Barcelona Cape Town Perth Water Security

Executive Summary

• No city can close its 2035 gap with a single technology. The most credible 2035 portfolios combine (a) funded climate-independent baseload already in construction, (b) low-cost leakage/pressure and targeted efficiency that delivers the cheapest drought-year cubic metre, and (c) one large reuse or desal increment whose schedule is the main risk.
• Perth is furthest ahead. ~145 GL/yr from two operating seawater plants plus 28 GL/yr groundwater replenishment already provide around half of drinking supply and baseload operation ^[5][31]. The funded Alkimos Stage 1 at 50 GL/yr desalination + 4.9 GL/yr groundwater for 2028 ^[4][5] covers roughly half of the modelled 75–125 GL/yr additional climate-resilient need by 2035 even with efficiency and recycling ^[63].
• Cape Town and Barcelona face delivery-risk gaps, not technology gaps. Cape Town's operating Table Mountain Steenbras boreholes at ~25 ML/d ^[35] and Phase-1 wellfields at ~80 ML/d ^[53] are an order of magnitude smaller than the 300 ML/d New Water Programme target for soon after 2030 ^[19]. Both Faure reuse at 70–100 ML/d and Paarden Eiland desalination at 50–70 ML/d target first water in 2030/31 ^[19][11] but require 2026–2027 PPP procurement ^[19], new or repeated environmental assessment, and a site change after the port-expansion conflict ^[30]. Barcelona already shifted from 3% desalination in 2021 to one-third in 2023 ^[39] on 80 hm³/yr from Llobregat at 60 hm³/yr plus Tordera at 20 hm³/yr ^[40], with a plan to reach 140 hm³/yr including Tordera expansion and a new 20 hm³/yr Foix plant within five years ^[40].
• Leakage and demand management are the cheapest and fastest drought insurance, but small relative to structural deficits. Perth can cost-effectively recover ~6.5 GL/yr in metro Perth for a net $2.5m/yr benefit by 2030 ^[22] against 20.3 GL/yr metro real losses and 19 GL/yr lost from ~273,000 leaky homes ^[22][8]. Cape Town targets 30 ML/d from pressure, leak detection and metering ^[53] and accelerated mains replacement to 50 km/yr water plus 100 km/yr sewer ^[11]. Barcelona greywater at city build-out saves only up to 0.3 Mm³/yr ^[51] versus desalination increments of tens of hm³/yr.
• Cost and energy comparisons are only valid within compatible boundaries. Llobregat desalinated water at €0.70/m³ versus €0.20/m³ for treated river water ^[39] cannot be directly compared to Cape Town provisional ZAR 2018 opex of R5/kL for reuse versus R9/kL for desalination ^[7] or Perth's $1.00/kL operating estimate that excludes renewals and membranes ^[4]. Seawater reverse osmosis typically uses 2.5–3.5 kWh/m³ with energy recovery, 4–6 kWh/m³ without ^[61], against 0.6–1.5 kWh/m³ for recycled water ^[57] and about one-quarter energy for Perth groundwater replenishment versus seawater desalination ^[27].

1. Current supply portfolios, 2035 demand projections and drought supply gaps

Perth – Integrated Water Supply Scheme shifting to manufactured water

Perth's Integrated Water Supply Scheme sourced 337 GL in 2024–25 for over 2 million people across Perth, Peel, Goldfields, Agricultural, South West and Upper Great Southern regions ^[9].

The portfolio uses streamflow, groundwater, desalinated seawater and recycled water ^[9]:

• 15 dams
• six groundwater treatment plants
• one groundwater replenishment plant
• two desalination plants ^[9].

The structural shift is stark. Only about 10% of Perth water now comes from surface water after more than four decades of decline ^[31].

Groundwater provides up to 40% of drinking supplies ^[27]. Two desalination plants provided around half of Perth homes and businesses drinking water on average over the past five years ^[5]. Jacobs reports the Kwinana and Binningup plants now supply 30–40% each year ^[63].

The Water Corporation states it has shifted away from streamflow and rainfall toward climate-independent sources ^[9].

Demand pressure comes from both climate and growth.

Climate:

• Winter rainfall in south-west Western Australia declined ~20% since 1970 ^[59]
• reduced rainfall plus hotter summers and drier winters since the 1970s significantly reduced surface and groundwater availability ^[4]
• annual rainfall runoff fell ~70% while population and demand rose almost 30% since 2000 ^[47]
• water tables fell up to 10 m in places, a 1,000 GL aquifer storage loss since 1980 ^[27]

Growth:

• Perth population at 2.12 million is projected to reach 2.9 million by 2031 and 3.5 million by 2050 ^[59]
• State modelling projects total Western Australia demand to 2,600 GL by 2030 and 3,600 GL by 2050 at ~1.8%/yr ^[43]
• urban use at 1–1.6%/yr as the community adapts to climate, efficiency and higher alternative-source costs ^[43].

Three mutually consistent gap statements define 2035:

• 75–125 GL/yr additional climate-resilient water needed by 2035 even with efficiency and recycling ^[63].
• More than 110 GL of new sources required into the 2030s in metro Perth alone ^[17].
• Up to an additional 125 GL/yr by 2035 even meeting reduction and recycling targets ^[5].

A mandated groundwater cut sharpens the near-term gap. Gnangara supplies almost half of all Perth water including scheme and direct extractions ^[4]. Under the 2022 allocation plan the Water Corporation must cut abstraction 27%, or 30 GL/yr, from 1 July 2028, with other sectors cut 10%, or 10.2 GL/yr, and garden-bore restrictions aligned to scheme users ^[4].

The Corporation must therefore plan for an initial 50 GL/yr replacement operational by 2028 with further expansion ^[4], and master planning indicates multiple new climate-independent sources through 2050 ^[4].

Cape Town – overallocated surface system with fast demand rebound

The Western Cape Water Supply System allocation totals 576 Mm³/yr: City of Cape Town 347 Mm³/yr or 60%, agriculture 186 Mm³/yr or 33%, other urban and industry 43 Mm³/yr or 7% ^[7]. That bulk system is only a small part of 3,037 Mm³/yr used in the Western Cape, three-quarters for agriculture ^[7]. Infrastructure comprises six major dams plus pipelines, weirs, diversions and supplements, with the City owning Steenbras and Wemmershoek while national Department of Water and Sanitation manages other dams and overall operation ^[7].

Yield assessments have tightened.

• Existing yield is quoted at 475 Mm³/yr at 1-in-50 assurance ^[25], rising to 556 Mm³/yr after the 81 Mm³/yr Berg Water Project due in 2007 ^[25].
• Updated hydrology including recent drought and invasive alien plants cut modelled yield 33–35 Mm³/yr ^[7].
• The system is already overallocated and without intervention becomes more so ^[7].
• Older planning expected the augmented 556 Mm³/yr yield to be fully utilised by 2011 under high growth, delayed only to 2015 with successful demand management ^[25], on 2–3%/yr long-term growth ^[25].

Drought history explains the planning conservatism. The 2015–2017 drought was the worst since 1904 approaching Day Zero and a disaster declaration ^[14], assessed as a 1-in-590-year runoff event for 2015–2017 and 1-in-330 for 2017 alone ^[35], emptying full 2014 dams to 38% in just over three years ^[34] and to 9.8% of usable storage at the height ^[35]. Dams offer only about two years security ^[34]. The City now assumes a 25% reduction in surface and groundwater availability over a 30-year horizon ^[34][35], and invasive plants currently cut yield 24 Mm³/yr with 85 Mm³/yr future potential if uncleared ^[34][35].

Current demand has rebounded sharply, creating the 2035 risk.

• Before 2018 the City used up to 1,200 ML/d in 2015, falling to ~500 ML/d when dams hovered at 20% ^[30].
• September 2017 use under Level 5 restrictions was ~570 ML/d versus ~850 ML/d recently, a 50% rise over nine years ^[30]. Average 2025 use was ~1,025 ML/d ^[53].
• The 2025/26 growth rate exceeds three times the ten-year WCWDM average and more than double bulk-water planning assumptions ^[30].
• Maximum dam allocation, set nationally with shared agricultural and town demands, is expected to be reached within two years ^[30]. Cumulative 2025 Wemmershoek catchment rainfall was below 2016 and only slightly above drought years 2015 and 2017 ^[30].

The policy response is:

• a 300 ML/d new-capacity target over 10 years from new supplies plus management ^[34][35], or 300 ML/d soon after 2030 ^[19]
• committed programme estimated to add 128 Mm³/yr, or 347 ML/d, rising to 154 Mm³/yr, or 417 ML/d, including demand management ^[7]
• 2019 Water Strategy split envisages ~40 ML/d surface, 100+ ML/d groundwater, 50–70 ML/d desalination and 70–100 ML/d reuse ^[34][35].

Barcelona – Ter-Llobregat system with shrinking rivers

Barcelona Metropolitan Area with 5.3 million inhabitants consumed 293 hm³ from all sources in 2019, projected to 329 hm³ in 2050 assuming all planned urban sectors develop, and 354 hm³ at long-term full build-out ^[51].

• Barcelona City with 1.62 million shows +4% demand by 2050 under a committed climate-action scenario ^[51].
• Domestic drinking use is already low at 98.7 litres per equivalent inhabitant per day ^[51], after population in the consumption analysis rose 9.44% from 1,503,451 to 1,660,122 ^[51].

Supply composition swung during drought. In 2021 drinking water was 63% surface, 34% groundwater and only 3% desalination ^[39]. In 2023 desalination provided one-third, wells less than one-quarter and shrinking rivers less than one-fifth ^[39]. North-eastern Catalonia resources are projected to fall 18% by 2050 ^[39], while Mediterranean warming at 20% faster than global average with 20–50% rainfall decline by 2100 implies more frequent severe droughts ^[40]. Drought emergency is triggered below 16% reserves, when northern-basin reservoirs were reported at only 25% ^[39]. Restrictions include a 230 litres per person per day allocation affecting agriculture, industry and some municipal uses ^[39].

No single 2035 gap volume is given in the evidence for Barcelona. The gap must be inferred from falling river contributions, the 18% resource decline to 2050 ^[39], and reliance on desalination rising from marginal to one-third of drinking supply within two years ^[39].

2. Desalination and wastewater reuse: operating, funded and proposed

Perth – operating baseload plus funded 2028 increment

Operating:

• Perth Seawater Desalination Plant at Kwinana started production in 2006 ^[3], described as Australia's first large-scale plant at 45 GL/yr ^[4], with Water Corporation citing 50 GL/yr output capacity ^[3] and on average 15% of Perth supply ^[3]. Other reporting cites 144 ML/d final product from brackish trains plus 160 ML/d seawater trains with 250 ML/d extended output ^[65], cost A$389m ^[65], and 17% of Perth supply ^[65]. The 45 versus 50 GL/yr difference reflects assessment-year versus nameplate reporting boundaries, not a physical change.
• Southern Seawater Desalination Plant at Binningup developed in 2012 at 100 GL/yr ^[4], initially 50 GL/yr expanded to 100 GL/yr or 33% of Perth requirements ^[58], estimated output 300 ML/d ^[58], cost A$955m ^[58], completed at full capacity January 2013 after September 2011 opening at reduced output ^[58], supplying Perth and Bunbury via a 30 km pipeline to Harvey storage ^[58].
• Groundwater Replenishment Scheme introduced in 2017 at 28 GL/yr ^[4], doubled by Stage 2 completed in 2022 to recharge up to 28 GL/yr ^[9]. Stage 1 capacity was 14 GL/yr operating since 2017 ^[27], recharging >54 GL in five years ^[27]. Stage 2 added a second advanced plant at Beenyup in Craigie, four recharge plus four monitoring bores at Neerabup and Wanneroo and a 13 km pipeline ^[27], cost $320m ^[27]. Beenyup treats up to 135 ML/d wastewater ^[41], purifying secondary effluent otherwise discharged to ocean via ultrafiltration, reverse osmosis and ultraviolet for aquifer recharge ^[64]. Stage 1 was designed for 14 GL/yr commissioned 2016 ^[64], Stage 2 commissioned 2019 ^[64] with full commissioning after off-site bore testing for the 2017-started build ^[27]. The 38-ML indirect potable reuse figure cited for Beenyup ^[34] is a daily rate consistent with ~14 GL/yr Stage 1, not a conflict.

Funded construction:

• Alkimos Stage 1 plus Eglinton at ~55 GL/yr expected to commence operations in 2028 ^[4]: 50 GL/yr seawater desalination co-developed with 4.9 GL/yr Eglinton groundwater integrated via Wanneroo Reservoir ^[4]. Stage 1 supplies up to 50 GL/yr, Stage 2 to 100 GL/yr ^[5] via a 33.5 km pipeline to Wanneroo feeding the IWSS ^[5]. Some earthworks, marine tunnels, tanks, trunk mains and power are sized for ultimate 100 GL/yr ^[4]. Construction commenced October 2024 for 2028 completion ^[29], or mid-2024 with 2028 delivery ^[59], with design nearing completion and construction ramping on track for 2028 ^[63]. Stage 1 delivers 150 ML/d ^[63] plus 4.9 GL/yr groundwater ^[63], with 6.6 km subsea tunnelling, 50 ML on-site storage and variable-speed pumping ^[37].
• Capital boundaries explain headline conflicts: $2.8 billion covers 34 scope elements plus program costs and risks for the recommended Alkimos plus Eglinton proposal ^[4], retained as $2.8 billion confirmed December 2023 within an overall program ^[29][63], while ~$1.4 billion AUD refers to the plant alliance delivery scope with Water Corporation, Acciona Agua, Acciona Construction and Jacobs ^[37]. Annual opex is estimated at $51.9m at $1.00/kL excluding major renewals and membrane replacement, which may understate lifecycle costs ^[4]. Manufactured desalination water is capital-intensive with substantially higher opex than historic sources, raising ongoing government costs ^[4].

Unfunded / future:

• Alkimos Stage 2 to 100 GL/yr is future-proofed but only 50 GL Stage 1 is confirmed ^[29], with design provision for a second 50 GL/yr expansion ^[37] and capacity designed to double as demand grows ^[63]. Projections that Kwinana, Binningup and Alkimos together supply 54% by 2028 and 64% with Alkimos Phase 2 from 2032 ^[47] depend on that unfunded second stage and demand assumptions.

Cape Town – small operating groundwater, large PPP reuse and desal still pre-construction

Operating:

• Table Mountain Group Aquifer Steenbras cluster completed at ~25 ML/d ^[35], within a ~50 ML/d fractured-rock system in three clusters ^[35]. Phase-1 wellfields are reported online at ~80 ML/d with full programme potentially ~120 ML/d depending on sustainable yield ^[53]. Only Table Mountain boreholes in upper Steenbras are described as currently delivering water among New Water initiatives ^[30]. Nuweberg and Groenland Berg clusters are post-2035, beyond the horizon ^[35].
• Cape Flats Managed Aquifer Recharge ultimate ~60 ML/d is under construction ^[35]: Strandfontein West completed, Phillipi and Hanover Park under construction, Strandfontein NE and Mitchells Plain by 2030 ^[35]. Atlantis MAR total ~26 ML/d due end August 2026 ^[35]. Atlantis abstraction was set back by cable theft and Cape Flats boreholes and treatment were still being built at last reporting ^[30].
• Three temporary Day Zero seawater plants at V&A Waterfront, Strandfontein and Monwabisi on short-term contracts were costly, contested and switched off when dams refilled ^[53].
• No permanent large desalination or potable reuse is yet operating. Identified effluent reuse potential is 50.2 Mm³/yr non-potable plus 74.6 Mm³/yr potable, total identified 124.7 Mm³/yr versus 11.6 Mm³/yr existing reuse ^[25].

Funded / committed but not yet delivering:

• Berg River–Voëlvlei Augmentation at 40 ML/d effective yield, or 15 Mm³/yr ^[7][30], City allocation ~40 ML/d ^[34][35], pumping Berg River water into Voëlvlei Dam ^[30]. Environmental approval was already given with detailed design by DWS ^[7], under a 20-year supply agreement approved October 2022 ^[34]. Completion estimates conflict by source year: 2023 originally, revised to 2027, now 2030 ^[30]; due 2028 after a one-year delay ^[11]; ~2026/7 to be confirmed ^[34]; ~2029/30 to be confirmed ^[35]. Cost is cited as ~R1.85 billion ^[30]. It is implemented by national DWS, not the City ^[30], fully funded off-budget with affordability of projected tariffs delaying municipal off-take agreements ^[42].
• Water Reuse Phase 1 effective yield 70 ML/d, or 26 Mm³/yr, and Desalination Phase 1 at 50 ML/d, or 18 Mm³/yr, were part of a committed 128 Mm³/yr programme at R7.5 billion capital in provisional ZAR 2018 costs ^[7]. Those phase definitions pre-date current Faure and Paarden Eiland scopes and cannot be directly substituted.

Proposed / PPP procurement:

• Faure New Water Scheme direct reuse at 70 ML/d initially to Faure Water Treatment Plant ^[14], ultimate 100 ML/d if required ^[14], or 70–100 ML/d ^{[11][19][34][35]}, expected to produce up to 100 ML/d potable ^[49]. Source is Zandvliet Wastewater Treatment Works effluent complying with General Authorisation under the National Water Act ^[34], piped 5 km to Faure for advanced purification ^[49], via Zone 1 multi-barrier, reservoirs and pump at Zandvliet plus Zone 2 advanced plant at Faure and connecting pipeline ^[14]. Treatment train is ozone, biologically and granular activated carbon, ultrafiltration and UV advanced oxidation with further treatment at Faure Water Treatment Plant ^[34], meeting SANS 241 and WHO guidelines ^[53]. Output is blended at 25% purified with dam water then retreated before distribution ^[14], also described as 20% reuse to 80% dam water ^[49][53]; the difference reflects design versus simplified reporting, with the feasibility blend at 25% ^[14]. Must be fully operational at 70 ML/d by 2031 to meet demand with construction started by 2028/29 at latest ^[14]. Cost estimates use incompatible scopes: R2.83 billion capital for 70–100 ML/d completion 2030/31 ^[11]; ~R3.3 billion total cost completion August 2029 with validation September–November 2029 ^[34]; R3.2 billion ^[35]. Opex in ZAR 2018 was R5/kL for Water Reuse Phase 1 ^[7]. Tariff impact is a once-off ~6–8.5% increase in commissioning year unless mitigated by subsidies ^[14].
• Paarden Eiland / Table Bay permanent desalination at 50–70 ML/d ^{[11][19][34][35]}, designed at 70 ML/d drinking water ^[53], using offshore intake, pre-treatment, high-pressure reverse osmosis, remineralisation, chlorination and pumping into the WCWSS network ^[53]. Cost estimates are incompatible: R2.38 billion for 50–70 ML/d completion 2030/31 ^[11]; ~R5.5 billion ^[35]; roughly R5 billion in 2023 prices for a single Atlantic plant ^[53]; ZAR 2018 opex for Desalination Phase 1 at R9/kL versus R5/kL for reuse ^[7]. Schedule estimates conflict: commencing 2026 fully operational by 2031 ^[11]; first water 2030/31 ^[19]; procurement starts 2026, first water ~2030, full operations end 2031 ^[53]; 2031/32 under review ^[35]; fully operational only by 2036 at best ^[30]; plus a five-year push from finding an alternative site and repeating environmental assessment after Transnet's port expansion scuttled Paarden Eiland after scoping was completed ^[30]. Desalination environmental and heritage assessment is still to be undertaken ^[35], with site selection, technical feasibility, finance-procurement and EIA tasks still in scoping and pre-feasibility as of March 2024 ^[34].

Barcelona – operating desalination with funded expansion to 140 hm³/yr

Operating:

Two plants provide 80 hm³/yr to the Aigües Ter-Llobregat system: Tordera 20 hm³/yr and Llobregat 60 hm³/yr ^[40].

Llobregat can produce 200 ML/d, or 60 hm³/yr ^[39], serving 4.5 million residents and up to 24% of metropolitan use ^[39].

Funded / planned within five years:

Extension of Tordera at €250m plus a new 20 hm³/yr plant on the Foix at hundreds of millions of euros to reach 140 hm³/yr total potential within five years ^[40]. Spain's €2.2 billion drought package pledged €220m to expand another desalination station north of Barcelona ^[39], likely overlapping the Tordera scope, plus €224m for treatment in southern Spain ^[39], but evidence does not confirm whether the €220m and €250m figures share the same boundary.

Llobregat production cost is €0.70 per 1,000 litres versus €0.20 for treated Llobregat River water ^[39]. Reverse osmosis requires large energy not yet entirely renewable ^[39], averaging ~4 kWh/m³ from 8 kWh/m³ in 1970s plants to 2.3 kWh/m³ in optimised systems ^[40]. Brine at ~0.55 litres per 0.45 litres fresh water ^[39] is mixed with Baix Llobregat treatment effluent and discharged 3 km offshore at 60 m depth ^[39], with atmospheric emissions, expensive conveyance with maintenance and limited life, and harm to marine flora cited as constraints ^[40].

Large-scale reclamation is judged not feasible besides treatment cost because full recycling would cut river flows and kill river life; production only reduces vulnerability if it substitutes for environmental abstraction ^[40]. This contrasts with Cape Town and Perth where indirect or direct potable reuse via aquifers or blending is central.

3. Leakage reduction and demand management credibly deliverable by 2035

Perth – metered network losses plus large customer-side opportunity

Statewide unbilled water in 2020–21 was 51.6 GL of 374 GL supplied, or 13.8%, up from 12.6% at the prior audit after peaking at 54 GL or 14.7% in 2018–19 ^[22].

Metro Perth unbilled was 30 GL or 11.8%, similar to 29.4 GL or 11.6% in 2013–14 after peaking at 32.7 GL or 13.3% in 2016–17 ^[22]. Target is 10% by 2030 for metro with 8% stretch, no equivalent regional or statewide target ^[22].

Industry standards suggest 15.6 GL statewide, about one-third of losses, is recoverable including ~6.7 GL in metro Perth ^[22]. The Corporation estimates ~6.5 GL in metro Perth could be cost-effectively recovered via pressure management, non-visible detection and meter replacement for $2.5m/yr net benefit by 2030 ^[22]. Had that been achieved in 2020–21 metro would have met the 2030 10% target, but meeting it while the network expands needs an effective plan the audit found lacking ^[22].

Real leaks and bursts statewide were ~36.7 GL or 10% of supply, up from 32.8 GL in 2014 and peaking at 39.5 GL or 11% in 2018–19 ^[22]. Metro leakage was 20.3 GL or ~8%, essentially unchanged from 20 GL in 2013–14, while non-metro losses rose except Great Southern ^[22]. About 30% of unbilled water is metering error, authorised use and theft, just over 70% leaks and bursts, of which 21.1 GL is deemed unavoidable for this network size ^[22]. Total metro losses equal 30% of Southern Desalination Plant annual capacity ^[22].

Metro Infrastructure Leakage Index 1.6–1.9 is mid-range nationally but higher than most capitals; Albany, Kalgoorlie-Boulder and Geraldton were 2.0–4.0 in 2019–20 ^[22]. Perth leak-burst rate 11.5 per 100 km in 2019–20 was lowest of any capital with trunk and distribution rates falling since 2016 ^[22].

The toolkit is pressure management, non-visible detection and repair, meter replacement and reservoir relining judged by Economic Level of Leakage where investment must be below marginal water cost ^[22]. Investment totals $464.7m since the last audit including ~$214m Pipes for Perth since 2016 ^[22], plus a $40.4m 2024–25 Budget two-year renewal of 30 km reticulation for completion by mid-2026 in Stirling, Joondalup, Belmont, Nedlands and Bayswater, targeting pipes over 70 years old ^[23].

Household leaks are material. Nearly 19 GL/yr is lost from ~273,000 leaky metro homes ^[8], based on a four-year study of 2,500 Perth and Peel homes ^[8].

• close to 15% of Perth and Peel homes leaking on any day and 30% experiencing a leak at least yearly ^[8]
• average leak loses just over 9 litres per hour or 218 litres per day ^[8]
• most commonly taps, toilets, hot water systems, air conditioners and garden reticulation ^[8]
• raising bills $153–$1,800/yr ^[8].

Demand programmes show verifiable savings.

• Since 2007 Waterwise Businesses saved >157 GL ^[17], with 287 participants saving 9.6 GL scheme water in 2023–24 ^[17]
• Data-logger analysis and leak identification saved >2.3 GL in 2023–24 exceeding a 1.3 GL target ^[17]
• Waterwise Towns saved >2.1 GL since 2013 ^[17], Aquatic Centres ~1.59 GL since 2008 ^[17], Great Southern initiatives 42 ML in 2024–25 ^[17], seven South West schools 13 ML/yr via plumbing ^[17], and comparative-use letters to >4,000 South West households saved 20 ML ^[17]
• In 2024/25 22,531 metro high-use customers received comparative letters ^[17]

More than 30% of scheme water is used on gardens in Perth and Peel ^[17], and over 36% of household use is outdoors ^[17], with >$3m to garden efficiency since 2021 ^[17]. Smart-meter rollout reached >1,000 Exmouth customers with near real-time data to detect leaks faster, piloted in Perth ^[17]. Modelling assumes a 15% per-person urban use cut by 2030 from 2014 levels with no further per-capita change after 2030 ^[43]; actual household growth at 1.6%/yr undershot 2.2% projected, likely reflecting awareness and scheme plus garden-bore restrictions ^[43].

Cape Town – pressure and behaviour proved in drought, now eroding

Cape Town unaccounted-for water was estimated at 23.3% or 186 ML/d, 67 Mm³/yr, with 20–35% minimum night flow indicating leakage and wastage ^[25]. More recent reporting puts ~18% of input lost as non-revenue through leaks and pipe failures ^[11].

With interventions total demand could be cut from 797 ML/d, 290 Mm³/yr, to 539 ML/d, 197 Mm³/yr ^[25].

The programme includes 30 ML/d from pressure management, leak detection and metering described as the cheapest megalitre, with 171 km of leaky main replaced in three years to end-2025 ^[53]. Replacement targets accelerated to 50 km water and 100 km sewer per year from 25 km ^[11], investment up 60% since 2023 ^[11].

Drought behaviour shows what is technically possible. Citywide use fell from ~1,200 ML/d in 2014 to ~555 ML/d by April 2018 and residential per-person from ~183 to 84 litres per day with no new plant, via progressive tariffs, devices, pressure management and campaigning ^[53].

Measures included aggressive pressure management, water-management devices and leak repair, progressive pricing across four restriction levels to January 2018 plus punitive drought pricing from February 2018 at Level 6 and an intensive two-year Day Zero campaign ^[35]. Invasive clearing ultimately saves 30 ML/d ^[35].

Recent rebound underlines fragility. Use at ~850–1,025 ML/d ^[30][53] with growth multiples above planning ^[30] suggests without sustained pricing, pressure control and communications the drought savings decay.

Evidence on behavioural magnitudes elsewhere cautions against assuming persistence:

• Costa Rica neighbour comparisons plus tips cut demand ~5% ^[1]
• South Africa social recognition and public-good appeals only ~1% on average ^[1]
• Indian goal-setting plus tracking cut 15–25% persisting over a year ^[1]
• Smart-meter feedback ranges 3–53.4% averaging 19.6% ^[33], Sydney in-home displays 7–10% ^[33], mid-coast NSW customised information ~8% ^[33], and metering studies 7–22% ^[50] with up to 46% customer savings via leak-finding ^[50]

Price elasticity averages –0.40, median –0.34, implying +25% price for –10% quantity ^[1]; California block-rate shifts cut demand ~3% controlling for price, reversing to +10% when returning to uniform rates ^[1].

Barcelona – already-efficient indoor use, small greywater increment

With domestic drinking use at 98.7 litres per equivalent inhabitant per day ^[51], large indoor savings require device and reuse shifts rather than behavioural slack. A four-person household generates ~200 litres per day reusable greywater, 50 litres per person in households and 80 litres in luxury hotels versus 35 litres per person for toilet flushing ^[51]. Greywater is defined as baths, showers plus excess pool water for cisterns and green areas, excluding sinks, kitchens, bidets, washing machines, dishwashers, industrial and contaminated streams ^[51]. At the last five years' development pace recycling could save up to 300,000 m³/yr drinking water ^[51], or 0.3 Mm³/yr — two orders below desalination increments. Modelling shows net saving per dwelling per year of –€90 for 16-dwelling buildings, –€124 for 26 and –€170 for 50 ^[51]. Draft rules would require recycling in new builds and renovations with 16+ dwellings, sports centres, hotels and buildings using >500 m³/yr, excluding hospitals, health centres, nursing homes, schools and nurseries ^[51]. No Catalan or Spanish greywater law exists; Barcelona made systems compulsory in 2022 for new and major renovations in Gràcia and 22@ with a city-wide compulsory byelaw programmed for 2025 ^[51]. Graywater laundry and bathing water for irrigation is an established demand-reduction pathway ^[1], and leak detection and repair in municipal lines can also cut consumption ^[1], but no Barcelona network leakage percentage is given in the evidence.

4. How the four families compare

Delivered-water cost — do not compare across price years without adjustment

Published benchmarks use incompatible price years including 2008 prices, 2015 dollars and 2020 dollars and require inflation and boundary adjustment before direct comparison ^[48]. The same applies across Barcelona euros, Perth Australian dollars and Cape Town rand.

• California reference points: Seawater desalination was most expensive in a 2019 analysis at ~$2.00/m³, versus reuse ~$1.50, brackish ~$1.00 and stormwater ~$0.50 ^[1]. Conservation and efficiency at $600–$1,800 per acre-foot is below indirect potable reuse at $2,000–$2,500, direct potable reuse at $2,400–$3,600 and seawater desalination at $3,100–$3,400 ^[46]. The state water-loss model defaults to $1,093 per acre-foot avoided cost by averaging alternatives: indirect potable reuse $1,800, brackish $1,100, imported $1,015 and stormwater $590 in 2015 dollars ^[28].
• Large-scale desalination ranges: Mostly $1–$2.50/m³ ^[55]; large municipal/industrial $0.5–$1.5, medium $1–$3, small island/offshore $2–$5 ^[61]; Sorek B contracted $0.41 with a 107-plant study minimum $0.27 ^[55]; Australian Water Association $1–$4/kL ^[31]. Economies of scale are strong: capex per m³ falls from $1,341 to $1,149 to $792 with size in 2008 prices, opex $0.43 to $0.22, unit product cost $1.401 at 1 MGD to $0.716 at 50 MGD ^[48]; large plants $1.53–$1.93 versus small $2.10–$3.31 median $2.13 ^[48]. Capital projections often understate true cost; Carlsbad final $1b was four times original ^[38]. Over lifecycle capex is only 20–30% with opex 70–80% ^[61].
• City-specific delivered costs:
  • Barcelona Llobregat €0.70/m³ versus €0.20/m³ treated river water ^[39].
  • Perth Alkimos opex $1.00/kL excluding major renewals and membranes, potentially understating lifecycle ^[4].
  • Cape Town provisional ZAR 2018: Reuse Phase 1 R1,360m capex, R5/kL opex versus Desalination Phase 1 R1,650m capex, R9/kL opex ^[7]. Faure capital R2.83 billion ^[11] versus ~R3.3 billion ^[34] versus R3.2 billion ^[35]; desalination R2.38 billion ^[11] versus ~R5.5 billion ^[35] versus ~R5 billion in 2023 prices ^[53]. These cannot be ranked without reconciling capacity, pipeline, land, contingency and price-year boundaries.
• Reuse versus desalination: Municipal reuse is cost-competitive with seawater desalination above 1,500 m³/d ^[48]; small non-potable median $0.48, small indirect potable $1.50, larger indirect potable $1.06 ^[48]; non-potable is cheaper than potable due to lower treatment ^[48]. Recycling or reusing treated wastewater is often less expensive than ocean desalination ^[38], and recycling reaches break-even faster especially with volatile energy ^[57], but potable reuse is almost as expensive as desalination where advanced treatment is required ^[44]. Financial levelized RO-treated wastewater falls from $0.81 at 50% utilisation to $0.68 at full utilisation ^[48].
• Leakage and efficiency: San Diego conservation marginal cost $150–$1,000 per acre-foot versus imported, recycled or desalinated supplies ^[1]; several California showerhead, washer and toilet measures had negative cost with bill savings exceeding purchase ^[1]; US manufacturing conserved water is at minimum five times cheaper than alternatives ^[1]; Sydney WaterFix Residential delivery $1.44/kL and economic $0.66/kL meeting the Economic Level of Water Conservation ^[21], levelised $1.47/kL in 2017–18 ^[26] against long-run marginal water $1.80–$2.20 depending on dam levels ^[26]. Active leak detection defaults are $595 per mile surveying plus $5,946 per main leak and $2,330 per service leak at 70% efficiency ^[28]; developing-country physical-leakage reduction costs $215–$500 per m³/d, ~$250 for efficient projects with 4–8 year payback at $0.20/m³ water value ^[16]; halving NRW at ~$500 per m³/d is the rule of thumb ^[13]. Unitywater analysis found every $1 on active detection saved ~$3 in bulk water at current leakage, with economic levels at ~37.2 litres/connection/day for detection alone versus ~33 for pressure management ^[36], and mains replacement purely for leakage uneconomic at ~$1,000/m or $43.6m for 40 km to cut 51 to 49 litres/connection/day ^[36].

Energy use per cubic metre

Removing salt requires large energy ^[1], largely determined by salt separated ^[44]. Brackish needs less than seawater ^[44]; sewage wastewater can save energy versus seawater but adds pretreatment complexity ^[44]. Energy is 30–50% of total variable production cost ^[48], 40–60% of SWRO opex ^[61], 50–60% without energy recovery ^[52], and up to half of salt-removal cost ^[38].

• Seawater RO: ~4 kWh/m³ average, 8 in 1970s to 2.3 optimised ^[40]; 2.5–3.5 today from ~20 in the 1970s ^[55]; 3.5–4.5 even with 2026 advanced recovery ^[57]; 4–6 without recovery, 2.5–3.5 with ^[61]; 3–5 modern with recovery, 8–10 older without ^[62]; as low as 3 at optimised industrial scale ^[62]; 3.3–3.6 land-based versus 1.8–2 Norwegian sub-sea trial, ~40% lower ^[47]; theoretical minimum ~1 ^[55][62]. Thermal uses 3–5 times RO ^[55], ~13 kWh/m³ ^[55].
• Reuse: 0.6–1.5 kWh/m³ ^[57]; Perth replenishment uses same RO as seawater but ~one-quarter energy ^[27]; seawater feed pressure is four times recycled-water pressure ^[48].
• System implications: At US power prices electricity alone is ~$0.45/m³ for 3.5 kWh/m³, ~$0.90 in California at double average power, implying ~$1.50 total at one-third energy share rising to ~$2 in California ^[55]. Desalinating full US household use at 1,135 litres per day needs ~1,450 kWh/yr raising domestic electricity ~13%, UK 349 litres per day ~15% ^[55]. A 10 m³/d system at 4 kWh/m³ needs ~40 kWh/d ^[62]. LAX desalination at 4,061 kWh per acre-foot versus 3,254 for imported Delta water excluding distribution narrows the gap ^[44]. Energy costs are 30–40% of opex in some treatment and pumping operations ^[60].
• City signals: Llobregat RO is large and not yet fully renewable ^[39]; Alkimos energy mitigated by design efficiencies and recovery to match or exceed benchmarks ^[5], with 132 kV supply integrating solar PV and battery ^[63], a 100% renewable commitment covered only by an assumed power purchase agreement at assessment ^[4] and an expected up-to-400 MW wind agreement also covering Kwinana and Binningup ^[29]; Kwinana uses 180 GWh/yr more than offset by 270 GWh/yr from the 80 MW Emu Downs Wind Farm ^[65], with the Corporation purchasing 66% of output at 24 MW, 185 GWh/yr ^[18], using isobaric exchangers up to 98% transfer for among the world's lowest specific energy ^[65]; Paarden Eiland 70 ML/d on a coal-leaning grid runs roughly a small town's power with an emissions footprint dam water lacks ^[53].

Construction and permitting lead times

Large supply implementation can take ten years or more from feasibility to completion ^[25]. Water and sanitation infrastructure generally has a 7–10 year lifecycle budgeted across the medium-term framework into outer years to 2034/35 ^[11].

• Alkimos: referral 2019 ^[29][5], assessment required June 2019 ^[29], environmental review for comment September–October 2022 ^[29][5], development approval August 2023 ^[29], state approval under Ministerial Statement 1207 consistent with EPA Report 1739 ^[5], federal approval November 2023 ^[29] after pending Commonwealth EPBC decision ^[5], alliance award to Acciona and Jacobs ^[5], construction October 2024 or mid-2024 to first water 2028 ^[29][59] — about nine years referral to water.
• Faure: feasibility complete, Council decision on competitive PPP by December 2025 ^[14], in-principle PPP approval December 2025 ^[11][35], qualification mid-2026 or second half 2026 ^[35][19], proposal second half 2027 ^[19], construction start 2028/29 at latest for 70 ML/d by 2031 ^[14], first water 2030/31 ^[19]. City-funded options would delay at least six years on capital constraints ^[14].
• Desalination at Paarden Eiland: still in scoping, pre-feasibility, feasibility and conceptual design with site screening, technical feasibility, finance-procurement and EIA tasks ^[34], EIA still to be undertaken ^[35], procurement from 2026 with first water ~2030 and full operations end-2031 in one timeline ^[53] but pushed to December 2031 programme-wide on procurement, approvals and costs ^[53] and to 2036 at best with five years added for a new site and EIA in independent reporting ^[30].
• Barcelona expansion to 140 hm³/yr is framed as within five years ^[40], but conveyance needs expensive new infrastructure with maintenance and limited life ^[40], and intake/brine permits add time and administrative cost ^[52].

Dependable yield during drought

Desalination, reuse and aquifer schemes offer almost 100% assurance ^[7] with higher opex and energy for desalination ^[7]. Seawater variability is essentially zero and can be operated out of phase with surface water, reuse mirrors demand and sewer inflow, and aquifer storage can be recharged wet months for dry-month use ^[48]. Desalination needs moderate capex, provides very reliable water with no precipitation correlation but energy-intensive relatively high opex ^[48]. On constant-reliability to guarantee 600 m³ in a 95% drought year, average-cost surface water at $1.00 costs $1.49, desalination $1.00 and counter-correlated outdoor conservation $0.86 ^[48]. Portfolio analysis should use the driest month, not average month, which would overstate reliability ^[48].

• Perth operates desalination as baseload reality and climate-independent solution ^[47], generally trimming Southern plant output when dam inflows are high because it is the most expensive desal water to minimise spill probability ^[47]. Both Perth plants remained in operation since built ^[31], unlike Gold Coast, Sydney and Adelaide plants placed in hot standby or standby when reservoirs were near capacity or river flows high ^[18]. Groundwater replenishment lets equivalent groundwater be abstracted later while reducing environmental and other-user impacts ^[27].
• Cape Town groundwater is still rain-fed; over-pumping drops levels, harms feeder-spring wetlands and induces coastal seawater intrusion, so sustainable yield not borehole maximum governs ^[53]. The City plans with 1-in-200 assurance and 50th-percentile climate impact including the New Water Programme ^[11], versus 1-in-50 for existing yield ^[25].
• Barcelona desalination's raw seawater is virtually unlimited and reliable ^[38][1], but large-scale affordable desalination remains limited ^[1] with major market-ready breakthroughs unlikely near-medium term ^[38] and costs possibly rising with energy ^[38].

Environmental constraints

Desalination externality costs driven primarily by salt-disposal ecological damage effectively double desalinated cost in a 2021 meta-analysis ^[1]. Brine with increased density, solids and pretreatment chemicals is a primary concern; thermal reject hotter than seawater causes significant impacts ^[2]. Intake impingement and entrainment injure or kill large organisms plus plankton, eggs and larvae ^[2]; intakes kill fish on screens and small organisms through treatment ^[38]. Inland discharge to rivers, lakes or wells adversely affects receiving waters ^[2]; withdrawing brackish groundwater can harm aquifer sustainability, cause subsidence or induce coastal saltwater intrusion ^[2]. Single-pass RO passes small fractions of sodium, chloride, boron, bromide and organics with boron health risks ^[2]; rising salinity near intakes cuts performance ^[2]. Brine twice as salty as the sea must be modelled to avoid near-outfall damage ^[53].

Mitigations in evidence: nearshore wells minimise impingement/entrainment but raise energy ^[2]; brine via miles-long undersea pipes with diffusers/rotors to near-ambient salinity reduces harm ^[44]; Alkimos brine limited to ~100 m from outfall diluted at least 30 times ^[5], targeting 45:1 dilution via innovative diffuser ^[63], meeting 99% species protection via whole-effluent toxicity testing ^[5], with 2.6 km intake and 4.1 km outfall tunnels protecting reef, seabed and marine life ^[63] and pipelines tunnelled deep beneath seabed avoiding reef ^[5]; Kwinana met all licence and Ministerial Statements 655 and 832 requirements with continuous real-time water-quality monitoring and Cockburn Sound marine plus diffuser monitoring ^[3], though shut twice in early 2008 on low dissolved oxygen ^[65]; Australian studies show minimal impact from returning excess salt ^[31]. Desalination needs environmental impact assessment of air, land and marine effects with mitigation ^[2]; Alkimos review was published on the EPA website with an Infrastructure Sustainability Council design and as-built rating commitment ^[4]. Integrating solar and wind could reduce high-energy impacts ^[2], but large energy use retains impacts until carbon-free supply arrives, still years off in most cases ^[38]. Desalination produces large greenhouse gases from high energy ^[2] and massive electricity raising emissions unless renewable ^[31].

Reuse and demand measures have different footprints. Wastewater reuse is more energy-efficient but needs advanced complex treatment ^[48]; recycling generates sludge and concentrate typically lower salinity than desalination reject ^[57]. Large-scale Barcelona reclamation would cut river flows and kill river life ^[40]; irrigation efficiency can raise total consumption via Jevons paradox as farmers intensify to recoup investment ^[40]. Cutting physical losses cuts embedded treatment and distribution energy, electricity bills and defers new sources ^[13], reduces energy footprint and emissions ^[56], and avoids new-source impacts on waterways, aquifers and ocean ecosystems near desalination plants ^[22]. Intermittent supply from high losses risks ingress of contaminated groundwater or sewerage during interruptions and low pressure ^[13]. Only 23% of major rivers still flow uninterrupted to the ocean ^[1].

Who bears the costs

Costs ultimately fall on users or public budgets, with different bearing arrangements by city and affordability risks for low-income households.

• Perth: Pricing is approved by Cabinet and uniform statewide regardless of supply cost; the Corporation operates commercially, pays a dividend and seeks subsidies where charges do not cover costs ^[22].
  • Funding for Alkimos blends retained dividends — $1.4 billion of forecast 2021–22 dividend ^[29][5] and $766m of forecast 2023–24 dividend including $656m for the plant ^[29] — plus $65m early enabling works ^[29], $597m allocated in 2024–25 ^[29], $1.4 billion in the 2021 Budget toward the next plant ^[5] and $2.8 billion confirmed December 2023 ^[29], with contributions listed as from user charges ^[29] and a minimum 10-year alliance operations contract ^[29][37].
  • Pipe renewals are budget-funded, e.g. $40.4m ^[23], while unbilled-water costs are passed to customers and government ^[22].
  • Desalination's higher per-litre cost versus dams makes system efficiencies more important ^[22].
• Cape Town: The City is expected to develop and finance most non-conventional augmentation while DWS implements Berg–Voëlvlei ^[7].
  • The PPP model has the private sector provide capital, skills and long-term operation while the City retains ownership and regulatory control ^[11], pays only if performance standards are met ^[19] via an agreed unitary payment only when fully operational covering construction and operation with a return for efficiency ^[14], without selling assets, handing long-term control, or letting operators set tariffs or run services independently ^[11][19].
  • The model must meet strict affordability, value-for-money and risk-transfer tests ^[19] and is supported as most suitable for complex Faure and desalination delivery ^[19].
  • It reduces the City's capital burden for long-term affordability ^[11] but extended timelines for Faure and Table Bay ^[11] and risks short-term resilience if demand outpaces plan ^[11].
  • Faure adds ~6–8.5% once-off to tariffs in commissioning unless subsidised ^[14].
  • Systemic affordability pressures include R7.5 billion committed capital in ZAR 2018 ^[7], off-budget Berg–Voëlvlei tariffs delaying municipal off-take agreements ^[42], R24 billion DWS debt at March 2024 delaying maintenance ^[42], agricultural tariffs capped at 50% constraining maintenance budgets ^[42], inadequate national funding for rapid growth ^[11], and construction-mafia 30% demands hindering tenders ^[42].
  • Tariff equity matters: raising rates disproportionately affects low-income households ^[1], South Africa operationalises the constitutional right via a free first block under increasing-block tariffs ^[1], annual water costs reach 24% of income in Uganda and 12% in Rwanda as a warning on burden ^[1], and conservation that lowers bills helps low-income households ^[46].
  • Under BOOT-style purchase agreements fixed capacity payments cover capital with variable payments for operations ^[6], private operators assume financing, construction and operation risks ^[6] while the public authority retains demand and out-of-envelope seawater-quality risks ^[6]; RO sensitivity to feedwater quality ^[6] and 2–5 year operation periods to verify membrane and filter replacement ^[6] with supplier guarantees conditional on pre-treated quality ^[6] allocate performance risk explicitly.
• Barcelona and general: Llobregat cost multiples (€0.70 versus €0.20/m³) ^[39] ultimately fall on users or drought-package subsidies such as the €220m desalination and €224m treatment pledges within Spain's €2.2 billion response ^[39].
  • Greywater recycling is net-saving to Barcelona dwellings (–€90 to –€170/yr by building size) ^[51], so building owners bear capex but gain bill savings.
  • Privatised US systems tend to higher prices with worst affordability in poorer older-infrastructure communities ^[1], and desalination proponents often do not proactively pursue low-income rate shields, small-community aid or consolidation ^[38].
  • Melbourne's parallel shows $10.8 billion total spend over 2009/10–2012/13 including $4.3 billion capex implying near-100% real price rises borne by customers ^[15].

5. Permits, environmental reviews and financing that threaten 2035 delivery

Perth approvals are largely secured, but renewable-power and market-capacity risks remain.

• Alkimos was referred to WA EPA and Commonwealth in 2019 ^[5].
• Required assessment June 2019 ^[29].
• Released review September 2022 to 25 October ^[29][5].
• Received EPA Report 1739 recommendation subject to Appendix A conditions ^[5].
• State decision under Ministerial Statement 1207 ^[5].
• Federal EPBC approval November 2023 after pending status ^[5][29].
• Development approval September by an independent panel ^[5].
• Offset land at least twice impacted environmental value ^[5].
• Aboriginal heritage survey over plant and Wanneroo pipeline with Whadjuk Noongar collaboration ^[5].
• Daylight-only marine piling for mammals despite 24-hour approval ^[5].
• Extended piling hours to meet 2028 ^[5].

Primary deliverability risks are early-stage planning for 100% renewable power with only a power purchase agreement assumed ^[4] and local market capacity to deliver on time ^[4].

Assessment also flags no do-nothing baseline preventing quantitative incremental comparison as required ^[4].

Cape Town faces a lengthy multi-layer approval and off-budget financing chain.

• Faure status:
  • Granted water-use licence under the National Water Act ^[14].
  • City-owned correctly zoned land for both zones ^[14].
  • Advanced peer-reviewed design ^[14].
  • Completed feasibility with December 2025 in-principle PPP decision ^[14][11][35].
  • 2025 public participation with >100 comments per project incorporated ^[19].
• Section 78 process for significantly expanded services ^[34]:
  • Requires investigation of financial, technical, operational and human resources and whether the City can internally construct, manage, operate and maintain the scheme ^[34].
  • In-principle internal versus external decision ^[34].
  • Then a compliant feasibility and business case with value-for-money modelling, extensive risk assessment, community and labour views and MFMA Chapter 11 Part 1 and Section 33 procurement compliance ^[34].
  • Plus an Independent Advisory Panel for transparent potable reuse review ^[34].
• Desalination still needs site selection, detailed technical feasibility, finance-procurement optimisation and environmental and heritage assessment ^[34][35].
• Financing depends on Trans-Caledon Tunnel Authority commercial loans requiring regulatory approvals and development-finance agreements ^[42]:
  • TCTA could not issue Berg–Voëlvlei or other construction tenders until loans were secured ^[42], with long-term funding procurement delaying Berg–Voëlvlei ^[42] and affordability of off-budget tariffs delaying municipal water-supply agreements ^[42].
• Municipal affordability concerns over funding and cost-recovery caused three-year delays on comparable uMkhomazi Phase 1 alongside environmental authorisation and EIA challenges ^[42], a template for Cape Town slippage.
• Only 39% of scheduled DWS maintenance was achieved versus 70% planned in 2023–24 ^[42].

Brine, energy and conveyance permitting adds time and cost, while greywater lacks overarching law.

• Reverse-osmosis brine and conveyance infrastructure with maintenance and limited life plus marine-flora harm ^[40] require intake and brine-discharge permits adding time and cost ^[52], with controlled sea discharge at depth and dilution as the common coastal approach subject to local rules ^[52][61].
• Greywater faces an absence of Catalan or Spanish law, with only district-level compulsion in Gràcia and 22@ since 2022 and a city-wide byelaw programmed for 2025 ^[51].

Limitations / Open Questions

Evidence limits prevent direct ranking and leave schedules, gaps and dependable yield uncertain.

• Incompatible cost boundaries prevent ranking:
  • Cape Town ZAR 2018 capex/opex ^[7].
  • Perth $1.00/kL excluding renewals/membranes ^[4] versus $1.4 billion plant versus $2.8 billion program ^[37][4].
  • Barcelona €0.70 versus €0.20/m³ without stated price year ^[39].
  • California 2015, 2019 and 2020 dollars ^[1][28][48] cannot be directly compared without inflation, exchange, energy-price, contingency and scope normalisation.
  • The report compares only within compatible sets and flags conflicts.
• Conflicting schedules and scopes are unresolved:
  • Faure at R2.83 billion ^[11] versus R3.2–3.3 billion ^[35][34].
  • Desalination at R2.38 billion ^[11] versus R5–5.5 billion ^[53][35].
  • Berg–Voëlvlei completion 2026/7 ^[34] versus 2028 ^[11] versus 2029/30 ^[35] versus 2030 ^[30].
  • Paarden Eiland first water 2030/31 ^[19] versus 2031/32 under review ^[35] versus 2036 at best after site loss ^[30] reflect different vintage, scope and optimism assumptions.
  • Alkimos 45 ^[4] versus 50 GL/yr ^[3] and GRS 38 ML/d ^[34] versus 14+14 GL/yr ^[27] are reconcilable as daily versus annual and Stage 1 versus expanded reporting, but should be confirmed against operating licences.
• Evidence is thin on Barcelona leakage, Cape Town desalination energy, and 2035 demand under combined climate-population scenarios:
  • No Barcelona network loss percentage, Faure or Paarden Eiland kWh/m³, or Barcelona 2035 gap volume is given.
  • Perth gap statements converge ^[5][17][63] but assume a 15% per-person cut by 2030 with no further change after 2030 ^[43] that may not hold if garden and industrial demands rebound.
  • Cape Town sustainable groundwater yields ^[53][35] and invasive-clearing savings at 30 ML/d ^[35] depend on rainfall, monitoring and continued Greater Cape Town Water Fund allocations ^[34][35].
• Drought-dependable yield assumes operations and water quality:
  • Almost-100% assurance for desalination, reuse and aquifers ^[7] and zero seawater variability ^[48] assume power reliability, feedwater within design envelope ^[6], membrane pretreatment performance ^[6], and for reuse, sustained wastewater flows and 20–25% blend acceptance ^[14][53].
  • Emergency fallbacks — temporary desalination, reuse and trucking — carry severe unacceptable economic implications if another 2015–2017 drought arrives before Faure ^[14], while trucked or bottled water at over $1 per gallon far exceeds even expensive desalinated seawater ^[38].

Sources

^[1] Water: Economics and Policy — https://www.bu.edu/eci/files/2025/07/Water-Module-2025-final.pdf · academic ^[2] Desalination Systems and Their Environmental Impacts — https://ask.ifas.ufl.edu/publication/SS746 · academic ^[3] Perth Seawater Desalination Plant — https://www.watercorporation.com.au/our-water/desalination/perth-seawater-desalination-plant · government ^[4] Major Infrastructure Proposal Assessment — https://preprod-iwa-public-files.s3.ap-southeast-2.amazonaws.com/public/2024-03/New%20Water%20Sources%20for%20Integrated%20Water%20Supply%20Scheme%20-%20Water%20Corporation%20-%20MIPA%20Summary%20Assessment%20Report.pdf · government ^[5] Alkimos Seawater Desalination Plant Project - Water Corporation — https://www.watercorporation.com.au/outages-and-works/ongoing-works/alkimos-seawater-desalination-plant · government ^[6] Desalination — https://blue-economy-observatory.ec.europa.eu/eu-blue-economy-sectors/desalination_en · government ^[7] Western Cape Water Supply System — https://documents1.worldbank.org/curated/en/099100002272330999/pdf/P17148306acd480fc0bfb504b0df294bfe8.pdf · government ^[8] Water Corporation reveals alarming number of Perth homes losing precious water each day — https://www.watercorporation.com.au/about-us/media-releases/2025/march/water-corporation-household-leaks · government ^[9] Integrated water supply scheme - Water Corporation — https://www.watercorporation.com.au/our-water/perths-water-supply/integrated-water-supply-scheme · government ^[11] Mayor's Infrastructure Report, 2025 — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/City%20research%20reports%20and%20review/CCT_Mayors_Infrastructure_Report_2025.pdf · government ^[13] The Issues and Challenges of Reducing Non-Revenue Water — https://www.adb.org/sites/default/files/publication/27473/reducing-nonrevenue-water.pdf · government ^[14] Faure-New-Water-Scheme-Feasibility-Study-Report- ... — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/Procedures,%20guidelines%20and%20regulations/Faure-New-Water-Scheme-Feasibility-Study-Report-Summary-Handbook.pdf · government ^[15] Yarra Valley Water Expenditure Review March 2009 — https://www.esc.vic.gov.au/sites/default/files/documents/f5725d43-1428-4399-8ab0-0989a76d8558.pdf · government ^[16] The Challenge of Reducing Non-Revenue Water (NRW) in ... — https://documents1.worldbank.org/curated/en/385761468330326484/pdf/394050Reducing1e0water0WSS81PUBLIC1.pdf · government ^[17] Operational report — https://pw-cdn.watercorporation.com.au/-/media/watercorp/documents/about-us/our-performance/annual-reports/2025-annual-report/annual-report-2025---operational-report.pdf?rev=837cd44abae04ed0a5a2641a48b20760&hash=052DA563B3AF3A4B83121D52C68E70F6 · government ^[18] National Centre of Excellence in Desalination — https://www.twdb.texas.gov/innovativewater/events/Roundtable/doc/DR%20Furukawa.pdf · government ^[19] Council gives in principle approval for Public Private Partnership to secure Cape Town’s new water projects — https://www.capetown.gov.za/Media-and-news/Council%20gives%20in%20principle%20approval%20for%20Public%20Private%20Partnership%20to%20secure%20Cape%20Town's%20new%20water%20projects · government ^[21] Water conservation report 2023–24 — https://www.sydneywater.com.au/content/dam/sydneywater/documents/water-conservation-report-2023-24.pdf · government ^[22] Water Corporation: Management of Water Pipes – Follow-Up - Office of the Auditor General — https://audit.wa.gov.au/reports-and-publications/reports/water-corporation-management-of-water-pipes-follow-up/ · government ^[23] Boosting reliability with $40 million injection into Perth water pipes — https://www.watercorporation.com.au/About-us/Media-releases/2024/September-2024/Boosting-reliability-with-$40-million-injection-into-Perth-water-pipes · government ^[25] western cape water supply system reconciliation strategy ... — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/City%20strategies,%20plans%20and%20frameworks/Water_Sanitation_Reconciliation_Strategy^[1].pdf · government ^[26] Water conservation in Greater Sydney — https://www.audit.nsw.gov.au/our-work/reports/water-conservation-in-greater-sydney · government ^[27] $320m investment doubles Perth’s rainfall-independent water source — https://www.watercorporation.com.au/about-us/media-releases/2022/august-2022/stage-two-groundwater-replenishment-scheme · government ^[28] Draft guidance for economic model (version 3.0) - water loss — https://www.waterboards.ca.gov/water_issues/programs/conservation_portal//docs/waterlosscontrol/2020/swrcb_waterlossmodel_guidance_1dec2020.pdf · government ^[29] Alkimos Seawater Desalination Plant - Infrastructure Pipeline — https://infrastructurepipeline.org/project/alkimos-seawater-desalination-plant · professional ^[30] Cape Town desalination plans delayed while water demand soars — https://groundup.org.za/article/cape-town-desalination-plans-delayed-as-water-demand-soars/ · professional ^[31] Cities turn to desalination for water security, but at what cost? — https://theconversation.com/cities-turn-to-desalination-for-water-security-but-at-what-cost-110972 · professional ^[33] Smart water metering technology for water management in urban areas — https://www.awa.asn.au/resources/latest-news/technology/innovation/smart-water-metering-technology-for-water-management-in-urban-areas · professional ^[34] Faure New Water Scheme Project Overview — https://swan-forum.com/wp-content/uploads/2024/03/Faure-New-Water-Scheme-Project-Overview_Swan-Conference.pdf · professional ^[35] Overview of the City of Cape Town's Water Strategy Director — https://swan-forum.com/wp-content/uploads/2026/03/Water-Strategy_-Presentation-to-Swan.pdf · professional ^[36] Reducing leakage to save bulk water costs — https://www.awa.asn.au/resources/latest-news/business/assets-and-operations/reducing-leakage-to-save-bulk-water-costs · professional ^[37] 21-05-2026 Alkimos Desalination Plant - Project Overview from Planning to Construction Stages — https://www.icheme.org/knowledge-networks/communities/special-interest-groups/water/events/21-05-2026-alkimos-desalination-plant-project-overview-from-planning-to-construction-stages/ · professional ^[38] Desalinating seawater sounds easy, but there are cheaper and more sustainable ways to meet people’s water needs — https://theconversation.com/desalinating-seawater-sounds-easy-but-there-are-cheaper-and-more-sustainable-ways-to-meet-peoples-water-needs-184919 · professional ^[39] Barcelona bets on water desalination to mitigate droughts — https://wodnesprawy.pl/en/barcelona-bets-on-water-desalination-to-mitigate-drought/ · professional ^[40] How to tackle the current water shortage, and the one on the horizon — https://www.creaf.cat/en/articles/how-tackle-current-water-shortage-and-one-horizon · professional ^[41] Beenyup Advanced Water Recycling Plant - Technical and Operational Reviews — https://www.stantec.com/en/projects/australia-projects/b/beenyup-advanced-water-recycling-plant-technical-operational-reviews · professional ^[42] Question to the Minister of Water and Sanitation - NW1243 | PMG — https://pmg.org.za/committee-question/26618/ · professional ^[43] Modelling Western Australia's Water Demand and Supplies — https://info.awa.asn.au/hubfs/Water%20e-journal/202002_005_001_004_Modelling%20Western%20Australia%E2%80%99s%20water%20demand%20and%20supplies.pdf · professional ^[44] About Desalination — https://lwvc.org/desalination/ · professional ^[46] Analysis: State Water Board’s Proposal to Water Down Urban Water Conservation Regulation Leaves California at Risk of Shortages — https://cleanwater.org/releases/analysis-state-water-boards-proposal-water-down-urban-water-conservation-regulation-leaves · professional ^[47] New Developments in Desalination — https://www.awa.asn.au/hubfs/Desalination%20Specialist%20Network/Desalination%20Fact%20Sheet_New%20Developments%20in%20Desalination.pdf · professional ^[48] NANCIAL AND ECONOMIC APPRAISAL OF WATER SUPPLY DESALINATION ... — https://cri-world.com/publications/qed_dp_4614.pdf · professional ^[49] Faure New Water Scheme Identified Water Reuse in Cape Town — https://www.water-reuse-europe.org/faure-new-water-scheme-identified-water-reuse-in-cape-town/ · professional ^[50] Smart Metering for Water Efficiency — https://www.gihub.org/infrastructure-technology-use-cases/case-studies/smart-metering-for-water-efficiency/ · professional ^[51] greywater — https://www.aquapublica.eu/sites/default/files/event/file/2024-06/Maria%20Jos%C3%A9%20Chesa%20PPT.pdf · professional ^[52] How expensive is it to desalinate ocean water? - Elemental Water Makers — https://www.elementalwatermakers.com/knowledge-base/desalination/how-expensive-is-it-to-desalinate-ocean-water/ · general ^[53] Cape Town desalination, explained — https://capetowndamlevels.co.za/articles/cape-town-desalination/ · general ^[55] How much energy does desalinisation use? Is it “absurdly cheap”? — https://hannahritchie.substack.com/p/how-much-energy-does-desalinisation · general ^[56] REDUCING NON-REVENUE WATER: — https://www.suez.com/-/media/suez-global/files/publication-docs/pdf-english/building-sustainable-cities-by-taking-action-against-water-losses-suez-en.pdf · general ^[57] Desalination vs Water Recycling: Which Is More Sustainable for Our Future? - Molewater — https://www.molewater.com/desalination-vs-water-recycling-which-is-more-sustainable-for-our-future · general ^[58] Southern Seawater Desalination Plant — https://en.wikipedia.org/wiki/Southern_Seawater_Desalination_Plant · general ^[59] The Alkimos Seawater Desalination Plant (ASDP) is a landmark water infrastructure project located in Alkimos, Perth, Western Australia. — https://www.acciona.us/projects/alkimos-seawater-desalination-plant · general ^[60] The Untapped Business Case for AMI in Australia’s Water Utilities — https://na.itron.com/w/the-untapped-business-case-for-ami-in-australia-s-water-utilities · general ^[61] Seawater Desalination Plant Cost: How Much Does It Cost per Cubic Meter? — https://www.newater.com/seawater-desalination-plant-cost-per-cubic-meter/ · general ^[62] How much energy does desalination use? - Elemental Water Makers — https://www.elementalwatermakers.com/knowledge-base/desalination/how-much-energy-does-desalination-use/ · general ^[63] Turning seawater into a reliable water supply for Western Australia’s drying climate — https://www.jacobs.com/node/4691 · general ^[64] THE BEENYUP ADVANCED WATER RECYCLING PLANT — https://membranes.com/wp-content/uploads/Documents/Technical-Papers/Application/Waste/IDA_WC2022-Keith-Andes-Beenyup-AWRP.pdf · general ^[65] Perth Seawater Desalination Plant — https://en.wikipedia.org/wiki/Perth_Seawater_Desalination_Plant · general

Source quality: 2 academic, 22 government, 21 professional, 13 general.

Verification

• Cape Town has zero support in evidence cards: no support for 576 Mm3/yr allocation, yield 475-556 Mm3/yr, demand 570-1025 ML/d, 300 ML/d New Water target, Table Mountain 25 ML/d, Phase-1 80 ML/d, Berg-Voelvllei 40 ML/d and dates 2026-2030, Faure 70-100 ML/d costs R2.83-3.3bn and 2030/31 first water, Paarden Eiland 50-70 ML/d costs R2.38-5.5bn and 2030-2036 dates, pressure/leakage 30 ML/d, or PPP/Section 78/EIA financing chain.
• Barcelona has zero support in evidence cards: no support for metro use 293 hm3 in 2019 to 329-354 hm3, 2021 mix 63% surface/34% groundwater/3% desal, 2023 desal one-third, Llobregat 60 hm3/yr + Tordera 20 hm3/yr, plan to 140 hm3/yr with Foix 20 hm3/yr, Euro 0.70/m3 vs Euro 0.20/m3, greywater 0.3 Mm3/yr and -Euro 90 to -170/dwelling/yr, or intake/brine and byelaw constraints.
• Perth 2035 gap and drivers unsupported: no evidence for 337 GL 2024-25, 15 dams/six groundwater plants, 10% surface/40% groundwater/~50% desal shares, rainfall -20% since 1970, runoff -70%, 1000 GL aquifer loss, population 2.12m to 2.9m/3.5m, demand 2600 GL by 2030, or gaps 75-125 GL/yr ^[63], >110 GL ^[17], up to 125 GL/yr ^[5] and 15% per-person cut assumption.
• Leakage and demand management savings unsupported: no evidence for statewide 51.6 GL/13.8% unbilled, metro 30 GL/11.8% and 20.3 GL real losses, 6.5 GL recoverable for $2.5m/yr net benefit, 19 GL/yr from 273,000 leaky homes, Waterwise >157 GL saved, Cape Town 23.3%/186 ML/d and 797 to 539 ML/d, 171 km mains replaced, 1200 to 555 ML/d drought fall, or elasticity -0.40 and smart-meter 3-53.4% effects.
• Delivered-water cost comparisons unsupported and incompatible: no evidence for Llobregat Euro 0.70 vs 0.20, Alkimos $1.00/kL excl. renewals/membranes, $2.8bn program vs $1.4bn plant, Cape Town ZAR 2018 R5/kL reuse vs R9/kL desal and R2.83-5.5bn capitals, California $2.00 desal/$1.50 reuse/$1.00 brackish/$0.50 stormwater, or leakage $215-500/m3/d and conservation $600-1800/AF benchmarks with price-year/boundary normalisation.
• Energy use per cubic metre unsupported: no evidence for seawater RO 2.5-3.5 kWh/m3 with recovery and 4-6 without, reuse 0.6-1.5 kWh/m3, Perth replenishment ~one-quarter of seawater, Llobregat ~4 kWh/m3, Alkimos renewables/Emu Downs 180 GWh/yr vs 270 GWh/yr, or Paarden Eiland coal-grid footprint; solar-powered 10-15 year TCO evidence card is unused in draft.
• Construction and permitting lead times unsupported: no evidence beyond 2028 Alkimos start for referral 2019, EPA Report 1739/Ministerial Statement 1207/EPBC Nov 2023, construction Oct 2024, Faure PPP decision Dec 2025/RFQ 2026/RFP 2027/construction 2028-29, desal EIA still to be done plus 5-year site-change delay, or Barcelona 5-year to 140 hm3/yr.
• Dependable yield during drought unsupported: no evidence for almost-100% assurance ^[7], zero seawater variability, baseload operation since built vs standby elsewhere, sustainable-yield limits from wetlands/seawater intrusion, 1-in-50 vs 1-in-200 assurance, or constant-reliability $1.49 surface vs $1.00 desal vs $0.86 conservation.
• Environmental constraints and mitigations unsupported beyond Kwinana general compliance: no evidence for brine doubling cost, 45:1 dilution/30x/100 m limit, 2.6 km intake/4.1 km outfall, 99% species protection, Cockburn Sound/diffuser specifics beyond general licence compliance and monitoring, river-flow kill from full Barcelona recycling, Jevons paradox, or NRW energy/emissions deferral.
• Who bears costs unsupported: no evidence for Perth Cabinet uniform pricing/dividends $1.4bn/$766m/user-charges/$40.4m renewals, Cape Town PPP unitary-payment-only-if-operational/6-8.5% tariff uplift/R7.5bn ZAR 2018/TCTA loans/R24bn DWS debt, Barcelona Euro 220m/224m drought pledges, or Melbourne $10.8bn/near-100% price rise analogy and low-income burden claims.