mathrm{I}}=E_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{L}\mathrm{E}\mathrm{I}}\times EF\times 1000$$

(20)

where $V_{\mathrm{loss},\mathrm{dis}}$ and ${SEC}_{\mathrm{d}\mathrm{i}\mathrm{s}}$ are the volume of water losses (MCM/yr) and the accumulated specific energy consumption (kWh/m³) of the final stage of system input (Distribution Process), respectively. To maintain a complete and closed energy and carbon balance in accordance with the IWA-LEI framework, these total values ($E_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{L}\mathrm{E}\mathrm{I}}$ and $C_{\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s},\mathrm{L}\mathrm{E}\mathrm{I}}$) are subsequently separated into real and apparent components based on their respective water volume ratios. As emphasized earlier, only the energy and carbon associated with real losses are classified as actual physical waste to the environment, whereas the values associated with apparent losses are calculated strictly to complete the overall balance calculation for unbilled consumption.

3. Results and Discussion

3.1. Water Component in WEC Nexus

Figure 4 illustrates the average annual water balance of the MWA system from 2017 to 2024. A total raw water volume of 2267.91 MCM/yr was abstracted, with the East Raw Water Canal and the West Raw Water Canal contributing 74.81% and 25.19%, respectively. Losses during raw water conveyance were minimal at 0.98%.

The raw water undergwent processing across 4 major treatment plants: Bangkhen (contributing 67.05%), Mahasawat (25.07%), Samsen (4.89%), and Thonburi (2.00%). The treatment process incurred losses of 122.77 MCM/yr (5.41%); consequently, the net treated water volume available for downstream supply stood at 2122.85 MCM/yr.

Regarding the transport of treated water, the flow bifurcated into two pathways, with 1537.26 MCM/yr entering the transmission pipeline system, while 585.59 MCM/yr was distributed directly to service areas situated near the water treatment plants. The transmission network, spanning 191 km, had volumetric losses of 55.27 MCM/yr (2.44%).

In contrast, the distribution network emerged as the primary source of water loss. Total distribution losses amounted to 615.92 MCM/yr (27.16%), comprising 81.21 MCM/yr (3.58%) in apparent losses and 534.71 MCM/yr (23.58%) in real losses. As a result, only 1451.66 MCM/yr—equivalent to 64.01% of the total raw water abstraction—was delivered as authorized consumption.

The analysis revealed a structural distinction between upstream operations and the downstream network. The abstraction, treatment, and transmission processes are relatively centralized and confined, resulting in minimal volumetric losses. Conversely, the distribution system represented a vast and complex infrastructure, spanning over 40,000 km of pipelines and serving more than 2.6 million active connections. Due to this extensive scale and high connectivity density, the distribution stage accounted for the highest proportion of total water loss.

Figure 5 presents a comparative analysis of average annual input water volumes across MWA subsystems during the pre-COVID-19, COVID-19 lockdown, and post-COVID-19 periods. Relative to the pre-COVID-19 baseline, the COVID-19 lockdown period had an increase in input volumes across all upstream and midstream processes; however, authorized consumption declined slightly to 1429 MCM/yr. This discrepancy suggested that the pandemic likely induced operational irregularities within the supply system. Furthermore, a compensatory demand dynamic was observed: the surge in residential consumption driven by ‘stay-at-home’ and ‘work-from-home’ measures, and unbilled authorized consumption being counterbalanced effectively by a substantial decline in non-residential demand.

In the post-COVID-19 period, a contrasting trend emerged, signaling economic recovery. While system input volumes decreased from their pandemic peaks—with raw water declining to 2275 MCM/yr and distribution input to 2048 MCM/yr—authorized consumption rose substantially to 1484 MCM/yr (a 3.85% increase). This inverse relationship, characterized by rising consumption alongside reduced system inputs, indicates improved operational efficiency, a resurgence in economic activity, and the restoration of commercial water demand, reflecting the normalization of socioeconomic dynamics following the pandemic.

Figure 6 illustrates the comparative analysis of average annual water losses across the MWA subsystems. Based on these data, the Distribution System (Real Loss) constituted the predominant component of total water loss and had the most substantial volatility in response to pandemic conditions. During the COVID-19 lockdown, real losses surged from a pre-pandemic baseline of 541 MCM/yr to a peak of 604 MCM/yr. This 11.6% increase likely resulted from potential disruptions to active leakage control activities during the lockdown.

However, in the post-COVID-19 period, real losses substantially declined to 482 MCM/yr, dropping below pre-pandemic levels. This substantial improvement suggested the successful resumption and intensification of leakage management strategies following the relaxation of restrictions. A divergent trend was observed in the Treatment and Transmission systems, where there was a continuous upward trajectory in losses, independent of the pandemic phases. Unlike distribution losses (which recovered), these continuous increases suggested underlying structural or operational challenges in the upstream infrastructure that persisted regardless of pandemic-induced operational irregularities. While the Infrastructure Leakage Index (ILI) is widely recognized as the optimal metric for benchmarking operational performance across different water utilities, this study utilizes the absolute volume and percentage of real water losses. This volumetric approach was specifically selected because the analysis evaluates a single utility’s performance over time, and the approach directly corresponds to the subsequent WEC calculations, allowing for a straightforward and mathematically consistent quantification of embedded energy and carbon emissions within the macroscopic balance framework.

Table 1. Comparison of Operational Characteristics of MWA and Incheon Water Systems.

System Category	MWA, Thailand (2017–2024) 1	Incheon, Korea (2021) 2
Population (people)	8,226,820	2,918,314
Service Area (km2)	3195	1067
Number of Customer Meters	2,491,678	N/A
Number of Residential Meters	2,035,890	N/A
Raw Water System Inflow (MCM/yr)	2268	386
Treatment System Inflow (MCM/yr)	2246	402
Transmission and Distribution System Inflow (MCM/yr)	2123	347
Total Authorized Consumption (MCM/yr)	1452	321
Residential Consumption (MCM/yr)	698	N/A
Total Authorized Consumption per Capita (L/d/cap)	483	301
Residential Consumption per Capita (L/d/cap)	232	N/A

¹ Data from MWA (Thailand) based on average operational records during 2017–2024 [35,43,44]. ² Data from Incheon (Republic of Korea) sourced from the study conducted by Min et al. [42].

presents a comparative profile of the operational characteristics of the MWA, Thailand, and the Incheon Water System, Republic of Korea [42]. The data highlights major disparities in demographic coverage and infrastructure scale between the two metropolitan utilities. The MWA serves a substantially larger population and manages a more extensive service area, approximately three times the size of Incheon’s. This structural difference is mirrored in the volumetric throughput, where MWA’s system flows—from raw water intake to authorized consumption—were substantially higher than those of Incheon. Notably, the Incheon dataset presented a statistical anomaly wherein the reported raw water inflow (386 MCM/yr) is lower than the treatment system inflow (402 MCM/yr). Despite this discrepancy, analysis of the data suggested that Incheon maintained a considerably lower percentage of water losses than the MWA.

Furthermore, the comparison revealed a distinct variation in water usage intensity: the total authorized consumption per capita for the MWA was 483 L/d/cap, notably exceeding Incheon’s 301 L/d/cap. These differences reflected varying urban water demand profiles influenced by socioeconomic conditions, household consumption behavior, and service delivery models. The vast scale of MWA’s service area underscores the operational challenges associated with managing a complex distribution network, whereas Incheon’s lower per capita consumption implies a more conservation-oriented utilization model driven by differentiated tariffs, policy interventions, and consumer behavior.

3.2. Energy Component in WEC Nexus

Figure 7 illustrates the average annual electrical energy flow within the MWA system during 2017–2024, visualizing the accumulated electrical energy embedded across the water supply chain, tracing the cumulative electricity consumption from raw water abstraction to customer delivery. Notably, the electrical energy demand of the Raw Water System was minimal. Since predominantly, the system utilized open canals for conveyance, the electricity consumption was low and localized at two specific points: the Samlae Raw Water Pumping Station (PS), which lifts water from the Chao Phraya River into the East Water Canal; and the Bang Sue Raw Water PS, which transmits water via pipeline to the Thonburi WTP. Consequently, the associated energy losses in this process were negligible, attributed to both the low volumetric loss and the low energy intensity of the raw water transport phase.

Regarding the water treatment process, the profile of energy consumption was dictated primarily by the scale of WTP. Naturally, the absolute volume of energy consumed varied in direct proportion to production capacity, meaning the largest facility, Bangkhen WTP, consumed the highest total energy, whereas the smallest, Thonburi WTP, consumed the least. However, a deeper analysis using SEC as a normalized efficiency metric revealed a contrasting insight. By dividing the point energy consumption (from Figure 7) by the corresponding inflow volume (from Figure 4), the efficiency disparity became evident. For the Bangkhen WTP, the SEC was approximately 0.037 kWh/m³ (derived from 56.95 GWh divided by 1520.65 MCM). In contrast, the Thonburi WTP had a much higher SEC of 0.211 kWh/m³ (derived from 9.58 GWh divided by 45.46 MCM). These findings demonstrated that energy efficiency is superior in larger, modernized facilities, confirming the existence of substantial economies of scale within the water treatment process.

The energy input profile highlighted the significant energy intensity of the transmission and distribution processes. The Transmission Pumping Stations (at Bangkhen and Mahasawat plants) and the Distribution Pumping Stations served as the primary energy consumers, injecting 154.83 GWh/yr and 140.71 GWh/yr of point energy into the system, respectively. Consequently, the water entering the distribution network carries a substantial accumulated energy load of 401.01 GWh/yr.

A critical disparity was observed in the energy losses analyzing. Energy losses in the upstream processes were relatively minor, with the Raw Water, Treatment, and Transmission systems accounting collectively for only 15.85 GWh/yr. In stark contrast, the Distribution System was the predominant source of energy wastage, with the real energy losses in this network amounting to 103.76 GWh/yr, constituting 24.89% of the total accumulated electrical energy supplied to the MWA system. Additionally, apparent energy losses contribute a substantial 15.75 GWh/yr (3.78%). These figures underscore a critical inefficiency: losses occurring at the furthest downstream stage—where water effectively carries the cumulative energy investment of all prior processes—resulting in the most severe impact on overall system energy waste.

Figure 8 illustrates the comparative analysis of the average annual input electrical energy across the MWA subsystems during the pre-COVID-19, COVID-19 lockdown, and post-COVID-19 periods. The treatment system displayed a continuous upward trajectory independent of pandemic phases, rising incrementally from 103.98 GWh/yr (Pre-COVID) to 110.06 GWh/yr (Post-COVID). This trend indicated that the energy demand in water treatment was governed predominantly by the raw water inflow volumes and quality parameters rather than external demand shocks.

Conversely, the transmission system emerged as the most energy-intensive component, with considerable volatility in response to the pandemic. During the COVID-19 period, transmission energy consumption surged from a pre-pandemic baseline of 149.30 GWh/yr to a peak of 164.96 GWh/yr. Since the operation of the MWA transmission system does not have a firm strategic operational rule curve for pressure management like the MWA distribution system, this spike likely reflected operational inefficiency due to shifting demand patterns in each type of service areas (residential, non-residential). In the post-COVID-19 phase, its energy use moderated to 155.44 GWh/yr showing a more efficient energy use.

Notably, the distribution system demonstrated a sustained reduction in energy consumption. Annual energy use declined from 146.02 GWh/yr in the pre-pandemic period to 137.75 GWh/yr during the lockdown, stabilizing at 137.37 GWh/yr in the post-COVID-19 phase. This stability could be attributable primarily to the strict adherence to an optimized operational rule curve for pressure management. Under this approach, the pressures at all pumping stations are continuously adjusted in real time to maintain established target pressure profiles at the downstream ends of the network. Furthermore, the energy embedded in authorized consumption—representing the net useful energy delivered to customers—mirrored the trajectory of economic recovery while benefiting from the simultaneous reduction in distribution losses.

Figure 9 illustrates the average annual electrical energy losses across the MWA subsystems. Compared with the water loss data in Figure 6, there is a clear relationship between water flow and energy efficiency. The trends in energy losses (Figure 9) follow the same pattern as the water losses observed in Figure 6, most notable in the distribution system, which is the main source of waste in both categories. During the COVID-19 period, real water losses rose to 604 MCM/yr (Figure 6) due to disrupted leakage control, directly causing real energy losses to peak at 117.03 GWh/yr (Figure 9).

Notably, energy loss in the distribution stage is severe because this water carries high cumulative energy. Unlike losses in the raw water or treatment stages (where energy input is low), water lost in the distribution network has already absorbed the energy inputs from abstraction, treatment, and transmission. Consequently, when real water losses dropped to 482 MCM/yr in the post-COVID-19 period (Figure 6), it resulted in a considerable reduction in the associated energy waste to 93.78 GWh/yr (Figure 9), proving that restoring active leakage control (ALC) and pressure management strategies after the pandemic were successful in reducing both water and energy waste.

Table 2. Comparison of Water Supply Energy Consumption and Intensity (MWA versus Incheon) with General Ranges.

System Category		MWA, Thailand (2017–2024) 1	Incheon, Republic of Korea (2021) 2	General 3
Energy Consumption (GWh/yr)	Raw Water System	12.94	68.15	N/A
	Treatment System	107.70	20.01	N/A
	Transmission and Distribution System	296.23	59.13	N/A
Total		416.87	147.29	N/A
SEC (kWh/m3)	Raw Water System	0.006	0.169	0.00–3.70
	Treatment System	0.052	0.050	0.03–4.23
	Transmission and Distribution System	0.143	0.147	0.03–0.58
Total		0.202	0.366	0.06–8.51
Total Electricity Consumption due to WSS per Capita (kWh/d/cap)		0.139	0.138	N/A
Residential Electricity Consumption due to WSS per Capita (kWh/d/cap)		0.067	N/A	N/A

¹ Data from MWA (Thailand) based on average operational records 2017–2024 [35,43,44]. ² Data from Incheon (Republic of Korea) sourced from study conducted by Min et al. [42]. ³ General reference range derived from global benchmark values reported in the comparative review of water–energy–carbon intensity in urban water systems [10].

presents a comparative analysis of energy consumption and intensity between the MWA (Thailand) and Incheon (Republic of Korea) water systems, benchmarked against global ranges. Consistent with the operational scale differences presented in Table 1, the MWA consumed a substantially higher amount of total electrical energy (416.87 GWh/yr) than Incheon (147.29 GWh/yr), a disparity expected given MWA’s threefold larger service population. However, when normalized by production volume, the analysis revealed a divergence in efficiency. The MWA demonstrated superior performance with a Total SEC of 0.202 kWh/m³, being one-half of Incheon’s 0.366 kWh/m³. Thus, while the MWA system required a larger absolute energy input, it maintained a considerably lower energy intensity per unit of water produced.

The most critical factor driving this efficiency gap lies within the Raw Water System. The MWA recorded an exceptionally low SEC of 0.006 kWh/m³, whereas Incheon’s figure was notably higher at 0.169 kWh/m³. This nearly 30-fold difference can be attributed primarily to the raw water conveyance method and geographical context. The MWA utilizes an open canal system situated on the Chao Phraya River Delta, where water is lifted only once at the source and conveyed via gravity. In contrast, Incheon relies on a long-distance pressurized pipeline system to transport water from the Paldang Dam across a complex metropolitan topography, necessitating substantially higher pumping energy. Conversely, the Treatment and Distribution systems had remarkable parity in energy intensity. The SEC values for treatment (~0.05 kWh/m³) and distribution (~0.14 kWh/m³) were closely aligned between the two utilities, suggesting that both cities utilized comparable technologies and operational standards for water purification and reticulation.

Notably, the Total Electricity Consumption per capita was virtually identical for both cities (~0.139 kWh/d/cap), as MWA’s higher per capita water consumption (Table 1) is offset by its lower energy intensity. Overall, MWA’s performance was at the highly efficient end of the global general value range (0.06–8.51 kWh/m³). While this confirms that the utility is effectively leveraging economies of scale and favorable geographical conditions, it is also largely attributable to the implementation of active pressure management strategies [14,16,27]. By maintaining distribution network pressures at optimized levels to mitigate leakage—potentially lower than the high-pressure standards typical of some developed nations—the MWA inherently reduces pumping head requirements, thereby contributing to one of the lowest energy intensities among major metropolitan water systems.

3.3. Carbon Component in WEC Nexus

Figure 10 illustrates the average annual flow of carbon emissions, expressed as CO₂e, within the MWA system. As the final component of the WEC nexus analysis, these emissions were derived exclusively from the electricity consumption presented in Figure 7, calculated by applying the national grid emission factor (Equations (12) and (13)). Consequently, the figure translates electricity use patterns into climate-relevant impacts, providing a carbon-based representation of system operations.

Based on this analysis, the transmission system was the largest contributor to operational emissions, totaling 74,192.49 tCO₂e/yr, followed by the distribution system at 67,170.33 tCO₂e/yr. Consistent with the energy analysis, downstream processes accumulate the carbon footprint of all upstream stages. While carbon losses in the raw water, treatment, and transmission systems were relatively minor (collectively 7544.58 tCO₂e/yr), the distribution system was the dominant source of carbon wastage. Real carbon losses in this network were 49,562.09 tCO₂e/yr, representing 24.89% of the total accumulated carbon supplied to the system. Additionally, apparent carbon losses contributed 7514.61 tCO₂e/yr (3.78%).

These findings highlighted a critical inefficiency in the urban water supply system, where carbon losses occurring in the downstream stage—where water effectively embodies the cumulative energy and carbon investments of all prior processes—resulted in the most severe impact on overall system performance. This pattern mirrored the electricity consumption trends and underscored that addressing distribution-level inefficiencies is the most effective strategy for reducing greenhouse gas emissions in large-scale water supply systems.

Figure 11 and Figure 12 extend the analysis of electrical energy consumption by translating energy use and losses in each MWA subsystem into their associated carbon emissions and GHG potential. As carbon emissions in the water supply system are directly proportional to electricity consumption under a fixed grid emission factor, the observed trends closely mirrored those discussed based on Figure 8 and Figure 9.

Figure 11 illustrates the annual total carbon emissions across the MWA subsystems during the pre-COVID-19, COVID-19 lockdown, and post-COVID-19 periods. Consistent with the energy analysis, the transmission system remained the dominant contributor to total carbon emissions throughout all three periods. Carbon emissions from transmission increased markedly during the COVID-19 period, rising from 72,197 tCO₂e/yr in the pre-pandemic phase to a peak of 79,307 tCO₂e/yr. This increase directly reflected the surge in transmission energy consumption observed in Figure 8, confirming that the lack of a firm operational rule curve in the transmission system led to energy- and carbon-intensive operations under fluctuating demand conditions.

In contrast, the treatment system exhibited a gradual increase in carbon emissions from 50,299 tCO₂e/yr (pre-COVID-19) to 51,531 tCO₂e/yr (post-COVID-19), following the same monotonic trend observed in the electrical energy input. This reinforced the conclusion that carbon emissions in water treatment were governed primarily by raw water inflow volumes and treatment requirements rather than short-term demand disruptions.

However, there was a sustained reduction in carbon emissions in the distribution system, decreasing from 70,653 tCO₂e/yr pre-COVID-19 to 64,317 tCO₂e/yr post-COVID-19. This trend corresponded directly to the decline in distribution energy consumption discussed earlier, highlighting the effectiveness of strict adherence to optimized operational rule curves, pressure management, and leakage control in reducing both energy use and carbon emissions simultaneously.

Figure 12 further links electrical energy losses to GHG potential by focusing on the carbon losses associated with inefficient water use. Based on those results, energy-related carbon losses closely tracked the patterns of real and apparent water losses presented in Figure 6. The distribution system, particularly real water losses, was the greatest source of carbon loss due to the high cumulative energy embedded in the distributed water. During the COVID-19 period, when real water losses increased to 604 MCM/yr, the associated carbon loss peaked at 56,272 tCO₂e/yr, directly reflecting the peak in real energy losses shown in Figure 9.

Following the restoration of active leakage control (ALC) and pressure management in the post-COVID-19 period, real water losses declined substantially, leading to a corresponding reduction in carbon losses to 43,910 tCO₂e/yr. This confirmed that measures aimed at reducing electrical energy losses in the distribution network deliver compounded benefits by simultaneously lowering water losses, energy waste, and GHG emissions. Apparent losses contributed a comparatively smaller share of carbon losses across all periods, emphasizing that physical leakage reduction remains the most effective pathway for carbon mitigation in urban water supply systems.

Table 3. Comparison of Carbon Emission and GHG Potential in Water Supply Systems (MWA versus Incheon) with General Ranges.

System Category		MWA, Thailand (2017–2024) 1	Incheon, Republic of Korea (2021) 2	General 3
EF (kg CO2e/kWh)		0.4788	0.4663	0.1734–2.5316
Carbon emission (tCO2e/yr)	Raw Water System	6195	9330	N/A
	Treatment System	51,565	27,569	N/A
	Transmission and Distribution System	141,829	68,676	N/A
Total		199,589	105,575	N/A
GHG potential (gCO2e/m3)	Raw Water System	3.00	78.80	N/A
	Treatment System	24.94	23.31	N/A
	Transmission and Distribution System	68.60	68.54	N/A
Total		96.53	170.65	54–4903
Total carbon emission due to WSS per Capita (gCO2e/d/cap)		66.47	64.47	N/A
Residential carbon emission due to WSS per Capita (gCO2e/d/cap)		31.96	N/A	N/A

¹ Data from MWA (Thailand) based on average operational records during 2017–2024 [35,43,44]. ² Data from Incheon (Republic of Korea) sourced from study conducted by Min et al. [42]. ³ General reference range derived from global benchmark values reported in comparative review of water–energy–carbon intensity in urban water systems [10]. EF is a variable.

provides a comparative assessment of carbon emissions and GHG potential between the MWA and Incheon water systems with general ranges. First, EF for Thailand (0.4788 kg CO₂e/kWh) and the Republic of Korea (0.4663 kg CO₂e/kWh) were very similar. Both values are quite low compared to the general range (0.1734–2.5316 kg CO₂e/kWh), showing that the cleanliness of the national power grids in both countries resulted in a lower impact on carbon emissions than in other countries. In addition, this similarity isolated the impact of energy source variation, implying that any differences in carbon performance were being driven primarily by operational efficiency and hydraulic characteristics rather than the cleanliness of the national power grids.

When analyzing the GHG intensity, a disparity is observed in the raw water system. MWA records an extremely low GHG potential of 3.00 g CO₂e/m³, whereas Incheon’s value is nearly 27 times higher at 78.80 CO₂e/m³. This validates the energy analysis in Table 2, confirming that the gravity-fed open canal system on the Chao Phraya River Delta considerably reduces the carbon footprint compared to Incheon’s energy-intensive raw water pumping.

In contrast to the raw water stage, there was consistency in the downstream processes, with the GHG potentials for the treatment system (24–25 g CO₂e/m³) and the transmission and distribution system (67–69 g CO₂e/m³) being almost the same for both utilities. This indicated that both cities used similar technologies for water production and distribution. Overall, the MWA had a considerably lower total GHG potential (96.53 g CO₂e/m³) than Incheon (170.65 g CO₂e/m³). Based on benchmarking against the general global range (54–4903 g CO₂e/m³), both systems performed at the highly efficient lower end of the spectrum.

Furthermore, there was an interesting trade-off in the per capita analysis. The total carbon emission per capita was nearly equal for both cities (66.47 g CO₂e/d/cap for the MWA versus 64.47 g CO₂e/d/cap for Incheon) because the MWA’s higher water consumption per person (Table 1) was effectively offset by its lower carbon intensity per unit.

3.4. Water–Energy–Carbon Balances According to IWA-LEI

Figure 13 presents the average WEC balance of the MWA distribution system (2017–2024), utilizing the Leakage Emissions Initiative framework developed by the International Water Association (IWA-LEI) [26]. This method applies a simple proportional allocation approach. Assuming that every cubic meter of water supplied to the distribution system carries the same embedded electricity and carbon intensity.

During the study period, the distribution system supplied an average of 2067.58 MCM/yr, consuming 416.87 GWh/yr of electricity and generating 199,597 tCO₂e/yr of carbon emissions. Based on the IWA-LEI calculation, the energy loss associated with water losses was 124.18 GWh/yr, consisting of 16.37 GWh/yr from apparent losses and 107.81 GWh/yr from real losses. Compared to the detailed WEC analysis in Figure 7 (which reported 15.75 GWh/yr for apparent and 103.76 GWh/yr for real losses), the IWA-LEI method was a slight overestimation; however, it remained consistent with the detailed findings.

Subsequently, carbon emissions were derived using the grid emission factor of 0.4788 kg CO₂e/kWh. Based on these results, there were total carbon losses of 59,457 tCO₂e/yr, comprising 7838 tCO₂e/yr from apparent losses and 51,619 tCO₂e/yr from real losses. Similar to the energy results, the comparison with the detailed analysis in Figure 10 (7515 and 49,562 tCO₂e/yr, respectively) showed that the LEI method resulted in a slight overestimation as well.

In conclusion, the IWA-LEI proportional allocation approach was a valuable tool. It provided a transparent, first-order estimation of energy and carbon losses, making it suitable for urban water supply systems that may lack detailed spatial or component-level data.

The quantitative analysis in

Table 4. Comparison of yearly real water loss, real energy loss, and real carbon loss in distribution process between WEC Nexus and IWA-LEI.

Year	Real Water Loss 1 (MCM/yr)	SEC (kWh/m3)	Real Energy Loss (GWh/yr)		Error (GWh/yr)	%Error	Real Carbon Loss (tCO2e/yr)		Error (tCO2e/yr)	%Error
	WEC Nexus/IWA-LEI		WEC Nexus	IWA- LEI			WEC Nexus	IWA-LEI
2017	567.19	0.203	112.17	115.28	3.11	2.78%	53,707	55,197	1490	2.78%
2018	511.62	0.196	96.78	100.03	3.25	3.36%	46,338	47,897	1558	3.36%
2019	543.30	0.202	105.71	109.49	3.78	3.57%	50,614	52,423	1809	3.57%
2020	598.14	0.199	114.93	119.15	4.22	3.67%	55,029	57,049	2020	3.67%
2021	610.07	0.204	119.12	124.25	5.13	4.30%	57,035	59,489	2454	4.30%
2022	560.27	0.211	110.40	118.20	7.80	7.07%	52,860	56,595	3736	7.07%
2023	453.28	0.200	86.86	90.87	4.01	4.62%	41,589	43,509	1920	4.62%
2024	433.81	0.203	84.09	88.04	3.95	4.70%	40,262	42,153	1891	4.70%

¹ Note: Data from MWA (Thailand) based on average operational records during 2017–2024 [35,43,44,45,46].

from 2017 to 2024 demonstrates a high degree of convergence between the WEC Nexus and IWA-LEI frameworks, despite their distinct methodological foundations. The annual real water loss data in the distribution process clearly demonstrates the magnitude of the operational trend shifts across the pre-COVID (2017–2019), during-COVID (2020–2021), and post-COVID (2022–2024) periods. While the WEC Nexus model is based on actual operational electricity consumption (Actual Energy), accounting for real-world systemic fluctuations, the IWA-LEI approach relies on theoretical estimations derived from the volume of real losses multiplied by SEC. However, to estimate the carbon emissions, both methods use the same approach by multiplying the energy consumption with EF.

The energy dimension analysis reveals a Root Mean Square Error (RMSE) of 4.63 GWh, with annual percentage errors (%error) ranging from 2.78% to 7.07%. These low error margins, relative to the total energy scale of 80–120 GWh, validate the high precision of the SEC-based model in forecasting energy dissipation from water leakage. Regarding carbon emissions, calculated using a Grid Emission Factor of 0.4788 kgCO_2e/kWh, the analysis yielded an RMSE of 2177 tCO_2e/yr. Although the absolute RMSE value appears elevated due to the large scale of the baseline data (40,000–59,000 tCO_2e/yr), the proportional accuracy remains perfectly aligned with the energy dimension. In conclusion, the IWA-LEI framework slightly overestimates both real energy and carbon losses compared to the WEC Nexus model.

To provide actionable insights for decision-makers,

Table 5. Scenario analysis of real water loss reduction and its corresponding impacts on real energy and carbon losses in the distribution process (baseline year: 2024).

Scenario	Real Water Loss 1 (MCM/yr)		Real Energy Loss (GWh/yr)		Real Carbon Loss (tCO2e/yr)
	Value	Reduction	Value	Reduction	Value	Reduction
Baseline (2024)	433.81	-	88.02	-	42,144	-
5% Reduction	412.12	21.69	83.62	4.40	40,037	2107
10% Reduction	390.43	43.38	79.22	8.80	37,930	4214
15% Reduction	368.74	65.07	74.82	13.20	35,823	6321
20% Reduction	347.05	86.76	70.42	17.60	33,715	8429

¹ Note: Data from MWA (Thailand) based on average operational records during 2024 [35,43].

presents a scenario analysis of carbon mitigation potential based on the 2024 baseline. The scenarios evaluate the environmental impacts of reducing real water losses by increments of 5%, 10%, 15%, and 20%. By applying the 2024 SEC and GHG emission values, the analysis reveals a clear marginal benefit: every cubic meter (m³) of real water saved yields a direct energy reduction of approximately 0.203 kWh and mitigates 97 gCO₂e of carbon emissions. Consequently, the results demonstrate a significant and linear environmental benefit. For instance, achieving a 20% reduction in real water losses would conserve 86.76 MCM/yr of water, subsequently saving 17.60 GWh/yr of embedded energy and mitigating 8429 tCO₂e/yr of carbon emissions. From a financial perspective, based on the MWA’s 2024 annual report [35], the direct operating cost for water production is approximately 4.73 THB/m³. Therefore, achieving this 20% reduction would yield estimated operational savings of 410.37 million THB annually. These quantifiable metrics offer crucial support for water utility managers conducting cost–benefit analyses for investments in active leakage control and other water loss reduction strategies. Furthermore, they help foster broader public policy development toward sustainable, low-carbon urban water management.

3.5. Water Loss Reduction Target

Table 6. Trends in target and actual water loss and corresponding reductions in electricity consumption and carbon emissions of the MWA (2017–2024).

Fiscal Year	Water Production (MCM/yr)	Water Loss Percentage (%)		Water Loss Reduction Beyond Target 1 (MCM/yr)	Electricity Saving Beyond Target 1 (GWh/yr)	Carbon Reduction Beyond Target 1 (tCO2/yr)
		Target Water Loss (%)	Actual Water Loss (%)
2017	2064	27.50	31.70	−86.68	−17.51	−8383.44
2018	1997	28.75	29.80	−20.97	−4.24	−2028.12
2019	2075	29.29	29.30	−0.21	−0.04	−20.07
2020	2121	27.29	31.25	−84.00	−16.97	−8123.85
2021	2117	30.20	33.10	−61.38	−12.40	−5936.38
2022	2080	30.50	31.60	−22.88	−4.62	−2213.22
2023	2042	27.50	27.10	8.17	1.65	790.15
2024	2038	26.00	25.68	6.52	1.32	630.79

¹ Note: Positive values indicate performance exceeding target (additional savings), while negative values indicate performance falling short of target (excess losses) [36].

presents the operational performance of the MWA from fiscal years 2017–2024 (based on October 1 to September 30 each year), sourced for the official annual reports [35,43,44]. The dataset includes total water production, organizational water loss targets, and actual water loss percentages. The current study applied the IWA-LEI framework to evaluate the environmental impact of these operations. Specifically, the metrics for water loss reduction beyond target, electricity saving beyond target, and carbon emission reduction beyond target were calculated by integrating the deviation between actual and targeted performance with SEC and EF from Figure 13.

The resulting analysis highlighted the trends in water loss management and their corresponding environmental impacts. According to the IWA-LEI framework, every unit of potable water carries embedded energy and embedded carbon. Consequently, any deviation between the actual water loss and the organizational target translated directly into either a waste or a saving of energy and carbon. The operational data from 2017 to 2022 highlighted a period of major challenge, where the actual water loss percentages consistently exceeded the set targets. This resulted in negative values for electricity and carbon reduction, representing avoidable inefficiencies. This prolonged underperformance, particularly during 2020–2022, was largely attributable to operational irregularities induced by the COVID-19 pandemic.

Subsequently, a distinct operational turnaround commenced in 2023, as the MWA implemented aggressive water loss reduction strategies to recover system efficiency in the post-pandemic phase. The utility successfully managed to suppress actual water losses below the established targets (such as 25.68% actual versus 26.00% target in 2024). This improved performance yielded positive values in the reduction columns, signifying energy and carbon savings that extended beyond the target. Specifically, in 2024, this large reduction contributed to a saving 630.79 tCO₂e/yr beyond the baseline expectation. This achievement aligned with the concept of Carbon Leakage Credits (CLCs) proposed by the IWA-LEI. CLCs are defined as environmental attributes generated annually when a utility successfully reduces its leakage below the initially set benchmark. Therefore, the surplus carbon savings achieved by the MWA in 2023 and 2024 could theoretically be recognized as CLCs, transforming technical efficiency into a quantifiable environmental asset.

4. Conclusions

This study provided a comprehensive WEC nexus analysis of the MWA, utilizing operational data from 2017 to 2024. In terms of energy performance, MWA operates with a total SEC of 0.202 kWh/m³ (Table 2), indicating high energy efficiency compared to the Incheon water system in the Republic of Korea (0.366 kWh/m³) [42] and the global reported range (0.06–8.51 kWh/m³) [10]. Based on these results, although the MWA required a larger absolute amount of energy due to its system scale, it operates with a much lower energy intensity per unit of water produced. This superior energy performance is directly reflected in the system’s carbon intensity. When benchmarked against Incheon, the MWA had a considerably lower total GHG potential of 96.53 gCO₂e/m³ (Table 3), compared to 170.65 gCO₂e/m³ [42] for Incheon. Primarily, this disparity was driven by the raw water system, where the MWA’s gravity-fed open canal network achieved a minimal carbon intensity of 3.00 gCO₂e/m³, whereas Incheon’s energy-intensive pumping system records 78.80 gCO₂e/m³ [42]. Furthermore, in the global context, the MWA’s performance lies at the highly efficient end of the general value range (54–4903 gCO₂e/m³) [10], confirming that the utility effectively leverages economies of scale and favorable geographical conditions.

Despite this inherent topographical advantage, the analysis revealed that real water loss was the dominant source of inefficiency in the downstream process, accounting for approximately 25% of the total accumulated carbon emissions supplied to the system and confirming that a substantial portion of the utility’s carbon footprint was effectively vented through pipeline leakages. From a methodological perspective, the application of the IWA-LEI framework proved successful, as the simplified proportional allocation approach produced results consistent with the detailed spatial WEC analysis, validating the LEI as a reliable and transparent tool for carbon accounting in large-scale urban water systems.

Finally, the current study demonstrated a viable pathway for decarbonization through operational recovery and system management. Following the disruptions caused by the COVID-19 pandemic (2020–2022), the MWA implemented aggressive pressure management and active leakage control strategies in 2023–2024, which were successful in reducing water losses below organizational targets and generating quantifiable environmental benefits. These surplus savings can theoretically be recognized as CLCs, thereby transforming technical efficiency into valuable environmental assets. Therefore, the conclusion for the current research was that for mega-cities in flat delta regions, while raw water transport may be inherently low-carbon, rigorous distribution network management generates the highest return on investment for climate change mitigation.

Regarding practical transferability, both WEC and IWA-LEI methodologies are universally applicable to water utilities worldwide. However, the magnitude of energy and carbon mitigation strongly depends on local topography and hydraulic contexts. Systems relying heavily on pumping, such as expansive flat networks, exhibit significantly higher marginal benefits from leakage reduction compared to gravity-fed mountainous systems, where the carbon footprint of physical water losses is inherently lower. To account for these diverse regional characteristics, this study highl4ights the critical need for standardized benchmarking. Similar to how reliable performance indicators for real water losses are typically normalized by the average operating pressure of the distribution system, future assessments of energy and carbon losses should be normalized by variables representing local topography and hydraulic conditions, such as the average pumping head. This approach will ensure fair and objective comparisons of carbon mitigation performance across different geographical and operational settings.