City of Tshwane Case Study

Economic benefit of ensuring TECHNICAL PAPER Journal of the South African uninterrupted water Institution of Civil Engineering ISSN 1021-2019 supply during prolonged Vol 61 No 4, December 2019, Pages 19–28, Paper 0469 electricity disruptions – CHRISTO POTGIETER (PrEng, MSAICE) works at Bigen in the company’s Water and Sanitation Sector. He graduated from the University of City of Tshwane case study Pretoria with a degree in Civil Engineer in 2011, obtained his BEng (Hons) degree in 2016 and completed his MEng degree in 2018 J C Potgieter, C Herold, M van Dijk, J N Bhagwan in Water Resources Engineering. Contact details: Bigen Group, PO Box 8994, Bloemfontein 9300, South Africa Mitigating the impact of electricity disruptions on water supply was investigated as a case T: +27 83 657 7504, E: [email protected] study on the City of Tshwane. This study found that current institutional arrangements Department of Civil Engineering, University of Pretoria between electricity suppliers (Eskom), water service providers and water service authorities Private Bag X20, Hatfield 0028, Pretoria, South Africa are insufficient. Therefore, a Risk Analysis and Mitigation Framework of Integrated Water DR CHRIS HEROLD (PrEng, FSAICE) specialises in and Electricity Systems, or RAMFIWES, was developed. Risks associated with water supply water resources and water quality model interruptions due to electricity disruption events were analysed. Risk categories that were development and assessment. His models have found wide practical application in catchment and addressed are short-term disruptions of less than one day, medium-term disruptions of up to a systems analysis studies that have saved South week and long-term electricity disruptions up to a month or even longer. The direct economic Africa billions of rands. He is the author of 44 benefits of ensuring uninterrupted water supply in the event of electricity disruption events technical papers published locally and overseas and is the principal author of over 130 unpublished technical reports. Chris has actively participated in over were analysed through cost versus benefit analyses. It was found that the direct benefit/cost a dozen Water Research Commission studies, most of which he has led. He ratio of supplying water during electricity disruption events is 5.6 for wet-industries and an obtained his BSc, MSc and PhD degrees at the University of the exceptional 117 for other economic sectors in Tshwane. The overall benefit/cost ratio is 15.5. Witwatersrand. He formed Umfula Wempilo Consulting in 2000. This benefit is possible at a low increase in the normal municipal bills of only 0.5%. Contact details: Umfula Wempilo Consulting PO Box 98578, Sloane Park 2152, Johannesburg, South Africa T: +27 82 459 5731, E: [email protected] INTRODUCTION Until recently in South Africa, the elec- MARCO VAN DIJK (PrEng, MSAICE) is a lecturer in The water supply sector is at the core of eco- tricity supply used in the water sector was the Department of Civil Engineering at the nomic growth and social well-being. Water considered safe, and the risk of electricity University of Pretoria (UP), and a Principal is indispensable to human survival – it is a supply failure did not play a significant role Researcher for the South African Water Research critical component required for the genera- in the design and operation of water supply Commission. He obtained a degree in Civil Engineering from UP in 1996, a BEng (Hons) in tion of power, it is required for the growing and distribution systems. Load-shedding 1998, an MEng in Water Resource Engineering in 2003, and is currently of crops and it is a basic natural resource for prompted the Water Research Commission completing his PhD. He has compiled numerous technical reports and daily existence. In the modern-day South (WRC) of South Africa in 2010 to conduct a journal publications, and has presented at various conferences in the field of pipelines, hydropower generation, and water distributions systems. He is African urban environment, without water high-level study of the effect of electricity dis- also a member of the Water Institute of Southern Africa. there will be insufficient electricity, very ruptions, specifically load-shedding, on water Contact details: little industry, no agriculture and no cities. supply (Winter 2011). The current study Department of Civil Engineering, University of Pretoria Electricity disruptions can cause water sup- (this paper), also funded and initiated by the Private Bag X20, Hatfield 0028, Pretoria, South Africa ply interruptions which will have dire social WRC, explores the implications in greater T: +27 12 420 3176, E: [email protected] and economic consequences for densely detail and takes account of new concerns that JAYANT BHAGWAN obtained a Master’s in Tropical populated urban areas (ADB 2009). have arisen since 2011. These new concerns Public Health Engineering from Leeds University, Water and electricity are intrinsi- relate to the high risk of prolonged electricity supported by degrees in Municipal and Civil Engineering. He is the Executive Manager of cally linked – the one cannot be supplied disruption events that the water supply sector Water Use and Waste Management at the South without the other. This is especially true is currently not sufficiently prepared to miti- African Water Research Commission, overseeing in South Africa where 85% of the country’s gate. In recent years there have been numer- around 150 research projects covering the management of drinking water electricity is supplied by coal power sta- ous water supply interruptions in Gauteng and wastewater in the domestic, mining and industrial sectors, with particular focus on water supply and sanitation for low-income areas. He tions (Pollet et al 2016), and urban water due to localised electricity disruptions not participated in the shaping of South Africa’s national water policy and supply, especially in-land, necessitates high- linked to scheduled load-shedding, the most legislation, and also drives the area of faecal sludge management (FSM), lift pumping systems over great distances. recent being the explosion at a City Power having pioneered an international series of FSM conferences. This link is often referred to as the Energy- substation that resulted in an electricity Contact details: Water Research Commission Water Nexus (Copeland & Carter 2017). disruption affecting Rand Water’s Eikenhof Private Bag X03, Gezina, Pretoria 0031, South Africa Electricity is used in the water sector for Pump Station (News24 2018). Subsequently, T: +27 12 761 9000, E: [email protected] pumping, treatment of raw water, distribu- large parts of Gauteng, and even parts of tion of potable water, collection and treat- North West, were affected by water supply Keywords: electricity disruptions, water supply, risk analysis, ment of wastewater, and water discharge. interruptions. This electricity disruption was load shedding

Potgieter JC, Herold C, Van Dijk M, Bhagwan JN. Economic benefit of ensuring uninterrupted water supply during prolonged electricity disruptions – City of Tshwane case study. J. S. Afr. Inst. Civ. Eng. 2019:61(4), Art. #0469, 10 pages. http://dx.doi.org/10.17159/2309-8775/2019/v61n4a2 19 Risk Analysis and Mitigation Framework of Integrated Water and Electricity Systems (RAMFIWES)

Test RAMFIWES on Electricity Water service Water service Tshwane case study suppliers providers authorities

• Define scope • Identify hazards Risk analysis Risk analysis Risk analysis • Estimate risks Document risk analysis • Risk acceptance criteria • Compare risks with criteria Risk evaluation Risk evaluation Risk evaluation Risk assessment • Identify risk reduction options

• Gather relevant data • Cost vs benefit analysis Document risk Risk mitigation Risk mitigation Risk mitigation Risk • Make decisions mitigation

mitigation • Implement decisions

• Consider the outcome of the Continuous Continuous Continuous risk analysis and mitigation Report development development development • DRMP focuses on water and Review of Disaster Risk of Disaster Risk of Disaster Risk electricity risks Communicate Management Management Management • Should be incorporated into Improve programme Disaster risk Disaster risk

management Programme Programme Programme existing DRMP if this exists

Figure 1 RAMFIWES outline and testing it on the City of Tshwane case study dealt with relatively quickly due to the nature QQ Review current institutional arrange- a duration of less than one day). There is of the electricity disruption event. If the event ments in place to mitigate the risk. a lack of preparation for medium- to long- had been more severe, there would have been QQ Propose institutional guidelines (pertain- term duration electricity disruption events. longer-duration water supply interruptions. ing to system management, operation Therefore, medium- to long-term duration To put the risk of water supply interrup- and design principles) for electricity sup- electricity disruption events pose a high tions due to electricity disruption events into pliers, Water Service Providers (WSPs) risk for Tshwane, from both economic and context: Rand Water’s current Number 1 and Water Service Authorities (WASs). public well-being points of view. strategic risk on its list of Top 10 Strategic This study found that the current institution- No guidelines, frameworks or government Risks, published in its 2017 Annual Report, al arrangements of South Africa’s electricity policies to aid electricity suppliers, WSPs and is the availability, reliability, reliance and suppliers, WSPs and WSAs are insufficient WSAs in the development of Disaster Risk quality of electricity supply, critical spares to mitigate the effect of long-term electric- Management Plans to mitigate the impact of and chemicals (Rand Water 2018). The ity disruptions on water supply. Inadequate electricity disruptions on water supply were electricity requirement of the Gauteng water preparation to deal with risks will have dire found (the study concluded that there are supply system is particularly high due to consequences, including prolonged water none). Subsequently the Risk Analysis and the elevation difference from the Vaal Dam supply interruptions if an electricity disrup- Mitigation Framework of Integrated Water and the Witwatersrand escarpment (the tion event occurs (especially medium- to and Electricity Systems (RAMFIWES) was maximum pump head from the Vereeniging long-term disruption events with durations developed. The Tshwane case study was also Wastewater Treatment Works (WTW) over of longer than a day). This is better illustrated used to test the part of RAMFIWES that the escarpment is roughly 320 m). Although through the following equation (WHO 2007): deals with WSPs and WSAs (refer to Figure 1 this is not necessarily the case for other for the outline of RAMFIWES). urban water supply systems in South Africa, Hazard × Vulnerability The scope of the case study was Risk = all supply systems require electricity for Level of Preparedness confined to the City of Tshwane, which treatment, supply and booster pumps. houses 3.1 million residents (CoT 2016). Risks related to electricity disruption As part of the Tshwane case study it was The city generates an annual Gross events were investigated using the City of found that the current level of prepared- Domestic Product (GDP) of R202 billion Tshwane as a case study. The main objec- ness to mitigate the impact of electricity (Tshwane Economic Development Agency tives of this study were to: disruptions on water supply is more suited 2015) and has an average water demand of QQ Evaluate the risk of electricity disruption to accommodate short-term electricity 843 Mℓ/‌day (GLS Consulting 2017) which and its impact on water supply failure. disruption events (e.g. load-shedding with includes 193 Mℓ/d losses (CoT 2015a). It

20 Volume 61 Number 4 December 2019 Journal of the South African Institution of Civil Engineering Tshwane’s water and electricity depart- Legend ments and from various readily available Rand Water Magalies Water data sources (such as Tshwane’s Integrated Tshwane (bulk) Development Plan and its annual financial Tshwane (network) reports) (CoT 2016; CoT 2015c). Tshwane (WTW) Magalies Water Various electricity supply disruption (Klipdrift WTW) events that would result in water supply interruptions in Tshwane were identified and grouped according to their areal extent, Temba WTW Magalies Water duration and probability of occurrence. (Wallmannsthal WTW) These risks were based on literature sources Roodeplaat WTW Magalies Water and discussions with representatives from (Cullinan WTW) Eskom, Tshwane and Rand Water. The impact of these events on water supply were assessed by considering three different Tshwane supply areas ranging in Rand Water Rand Water size from: (Mapleton Pump (Palmiet Pump QQ a small residential area Station) Station) QQ one of Tshwane’s six regions compris- Rietvlei WTW Bronkhorstspruit WTW ing mixed residential, commercial and industrial water uses Figure 2 Tshwane water supply sources and bulk distribution QQ the whole of Tshwane. One-day, seven-day and thirty-day dura- goes without saying that addressing high springs and wells (GLS Consulting 2017). tions were examined for each selected real-water losses in the distribution system Figure 2 illustrates the water network, size of supply area, giving a total of nine to decrease water demand is an essential including the WSPs’ water infrastructure. scenarios. part of ensuring uninterrupted water sup- Mitigation options to sustain a mini- ply during electricity disruptions. mum domestic water supply and protect Some 81% to 86% of Tshwane’s water METHODOLOGY most of the economic activity were supply is derived from Rand Water and Background information on Tshwane’s identified and costed. Effective mitigation Magalies Water (CoT 2015b); the rest being demographics, economic activity, and water requires infrastructure development and supplied from own sources at Rietvlei Dam, and electricity infrastructure was obtained operation, as well as effective implementa- Roodeplaat Dam and various dolomitic through consultation with officials from tion of institutional guidelines.

Table 1 Summary of electricity disruption scenarios

Scenario 1: Short-term (1 day) Scenario 2: Short-term (1 day) Scenario 3: Short-term (1 day) Scenario 4: Medium-term (7 days) Scenario 5: Medium-term (7 days) Scenario 6: Medium-term (7 days) Scenario 7: Long-term (1 month) Scenario 8: Long-term (1 month) Scenario 9: Long-term (1 month) Description: Description: Description: Constantia Park Tower’s water supply interrupted Tshwane bulk water supply to region 6 The entire City of Tshwane. Population affected: Population affected: Population affected: 140 households (± 490 people) 173 000 households (± 600 000 people) 911 550 households (± 3.1 million people) Area Annual Average Daily Demand (AADD): Area AADD: Area AADD: Residential – 355 kl/day Residential – 120 Mℓ/day Residential – 632 Mℓ/day Industrial – 10 Mℓ/day Industrial – 55 Mℓ/day Other – 30 Mℓ/day Other – 157 Mℓ/day

Journal of the South African Institution of Civil Engineering Volume 61 Number 4 December 2019 21 Table 2 Steps followed for the electricity disruption scenarios analyses

Step 1: Scenario description Describe the scenario in terms of the electricity disruption event duration and area of effect.

Step 2: System description Describe the electricity and water infrastructure and the number of end-users affected by the electricity disruption event.

Step 3: Risk analysis Discuss the risks that result from the electricity disruption event.

Step 4: Risk mitigation options Identify and describe risk mitigation options for the scenario.

Estimate the costs of risk mitigation options and compare the cost of the most economical option with the estimated Step 5: Cost estimate benefit of mitigation.

Summarise the scenario, identify shortcomings of the risk mitigation approach, and develop alternative solutions where Step 6: Scenario conclusion necessary.

The benefit of mitigation was defined by supply for the nine scenarios is described more options). The latter (standby genera- the reduction in the loss of GDP generation in Table 2. tors) proved to be the cheapest and most achieved. A cost versus benefit analysis was effective. This despite the fact that the done to determine the economic viability already high water losses, increased by the of mitigating the risks. MITIGATION OPTIONS elevated water pressure during periods of Loss of electricity supply has an immediate restriction, have to be fulfilled before water impact on the functioning of wastewater can be delivered to end users. Another ELECTRICITY DISRUPTION treatment works and raw sewage pump- important factor is that it is virtually SCENARIOS ANALYSED ing systems, with consequent overflow of impossible to provide a minimum water Different types of electricity disruptions untreated effluent. There is also an impact supply (25 ℓ/capita per day) to all domestic and their impact on water supply were on the ability to deliver water to elevated users without affording upstream users the analysed in order to determine if the tanks supplying water to high-lying areas. opportunity to withdraw far more than options identified to mitigate the risks Longer-term electricity supply disruption their quota. Even in the highly unlikely would be practically feasible and economi- results in reduction or total loss of water event of 100% cooperation, it is impossible cally viable. supply even to lower-lying areas fed by to implement effective water restrictions The three electricity disruption areas gravity from reservoirs. within the time spans of typical electricity and electricity disruption durations were Mitigation options range from water disruption events. analysed. A summary of the electricity supply by road tankers, provision of The costs associated with mitigating the disruption scenarios is given in Table 1 on additional storage, provision of additional effect of each of the electricity disruption page 21. interconnectivity of pipe systems and the scenarios were scaled up (for scenarios The approach used to analyse the use of standby generators to make use of that only considered smaller sections impact of electricity disruption on water existing infrastructure (see Table 3 for of Tshwane) and applied to the entire

Table 3 Implementation options to mitigate the effects of a water and power outage (adapted from Mank 2015)

Water Electricity Communication

Water sources: Emergency electricity provision: Internal communication: QQ pool water QQ backup generators plus fuel QQ radio-relay systems QQ rainwater QQ shared backup generators QQ field wire QQ domestic water wells QQ rent backup generators QQ satellite communication systems with batteries, QQ artificial static water supply sources QQ energy self-sufficient systems – photovoltaic rechargeable batteries or solar panels QQ open sources panels, wind parks, sludge fermentation QQ short-wave radio gadgets with car batteries QQ hand-turned pumps QQ high-level water storage tanks External communication: QQ mobile water tanks QQ alarm systems QQ water from neighbouring cities QQ radio broadcasting QQ small water channels QQ flyers and brochures QQ bottled water QQ personal communication QQ interconnectivity of pipe systems QQ loudspeaker announcements

Water treatment: External communication places: QQ water filtration QQ city halls QQ silver chloride pills QQ fire brigade houses QQ UV light irradiation QQ municipality houses QQ boiling Risk communication channels: QQ distillation QQ news QQ chlorination QQ specific events – change of the millennium Hygiene: QQ seminars, workshops QQ plastic bags QQ environmental and political actions QQ manholes in public places connected to the QQ public incentives sewage system QQ platforms QQ books

22 Volume 61 Number 4 December 2019 Journal of the South African Institution of Civil Engineering Table 4 Recurrence intervals for each scenario

Duration RI Scenario Event Cause Extent (day) (year)

1 Local substation Maintenance, old equipment, cable theft Small 1 5

4 Local substation As above Small 1 20

7 Local substation As above Small 30 50

2 One main sub-station As above Medium 1 20

5 One main sub-station As above Medium 7 50

8 One main sub-station As above Medium 30 100

Regional blackout, islanding successful, 3 Operating error Large 1 38 no serious damage

Blackout, no islanding, failed cold start, 6a As above, operator strike, lower intensity solar flare Large 7 44 limited damage

Blackout, no islanding, infrastructure High intensity solar flare with inadequate warning, 6b Large 7* 100 damage computer attack, etc

6 As above As above Large 7 30**

High intensity solar flare with inadequate warning, Blackout, no islanding, Black-Start facilities 9 sabotage, attack on control centre and/or operating staff, Large 30 155 damaged high altitude EMP device, war, natural disaster

* Option 6b is estimated to have a duration of 10–15 days, but has been simplified to seven days to conform to the assumptions of Scenario 6 Notes: ** The combined RI of Scenarios 6a and 6b is calculated as (1/44 + 1/100)–1

Tshwane to obtain the total cost of mitigat- facilitates direct comparison with normal should reflect the net rate (i.e. after ing each of the risks identified. water accounts to place decision-makers in inflation) that Tshwane can expect to Finally, the scaled up costs were com- a position to assess the implications. earn on a similar investment if it did not pared to the economic benefit of ensuring The following were considered for the invest in the infrastructure required to uninterrupted water supply to Tshwane. cost-benefit analysis: ensure uninterrupted water supply dur- QQ The city’s water demand ing electricity disruption events. QQ The city’s water supply QQ The economic outlook period has been COST / BENEFIT ANALYSIS QQ The city’s wastewater treatment taken as the recurrence interval of each The cost of meeting minimum water QQ The type, probability and duration of power outage event. Where necessary, supply requirements during electricity electricity disruption events the remaining value of capital works at disruption events was compared to the QQ The direct and indirect cost of the elec- the end of the outlook period has been economic and other benefits. The cost of tricity disruption event. credited before being discounted to a mitigating each electricity disruption event Mitigation options were aimed at meeting present value. A straight-line deprecia- type (load-shedding, distribution failure the minimum requirements for the follow- tion of capital over the life of the works and blackout) was compared to the benefit, ing water uses: has been used for this purpose. taking account of the probability of occur- QQ Basic minimum supply for domestic In instances where there are no quantifi- rence of the event. water use able direct economic benefits, such as The costs that will result from an elec- QQ Prevention of spillage of untreated meeting the minimum water supply tricity disruption event that causes water sewage requirement for domestic users, or pre- supply interruptions can be either direct or QQ Sustaining GDP-generating activities. vention of sewage overflow, the outlook indirect. The following assumptions were made: period has been set equal to the life of the Direct costs are due to interruption QQ A 30-year life has been assumed for the capital works. of the economic activity of the City of standby generators. Although the life Costs have been expressed as cents Tshwane. The benefit of mitigation is then of mechanical/electrical plant is nor- per kℓ of normal billed consumption for the reduction in these costs. In such cases, mally taken as 15 years (DPLG 2009), comparative purposes. In the cases of sce- it is possible to calculate the net benefit, a longer life is feasible, since the plant narios covering power outages in small and benefit/cost ratio and the present value will seldom be used, and then only for a medium-sized areas, the costs were scaled net benefit. relatively short duration. up to cover the whole of Pretoria. For scenarios that do not have a direct QQ For similar reasons, the annual main- Table 4 shows the estimated recurrence benefit with which to compare costs, the tenance cost for generators has been intervals for each scenario. economic analyses have been confined to reduced to 1% of capital. Interestingly, some of the severest calculating the levelised annual cost and QQ An annual net discount rate of 3% has blackout events are the easiest to assign comparing this with the volume of water been assumed. This value is consid- RIs to. For example, the massive solar normally supplied to all Tshwane consum- ered appropriate for a large entity like flare (the Carrington event) that occurred ers to calculate a unit cost per kℓ. This Tshwane, since the net discount rate in 1859 (Cliver & Dietrich 2013) is well

Journal of the South African Institution of Civil Engineering Volume 61 Number 4 December 2019 23 Table 5 Cost requirements for minimum domestic supply

Scenario 3% NDR* RI Capital Maintenance Operation # Ann. Cost ∆ billing+ No Description (year) (mill R) (mill R) (mill R) (mill R) (c/kℓ)

1 Small, 1 day 1 0.000 0.000 0.000 0.000 0.000

2 Medium, 1 day 10 1.856 0.019 0.023 0.116 0.070

3 Large, 1 day 19 1.856 0.019 0.023 0.114 0.069

4 Small, 7 day 10 1.856 0.019 0.158 0.129 0.078

5 Medium, 7 day 30 1.856 0.019 0.158 0.119 0.072

6 Large, 7 day 25 108.329 1.083 9.205 6.978 4.236

7 Small, 30 day 50 1.856 0.019 0.676 0.127 0.077

8 Medium, 30 day 100 1.856 0.019 0.676 0.120 0.073

9 Large, 30 day 155 108.329 1.083 39.452 6.877 4.175

Combined 108.329 1.083 50.370 7.290 4.425

* Net discount rate. + Required increase in normal billing to customers based on average supply to paying domestic customers of 451 Mℓ/day (after deducting NRW Notes: water). # The capital included in the cost estimate is to mitigate the risks associated with the specific electricity disruption event (i.e. R1.856 million to ensure uninterrupted water supply for Scenario 2 includes backup electricity generation to high lying areas in the scenario area). documented, and two lesser events that are water to be supplied to domestic con- probability of occurrence in any one still considered to be of sufficient magni- sumers during electricity disruptions is year and nearly a one in six chance of tude to hold grave consequences for mod- 451 M‌ℓ/‌day at 50 ℓ per capita per day (this occurrence within the term of office of ern electricity supply systems also occurred includes real-water losses which will have a politician. The small economic cost of in the late 19th and early 20th centuries, to be supplied additionally). protecting society against such a calam- when electricity generation and supply Pumping water into elevated towers and ity pales into insignificance against such systems were still in their infancy. The reservoirs is considered to be part of the an enormous risk, which has an almost possibility of the detonation of an Electro- cost of meeting the basic minimum domes- incalculable associated cost and high Magnetic Pulse (EMP) device above the tic water supply. probability of occurrence. ionosphere is also not implausible, since The cost requirements for each scenario both the United States and the USSR tested are shown in Table 5. Maintaining industrial activity such devices in 1962, with devastating It is important to note that, once the R22.6 billion of Tshwane’s GDP is derived results. Considering that between 1945 most severe event has been catered for from the manufacturing sector, much of and 1980 over 500 nuclear weapons tests (Scenario 9), all of the capital and capital which is from wet industries (Tshwane were conducted (Beck & Burton 2002), the maintenance costs for the lesser events Economic Development Agency 2015). prospect of any nation which possesses are automatically covered. However, the Since Rand Water supplies 76% of nuclear weapons detonating such a device operating cost for each of the lesser events Tshwane’s water supply (CoT 2015b), the in the upper atmosphere is an all too real must be divided by its RI and accumulated industrial water supply is only threatened possibility. to arrive at the total annual operating cost, by an event that shuts down power supply which is added to the discounted capital to both Tshwane and Rand Water. Since Basic minimum domestic and capital maintenance cost. Hence the wet industries are highly dependent on water supply additional charge to be borne by domestic water supply, process utilisation would be Meeting the basic minimum water require- water users paying for their services would the dominant water use and there is little ment of residents is a non-negotiable cost. come to 4.4 c/kℓ, which amounts to an flexibility to reduce demand. Hence the The cost, in terms of both human suffering increase of 0.44% (assuming a water tariff simplifying assumption has been made that and economic collapse, is simply too mas- of R10/kℓ (Province of Gauteng 2017)). This the full industrial water demand would sive to ignore. A basic minimum domestic is a small price to pay for protection against have to be met. However, it is unreasonable supply of 25 ℓ per capita per day has been the social and political consequences of a to expect that, even after providing a full used (CSIR 2005). However, achievement national blackout. water supply, industrial activity will not of this is assumed to require a supply at the There is a strong likelihood of violent be constrained by labour stay-aways and top end of the distribution system of twice social upheaval inherent in a national late arrivals due to transport difficulties this amount to account for users higher up blackout. Moreover, such an event also during an electricity disruption event. Also, the system being able (even unconsciously) has a high probability of occurrence (1:25 some users may not have enough power- to abstract well above their quota, which year RI for a seven-day outage and 1:155 generating capacity to maintain full opera- would leave downstream residents with RI for a 30-day outage, giving a combined tion. Accordingly, the assumption has been no water at all. The calculated minimum RI of 1:22 years). That represents a 4.5% made that only 50% of industrial output

24 Volume 61 Number 4 December 2019 Journal of the South African Institution of Civil Engineering Table 6 Costs and benefits of maintaining 50% of industrial output

3% net discount rate annual Mainte Opera Benefit/ Scenario RI Capital ∆ nance tion event Benefit Cost B-C B/C ratio billing+ No Description Year mill R mill R mill R mill R mill R mill R mill R – c/kℓ

6 Large, 7 day 30 28.303 0.283 2.405 216.600 7.220 1.807 5.413 4.00 12.69

9 Large, 30 days 155 28.303 0.283 10.308 928.100 5.988 2.298 3.689 2.61 16.13

Combined 28.303 0.283 – – 13.208 2.378 10.929 5.55 16.70

Note: + Required increase in normal billing to customers, based on average supply to paying domestic customers of 39 Mℓ/day (after deducting NRW water).

could be maintained. The calculated The benefit/cost ratio of 5.6 is very can be reduced by 75% before serious human minimum water to be supplied to maintain attractive. Moreover, aside from the eco- impact affects business. These sectors 50% of industrial output during electricity nomic advantage, maintaining the employ- are less dependent on labourers, and their disruptions is 39 Mℓ/day (including real- ment of labourers working in this sector wealthier employees would be better able to water losses). would hold the advantage of sustaining overcome transport difficulties than is the The most feasible cause for a disruption their families. case for industries. Moreover, the electricity of this areal extent and duration would be Naturally, the provision of a basic mini- requirements to sustain core operations are a national blackout, i.e., Scenarios 3, 6 or 9. mum water supply to all domestic users is more easily met by standby generators, many Moreover, in view of the available reservoir essential, since otherwise, at best, absen- of which would already be in place after the storage such an event would have to persist teeism due to people desperately looking long sequence of rolling blackouts. Hence, for longer than one day. This rules out for water will temporarily shut down all the assumption has been made that 75% of Scenario 3, leaving Scenarios 6 and 9. The economic activity. the GDP-generating activities of these sectors cost requirements for each scenario are could be maintained. The calculated mini- shown in Table 6. Maintaining other sectors’ mum water to be supplied to maintain 75% The combined increase in the cost per economic activity of other sectors’ activity during electricity kℓ of water supplied to paying industrial R179.4 billion of Tshwane’s annual GDP disruptions is 45 Mℓ/day (including real- water users comes to 16.70 c/kℓ, which is attributable to sectors such as finances, water losses). is 1.7% of the normal water charge. This commerce and other services that for The capital, annual capital maintenance is higher than was the case for domestic most of the time are impervious to and operating costs for the standby power users, since the assumption here is that restrictions in water supply (CoT 2015c). generation plant required to maintain 75% the full industrial demand would have However, below a critical level the work of the GDP output of other economic sec- to be provided to preserve 50% of the force of these enterprises will be affected, tors (excluding industries) for each scenario industrial output. Since the capital works resulting in absenteeism and impairment are shown in Table 7. will have to be paid off over a shorter of production. It is reasonable to assume The combined increase in the cost per period than used in the longer discounting that this point will be reached once kℓ of water supplied to paying water users period used to make valid comparisons, domestic water supply falls below a basic comes to 0.83 c/kℓ, which represents a the actual change in the billing will be minimum requirement. negligible increase of 0.08% of the normal greater than 16.7 c/kℓ while the capital is For the same reasons as before, water sup- paid water charge. being paid off, but thereafter for the rest ply to these sectors would only be disrupted The benefit/cost ratio of 117 and the of the life of the works the capital redemp- by a wide area power blackout affecting both annual net benefit of R156 million are tion cost will drop to zero. Nevertheless, Tshwane and Rand Water. It is assumed both huge, indicating that the small cost at 1.7% the increase in the billing will that water supply to these sectors, for which of protecting the other sectors of the remain small. water is not a part of their product stream, economy is well worth the investment.

Table 7 Costs and benefits of maintaining 75% of other sectors’ output

3% net discount rate annual Mainte Opera- Benefit/ Scenario RI Capital ∆ nance tion event Benefit Cost B-C B/C billing+

No Description Year mill R mill R mill R mill R mill R mill R mill R ratio c/kℓ

6 Large, 7 day 30 20.216 0.202 1.718 2 578.6 85.953 1.291 84.663 66.6 0.813

9 Large, 30 day 155 20.216 0.202 7.363 11 051.3 71.299 1.281 70.018 55.6 0.778

Combined 20.216 0.202 – – 157.252 1.339 155.913 117.4 0.833

Note: + Required increase in normal billing to customers, based on average supply to paying domestic customers of 45 Mℓ/day (after deducting NRW water).

Journal of the South African Institution of Civil Engineering Volume 61 Number 4 December 2019 25 Table 8 Cost requirements for preventing sewage overflow

3% NDR* Scenario RI Capital Maintenance Operation Ann. cost ∆ cost+

No Description (Year) (mill R) (mill R) (mill R) (mill R) (c/kℓ)

1 Small, 1 day 1 0.000 0.000 0.000 0.000 0.000

2 Medium, 1 day 10 119.150 1.192 1.446 8.375 3.926

3 Large, 1 day 19 119.150 1.192 1.446 7.847 3.679

4 Small, 7 day 10 0.000 0.000 0.000 0.000 0.000

5 Medium, 7 day 30 119.150 1.192 4.484 7.420 3.479

6 Large, 7 day 25 119.150 1.192 4.484 7.669 3.595

7 Small, 30 day 50 0.000 0.000 0.000 0.000 0.000

8 Medium, 30 day 100 119.150 1.192 16.128 7.450 3.493

9 Large, 30 day 155 119.150 1.192 16.128 7.377 3.458

Combined 119.150 1.192 44.116 8.088 3.792

* Net discount rate. Note: + Required increase cost based on 584 Mℓ/day sewage effluent discharge.

The large disparity between the benefits The associated overall annual cost raw sewage is considered necessary to pro- derived from these economic sectors, is estimated at R11.0 million, yielding tect the natural environment and informal compared with those of the wet industries, an annual benefit of R170.46 million. users, and to obviate biological overloading is immediately apparent. Moreover, at only Hence the annual net benefit comes to of downstream water treatment works. 30 Mℓ/day (25% of normal demand), the R159.45 million, with an extremely high The capital, annual capital maintenance estimated minimum water requirement to benefit/cost ratio of 15.5. and operating costs for the standby power sustain these activities is frugal. In terms The very high benefit/cost ratio indi- generation plant required for each scenario of minimum water use these sectors con- cates rare resilience against variations are shown in Table 8. tribute 19 times the economic contribution in the estimates of cost and benefit. A The increase in cost has been per unit of water used. Under normal cir- sensitivity analysis was done to determine expressed in c/kℓ of effluent treated. This cumstances this comparison is immaterial the robustness of the cost versus benefit has not been expressed as a percentage and the disparity in the overall economic analysis outcome; the following examples of normal billing for sanitation services, contribution much smaller. But when are given to illustrate this: since the requisite data was not at hand. water supply is severely constrained it is a However, if it is assumed that the sanita- much more important consideration. It is Doubling the cost or reducing tion billing is about half of that for water therefore considered extremely important the benefit by half supply, then the percentage increase in to provide enough emergency water supply Doubling the cost estimate (to R22 million sanitation billing to paying consumers to sustain these sectors. annually) or reducing the estimated benefit would be roughly 0.8%. Underpinning this is the imperative to by half (to R85 million annually) would still If a decision is taken to make the ensure a basic minimum water supply to yield a benefit/cost ratio of 7.7. R119.2 million capital investments in domestic users. Without this the fabric of standby power generation plant to prevent society will collapse, and with it all eco- Doubling the cost and reducing raw sewage overflows, then all nine sce- nomic activity. the benefit by half narios would be covered. If the estimated cost of mitigation were Overall impact on potable doubled and the benefit were halved, the water billing benefit/cost ratio of 3.9 would still remain INDIRECT COSTS OF ELECTRICTY The overall average increase in potable very attractive. DISRUPTION EVENTS water costs, taking account of the Indirect costs are expected to be more than proportions of water supplied to each Benefit/cost ratio break-even point the direct costs. For instance, if an electric- of the previously discussed three broad For the benefit/cost ratio to break even (i.e. ity disruption event that causes water sup- groupings (domestic, industrial and other benefit/cost = 1), the annual benefit will ply interruptions triggers major city-wide economic sectors), comes to 5.0 c/kℓ, or have to be decreased by a factor of 4 and (or nation-wide) civil unrest, it can result in 0.5% of normal paid billing. This is the the annual cost estimate will have to be the following indirect costs: estimated cost to secure a minimum basic increased by a factor of 3.87. QQ Halting of all economic activity during water supply to domestic users, sustain and after the water supply interruption 50% of industrial economic output and Prevention of raw sewage spillage QQ Violent uprising 75% of the output of other economically Standby power generation to prevent the QQ Loss of human life active sectors. spillage of up to 584 Mℓ/day (DWS 2014) of QQ Spreading of disease

26 Volume 61 Number 4 December 2019 Journal of the South African Institution of Civil Engineering QQ Loss of infrastructure The following conclusions are drawn long-term risks will also address the QQ Economic vulnerability due to global from this case study’s review of the City short-term risks. uncertainty in South Africa’s economy of Tshwane’s water infrastructure and the QQ Backup power generators (both mobile and emigration from the country. outcome of the scenario analyses: and permanent) will require ongoing Violent regime change might shut down QQ The benefits of ensuring uninterrupted servicing and maintenance – this will Tshwane’s economy for a year or more, minimum water supply greatly outweigh have to be incorporated into the opera- resulting in an economic cost equal to its the costs of ensuring uninterrupted tional and maintenance schedules of the entire GDP of R202 billion. water supply purely from a direct eco- city’s water department. nomic cost-benefit analysis perspective. QQ Alternative energy sources (such as The benefit/cost ratio of mitigating the solar panels or batteries) should be CONCLUSIONS risks posed by electricity disruption considered as part of further investiga- The main conclusions drawn from this events on water supply is approximately tions if it is decided to provide backup study are as follows: 5.6 for the city’s wet industries and 117 power to mitigate the risk of electricity QQ There are currently no frameworks, for the city’s other economic sectors. disruptions on water supply – this regulations or government policies that QQ Provision of a basic domestic water will have to be investigated in separate guide electricity suppliers (i.e. Eskom), supply, retaining 50% of industrial cost-comparisons between various Water Service Authorities and Water production and 75% of the output of backup power supply options during the Service Providers on how to mitigate other economic sectors, would increase preliminary design stage for providing the impact of electricity disruptions on the current normal billing to paying backup power generators. water supply. consumers by 0.5%. QQ Providing emergency storage capacity QQ There appear to be insufficient arrange- QQ The intangible risks associated with pro- for sewage inflow in wastewater treat- ments between key role players (Eskom, longed water supply interruptions (socio- ment works is more expensive than Water Service Providers and Water economic impacts) will probably be of providing backup power generation Service Authorities) regarding the roles greater concern than economic inactivity at wastewater treatment works, hence and responsibilities of each role player due to water supply interruptions. emergency storage will not be practical in the event of an electricity disruption. QQ Standby generators to prevent large- for medium- to long-term duration Cooperation and coordination between scale spillage of raw sewage would add electricity disruption events. these role-players during an electricity 0.8% to the normal sanitation bills QQ The supply and delivery of fuel to the disruption event would be very limited received by consumers. city’s water and sewer pump stations, as a result of this. QQ Reducing the risk of damage to Eskom’s and its water and wastewater treat- QQ In accordance with South Africa’s consti- power generating facilities and distribu- ment works will have to be planned tution, measures should be put in place tion network during a blackout is highly (and secured via a contract or formal to ensure that the country’s citizens are desirable. arrangement) to ensure fuel gets supplied with water to meet the mini- QQ For short-term electricity disruption delivered in the event of an electricity mum needs in the event of an electricity events it is crucial to ensure, firstly, that disruption event. On-site fuel storage disruption. This is a joint responsibility reservoirs and elevated towers are large will need to be considered, since lengthy between Water Service Authorities and enough to be able to supply at least two electricity outages will probably also the Water Service Providers that supply days’ AADD, and secondly, that operat- disrupt fuel supplies and dislocate them with water. This action is necessary ing rules for reservoirs and towers are transportation. regardless of the economic benefit of adhered to in order to ensure that water QQ Water restrictions implementation and ensuring continuous water supply during levels are maintained within the fluc- end-user buy-in will be critical to ensure electricity disruption events. tuation volume of the reservoirs/towers. that water supply to the city is not QQ The City of Tshwane has very few QQ For medium- to long-term electricity interrupted in the event of an electricity mitigation measures in place to ensure disruption events it is concluded that disruption event. The most effective way continuous water supply during elec- the volume of water stored in the city’s to restrict water use to domestic and tricity disruption events (specifically reservoirs and elevated water towers as commercial users during electricity dis- medium- to long-term events). Given the a measure to mitigate risks posed by ruption events will probably be to close effect of medium- to long-term electric- electricity disruptions is less important. reservoir and elevated tank outlet pipes, ity disruption events on the city’s water This is due to the fact that the volume and only open the pipes at certain times infrastructure, these events would have of water stored in the city’s reservoirs of day (after getting community buy-in). devastating effects on the city’s economy and elevated towers will almost certain- QQ Public buy-in and acceptance of all and, more importantly, its citizens. ly run out during medium- to long-term other water supply mitigation options QQ The Risk Analysis Mitigation electricity disruption events if water opted for by the City of Tshwane will be Framework of Integrated Water and supply cannot be sustained. crucial to avoid intangible risks associ- Electricity Systems, or RAMFIWES, was QQ The City of Tshwane will, in address- ated with water supply interruptions developed to address the need identified ing medium- to long-term electricity (such as wide-spread civil unrest, loss for such a framework. RAMFIWES disruption events, mitigate all risks of human life, economic meltdown and proposes a structured approach which associated with short-term electricity civil war). The public has to believe that can be used to mitigate the impact of disruption events, which means that the the mitigations opted for are, firstly, electricity disruption on water supply. capital cost of mitigating medium- to put in place by the municipality in the

Journal of the South African Institution of Civil Engineering Volume 61 Number 4 December 2019 27 public’s best interest, secondly, not CoT (City of Tshwane) 2015b. City of Tshwane parts-of-gauteng-and-nwest-without-water-after- occurring unnecessarily as the result of Newsletters. City of Tshwane Website. Available at: pump-station-explodes-20180827 [accessed on negligence by any of the parties involved http://www.tshwane.gov.za/sites/Departments/ 5 November 2017]. (e.g. Eskom, Tshwane or Rand Water), Public-works-and-infrastructure/Newsletters/ Mank, I 2015. Energy blackouts and water outages: and thirdly, that the implementation of Water%20and%20Sanitation%20Newsletter%20 A risk management approach towards raising the mitigating options is the best way Issue% 205.pdf [accessed on 24 July 2017]. awareness and assuming responsibility. Master’s to ensure continued economic activity CoT (City of Tshwane) 2015c. City of Tshwane Annual dissertation. Vienna, Austria: Vienna School of in Tshwane. Report 2014/2015. Pretoria: CoT. International Studies. CoT (City of Tshwane) 2016. City of Tshwane 2016/21 Pollet, B G, Staffel, I & Adamson, K 2016. Current Integrated Development Plan. Pretoria: CoT. energy landscape in the Republic of South AcKNOWLEDGEMENTS Cliver, E W & Dietrich, W F 2013. The 1859 space Africa. International Journal of Hydrogen Energy, This study was funded by the Water weather event revisited: Limits of extreme activity. 40: 16685–16701. Research Commission whose support is Journal of Space Weather and Space Climate, 3. Province of Gauteng 2017. New Tshwane Municipal acknowledged with gratitude. The study Copeland, C & Carter, N T 2017. Energy-water nexus: Tariffs as per Government Gazette 293 of 28 June 2017. team also wishes to express their gratitude The water sector’s energy use. Washington, DC: Rand Water 2018. 2017 Annual Report. Available to the City of Tshwane’s Electricity and Congressional Research Service. at: http://www.randwater.co.za/Annual%20 Water & Sanitation Departments for mak- CSIR (Council for Scientific and Industrial Research) Reports/2016-2017%20Annual%20Report/ ing the necessary information available for 2005. Guidelines for human settlement, planning Final%20Annual%20Report%2014Nov2017.pdf the Tshwane case study. and design. Pretoria: CSIR. [accessed on 5 November 2018]. DPLG (Department of Provincial and Local Tshwane Economic Development Agency 2015. Government) 2009. Guidelines for infrastructure Economic overview of Tshwane. Available at: REFERENCES asset management in local government (2006–2009). http://www.teda.org.za/Documents/SubMenu/ ADB (Asian Development Bank) 2009. Guidance Pretoria: DPLG. ExploreTshwane/Economic%20Overview%20 Note: Urban water supply sector risk assessment. DWS (Department of Water and Sanitation) 2014. of%20Tshwane.pdf [accessed on 30 July 2017]. Philippines: Asian Development Bank. Masterplan for Gauteng Waste Water Treatment WHO (World Health Organization) 2007. Risk Beck, H L & Burton, G 2002. Historical overview of Works: Book B – City of Tshwane. Pretoria: DWS. reduction and emergency preparedness: WHO six- atmospheric nuclear weapons testing and estimates GLS Consulting 2017. City of Tshwane: Master Plan – year strategy for the health sector and community of fallout in the continental united states. Health Water System GIS Database. Obtained from City of capacity development. Geneva, Switzerland: WHO. Physics, 82(5): 591–608. Tshwane Water Department. Winter, D 2011. Power outages and their impact on CoT (City of Tshwane) 2015a. Water Conservation News24 2018. Parts of Gauteng and NWest without South Africa’s water and wastewater sector. WRC and Demand Management System Strategy and water after pump station explodes. Available at: Report No. KV 267/11. Pretoria: Water Research Activities. Pretoria: CoT. https://www.news24.com/SouthAfrica/News/ Commission.

28 Volume 61 Number 4 December 2019 Journal of the South African Institution of Civil Engineering