Many countries, like South Africa, Australia, India, China and the United States, are highly dependent on coal fired power stations for energy generation. These power stations require significant amounts of water, particularly when fitted with technology to reduce pollution and climate change impacts. As water resources come under stress it is important that spatial variability in water availability is taken into consideration for future energy planning particularly with regards to motivating for a switch from coal fired power stations to renewable technologies. This is particularly true in developing countries where there is a need for increased power production and associated increasing water demands for energy. Typically future energy supply options are modelled using a least cost optimization model such as TIMES that considers water supply as an input cost, but is generally constant for all technologies. Different energy technologies are located in different regions of the country with different levels of water availability and associated infrastructure development and supply costs. In this study we develop marginal cost curves for future water supply options in different regions of a country where different energy technologies are planned for development. These water supply cost curves are then used in an expanded version of the South Africa TIMES model called SATIM-W that explicitly models the water-energy nexus by taking into account the regional nature of water supply availability associated with different energy supply technologies. The results show a significant difference in the optimal future energy mix and in particular an increase in renewables and a demand for dry-cooling technologies that would not have been the case if the regional variability of water availability had not been taken into account. Choices in energy policy, such as the introduction of a carbon tax, will also significantly impact on future water resources, placing additional water demands in some regions and making water available for other users in other regions with a declining future energy demand. This study presents a methodology for modelling the water-energy nexus that could be used to inform the sustainable development planning process in the water and energy sectors for both developed and developing countries.
Both energy and water are critical aspects of any economy, and yet despite their strong interdependence the two sectors are often managed independently (Hussey and Pittock, 2012). In resource use terms water is used to generate energy and energy is used to deliver, treat and supply water (Scott, 2011) (see Fig. 1) these links are referred as the water-energy nexus.
Developing an integrated approach to modelling the water-energy nexus is critical to supporting the development of effective national policies and regulations to ensure continued economic development and growth in a sustainable way (Bazillian et al., 2011).
Many countries, like South Africa, Australia, India, China and the United States, are highly dependent on coal fired power stations for energy generation. These power stations, however require significant amounts of water, particularly for cooling and when fitted with technologies to reduce pollution and climate change impacts.
Defining the water energy nexus (after WEC, 2010).
Water availability is becoming more constrained in almost all countries through the combined effects of increasing demands, reducing water quality, and land use change. This presents a significant threat to future energy production (WEC, 2010). Similarly ever increasing water demands require consideration for more energy intensive technologies such as inter basin transfers (IBT), desalination and re-use of wastewater (Pittock, 2011).
As water resources come under ever increasing stress it is important that spatial variability in water availability is taken into consideration for future energy planning. Similarly, changes in energy policy are also likely to impact on water availability and also need to be taken into account for water resources planning and decision making.
Regional variability in potential power generation activities in each water supply region and associated WMAs.
This study seeks to make the link between energy and water resources policy and decision making through developing a linked water-energy nexus model. This study focuses on the water supply for energy generation part of the nexus. Specifically the objectives of this study are to incorporate regional variability in water availability and supply costs into an existing energy model so as to account for (1) a more representative cost of water supply to different technologies, (2) the spatial mismatch between water supply and the location of power plants, (3) the full cost of water supply to the energy sector including water supplied to mines, and (4) the opportunity costs of water use for energy production in a country with limited water resources and increasing demands.
South Africa was identified as a case study given its well documented water scarcity (DWAF, 2004), the importance of water for energy production, the extensive knowledge and strong analytical capacity for addressing the water-energy issue in the country, and the fact that the country is starting to plan water and energy in an integrated manner.
This South Africa case study is part of the World Bank's Thirsty Energy program and provides valuable knowledge and a framework for a water-energy nexus model that can be applied in other countries facing similar challenges relating to the water energy nexus (Rodríguez et al., 2013).
Electricity supply in South Africa is dominated by a fleet of coal fired power stations operated by the state owned entity, Eskom. Eskom also functions as the system operator and owns and operates the transmission and distribution networks outside those owned and managed by the large metropoles. Eskom operates 22 power stations with a total nominal capacity of 41.9 GW, of which 85 % of the capacity is coal-fired. The balance of capacity is provided by nuclear, open-cycle gas turbine, hydro and pumped-storage power plants (ESKOM, 2013). In an attempt to address energy diversification, environmental concerns, and economic growth aspirations, energy sources such as nuclear, gas and renewables are being examined as alternatives by the Department of Energy (DoE). Of particular concern is the large volumes of water used for cooling of existing coal fired power stations and the benefits that renewable and alternative energy supply options might also provide in terms of improved water use efficiency in addition to reduced carbon impacts.
South Africa is a water-scarce country (annual freshwater availability is
less than 1700 m
The long-term infrastructure planning process for the supply of both energy and water is well established through government departments. The planning of both resources has taken into account cost and scarcity of the other to various degrees, but to date integrated modelling of the bulk supply infrastructure of both systems has not been undertaken. Investigating the significance of these linkages and how they affect future water and energy planning requires the integration of water constraints into energy models and energy constraints into water supply models.
The South African TIMES model (SATIM), a public domain energy systems model
developed by the University of Cape Town's Energy Research Centre (ERC), is a
suitable base model for integrated water-energy planning More
detailed documentation of SATIM can be found at:
Typically future energy supply options are modelled using an optimization model such as SATIM that considers water supply as an input costs, but is generally constant for all technologies. It is however important to note that different technologies are located in different regions of the country with varying levels of water availability and associated infrastructure development and supply costs. In order to address this limitation regional variability was introduced in to SATIM-W and individual water supply options, include major investments in dams and transfer projects and water supply energy needs, were incorporated into the SATIM-W model so as to capture the water-energy interplay. Incorporating a regional cost and quality for water allows the model to examine potential trade-offs within the supply sector e.g. fuel extraction and processing, treatment of water, cleaning and flue gas desulphurization. Regional variability in the non-energy water demands were also included in the model as this affects the relative opportunity cost for water supply to energy in each area. The updated model, SATIM-W, allows these activities to be represented so that the model is responsive to the regional cost and availability of water and energy supply, connected to a single national demand-side representation.
South Africa is divided into 19 Water Management Areas (WMAs). The location of these WMAs, relative to the different energy producing regions of South Africa are shown in Fig. 2. These regions include: A: Waterberg (Lephalale); B: Mpumalanga, Witbank; C: Mpumalanga, Secunda; D1: Northern Cape, Upington; D2: Northern Cape, Karoo (Fig. 2). The nature of the specific energy activity in each of these regions is given in Table 1. In each region the Department of Water and Sanitation (DWS), formally known as the Department of Water Affairs (DWA) has assessed all future water supply options and determined the available yield, estimated capital and operational costs, and average energy requirements (e.g. for the pumping of water or desalination or water treatment) (DWA, 2010a).
South African Water Management Areas and Energy Producing Regions (A: Waterberg (Lephalale); B: Mpumalanga, Witbank; C: Mpumalanga, Secunda; D1: Northern Cape, Upington; D2: Northern Cape, Karoo. The WMAs are numbers as per the original definition given in the National Water Resources Strategy (DWAF, 2004).
The regional distribution of water sources and consumers varies greatly in
South Africa and as a result, the demand for water and supply is highly
regionalized. For example, in the Waterberg district municipality (Lephalale)
in Limpopo province, where the Waterberg coal deposits occur, the demand for
water is dominated by water needs of the dry-cooled Matimba coal-fired power
station for make-up water and other plant needs
(4.3 million m
A representative cost for water supply to the different energy producing
regions was determined according to the
The marginal cost for water supply also increases with increasing demands and hence it was important to consider the impact of other non-energy water demands in each region and to develop a regional marginal cost curve for water supply as a function of the total demand in the region. This was also important in that the cost of water could also vary depending on when the proposed energy developments take place as future water supply options are generally more expensive than the current supply costs.
The resulting regional marginal cost curves for water supply are then included in the model associated with each region and SATIM-W then weights each water supply and delivery option (or scheme) and chooses the combination that delivers the needed water at least-cost, resulting in the determination of the total marginal water supply cost.
The costs for delivery of water to power plants is based upon estimates for deploying and managing major water supply and transfer schemes, but does not capture final details (and associated costs) that can only be determined when a specific site is identified. This is also true for hydraulic fracturing and Concentrated Solar Power where the exact locations and method of water delivery have not been determined. But in both cases these are rather small compared to the other costs characterizing each scheme.
In order to assess the benefits of integrating the regional variability of
water supply costs into future energy supply planning, two model scenarios
were considered. The first scenario considered water supply as a uniform cost
applied to all future energy supply options. The second scenario included the
regional marginal costs of water supply as well as the other non-energy water
demands relative to the locations in which the different future energy supply
options where located. The results of these scenarios are given in Figs. 3
and 4 respectively. In order to model the potential impacts of climate change
policy on future energy and water supply, a number of other scenarios were
considered, including a limit on the total production of CO
Optimal future energy mix under the reference scenario with no regional variability in water supply costs.
Optimal future energy mix under the reference scenario but with regional variability in water supply costs.
Resulting power sector water consumption and efficiency (l kWh
In the absence of taking into account the regional variability of water
supply and associated infrastructure costs (the No Water Cost scenario –
Fig. 3), wet-cooled coal power plants are the preferred choice due to their
lower investment costs and higher net generation efficiencies. However, when
consideration is given to the regional variability in water supply (the Water
Cost scenario in Fig. 4), dry-cooling is the preferred option for new coal
power plants, particularly in the Waterberg region where the remaining
economically viable coal reserves are located. New dry-cooled capacity of
approximately 40 GW is commissioned by 2050 and includes the replacement of
the existing stock of 37 GW which will mostly be retired by then. This
SATIM-W result indicates that Eskom's dry cooling policy is really in the
economic interests of the country, even though it increases the cost of
electricity from coal power plants. This has a significant impact in terms of
future water use efficiency in the power sector, which could either reach a
peak of 1.65 L kWh
Other than water consumed by power plants, the two scenarios have similar
total system cost, energy supply expenditures, and primary and final energy
consumption. Interestingly, the Water Cost scenario produces slightly more
CO
The impact of climate change policy, such as imposing a carbon limit on the energy sector, results in a significant shift in the optimal future energy mix. In particular this results in increase demand for renewables, but particularly concentrated solar power (CSP). The result of this is surprisingly an increase in the overall water use efficiency for energy production as shown in Fig. 5 as compared to the reference case Water Cost scenario which is dominated by dry-cooled coal fired power stations. The imposing of a carbon cap also has a significant impact on future water resources planning as it looks to develop more CSP plants, which are located in along the Orange River at the expensive of existing coal fired power stations in the Olifants and Upper Vaal catchments or new coal in the Lephalele area. This could result in a shift in water demand leading not only to stranded assets in terms of decommission power stations, but also the existing bulk water distribution systems that have been developed over the year to supply these power stations with water.
These findings illustrate the insights gained from integrated and regionally
disaggregated water-energy modelling and in particular the importance of
taking in to consideration the regional variability in water supply costs
and associated locations for different energy technologies. The main message
from this study is that water and energy planning should be integrated. The
analysis also shows the ability to identify major water infrastructure
investments that could become stranded (or sub-optimal from an economic
perspective) in light of possible future energy policy changes and can
thereby help to formulate hedging strategies aimed at minimizing the
likelihood of these stranded assets. Alternatively however, the opportunity
to utilise these “stranded” water infrastructure investments could
represent a significant economic opportunity as these could be repurposed to
provide water needed to support other activities in the region at a
relatively low costs. Hence it is important not only that water and energy
policy be considered together in the context of the water-energy nexus, but
that this should also be integrated with regional economic development and
planning particularly in the context of future climate change and other
uncertainties. This study also does not include additional externalities
such as impacts on water quality or the total impact on CO
UCT, the custodians of the model and model results
can be accessed:
The authors declare that they have no conflict of interest.
This article is part of the special issue “Water security and the food–water–energy nexus: drivers, responses and feedbacks at local to global scales”. It is a result of the IAHS Scientific Assembly 2017, Port Elizabeth, South Africa, 10–14 July 2017.
This work was made possible by the financial contribution of the Water Partnership Program of the Water Global Practice, World Bank Group which supports the Thirsty Energy Initiative The authors would also like acknowledge the support of the members of the Thirsty Energy program steering committee as well as the anonymous reviewers for providing additional guidance.Edited by: Barry Croke Reviewed by: Fortune Faith Gomo and one anonymous referee