An analysis of indirect water withdrawal and consumption due to electricity and workers in automotive manufacturing facilities

Abstract

Water is a key resource for all life. Recent droughts have also exemplified the importance for manufacturers to understand their impact on water resources. While most manufacturers typically know the amount of water they use in their facilities, they have little knowledge about the indirect effects they have on water resources. In this paper, the indirect impact from the automakers’ electricity use and work force on water resources is examined. Water withdrawal and consumption from electricity use by hypothetical but representative facilities around the world is quantified and analysed. Water withdrawal by the workers is also quantified and analysed. The results indicate that the water withdrawal and consumption by the workforce and caused by the use of electricity is larger than the direct water use and consumption in the facilities themselves.

Keywords:

• Water withdrawal
• consumption
• automotive
• electricity
• workforce

1. Introduction

For automotive manufacturing in general, the water use is an important resource because it impacts cost, brand image and the relationship with the local actors near facilities (CDP 2014). Many automakers are tracking direct facility use of water. This direct water use or withdrawal is, however, only a partial accounting of the impact auto manufacturing has on water resources. Analogous to scope 1, 2 and 3 greenhouse gas emissions, manufacturers also have indirect water use, withdrawal and consumption. Having a complete understanding of the indirect water use is important, because even if the water is not directly withdrawn for a facility, that water is removed from the source and is unavailable for other purposes, which can increase the stress in a location (Schornagela et al. 2012). For example, studies of the amount of water use (withdrawal) for different industrial sectors in the US have shown that a large majority of water withdrawal is used for power generation, followed by irrigation, non-domestic public supply and industrial water use (Blackhurst, Hendrickson, and Vidal 2010). Many other authors identify the importance of water in energy generation, e.g. (Feeley Iii et al. 2008; Fthenakis and Kim 2010; Gleick 1994). Fossil fuels used in energy production require large quantities of water for extraction or mining and may be produced in water stressed localities. Water is also required for processing, refining and distributing these fuels as well as for cooling and maintenance of thermoelectric power plants where they are burned (Yen and Bras 2012).

Although many studies have been performed that focus on energy use or carbon emissions during an automobile’s life cycle, few have focused on water-related issues. Those that do mention water consumption are generally limited to case studies of a specific plant, location or vehicle. According to (Kim et al. 2015), most studies of water use associated with vehicles have evaluated only the vehicle operation phase (oil extraction and fuel refining) and that there are only three published studies of life cycle water consumption of vehicles that include material production and vehicle manufacturing.

In (Tejada et al. 2012b), a life cycle inventory for water consumption for a given vehicle was performed. Large differences in water use vs. consumption were noted. A report by Volkswagen detailing a 10-year life cycle inventory of four VW Golf A4 variants echoed these inventory results (Schweimer and Levin 2000).

In (Tejada, Bras, and Guldberg 2012a), indirect water consumption due to electricity generation was added to the material and manufacturing phases presented in (Tejada et al. 2012b), and it was noted that the water consumption of electricity generation was significant in this case. The importance of water in electricity production was also noted in (Kim et al. 2015) where the authors compare two specific vehicles and state that the battery electric vehicle withdraws seven times more water over its life cycle than the internal combustion engine vehicle because of the large volume of cooling water used in thermoelectric power plants. The water consumption, however, is comparable.

Another study of three Volkswagen models followed an impact-oriented water footprint method to quantify freshwater consumption (Berger et al. 2012). This study, however, indicated that 95% of water was consumed during the production phase, predominantly due to material production. This is contrary to the findings in (Tejada et al. 2012b) which cited the use phase as the primary consumer of water because of the water needed for oil exploration and fuel production. These differences highlight the lack of standardisation inherent in current water focused life cycle inventory and analysis data and studies. This was exemplified in (Semmens, Bras, and Guldberg 2014) where a comparison was made between water use and consumption for vehicle production based on a historical data from corporate sustainability reports from select automotive companies. Inconsistencies in data reporting and definitions were noted even with data from single automotive companies.

These prior studies indicate that the amount of water consumed and used directly and indirectly in vehicle manufacturing is not insignificant and requires further study. This paper examines the indirect water impact that can occur from (a) electricity generation and (b) the workers in their daily life for an automotive company with different manufacturing locations around the world.

2. Hypothetical Automotive Company with global production

2.1. Automotive manufacturing facility location and water data

For the analyses in this paper, a Hypothetical Automotive Company (HAC) was created based on public data collected about current automotive manufacturers. For example, Volkswagen lists their facilities and many production and worker figures publically (Bentley 2013; Volkswagen 2014). Ford (2014b), Subaru (Domestic, FHI 2014; Overseas, FHI 2014), Hyundai (2014), BMW (2014a), and Fiat-Chrysler (FCA) (2014) have similar public references. All of the HAC locations are located close to, or exactly at, a location from one of these automakers. For example, the USA Car and Truck locations are located near Detroit where Ford, GM and FCA all have substantial factories. The India Car and Truck are located near Chennai, India, which is a very industrial area with automotive production. The Germany Car location is located near numerous vehicle factories. Mexico Car, China Car and Truck, Japan Car and Truck, South Korea Car, and Brazil Car and Truck are all located close to other automotive manufacturing locations as well. A notable location is the UK Super Luxury Car location, which is loosely modelled from Bentley (Bentley 2013).

Water data were collected from Corporate Sustainability Reports (CSRs) and web searches for various automakers in order to create hypothetical but representative automotive production facilities in varies parts of the world. Both car and truck production facilities are included, as well as a ‘luxury car’ production facility. These facilities and their representative of water data are presented in Table 1. The country of location of each facility is given in Table 2 as well as each facilities production. The Super Luxury car plant has the highest per vehicle impact, which is not surprising due to the low production volume and resources spent in producing these super luxury cars. In Table 1, total withdrawal is the sum of surface water, ground water and municipal water. Total consumption is total withdrawal minus discharge and recycled water.

Table 1. Water data for representative global facilities for hypothetical automotive company (all water values are in m³/year).

Facility name	Surface water	Ground water	Municipal water	Total withdrawal	Discharge	Recycled	Total consumption
USA Car			400,000	400,000	200,000	100,000	100,000
USA Truck			800,000	800,000	400,000	200,000	200,000
India Car	200,000			200,000	100,000	50,000	50,000
India Truck	400,000			400,000	200,000	100,000	100,000
Germany Car			500,000	500,000	250,000	125,000	125,000
Mexico Car	200,000			200,000	100,000	50,000	50,000
China Car		200,000		200,000	100,000	50,000	50,000
China Truck			400,000	400,000	200,000	100,000	100,000
Japan Car			400,000	400,000	200,000	100,000	100,000
Japan Truck			800,000	800,000	400,000	200,000	200,000
South Korea Car			2,000,000	2,000,000	1,000,000	500,000	500,000
Brazil Car	200,000			200,000	100,000	50,000	50,000
Brazil Truck	400,000			400,000	200,000	100,000	100,000
UK Super Luxury			110,000	110,000	55,000	27,500	27,500

Table 2. Water data for representative global facilities per unit production.

Facility name	Country	Production [vehicles/yr]	Withdrawal per vehicle [m^3/vehicle]	Consumption per vehicle [m^3/vehicle]
USA car	USA	100,000	4.00	1.00
USA Truck	USA	200,000	4.00	1.00
India car	India	50,000	4.00	1.00
India Truck	India	100,000	4.00	1.00
Germany Car	Germany	200,000	2.50	0.63
Mexico Car	Mexico	50,000	4.00	1.00
China Car	China	50,000	4.00	1.00
China Truck	China	100,000	4.00	1.00
Japan Car	Japan	100,000	4.00	1.00
Japan Truck	Japan	200,000	4.00	1.00
South Korea Car	South Korea	500,000	4.00	1.00
Brazil Car	Brazil	50,000	4.00	1.00
Brazil Truck	Brazil	100,000	4.00	1.00
UK Super Luxury	United Kingdom	2000	55.00	13.75

2.2. Electricity consumption for vehicle production

The manufacturing of vehicles requires the use of electricity, and that electricity usage also contributes to indirect water use (Semmens, Bras, and Guldberg 2014). Automotive manufacturing companies track their energy use and disclose it in their corporate sustainability reports (CSR). Table 3 is a collection of these values from six different major automakers. Not all automakers report electricity use in their corporate sustainability reports, but averages of 2.15 MWh of total energy needed to produce a vehicle of which 1 MWh is electricity seem to be good assumptions. Not all energy needed to produce a vehicle is in the form of electricity. A large portion of the rest of the total energy is mostly natural gas, which is typically used to heat facilities and also for powering so-called powerhouses that generate steam, compressed air and sometimes even some on-site electricity for the facilities. For the analysis in this paper, the average value of 0.99 MWh/vehicle will be used in conjunction with the automotive facility profiles for the indirect usage calculation of total electricity. The average electricity needed to produce a vehicle can be used with the facility production amounts to examine the electricity use for the facilities.

Table 3. Energy Intensities from various automakers CSR’s in MWh/Vehicle.

Automaker	Total energy consumption per vehicle [MWh/vehicle]	Electricity consumption per vehicle [MWh/vehicle]	Reference
GM	2.19		GM (2014a)
Volkswagen	2.21	1.08	VW (2014) and Dooley, Kyle, and Davies (2013)
Ford	2.44		Ford (2014a)
Nissan/Renault	2.19	1.00	Nissan-Renault (2014)
Peugeot SA	1.46	0.80	Peugeot (2014)
BMW	2.40	1.06	BMW (2014b)
Average	2.15	0.99

3. Water impact of electricity generation

3.1. Water impact of different generation technologies

Many different researchers and organisations have calculated values of water withdrawal and consumption by energy generation type for the use in life cycle assessments. Two papers that aggregate some of these results are by Semmens, Bras, and Guldberg (2014) and Dooley, Kyle, and Davies (2013). Table 4 is a collection of the information used to calculate the indirect water withdrawal and consumption. These are some averages for current technologies. Some of these numbers can change significantly if, e.g. different state-of-the-art cooling technologies are used in the future. Notable are the low consumption and withdrawal amounts for photovoltaics (PV). The only water really needed is for cleaning PV panels, making this renewable energy source not only excellent option for reducing greenhouse gas emissions, but also for reducing water consumption and withdrawal. The other extreme is hydroelectric power generation which has large consumption amounts due to water evaporation from its lakes. Withdrawal is zero because water is not altered when it passes through hydroelectric turbines.

Table 4. Water consumption and withdrawal values for electricity sources in m³/MWh.

Electricity source	Water consumption (Semmens, Bras, and Guldberg 2014)	Water withdrawal (Dooley, Kyle, and Davies 2013)
Coal	0.871	80.90
Natural gas	0.547	25.23
Nuclear	1.643	98.59
Hydroelectric	17.000	0
Solar (PV)	0.023	0.02

3.2. Electricity generation grid mix by source for different countries

With the water use information for different sources, the other piece of information needed to calculate the indirect water use by electricity is to find the electricity resource profiles for the countries where the HAC operates. The IEA (IEA 2012) tracks the electricity resource profiles of a majority of countries worldwide. The IEA statistics were in GWh for each type of resource, so to create a percentage profile, the GWh for each resource was divided by the total GWh for that country, and included in Table 5.

Table 5. Electricity resource profile for selected countries (IEA 2012).

	Coal (%)	Natural Gas (%)	Nuclear (%)	Solar (PV) (%)	Wind (%)	Hydro (%)
Brazil	2.6	8.5	2.9	0.0	0.9	75.2
China	75.8	1.7	2.0	0.1	1.9	17.5
Germany	45.6	12.3	15.8	4.2	8.0	4.4
India	71.1	8.3	2.9	0.2	2.5	11.2
Japan	29.3	38.4	1.5	0.7	0.5	8.1
Mexico	11.7	51.4	3.0	0.0	1.2	10.8
South Korea	44.8	20.9	28.1	0.2	0.2	1.4
UK	39.6	27.5	19.4	0.3	5.4	2.3
USA	38.3	29.5	18.7	0.2	3.3	7.0

With the average electricity use for a vehicle, water withdrawal and consumption for different types of electricity generation, and the resource profiles of the countries in which the HAC operates, it is now possible to calculate the indirect water withdrawal and consumption for the HAC.

3.3. Calculation of indirect water withdrawal and consumption by energy

The HAC profile established a realistic production number for the different facilities with respect to their location. With realistic production information and the average energy intensity use from automakers CSR’s, it is possible to calculate the electricity usage for each facility. With the country-based electricity resource profile, the MWh usage by each facility can be broken down by electricity source, which can be multiplied by the water withdrawal or consumption information to find the indirect withdrawal or consumption for each facility by means of Equations (1(1) ) and (2(2) ).(1) (2)

These equations follow the standard practice used to calculate the water consumption and withdrawal from electricity sources (Dooley, Kyle, and Davies 2013; Semmens, Bras, and Guldberg 2014). The novelty of calculating both types of water use is that withdrawal is typically not calculated as an indirect use. Many CSR’s from automakers include indirect CO₂ emissions, indirect waste or indirect water consumption, but none documented include indirect water withdrawal due to energy (BMW 2014b; Fiat-Chrysler 2014; Ford 2014a; GM 2014a; Nissan-Renault 2014; Peugeot 2014; VW 2014).

4. Indirect water withdrawal and consumption due to electricity

4.1. Results

The indirect withdrawal and consumption from electricity are shown in Table 6. The results also include total electricity per facility. The withdrawal and consumptions are also normalised by vehicle production in cubic metres per vehicle.

Table 6. HAC facilities indirect water withdrawal and consumption due to electricity.

Facility	Total electricity per facility [MWh/year]	Total indirect withdrawal per facility [m^3/year]	Total indirect withdrawal per vehicle [m^3/vehicle]	Total indirect consumption per facility [m^3/year]	Total indirect consumption per vehicle [m^3/vehicle]
USA Car	98,500	5,597,981	55.98	198,915	1.99
USA Truck	197,000	11,195,962	55.98	397,830	1.99
India Car	49,250	3,076,605	61.53	130,078	2.60
India Truck	98,500	6,153,210	61.53	260,156	2.60
Germany Car	197,000	10,942,563	54.71	430,950	2.15
Mexico Car	49,250	1,248,333	24.97	112,220	2.24
China Car	49,250	3,135,674	62.71	181,830	3.64
China Truck	98,500	6,271,349	62.71	363,655	3.64
Japan Car	98,500	3,440,265	34.40	195,037	1.95
Japan Truck	197,000	6,880,531	34.40	390,067	1.95
South Korea Car	492,500	34,092,550	68.19	613,102	1.23
Brazil Car	49,250	348,295	6.97	635,178	12.70
Brazil Truck	98,500	696,590	6.97	1,270,347	12.70
UK Luxury Car	1,970	114,414	57.21	2,485	1.24

As can be seen, the indirect withdrawal from electricity for the HAC facilities is substantially higher than the direct withdrawal by the facility. The per vehicle direct withdrawal numbers are typically from 2 to 6 m³ (with UK Super Luxury being 55 m³ which is not surprising due to the low production volume and high resource use for very expensive super luxury cars like Bentleys.

Indirect consumption due to electricity use is substantially lower than withdrawal. This is due to most types of electricity generation having low evaporative losses. Intuitively, the indirect consumption will be orders of magnitude lower for all of the electricity sources except hydroelectric, which has an average water consumption of 17 m³/MWh due to the high evaporation rates for standing water in lakes (Dooley, Kyle, and Davies 2013; Semmens, Bras, and Guldberg 2014).

The indirect water consumption values calculated for the HAC facilities not located in Brazil are all within the range of 1.23–3.64 m³/vehicle. Facilities in Brazil have substantially higher consumption due to the increased use of hydroelectric power in that country. The consumption by Brazilian facilities is in the range of high 12s m³/vehicle. The average consumption for the HAC facilities overall is 4.22 m³/production, but without the Brazilian facilities it is calculated as 2.26 m³/production. These values broadly agree with the results by (Semmens, Bras, and Guldberg 2014) shown in Table 7. The average of the automakers from Semmens’ calculation was 2.21 m³/vehicle which is below the average of the HAC facilities (not located in Brazil), which tended to be close to the Daimler value of 3.69 m³/year (Semmens, Bras, and Guldberg 2014). Given that the HAC profile is hypothetical and different sources from Semmens were used, the consumption values are reasonable.

Table 7. Results for indirect water consumption by electricity from (Semmens, Bras, and Guldberg 2014).

Automaker	[m^3/vehicle]
BMW	1.91
Chrysler	3.32
Daimler	3.69
Ford	2.41
GM	1.80
Honda	2.27
Hyundai	2.25
Kia	1.24
Mazda	2.12
Nissan	1.54
Volkswagen	1.77
Average	2.21

4.2. Comparison with direct withdrawal and consumption

In order to compare the ratio of indirect water withdrawal from electricity to the direct withdrawal from the facilities themselves, a factor called Indirect Electricity Withdrawal Factor (IEWF) is created to show this ratio. The IEWF is simply the Total Indirect Withdrawal per Vehicle divided by the Direct Withdrawal per Vehicle.

This factor is a measure of how much more water is being withdrawn by the indirect use than the direct use. For Example, the Germany Car facility is the most water efficient of all the car assembly facilities, but it has an IEWF of 21.9. This is due to Germany having an electricity resource profile that is heavy on coal and nuclear (80.9 and 98.59 m³/MWh, respectively). The Brazilian facilities have very low IEWF’s (1.74) because Brazil uses a substantial amount of hydroelectric which has a negligible water withdrawal. However, in the consumption calculations the reverse is true.

The IEWF may be a useful concept because it enables automakers (or other manufactures’) to prioritise which impacts from their facilities are causing constraints on the supply to a location. Although water withdrawn for electricity generation typically has a very high return ratio (the vast majority of the water returns to the source for other users (Dooley, Kyle, and Davies 2013)) it can still cause availability problems (UN 2012).

For the indirect consumption of water due to electricity generation, no factor relating to the direct withdrawal will be made for two reasons: first, the overwhelming reporting of water use by automakers is direct withdrawal. Second, it does not appear to vary much between facilities based on the calculation.

4.3. Discussion

The indirect water either withdrawn or consumed by the electricity generation for automobile production is an impact that is not currently covered in CDP Water Disclosures (CDP 2014) of CSR’s electricity (BMW 2014b; Fiat-Chrysler 2014; Ford 2014a; GM 2014a; Nissan-Renault 2014; Peugeot 2014; VW 2014). Despite this, the indirect use in electricity generation can be orders of magnitude larger than the direct withdrawal by automotive manufacturing facilities.

The use of indirect withdrawal and consumption for electricity generation is also not typically included when calculating life cycle assessments of the impact of vehicles. Only very few studies include indirect water consumption for a life cycle assessment (Semmens, Bras, and Guldberg 2014).

Including the indirect withdrawal due to electricity generation can potentially be used to show the benefits of switching from non-renewable sources to renewable sources, excluding hydroelectric due to the dramatic consumption (17 m³/MWH). Solar and wind power can dramatically reduce the indirect water withdrawal and consumption, but the effect is more pronounced in the indirect withdrawal. In Table 8, the HAC USA Car facility is compared with a facility called Solar Country Car. The only difference between the two facilities is the USA Car facility uses the USA electricity source profile from IEA (2012) and the Solar Country Car facility uses entirely photovoltaic (PV) solar power.

Table 8. USA car compared with solar powered car facility.

Facility	Total indirect withdrawal for electricity [m^3/year]	Total indirect consumption for electricity [m^3/year]	Total indirect withdrawal per vehicle [m^3/vehicle]	Total indirect consumption per vehicle [m^3/vehicle]
USA Car	5,597,981	198,915	55.98	1.99
Solar Country Car	1,980	1980	0.02	0.02
Difference	5,596,001	196,935	55.96	1.97

The difference in indirect water withdrawal and consumption is dramatic, particularly the withdrawal, which is over 2000× times greater for the USA Car facility. Even the indirect consumption by USA Car is over 80× the indirect consumption by the facility that uses exclusive solar PV power. The electricity usage to manufacture cars and the energy profile of the countries in which the facilities are located is an underrated aspect of the water impacts of automotive manufacturing. Due to their low water withdrawal and consumption, car companies should pursue implementation of photovoltaics renewable energy on-site. For example, Bentley Motors constructed a 5 MW rooftop photovoltaic system at its site in Crewe, UK which covers about 40% of the factory’s power demand (Dooley, Kyle, and Davies 2013; VW 2014). Although the primary focus for this installation seems to have been related to greenhouse gas emissions savings, this PV array would also reduce the factory’s indirect water withdrawal and consumption dramatically due to the low impact of electricity from PV.

5. Indirect water withdrawal by workers

Electricity use is one source of indirect water use, but how much water do the workers of each facility use in their personal life, and what is the magnitude compared to direct and indirect water use of an automotive facility? In order to answer these questions and calculate the indirect water usage of workers, we expanded the profiles of the (hypothetical) automotive facilities to give reasonable values for number of workers and their production intensity. Those values are based on worker information from BMW, Hyundai, GM and VW’s public reporting (BMW 2014a; GM 2014a, 2014b; Hyundai 2014; USA, Hyundai 2014; Volkswagen 2014). Those companies provide enough information about their operations publically to estimate the worker and water intensities of production. Per capita withdrawal amounts can be found for each country using the AquaStat database from the Food and Agriculture Organization (FOA) of the United Nations. Given the number of workers and the per capita withdrawal of each, the indirect water consumption of the automotive facilities’ workforce can be calculated, as will be discussed in the following sections.

5.1. Results for indirect water withdrawal by workers

In order to quantify the indirect water use by the work force, we need to know the water use of workers outside the factory. Per capita withdrawal amounts for each country where the HAC facilities are located are given in Table 9 from the FOA AquaStat database. Only water withdrawal will be used for the purpose of comparing the direct water use by the production facilities and the workers’ indirect use, because the FAO Aquastat database only contains data on withdrawal. Despite not having consumption data, withdrawal is still a useful value to examine because withdrawal is what limits the availability of water in a region or location (Joost Schornagela et al. 2012).

Table 9. Average water withdrawal in m³ per capita per year.

Country	[m^3/capita]
Brazil	376.7
China	406.0
Germany	386.5
India	615.4
Japan	713.2
Korea, Republic of	549.0
Mexico	664.5
United Kingdom	212.9
USA	1575.0

Using the withdrawal by country data from Table 9, the total withdrawal by employees for the HAC facilities can be calculated using Equation 3(3) . The ‘# of Employees’ for the HAC facilities used in this paper is based on actual public worker profiles from automotive manufacturing companies and is listed in Table 10.(3)

Table 10. Facility production, workers, withdrawals and employee water factor.

Facility name	Production [vehicles]	Total withdrawal [m³]	Withdrawal per vehicle [m³]	No. of workers	Production/ year/worker	Withdrawal by workers [m³]	Employee water factor
USA Car	100,000	400,000	4.00	2222	45.0	3,500,000	8.75
USA Truck	200,000	800,000	4.00	4444	45.0	7,000,000	8.75
India Car	50,000	200,000	4.00	833	60.0	512,833	2.56
India Truck	100,000	400,000	4.00	1667	60.0	1,025,667	2.56
Germany Car	200,000	500,000	2.50	5714	35.0	2,208,571	4.42
Mexico Car	50,000	200,000	4.00	1111	45.0	738,333	3.69
China Car	50,000	200,000	4.00	833	60.0	338,333	1.69
China Truck	100,000	400,000	4.00	1667	60.0	676,667	1.69
Japan Car	100,000	400,000	4.00	1667	60.0	1,188,667	2.97
Japan Truck	200,000	800,000	4.00	3333	60.0	1,188,667	2.97
South Korea Car	500,000	2,000,000	4.00	8333	60.0	4,575,000	2.29
Brazil Car	50,000	200,000	4.00	3333	15.0	1,255,667	6.28
Brazil Truck	100,000	400,000	4.00	6667	15.0	2,511,333	6.28
UK Super Luxury	2000	110,000	55.00	4000	3.0	851,600	7.74

It is important to note that the absolute value of the withdrawal for each facility is not necessarily a good indicator. For example, the South Korea Car facility has substantially higher production than any other facility (2.5× more than any other facility) and comparing the total withdrawal of the employees of that facility is not appropriate because it employs by far the most employees (over 1000 more than any other). An ‘Employee Water Factor’ (EWF) was created to understand the relationship between the water withdrawn by a facility and the water withdrawn by the employees, and is shown in Equation (4(4) ). This equation takes into account the values for per person water withdrawal and the total employees from Equation (3(3) ) and essentially scales that value in relation to the facility withdrawal, which correlates with the production of the facility.(4)

From Equation (4(4) ), it becomes apparent that the facilities with the highest EWF have the greatest amount of indirect water use by employees compared to their facilities’ direct withdrawal. The results of the EWF calculation and the relevant facility profile information are shown in Table 10.

5.2. Discussion

As can be seen in Table 10, the EWF values range from 1.69 to 8.75. The latter value means that the indirect withdrawal by a worker is 8.75 times higher than the direct withdrawal by the facility for producing a vehicle. These values are in the same order of magnitude as the indirect water withdrawal from electricity use (see Table 11).

Table 11. HAC facilities indirect electricity withdrawal factor results.

Facility	Indirect withdrawal per vehicle [m^3/vehicle]	Direct withdrawal per vehicle [m^3/vehicle]	Indirect electricity withdrawal factor (IEWF)
USA Car	55.98	4.00	13.99
USA Truck	55.98	4.00	13.99
India Car	61.53	4.00	15.38
India Truck	61.53	4.00	15.38
Germany Car	54.71	2.50	21.89
Mexico Car	24.97	4.00	6.24
China Car	62.71	4.00	15.68
China Truck	62.71	4.00	15.68
Japan Car	34.40	4.00	8.60
Japan Truck	34.40	4.00	8.60
South Korea Car	68.19	4.00	17.05
Brazil Car	6.97	4.00	1.74
Brazil Truck	6.97	4.00	1.74
UK Luxury Car	57.21	55.00	1.04

The lower EWF values for facilities were for China, India, Japan and South Korea and were around 2. This means that the workers are only withdrawing about 2× the water the facility uses for the personal use. This corresponds with very low per capita use numbers from those countries and having a high production per worker per year figure. The facilities with the highest values of EWF are USA Car, USA Truck, UK Super Luxury, Brazil Car and Brazil Truck, with values ranging from 4 to 9. These facilities’ workers are withdrawing 4–9× as much water for personal use compared to the amount the facility uses for vehicle production. Because of this, these facilities would benefit most from employee engagement about water use outside of the facility; the facilities may not benefit as much from direct investment in the reduction of water use from the facility. A reduction in the withdrawal in the local area of a facility can inherently improve the water risk and stress situation (UN 2012). This raises an important issue: to what extent are workers are directly responsible for local water withdrawals in their own watershed? To answer this question, we need to look at what withdrawn water is actually used for.

Water withdrawals globally are primarily for agricultural purposes. Globally, agriculture uses 69% of the total water withdrawn (Aquastat 2016). Industrial uses account for 19% of the total and municipal withdrawal is only 12% (Aquastat 2016). The percentages for sector withdrawals vary greatly by region (Table 12) and by country (Figure 1). This is significant because for industrial operations, it is important to know what other activities are using water in the region. For example, in North America and Europe, about half of the water withdrawn is for industrial purposes.

Table 12. Withdrawal by sector by region (Aquastat 2016).

Region	Municipal (%)	Industrial (%)	Agriculture (%)
Africa	15	4	81
North America	13	47	40
Central America	23	18	59
South America	17	12	71
Asia	9	10	81
Europe	21	54	25
Oceania	20	15	65

Figure 1. Water Withdrawal Profiles by Sector for Selected Countries (from FAO Aquastat (FAO 2016))

Figure 1 shows that some countries have radically different water use profiles. For example, Germany uses 83.9% of its’ total water withdrawal for industrial purposes and only 0.3% for agriculture. Conversely, India uses 90.4% of its’ total water withdrawal for agriculture, and only 2.2% for industrial purposes. Combining this information with the EWF, it is possible to calculate EWF for agricultural, municipal and industrial purposes, shown in Table 13. The values for each type of EWF are rounded to one decimal place, and each factor is the percentage of the EWF that goes to the sector described. For example, the EWF for USA Car is 8.8. The EWF Agricultural is 40.2% of the original EWF because of the per capita withdrawal for the USA, 40.2% is for agricultural purposes.

Table 13. EWF values for different sectors of withdrawal for HAC facilities.

Facility name	Employee Water Factor (EWF)	EWF agricultural	EWF municipal	EWF industrial
USA Car	8.75	3.5	1.2	4.0
USA Truck	8.75	3.5	1.2	4.0
India Car	2.56	2.3	2.1	1.9
India Truck	2.56	2.3	0.2	0.1
Germany Car	4.42	0.0	0.7	3.7
Mexico Car	3.69	2.8	0.5	0.3
China Car	1.69	1.1	0.2	0.4
China Truck	1.69	1.1	0.2	0.4
Japan Car	2.97	1.8	0.6	0.5
Japan Truck	2.97	1.8	0.6	0.5
South Korea Car	2.29	1.4	0.6	0.3
Brazil Car	6.28	3.8	1.4	1.1
Brazil Truck	6.28	3.8	1.4	1.1
UK Super Luxury	7.74	0.7	4.5	2.6

Dividing the EWF into the three different withdrawal sectors for the automotive facility locations allows the EWF to have further meaning. For example, in the USA, the municipal withdrawal is not significant compared to the agricultural and industrial withdrawal. For mitigating water stress near the USA Car and Truck facilities, engaging employees about their own use will only be able to help with the municipal withdrawal, which is not as large as the agricultural or industrial EWF. From Table 13, it is apparent that the ideal facility to engage employees about their personal water use is the UK Luxury Car facility. Its EWF municipal value of 4.5 is much greater than any other HAC facility. This means that employee engagement could significantly reduce the indirect water withdrawal by this facility within their own watershed. Similarly, in countries with high withdrawals from agriculture, engagement programmes could focus on educating employees on (a) the impact of agriculture on water withdrawals and (b) food sources that have low water needs for growing and cultivating. It might be possible to reduce water withdrawals merely by offering different food choices in the factory cafeteria. Similar efforts can be made by educating employees on the amount of water withdrawn by the industrial sector, which includes electricity generation. Reducing electricity at home also has a beneficial effect on water withdrawals.

As shown, the EWF as a standalone metric helps automotive companies identify facilities where the local water supply will benefit more by employee engagement for water reduction than reducing the facilities’ usage directly. In facilities with high EWFs, the workers’ indirect withdrawal dwarfs the direct usage of the facility. This finding validates community engagement activities documented in various automakers’ corporate sustainability reports. Automakers engaging with employees about water use in the locations the automakers operate is a growing area of emphasis. This engagement will help the different companies brand value as well as help lower the stress in those localities (CDP 2014).

6. Conclusion

Although many manufacturers are familiar with their direct water use, few have any idea about their indirect impact on water resources. As shown, the amounts of water withdrawn and consumed as a result of electricity generation for automotive production facilities surpass the amount of water directly used and consumed in facilities. Thus, manufacturers should not only focus on reducing direct water use in their facilities, but also electricity use. This has a synergistic benefit from an environmental and financial perspective. It also highlights the need of researches to focus on the water–energy nexus that is becoming more and more important. For example, on-site installations of photovoltaic panels for electricity generation not only reduce greenhouse gas emissions, but also have dramatic effects on indirect water withdrawal and consumption due to PV’s very low need for water in its operation.

It was also shown that workers can have an equally large indirect water withdrawal. Engagement and education programmes focused on reducing water withdrawal by employees in their personal lives can arguably have a greater benefit than any technology focusing on reducing direct water withdrawal of automotive facilities. It is important to know, however, where the largest water withdrawals occur because municipal withdrawals which are in direct control of workers are often lower than agricultural and industrial sector withdrawals. Nevertheless, changes in worker habits may have significant sustainability benefits that reach far beyond the automotive facilities where they work.

These indirect water factors are not typically shown in the Corporate Sustainability Reports of automakers or other industrial companies but their inclusion would be advisable because it emphasises a holistic approach to water stewardship. Furthermore, as shown in this paper, the indirect water impact that can occur from electricity generation and workers in their daily life is significant and merits further study.

It should be noted that the data and results presented here represent an inventory analysis. The next step is an impact analysis of water consumption by locality. In addition, the quality of water (input and output) is important: obviously the consumption of drinkable water in water scarce regions is much more important than the consumption of non-drinkable water. Thus, a logical next step would be to include an impact assessment of the indirect water consumption and withdrawal by locality.

Notes on contributors

Bert Bras, PhD, is a professor at the George W. Woodruff School of Mechanical Engineering at the Georgia Institute of Technology since September 1992. His research focus is on sustainable design and manufacturing, including design for recycling and remanufacture, bio-inspired design, and life-cycle analysis with applications in automotive and energy systems. He has authored and co-authored over a 150 publications. From 2001 to 2004, he served as Director of Georgia Tech’s Institute for Sustainable Development, and he received the 2007 Georgia Tech Outstanding Interdisciplinary Activities Award. In 2014, he was named a Brook Byers Professor of Sustainability. In 2016, he was appointed as the Associate Chair for Administration in the G.W. Woodruff School of Mechanical Engineering.

Andrew Carlile was a Graduate Research Assistant in the George W. Woodruff School of Mechanical Engineering at Georgia Tech where he worked on water risk analyses and strategies for automotive manufacturing and received his Master of Science degree in 2015. He is currently employed by General Electric Analytical Instruments, a division of GE Water & Process Technologies, making water quality testing equipment. He received a Bachelor of Science in Aerospace from the University of Oklahoma in 2011.

Disclosure statement

No potential conflict of interest was reported by the authors.

Acknowledgements

The material presented in this manuscript is based on research done within the Sustainable Design and Manufacturing group at the Georgia Institute of Technology. The authors would like to thank all who have contributed invaluable input and support, including but not limited to Thomas Niemann, Sherry Mueller, Sue Rokosz, Heidi McKenzie, Hyung Chul Kim and Tim Wallington from the Ford Motor Company. Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of research sponsors and/or the authors’ parent institutions.