Closing Urban Water Gaps by 2035

Executive Summary

• The credible 2035 portfolios are different by city, but share the same logic: lock in low-regret leakage and efficiency first, then add one large climate-independent anchor. Perth has the firmest path because Alkimos Stage 1 at about 55 GL/year — 50 GL/year desalination plus 4.9 GL/year Eglinton groundwater — is approved, contracted and scheduled for first water in 2028 ^[6][9]. Cape Town can credibly reach about 305 Ml/d peak and 101 Mm³/year by December 2031 only if Faure reuse Phase 1 at 70 Ml/d and desalination Phase 1 at 70 Ml/d both survive PPP procurement, EIA and construction starting by 2028/2029 ^[32][24][61]. Barcelona already operated in 2023-2024 with 33-58% of supply from desalination or reuse and reservoirs below 16%, so its gap is met by running existing 80 Hm³/year desalination hard plus the Tordera extension and a new 20 Hm³/year Foix plant to reach 140 Hm³/year, alongside Besòs regeneration and greywater rules ^{[49][33][41][36]}.
• Costs and energy cannot be ranked on a single €/R/$ per m³ league table because boundaries and price years differ. Cape Town committed-programme estimates in 2018 Rand put operating cost at about R5/kl for Water Reuse P1, R9/kl for Desalination Phase 1 and R3-5/kl for Berg augmentation ^[19], while 2025-Rand Unit Reference Values had risen to R30/m³ for Faure, R50/m³ for desalination and R7.2/m³ for Berg-Voëlvlei ^[32]. Perth reports Alkimos Stage 1 at $2.8 billion capital with $51.9 million/year operating cost at a $1.00/kL benchmark that excludes major renewals and membrane replacement and may understate lifecycle cost ^[6]. Barcelona reports €250 million for Tordera extension plus hundreds of millions for Foix ^[49], a €467 million credit for desalination expansion ^[41], and a €2.4 billion Catalan Water Agency programme to 2027 ^[33].
• Energy intensity consistently favours reuse over seawater desalination, but whole-system boundaries matter. Average seawater desalination is reported at around 4 kWh/m³, down from 8 kWh/m³ in 1970s plants to 2.3 kWh/m³ in optimised systems ^[49], with modern reverse osmosis at 2.5-3.5 kWh/m³ ^[56]. Cape Town planning assumes about 2 kWh per thousand litres for reuse versus 3.5-4 kWh for desalination ^[23][40], and reclaimed water is estimated to use one-third the energy of desalination ^[41]. Generic comparisons that include conveyance, treatment and distribution widen ranges substantially and should not be mixed with fenceline membrane numbers ^[47][55].
• Lead time is the binding 2035 constraint, not technology. Alkimos shows the full cycle: engagement 2017-2019, referral 2019, EPA assessment May 2023, Ministerial Statement August 2023, alliance award April 2024 and completion 2028 ^[9]. Cape Town notes standard South African owner design-build takes four to five years versus two to three years internationally for reuse and desalination ^[23][40], which is why Faure must start construction in 2028/2029 at the latest for 70 Ml/d by 2031 and would slip at least six years if it waits for City capital budget ^[24]. Desalination permitting generically needs all local, state and federal permits before construction ^[11], typically three months to two years but longer near sensitive habitats or with multiple agencies ^[58], illustrated by a Texas brackish plant that ran from planning in 2001 to completion in 2007 ^[11].
• Recommendation: fund pressure management, meter replacement and targeted efficiency now; procure reuse before desalination where potable standards and river-flow rules allow; reserve desalination as drought-dependable baseload; and make tariffs, taxes and developer charges explicit. Desalination, reuse and aquifers are all described as offering almost 100% assurance of supply versus climate-vulnerable surface storage ^[19], but aquifers carry uncertain climate resilience and environmental impacts ^[19], desalination has higher operating costs and energy needs ^[19], and reuse reduces flows to rivers and wetlands ^[40]. Robust-planning evidence warns that central-projection optima carry up to 12% maximum regret, that explicit robustness cuts regret by 23% at 8% expected-NPV cost ^[1], and that softer reversible measures are generally preferable but large supply beats them in about 15% of less-favourable scenarios ^[3].

1. Supply-demand gaps and what capacity is operating, funded or proposed

1.1 Barcelona and the Ter-Llobregat system: already in manufactured-water operation

Barcelona's bulk system is managed through ATLL, which distributes bulk water to suppliers in the Metropolitan Region, while AGBAR distributes to domestic customers in Barcelona and the metropolitan area ^[4]. Reported physical efficiency is high at 97% for ATLL and 93% for AGBAR, with 98% domestic metering, but that excludes about 10% apparent losses from faulty meters or fraud ^[4]. This leaves limited easy leakage headroom compared with high-loss utilities.

Demand is already low by European standards and still growing in aggregate. Dense-city use is reported around 130 litres per person per day versus over 200 and cases over 500 in suburbs ^[4], with 2004-2009 average at 117.42 falling to 108.44 in 2009 ^[4]. More recent accounts give Barcelona total drinking-water consumption of 88.04 Hm³ in 2021, down 26.4 Hm³ since 1999, split 72% domestic, 22% industry and commerce and 6% municipal services ^[33]. Average use is put at approximately 106 litres per person per day ^[33] and domestic drinking-water at 98.7 litres per equivalent inhabitant per day ^[35], with Sarrià-Sant Gervasi at 128 versus Nou Barris at 92 ^[33]. The City target is below 100 litres per person per day by 2030 ^[33]. Population in the consumption analysis rose from 1,503,451 to 1,660,122, up 9.44% ^[35]. For 1.62 million city inhabitants demand is forecast up 4% by 2050 under a committed climate-action scenario ^[35]. For 5.3 million metropolitan inhabitants, total consumption from all sources was 293 Hm³ in 2019, estimated at 329 Hm³ for 2050 including plans in progress and 354 Hm³ long-term including full urban development ^[35].

The drought gap is stark. In early 2024 Ter and Llobregat reserves fell below 100 Hm³, just 16% of capacity, triggering drought emergency ^[30]. Reservoirs on the Ter and Llobregat supplying about 6 million people in northern Catalonia were less than 16% full, a record low, after nearly 40 months of drought described as the worst since monitoring began ^[36].

The 2023 metropolitan strategic plan foresees 12% less surface water and 9% less groundwater by 2050 ^[33]. Climate framing points to Mediterranean warming 20% faster than global average and rainfall down 20-50% by 2100 ^[49].

Table 1 — 2035-relevant capacity status. Units as reported; do not convert across Hm³, GL and Ml/d without stated boundaries.

Barcelona's operating mix explains why desalination energy now drives tariffs.

• In 2021 before the drought desalination was just 3% of Barcelona drinking water, but by 2023 it was 33% according to Aigües de Barcelona ^[41].
• At the drought depth in August 2022 plants ran at 100% ^[41] and by September 2025, six months after the declared end, still over 80% ^[41]. Normal intent had been about 20% to protect reservoirs and reduce rainfall dependence ^[41], with elevated operation planned until other plants operate in 2029 ^[41].
• One account says desalination and reclamation supplied 55% of total regional consumption, allowing avoidance of stricter restrictions ^[36], while another says 58% of Barcelona consumption is now desalination or reuse versus 42% rivers and wells, with the latter to be reduced ^[33].
• These figures differ in year and boundary — city versus region, drinking water versus total consumption — so they should not be averaged.

Cape Town's gap must be read against over-allocation and shifting assurance. Total current allocation from the Western Cape Water Supply System is 576 Mm³/year, split Cape Town 347, other urban and industry 43, and agriculture 186 ^[19]. That supply is a small share of 3,037 Mm³/year Western Cape use, three-quarters agricultural ^[19]. Another assessment puts allocation at ~590 Mm³/year against 2018 revised yield of 547 Mm³/year even before ecological reserve, with average combined demand of 538 Mm³ in 2009/10-2018/19 ^[54]. Updating hydrology for drought and aliens cut modelled yield by 33-35 Mm³/year ^[19]. National planning used 98% assurance with restrictions only beyond a 1-in-50-year drought ^[23], but Cape Town now retains 1-in-200-year assurance while moving from 95th-percentile climate impact with five-years-early delivery to 50th-percentile impact with schemes online as required by end-2031 ^[32]. Under the tougher assumptions the low demand curve sits just above supply with two potential shortfall periods if demand exceeds it; under 50th-percentile assumptions the system is in approximate equilibrium tracking the low curve ^[32]. Bulk capacity must still rise 94-185 MLD every five years to meet peak-week demands to 2050 ^[32]. Demand rebound ran to 2023/24 then inflected to 2-2.5% growth consistent with the 2019 Water Strategy and 2025 Bulk Masterplan and attributed to population growth ^[32], with per capita settled at ~160 litres per capita per day versus 200-225 before drought, implying embedded outdoor efficiencies ^[32].

Perth's gap is driven less by a single drought than by a permanent groundwater cut plus growth.

• The Water Corporation must cut Gnangara abstraction by 27%, or 30 GL/year, with reduced entitlements from 1 July 2028 ^[6], while other industry cuts 10%, or 10.2 GL/year, and garden bores face aligned restrictions ^[6]. Gnangara supplies almost half of all Perth water including scheme and direct extraction ^[6].
• Master planning sees multiple new climate-independent sources needed to 2050 from population, climate and reduced groundwater reliance ^[6].
• Estimates of the residual need even if reduction and recycling targets are met cluster at 75-125 billion litres/year by 2035 ^[10], up to 125 billion litres/year ^[9], and more than 110 billion litres for metro alone into the 2030s ^[29].
• An audit statement that desalination and aquifer recycling have ensured reliable supply to at least 2035 ^[28] is therefore best read as conditional on Alkimos and efficiency being delivered, not as absence of a gap.

2. Delivered-water cost and energy intensity on compatible boundaries

Directly ranking Barcelona euros, Cape Town Rand and Perth dollars per cubic metre would violate the requirement to compare only compatible boundaries and price years. The evidence instead supports within-programme comparisons and careful cross-reading.

2.1 What the city programmes themselves report

For Cape Town's committed programme in 2018 Rand, provisional capital totals R6,746 million including demand management and R6,336 million for new supply, also narrated as R7.5 billion for 128 Mm³/year ^[19].

• Within that same table, Desalination Phase 1 capital is R1,650 million with operating cost R9/kl, Water Reuse P1 is R1,360 million with R5/kl, and Berg augmentation operating cost is ~R3-5/kl ^[19].
• An earlier GreenCape tabulation lists Faure reuse at R1,882 million for 70 MLD at R5.7/kl and permanent desalination at R1,800 million for 50 MLD at R9.0/kl, with Groenlandberg groundwater lowest at R2.2/kl and Berg-Voëlvlei at R4.62/kl ^[54].
• Those two vintages should not be blended because capacities, scopes and price years differ — for example Paarden Eiland is later costed at roughly R5 billion in 2023 prices for 70 Ml/d ^[61], versus R1,800 million for 50 MLD in the earlier table ^[54].

By 2025 the same Cape Town schemes look markedly more expensive in 2025 Rand Unit Reference Values: Berg-Voëlvlei from R5/m³ in 2019 to R7.2/m³ by 2028, desalination Phase 1 from R24/m³ in 2019 to R50/m³ in 2025, and Faure from R12/m³ in 2019 to R30/m³ by 2025 ^[32]. Current bulk storage, purchase, purification and conveyance is R7.38/m³ excluding distribution, while New Water Programme water due by 2031 is estimated at about R18 in 2025 Rand despite 75% reliance on cheap surface water ^[32]. Faure alone is calculated to cause a once-off 6-8.5% tariff increase in its commissioning year ^[24]. The direction is consistent — reuse cheaper than desalination, surface cheapest — but the levels have moved with procurement, energy and financing assumptions.

Perth reports Alkimos Stage 1 capital at $2.8 billion across 34 scope elements plus program and risk ^[6], against $1.4 billion included in the 2021 State Budget as down payment ^[9][31]. Annual operating cost is estimated at $51.9 million at a $1.00/kL benchmark from existing operations, explicitly excluding major renewals and membrane replacement which may understate lifecycle cost ^[6]. The assessment notes manufactured water is capital-intensive with substantially higher operating costs than historical sources, raising ongoing costs to Government ^[6]. No Do Minimum base case was developed, so no quantitative incremental comparison of shortlist options as required is possible ^[6].

Barcelona cost evidence is tariff- and budget-led rather than scheme-levelized.

• Average domestic price is 1.97 €/m³ without VAT, split 1.09 supply and 0.55 sanitation in one vintage and 0.88 sanitation in another, reflecting different years ^[39][4].
• For 175 m³ household use, Barcelona at €520/year is more than triple Guadalajara at €164, with variation attributed mainly to lack of structured cost rules and political influence rather than pure physical cost ^[39].
• ATL approved a 33% increase after a freeze since 2017 to cover energy and drought cycle costs and guarantee supply ^[39], and end-2023 prices rose more than 33% partly from running desalination ^[41].
• Spanish tariffs are 45% below European average and, per the 2022 AEAS-AGA study, do not adequately cover service costs, sometimes covering only operations without depreciation, renovation and improvements ^[39].

Table 2 — Cost and energy signals. Keep rows within their stated scope; cross-row arithmetic is not valid.

Delivered-water cost and energy signals use different boundaries and years, so rows are not directly comparable.

• Cape Town surface / Berg augmentation: Opex ~R3-5/kl in 2018 Rand ^[19]; URV R5/m³ in 2019 to R7.2/m³ by 2028 ^[32]; average surface ~R5 per thousand litres, dam water R5.20/kl in older planning ^[23]. Energy: lowest among new supplies; surface storage is backbone but vulnerable to variability ^[19].
• Cape Town reuse: Opex R5/kl in 2018 Rand ^[19] and R5.7/kl in GreenCape table ^[54]; URV R12/m³ in 2019 to R30/m³ by 2025 ^[32]; potable reuse ~R7.50 per thousand litres, just over half efficient desalination ^[40]. Energy: ~2 kWh per thousand litres versus 3.5-4 for desalination ^[23][40].
• Cape Town desalination: Opex R9/kl in 2018 Rand ^[19] and GreenCape R9.0/kl ^[54]; URV R24/m³ in 2019 to R50/m³ in 2025 ^[32]; efficient BOT <$1 (R12) per thousand litres versus $2-3 (R24-36) owner-build ^[23]; 150 MLD inefficiency gap R0.65-1.3 billion/year or R6.5-13 billion over ten years ^[40]. Energy: 3.5-4 kWh per thousand litres ^[23][40]; 70 Ml/d plant described as small-town power load on coal-leaning grid with brine twice seawater salinity ^[61].
• Cape Town groundwater: Groenlandberg R2.2/kl operating cost in GreenCape table ^[54]; Cape Flats R2.6 billion for 50-60 Ml/d and R4.7 billion for ~105 Ml/d by 2040 ^[7]. Timing and energy: groundwater fastest then reuse then desalination by complexity ^[23]; pumping energy small versus RO but recharge-dependent ^[40].
• Perth Alkimos: $2.8 billion capital, $51.9 million/year at $1.00/kL excluding renewals/membranes ^[6]; $1.4 billion budgeted down payment ^[9]. Energy: desalination energy-intensive, hence net-zero-by-2035 with up to 400 MW additional wind within $3.8 billion renewable investment ^[31]; up to 1.3 TWh/year wind procurement in progress ^[29].
• Barcelona desalination: Tordera extension €250 million, Foix 20 Hm³/year at hundreds of millions ^[49]; €467 million credit, later $590 million, for new facilities over five years ^[41]. Energy: average ~4 kWh/m³, 8 in 1970s to 2.3 optimised ^[49]; reclaimed estimated at one-third desalination energy ^[41].
• Generic benchmarks: Texas brackish $1.25-2.60 per 1,000 gallons versus seawater $3.60-5.80 as function of capital, debt and opex ^[11]; large seawater $1.53-1.93/m³ versus small $2.10-3.31 median $2.13 ^[48]; small indirect potable reuse median $2,300/AF versus large $1,800/AF; small seawater median $2,800/AF versus large $2,100/AF in 2015 dollars at 6% ^[52]. Energy: SWRO process with recovery 6.8-8.2 kWh/kgal or 2,216-2,672 kWh/AF by membrane ^[47]; brackish 3.0-5.0 kWh/kgal or 980-1,630 kWh/AF; indirect potable reuse total 7.0-11.5 kWh/kgal or 2,280-3,740 kWh/AF; Pacific seawater 10.0-14.0 kWh/kgal or 3,260-4,560 kWh/AF ^[47]; 1 kWh/kgal = 325.8 kWh/AF for boundary conversion ^[47].

Main point: electricity prices dominate desalination economics.

• Cost shares: energy is 28-50% of total seawater facility cost and can approach or exceed capital recovery ^[47], and separately 30-50% of total variable cost ^[48], with a small-system vendor range of 60-70% for conventional desalination ^[59].
• Salinity and temperature: within the fenceline salinity has the most significant impact on power ^[47], with net driving pressure up about 11 psi or 0.76 bar per 1,000 mg/L salinity change ^[47], Middle East needs 15-20% above US baseline and river-influenced embayments at least 15-20% below ^[47], and 1°C feed change shifting membrane throughput 3% ^[47].
• Component splits: membrane at 65-85% excluding conveyance and distribution, nearby intake at 15-20%, ancillary at 10-15% and post-treatment under 2% ^[47].
• Imported-water context: California State Water Project total is 9.8-12.1 kWh/kgal or 3,190-3,940 kWh/AF and Colorado River 6.8-9.5 kWh/kgal or 2,210-3,090 kWh/AF, so a Gulf seawater plant uses about the same power as California imported water even before treatment ^[47]. At Pearblossom intensity reaches 4,444 kWh/AF, roughly equivalent to or higher than ocean desalination estimates ^[55].

Leakage-reduction cost evidence is sparse for the three cities and must be borrowed with caveats. In developing countries physical-leakage reduction is $215-500 per m³/day saved, with ~$250 achievable efficiently ^[25]. Bangkok's three zones averaged $340, split $246 versus $408 and $518 by contractor ^[25]. Selangor Phase 1 targeted $243 and achieved $215, while Phase 2 averaged $528 including maintenance ^[25]. Dublin cost about $750 per m³/day saved ^[25]. Valuing physical losses at $0.20/m³ marginal cost and commercial at $0.25 tariff, payback is 2 to over 30 years across $200-800 unit cost at 10% discount, or about 4-8 years for high-loss cases at $250-400 and $0.20 value ^[25]. Wisconsin, in 2010 dollars at 6.5% to 2030, finds water-loss control at $0.46 per 1,000 gallons, less than half efficiency measures at $1.59-2.19, against a $1.86 per 1,000 gallons alternative-supply benchmark ^[34]. Those boundaries — dollars per flow saved versus dollars per volume saved — cannot be collapsed into one €/m³ without project-level loss duration and discount assumptions.

3. Construction and permitting lead times: what can credibly deliver by 2035

Main point: Perth Alkimos is the only large seawater scheme among the three that has passed approvals and entered major construction, implying roughly a decade from engagement to water with about four years heavy construction.

• Public engagement ran 2017-2019 ^[9].
• Proposal referred in 2019, public comment late 2022, EPA Report 1739 in May 2023 recommending implementation subject to conditions, Ministerial Statement 1207 in August 2023, and State and Commonwealth approval by November 2023 plus September development approval ^[9].
• Enabling earthworks November 2023-January 2025, power upgrades October 2024-late 2025, marine tunnels mid-2025-late 2026, seabed structures December 2025-July 2026, Eglinton bore and pipeline plus Alkimos-Wanneroo pipeline mid-2025-late 2027, and first water in 2028 ^[9].
• Alliance contract to Acciona and Jacobs in April 2024 for design, construction and operation with plant start mid-2024 ^[9].
• Tunnelling finished after 15 months with machines reaching intake in May and outfall in late August 2026, and marine phase closed when barge departed 28 July ^[9].
• Implication: timeline consistent with assessment that market capacity to resource project on time is key risk alongside early-stage renewable planning ^[6].

Main point: Cape Town dates keep moving for governance and finance reasons, not technical infeasibility.

• Committed-table pushes after drought recovery and COVID-19 suppressed demand and finances: reuse P1 2024 to 2027, desalination 2026 to 2030, Berg 2023 to 2024, Cape Flats P2 to 2025 and Atlantis to 2024 ^[19].
• June 2025 outlook pushes further: Berg to June 2028 with DWS now to June 2029, five to six years later than the 2019 Water Strategy; Faure P1 to March/mid-2031; desalination to end-2031 ^[32].
• Earlier vintages had Faure first water July 2025 and desalination December 2026 ^[54], and a March 2024 overview had Faure completion August 2029 with validation September-November 2029 and Berg 2026/27 to be confirmed ^[42].
• Governance: Faure needs a multi-stage Section 78 plus Municipal Finance Management Act procurement with value-for-money modelling, risk assessment and community and labour views ^[42]; it was at Section 78(1) in March 2024 ^[42], reached Section 78(2) approval in March 2024 to study options including PPP with value and due-diligence assessments ^[5], registered as potential PPP in October 2024 with Council PPP-in-principle targeted December 2025 ^[24], then two-stage procurement with RFQ and RFP in second half of 2027 ^[22], and procurement formally begun 25 September 2026 for financing, design, construction, operation and maintenance ^[22].
• Financing contrast: construction must start 2028/2029 at the latest for 70 Ml/d by 2031, while City-funded options cannot be accommodated short-to-medium term and would delay at least six years, whereas PPP has no financing delay because the private party funds from the start ^[24].
• Paarden Eiland parallel path: feasibility studies in September 2024, EIA underway ^[5], technical feasibility described as complete in one account ^[7], but still at scoping/pre-feasibility and feasibility/conceptual design in March 2024 with site screening, finance-procurement method and environmental and heritage assessment still required ^[42].
• Procurement from 2026 to first water ~2030 and full end-2031 ^[61] is therefore tight and exposed to generic risks that incomplete submissions are the most common delay cause and multi-body parallel or sequential processes add time ^[58].

Main point: Barcelona points to a five-year build window layered on a decade of delay.

• Build plans: plans to spend €250 million on Tordera and hundreds of millions on a 20 Hm³/year Foix plant are framed as all in the next five years for 140 Hm³/year total ^[49].
• Delay context: a separate strategy note says La Tordera is due to be extended and a Foix-basin plant built that would already run had it not been for 2009 cuts ^[33]; ATL plans elevated desalination operation until other plants operate in 2029 ^[41], matching that five-year horizon; 2009-2017 investments were not made, leaving policy almost a decade behind despite resumption in 2017 ^[33].
• Permitting and sequencing risk: permitting for desalination in Spain requires EIA before acceptance, but critics say few environmental regulations guide building and operation and a September 2025 study warned full impacts remain unknown ^[41]; generic permitting of three months to two years depending on country, size and agencies ^[58], with construction barred before all permits and violations risking fines, removal or shutdown ^[58]; this suggests Barcelona risk is less greenfield EIA than grid, brine and budget sequencing: operating enough water-making capacity at Barcelona scale would strain local grids to breaking point in one journalistic account ^[41].

Main point: demand and leakage measures are fastest where institutions allow.

• Order and procurement: groundwater is in principle fastest then reuse then desalination ^[23][40], but standard South African procurement stretches moderate-large projects to four to five years versus two to three years for reuse and desalination efficiently procured ^[23][40].
• Emergency precedent: emergency desalination can use short-term licences in days or weeks ^[58], as shown by Cape Town temporary plants reaching full production in early-to-mid 2018 ^[40][23] and a 2.5 ML/day pilot targeted to run by July 2017 at R30 million ^[26].
• Smart metering: can be implemented immediately subject to supply with results as soon as installed ^[45], and studies show 7-22% consumption fall after metering and up to 46% savings through leak-finding in cited cases ^[45].
• Greywater scale: Barcelona greywater at 136,000 m³/year ^[30] and 300,000 m³/year potential at recent development pace ^[35] is small versus Hm³-scale desalination but inherits building-cycle lead times, not water-plant EIA.

Table 3 — Lead-time credibility to 2035.

Lead-time credibility to 2035 by measure.

• Leakage / pressure / meters:
  • Barcelona: High efficiency already; targeted renewal is incremental and fundable within tariff/transfer mix, but Catalonia-wide 25% loss versus <10% best practice ^[41] indicates distribution outside Barcelona needs capital.
  • Cape Town: Pressure savings 55 MLD proven ^[23]; R836.6m pipe replacement ^[7], R170m pressure over five years and R1,261 billion meters at ~R252m/year 2017-2023 vintages ^[26]; Water Conservation/Demand Strategy due late 2025 to limit ten-year growth via loss reduction ^[32].
  • Perth: 6.5 billion litres metro cost-effectively recoverable via pressure and meters of 6.7 recoverable ^[28]; programme since 2016-17 of pressure, detection, meters and reservoir relining described as on track for 10% by 2030 ^[28], but auditor finds no clear plan for 10% nor for ~22 billion litres rest-of-state ^[28].
• Efficiency / greywater / Waterwise:
  • Barcelona: Greywater ordinance 2025 ^[30]; fixtures save 50% taps/showers, 30% dual-flush, 20-40% outdoor irrigation ^[4]; 2008 home actions saved 20% ^[33].
  • Cape Town: Demand cut 1200 to 500 MLD Feb 2015-2018 ^[23]; per capita now ~160 versus 200-225 ^[32]; restricting >20% hits households, >30% believed no longer possible ^[32].
  • Perth: 2023-24 logger savings 2.3 billion litres versus 1.3 target; 287 businesses saved 9.6 billion litres; Waterwise business 157 billion litres since 2007 ^[29]; 35% wastewater recycling target by 2035 ^[29].
• Reuse:
  • Barcelona: Llobregat 2,000 l/s in use; Besòs plus 25 stations planned ^[33]; health/environment legislation bounds scale ^[33].
  • Cape Town: Faure 70-100 Ml/d needs construction start 2028/29 for 2031 ^[24]; Potsdam 47→100 Ml/d by late 2027 expands non-potable reuse ^[21]; Zandvliet temporary 10 MLD precedent ^[40].
  • Perth: Beenyup 28 billion litres/year ^[10]; optimisation +2.7 billion litres/year ^[29]; Kwinana +10 ML/day early works ^[29].
• Desalination:
  • Barcelona: Existing 80 Hm³/year run at 100% Aug 2022 and >80% Sep 2025 ^[41]; Tordera/Foix within five years to 140 ^[49]; full operation of others by 2029 ^[41].
  • Cape Town: Paarden Eiland 50-70 Ml/d first water ~2030 full end-2031 ^[61][32]; EIA/feasibility/PPP still gating ^[5].
  • Perth: Alkimos Stage 1 50 billion litres/year first water 2028, expandable to 100 ^[9][10]; approvals and alliance complete ^[9].

4. Drought-dependable yield and environmental constraints

Dependability ranking is consistent across sources even as numbers move. Desalination, reuse and aquifer schemes are all described as offering almost 100% assurance and able to support the entire Western Cape system ^[19]. Existing and new deep and shallow aquifers buffer variability but their climate resilience and environmental impacts remain uncertain ^[19]. Planned large direct potable reuse and aquifer recharge supplement with high-assurance climate-resilient water ^[19], while desalination has high assurance but higher operating costs and energy ^[19]. Only desalination is described as totally independent of rainfall with otherwise unlimited augmentation except technical and financial limits ^[23][40]. That is why Cape Town's optimal desalination scale is put at 120-150 Ml/d in one plant with guidance not to exceed 200 MLD ^[40][23], and why Tampa Bay's choice of seawater desalination after groundwater withdrawals were cut from 192 to 90 mgd for environmental reasons is cited as a drought-proof alternative ^[47].

Main point: Barcelona drought test shows desalination plus reuse held city supplies but not agriculture and ecosystems.

• Context: desalination at full capacity plus indirect reuse emerged as the new frontier during the most severe meteorological drought in over a century ^[30].
• Compliance: household and Drought Plan compliance kept Barcelona and most of metro within 200 litres per person per day including domestic, urban, economic and commercial uses ^[30], after February 2024 cuts to 200 litres per person per day, agriculture by 80%, animal husbandry by half, and industry and urban by 25%, with car-washing, garden-watering and pool-filling bans and closures of water parks, ice rinks and most pools ^[36].
• Emergency import: emergency import of up to 20,000 m³/day by ship from June with at least one ship per day was envisaged ^[36].
• Limits: alternatives are unlikely to solve agriculture, livestock, forest and river-ecosystem impacts hardest hit by drought ^[30], farmer insurance compensation reached €110 million in 2023 ^[36], and cruise lines covering 90% of Barcelona calls voluntarily stopped taking water there ^[36].

Cape Town's yield arithmetic shows why demand management counts as supply. The committed programme counts demand management at 70 Ml/d or 26 Mm³/year, alien clearing at 55 Ml/d or 20 Mm³/year, and system management at 27 Ml/d or 10 Mm³/year alongside aquifer, Berg, reuse and desalination increments totalling 347 Ml/d or 128 Mm³/year new supply and 417 Ml/d or 154 Mm³/year including demand management ^[19]. Updated outlook counts alien clearing lower at 30 Ml/d peak and 11 Mm³/year by June 2026 ^[32]. Aquifer yields illustrate licensing versus infrastructure: Cape Flats Phase 1 licence at 20 Mm³/year equals about 55 MLD year-round with peak infrastructure ~80 MLD for summer drought, while Table Mountain Group Phase 1 totals 42.2 Mm³/year in one account ^[23]. Planned sustainable year-round yields are put at ~55 MLD Cape Flats Phase 1, ~32 MLD final Atlantis with 12 already in system, and ~50 MLD Table Mountain Group with Steenbras 12 Mm³ equal to 33 MLD ^[23]. The City assumes 25% less surface and groundwater availability over 30 years from climate change, marginally more than in the 2019 strategy ^[42], consistent with modelling that 15% mean-runoff fall cuts system yield ~160 MLD by 2050 within a -2% to -17% runoff range ^[23][40]. At 1-in-200 assurance, 50-100 MLD sufficed ignoring climate, but 200-250 MLD is needed if very low rainfall continues to keep dams above 25% at end of summer, hence the interim 350 MLD risk-averse target ^[40].

Environmental constraints differ by option.

Large surface schemes affect sensitive mountain wilderness with dams, pipelines and pump stations, and aliens substantially cut yield ^[23]. Clearing benefits surface water by preventing long-term losses ^[19]. Current alien impact of 24 Mm³/year could grow to 85 Mm³/year ^[42]. Theewaterskloof alone is estimated to lose 30 million m³/year to aliens ^[57]. Funding depends on engaging DWS on an ongoing mechanism ^[32] plus R125 million City funding 2022-2026 matched privately ^[32] and R25 million in 2023/24 for four catchments ^[7].

Groundwater, if not over-abstracted, has least impact with very low footprint, regulated by licensing, monitoring and recharge from rainfall, stormwater and treated wastewater ^[23]. Siting near dams and sensitive areas constrains Table Mountain Group, with review threatening medium-term yield, while Cape Flats Phase 1 needs 12 Mm³ or 33 MLD recharge rising to 25 Mm³ or 68 MLD in Phase 3 and Table Mountain Group needs no artificial recharge ^[23]. Full groundwater programme to ~120 Ml/d depends on sustainable yield without dropping levels, losing wetland springs or inducing coastal intrusion ^[61].

Private drilling during Day Zero lowered public demand 5.3% in crisis and 4.9% after ^[38]. Levels near above-median parcels fell ~1 m/year ^[38] and 0.89 m deeper than below-median areas ^[38]. Extraction was largely unregulated open-access, exempt from outdoor restrictions, with minimal monitoring and roughly one-in-ten boreholes licensed in wealthy Newlands ^[38]. Aquifer recharge rates are generally high but heavily rainfall dependent ^[38]. Unmanaged pumping creates stock and pumping-cost externalities ^[38].

Reuse constraints are health, flow and energy. Faure advanced treatment of ozone, biologically activated carbon, granular activated carbon, ultrafiltration to 0.01 microns and ultraviolet advanced oxidation with further Faure Water Treatment Plant treatment ^[42] must meet South African and international drinking standards ^[22] under SANS 241 with testing for over 700 contaminants including pathogens, PFAS, pharmaceuticals and disinfection by-products ^[32]. Purified water makes up 25% of the blend with dam water before further treatment ^[24], elsewhere described as roughly 20% reuse to 80% dam water ^[61]. Health risks from wastewater or polluted groundwater are described as manageable via tertiary multi-barrier treatment, with Faure designed to avoid reverse osmosis in one vintage ^[23]. Reducing wastewater flows to rivers and wetlands is an explicit trade-off ^[23][40], and Barcelona analysis warns large-scale reclamation is constrained because recycling all treated water would cut river flows so some rivers stop functioning normally ^[49]. Source water must comply with General Authorisation under the National Water Act ^[42], and Cape Town holds a water use licence for Faure ^[24] on City-owned correctly zoned land for Zones 1 and 2 with peer-reviewed advanced design ^[24].

Desalination constraints are energy, brine and coast.

Desalination is energy-intensive with large carbon footprint on coal electricity, and brine plus marine works may affect sensitive coastal areas ^[23]. Perth's response is to power desalination with renewables toward net zero by 2035 ^[31]. Cape Town's EIA is underway ^[5] and Barcelona requires EIA but is said to lack detailed operating rules ^[41].

Perth reports 40% RO recovery with 60% returned as brine via a diffuser for rapid mixing and dilution ^[9], with sunken design behind dunes and tunnel boring to limit seabed and beach disturbance ^[31]. Generic brine is highly concentrated with chemicals, potentially degrading coastal quality via salinity, turbidity and temperature, with impingement and entrainment at intakes ^[48]. Barcelona brine is described as oxygen-depleting with ferric chloride toxicity concerns ^[41]. Perth audit notes supply failures risk harm to waterways, aquifers and ocean ecosystems near desalination plants ^[28].

Concentrate management can determine feasibility, with seawater via deep-well injection or controlled ocean return and inland via surface water, evaporation ponds, wastewater plants or industrial reuse ^[11]. Sites near marine protected areas face more detailed review with possible intake or discharge restrictions, additional surveys, subsurface intakes or stricter dilution ^[58]. Brackish groundwater desalination needs less energy, higher recovery and smaller brine volumes than seawater ^[14], but brine regulations remain a key environmental concern plus variability, pre-treatment for high pH and magnesium, and costly remote infrastructure ^[14].

5. Who pays and which portfolios stay robust when assumptions vary

5.1 Tariffs, taxes, transfers and developers

No city funds new water from volumetric tariffs alone, consistent with the finding that tariffs are essential but not the only pathway and that equating tariffs with full economic recovery should give way to recognising taxes and transfers while conserving water ^[15]. Price rarely reflects treatment and distribution costs ^[15], and assuming full recovery while ensuring affordability leaves many utilities chronically underfunded ^[15].

Barcelona uses a binomial fixed-plus-variable tariff with increasing blocks ^[39], where the structure is explicitly to promote responsible consumption ^[8].

• Fixed service fee depends on housing type and rises with water points ^[8].
• Volumetric price rises per tier ^[8].
• Larger households over three residents can request larger tier volumes ^[8].
• Catalan Water Agency canon is a tiered tax on the same basis ^[8].
• Sewerage and waste collection are Barcelona City Council taxes ^[8], with 10% VAT on consumption and canon ^[8].
• Bills arrive every two months with meter readings and history histograms ^[8].

Protection is via a social tariff giving 100% discount on service fee and first- and second-tier consumption, 100% canon discount if the first tier is not exceeded else 50% per tier, 50-100% waste-tax discount and 100% sewerage exemption, with Energy Advisory Points supporting access ^[8].

Yet the transition is expected to raise prices and charges significantly, with funding and management of new infrastructure still to be decided and cost distribution required not to hit vulnerable groups or working-class households that already consume very little ^[30]. Efficiency is framed as a right that could offset bills for the most vulnerable ^[30].

The tension is visible in 12.8% of metro population reporting bill-payment problems in 2024 Institut Metròpoli data ^[41] and a history where 1987-1993 domestic prices rose 108% from 0.59 to 1.23 €/m³ to pay for EU wastewater plants with domestic taxes four times industrial ^[4], triggering non-payment by over 80,000 families across 23 municipalities and 2.6 million people and leading to the Catalan Water Agency in 1999 and a single water tax ^[4].

Cape Town blends inclining blocks, fixed charges, free basic water and developer charges.

Monthly step allocations use an average month of 365/12 days and balance day-by-day ^[13]. Every non-indigent household pays a monthly fixed basic charge by meter size covering about a quarter of fixed supply costs, with lower rates for registered indigents ^[13]. Registered indigents with direct connections receive first 6 kl water and 4.2 kl sewerage free, with excess at normal tariffs, extended to approved cluster developments ^[13]. Standalone sanitation is 70% of metered water to 35 kl/month and cluster 90% to 35 kl per household ^[13]. Informal settlements use communal taps and toilets ^[13].

Drought pricing showed the limits of volumetric-only recovery: average real price for non-indigents rose 230% March 2015-March 2018 with highest-block marginal up over 2000% ^[38]. Punitive rates reached R768.64 per 1,000 litres over 35,000 litres/month ^[57]. Non-compliance could bring a device limiting use to 350 litres/day at R4,500 homeowner cost ^[57]. To recover about R3.35 billion, average tariff needed to be ~R21/kl at 178 million kl Level 1 sales versus ~R32/kl at 93 million kl Level 6 sales, with fixed charges covering about a quarter of fixed costs ^[23].

After normalisation, volumetric charges were 22% below pre-drought mean while total charges rose ~4% after July 2018 fixed charges of mostly 56 or 100 ZAR for non-indigents, with indigents exempt and free block raised to 10.5 kL ^[38]. Without reforms richest deciles would have paid 51.8% less, while property-differentiated fixed charges offset some regressive shift from wealthy drilling and tanks ^[38].

Developers pay a once-off Development Charge at application as pro-rata actual capital cost of engineering services for land-use intensification ^[17], with a 27 October 2022 exemption only for the sewer portion where temporary package plants are approved in at-capacity treatment catchments with national authorisation ^[17]. Financing of non-conventional schemes under the national Raw Water Pricing Strategy was unclear with a 2022 revision expected, while National Treasury promotes crowding in private finance ^[19]. Faure illustrates the shift: under PPP the City pays an agreed unitary payment covering private financing, construction and operation with first payment only when fully operational, which is contracting not privatisation ^[24].

Affordability echoes African patterns where increasing block tariffs are most common ^[12][16][27]. Most utilities achieve operations and maintenance recovery only at highest blocks not first blocks for poor low-volume users, few recover even small capital shares even at top blocks, lowest-block subsidies do not exclusively reach the poor, and minimum charges burden the poorest while connection charges averaging 28% of GNI block access ^[12][16][27].

Perth socialises cost through uniform statewide pricing set by Cabinet regardless of supply cost, with the Corporation operating commercially, paying a dividend and seeking subsidies where charges do not cover costs ^[28].

Reducing unbilled water would cut costs passed to customers and Government from sourcing extra water such as desalination and treating and distributing it ^[28]. Poor pipe management would raise need for desalination investment ^[28]. Available loss reduction is estimated to net $2.5 million/year in lower operating costs by 2030 ^[28].

Housing growth is partly enabled by a $200 million Housing Enabling Infrastructure Fund for over 60,000 future homes in corridors ^[29]. Industrial reuse incentives are discussed elsewhere as tax credits for retrofitting to municipal recycled or onsite recycling to ease municipal demand and taxpayer burden ^[51], but Perth evidence does not commit to that mechanism.

5.2 Robust versus assumption-dependent interventions

Robust across drought scenarios: leakage, pressure management and efficient fixtures are consistently low-regret because they save already-treated water and defer supply.

• Cape Town pressure savings of 55 MLD ^[23] and 8.24 million m³/year in 2015/16 zones ^[26].
• Perth cost-effective 6.5 of 6.7 billion litres recoverable metro ^[28].
• Wisconsin loss control at less than half efficiency cost ^[34].
• Loss control offers best return even as efficiency offers up to ten times savings ^[34].

Demand behaviour proved it can move in weeks while supply takes years ^[61]:

• Cape Town cut city use from ~1,200 MLD in 2014 to ~555 by April 2018 ^[61], from 1,200 in February 2015 to 500 in February 2018 ^[23], and aggregate public use fell ~50% ^[38].
• Barcelona home actions saved 20% in 2008 ^[33].

These gains are robust but bounded: Cape Town notes >20% restriction hits households and >30% is believed no longer possible ^[32]. Barcelona greywater saves only up to 136,000-300,000 m³/year ^[30][35] versus Hm³-scale needs. Perth metro leakage at 20.3 billion litres is ~8% with unavoidable share large ^[28].

Dependable but cost- and energy-exposed. Seawater desalination provides very reliable water with no precipitation correlation and essentially zero variability manageable out of phase with surface sources ^[48]. It is therefore held at 100% assurance ^[19] and was run at 100% in Barcelona drought ^[41].

Its robustness fails on energy-price, grid and brine assumptions. For a 150 MLD plant, the inefficient-versus-efficient procurement gap is R0.65-1.3 billion/year ^[40], energy is up to half of cost ^[47][48], and Barcelona price rises are already linked to desalination running costs ^[41]. Climate-trajectory uncertainty is the second-largest net-present-value driver after discount rates in adaptation economics ^[1], and capital-intensive megaprojects show wider benefit-cost confidence intervals from 2.1-7.8 for low-capital to 1.2-4.1 for high-capital scenarios ^[1].

Mid-cost, mid-energy, but institutionally gated. Reuse is cost-competitive with seawater when flows exceed 1,500 m³/day ^[48].

RO-treated wastewater levelized cost falls from $0.81/m³ at 50% use to $0.68/m³ at 100% ^[48]. Larger indirect potable reuse (median $1.06/m³, range $0.91-1.28) costs less than small indirect potable reuse (median $1.50) ^[48].

Its yield is high-assurance and climate-resilient ^[19] at about half desalination energy ^[23]. It depends on wastewater availability, public acceptance, SANS 241 and Water Safety Plan compliance for 700+ contaminants ^[32], and environmental rules on river and wetland flows ^[23][40][49]. Faure's 25% blend limit ^[24], Zandvliet/Macassar sourcing ^[40], and Potsdam's shift to non-potable reuse for irrigation and industry ^[21] show how reuse portfolios split potable and non-potable to stay within those constraints.

Buffer value with sustainability risk. Aquifers and managed recharge can be recharged in wet months and recovered dry, managed out of phase with surface water ^[48]. Aquifer storage levelized cost falls from $0.25/m³ at 1 Mm³/year to $0.10/m³ at 5 Mm³/year ^[48].

Cape Town groundwater storage exceeds dam storage — dams ~900 million m³ versus Cape Flats >600 plus Table Mountain Group >1,000 ^[40] — and sustainable yield with recharge far exceeds 200 MLD in older planning ^[40].

Yet resilience and impacts remain uncertain ^[19], recharge is rainfall-dependent ^[38], private open-access pumping already shows metre-scale declines ^[38], and licensing, monitoring committees and DWS funding mechanisms gate dependability ^[32].

Portfolio implications follow the robustness literature rather than single-point optima. Coordinated water-energy planning saves 18-34% versus siloed planning ^[1]. This is relevant where desalination and reuse add small-town-scale loads ^[61] and where Perth pairs desalination with 400 MW wind and net zero ^[31].

Benefit-cost ratios for proactive adaptation are 2.4-11.8 with high pay-offs in drought-prone water ^[1]. Mediterranean water is at 4.6 and arid at 5.8 ^[1], and Spain's 62% water-resilience portfolio is at 5.2 ^[1]. This supports overweighting drought-resistant storage and demand management in all three cities.

But discounting from 3% to 7% swings scenario net present value from $118.6 billion to $68.4 billion as the largest variance driver ^[1], and robust optimisation that cuts maximum regret 23% costs 8% expected net present value in redundancy ^[1]. In supply-option terms, softer reversible measures are optimal on robustness when feasible, yet larger supply beats them in ~15% of less-favourable availability-demand scenarios hidden by central-scenario analysis ^[3].

For 2035 that means:

• do not trade away Alkimos, Faure or Tordera/Foix redundancy to hit a low central demand path;
• sequence reuse ahead of desalination on cost and energy, but keep desalination procurement moving because only it is fully rainfall-independent ^[23];
• treat Cape Town alien clearing, Barcelona river-flow limits and Perth Gnangara licence cuts as hard environmental constraints, not tuning parameters.

6. Limitations / Open Questions

• Incompatible cost boundaries prevent a true delivered-water ranking. Cape Town mixes 2018-Rand capital-plus-opex, later Unit Reference Values in 2025 Rand, and older R/kl planning numbers ^[19][32][54]; Perth reports $2.8 billion capital plus $1.00/kL opex excluding renewals and membranes ^[6]; Barcelona reports scheme budgets and tariff impacts without levelized €/m³ ^[49][41][33]. Discount rates, lifetimes and whether distribution, recharge, brine or grid upgrades are included are generally unstated.
• Yield definitions shift. Peak Ml/d, annual Mm³/year, licensed abstraction and sustainable recharge are used interchangeably across vintages. Examples include Cape Flats 20 Mm³/year equalling 55 MLD year-round with 80 MLD peak infrastructure ^[23], Faure described as 70 Ml/d, 70-100 Ml/d and 100 Ml/d ultimate in different documents ^[32][42][24], and desalination as 50, 50-70 and 70 Ml/d across the same programme ^[19][5][32]. Without capacity factors these cannot be summed precisely.
• Leakage economics are thinnest for Barcelona. High reported physical efficiency ^[4] coexists with a Catalonia-wide 25% loss estimate versus <10% best practice ^[41], but no city-specific $ or € per m³/day-saved curve is provided. Perth quantifies recoverable volumes and net operating benefit ^[28] and Cape Town quantifies pressure savings and pipe/meter budgets ^[23][26], yet neither provides full marginal-cost curves to the Economic Level of Leakage where further reduction is uneconomic ^[28].
• Groundwater sustainability is monitored but not closed. Cape Town private extraction is largely unregulated with limited monitoring ^[38], Table Mountain Group further phases are deferred to 2040 to monitor impacts ^[32], and Perth brackish potential of ~60 plus >110 GL ^[14] depends on acceptance, licences, pre-treatment and brine rules that are not yet resolved.
• Environmental unknowns remain material. A September 2025 study warns full desalination impacts remain unknown with few operating regulations despite mandatory EIA in Spain ^[41]; brine oxygen, chemical and mixing risks are noted in all three cities ^[41][9][48]; and reuse scale is bounded by river and wetland flow needs in both Cape Town and Barcelona ^[23][49].
• Who-bears-cost incidence is incomplete. Faure's 6-8.5% once-off tariff impact ^[24] and Perth's uniform pricing with subsidies ^[28] describe averages, but the evidence also shows lowest-block subsidies often miss the poor, minimum charges burden the poorest, and informal resale can be two to five times formal prices ^[12][27], while Barcelona working-class use is already very low ^[30]. Property-differentiated fixed charges partly offset regressive shifts in Cape Town ^[38], but distributional modelling for the 2031-2035 tariff path is not provided.
• Forward demand is assumption-sensitive. Cape Town moved from 95th- to 50th-percentile climate impact and from five-years-early to as-required delivery while retaining 1-in-200 assurance ^[32]; Barcelona demand is +4% to 2050 in a committed-action scenario ^[35] against -12% surface and -9% groundwater resources ^[33]; Perth assumes -15% per-person use by 2030 then flat ^[50]. Small changes in population, outdoor use, private drilling and industrial uptake therefore determine whether 2035 balances hold.

Sources

^[1] Economic optimization of climate adaptation for water security: an integrated water–energy nexus risk assessment framework — https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2026.1817765/full · academic ^[3] Insights into Robustness as a Criterion to Support Cost-Benefit Analysis of Supply-Enhancing Investments under Uncertainty. Part II – Investment Evaluation and Robustness Appraisal — https://link.springer.com/article/10.1007/s11269-025-04314-3 · academic ^[4] Strategies for sustainable water management in the face of ... — https://eau3e.hypotheses.org/files/2010/11/ATHENS_Sauri_10.pdf · academic ^[5] City progresses with plans for permanent desalination plant — https://www.capetown.gov.za/Media-and-news/City%20progresses%20with%20plans%20for%20permanent%20desalination%20plant · government ^[6] 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 ^[7] New water to secure more water supply for a growing city — https://www.capetown.gov.za/Media-and-news/New%20water%20to%20secure%20more%20water%20supply%20for%20a%20growing%20city · government ^[8] Do you want to understand your water bill? We’re here to help! — https://www.barcelona.cat/habitatge/en/actualitat/noticies/vols-entendre-la-factura-de-laigua-thi-ajudem-1361800 · government ^[9] Alkimos Seawater Desalination Plant Project - Water Corporation — https://www.watercorporation.com.au/outages-and-works/ongoing-works/alkimos-seawater-desalination-plant · government ^[10] Securing Perth's water supply - Water Corporation — https://www.watercorporation.com.au/our-water/climate-change-and-wa/climate-and-perth/securing-supply · government ^[11] Texas Water Development Board — https://www.twdb.texas.gov/innovativewater/desal/faq.asp · government ^[12] Open Knowledge Repository — https://openknowledge.worldbank.org/entities/publication/1e6c00ad-52c5-5a22-a6eb-8f1057bc8cb2 · government ^[13] Water and sanitation services and costs in formal housing — https://www.capetown.gov.za/Family%20and%20home/residential-utility-services/residential-water-and-sanitation-services/water-and-sanitation-services-and-costs-for-formal-housing · government ^[14] Brackish Groundwater - Brackish Groundwater Desalination — https://research.csiro.au/bgwd/brackish-groundwater/ · government ^[15] Can Tariffs Support Affordable and Sustainable Water and Sanitation Services in Times of Climate Change and other Crisis? — https://blogs.worldbank.org/en/water/can-tariffs-support-affordable-and-sustainable-water-and-sanitation-services-times-climate · government ^[16] Open Knowledge Repository — https://openknowledge.worldbank.org/entities/publication/1d133581-bf38-5a4e-9d66-ed8812544b87 · government ^[17] Development charges — https://www.capetown.gov.za/work%20and%20business/planning-portal/tariffs-and-charges/development-charges · government ^[18] Integrated water supply scheme - Water Corporation — https://www.watercorporation.com.au/our-water/perths-water-supply/integrated-water-supply-scheme · government ^[19] Western Cape Water Supply System — https://documents1.worldbank.org/curated/en/099100002272330999/pdf/P17148306acd480fc0bfb504b0df294bfe8.pdf · government ^[21] Potsdam wastewater plant upgrade reaches 60% mark — https://www.capetown.gov.za/Media-and-news/Potsdam%20wastewater%20plant%20upgrade%20reaches%2060%20mark · government ^[22] City begins procurement for Faure New Water Scheme - News and Notices — https://web1.capetown.gov.za/web1/newsandnotices/Home/Release/City-begins-procurement-for-Faure-New-Water-Scheme · government ^[23] WATER OUTLOOK 2018 REPORT — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/City%20research%20reports%20and%20review/Water%20Outlook%202018%20-%20Summary.pdf · government ^[24] 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 ^[25] The Challenge of Reducing Non-Revenue Water (NRW) in ... — https://documents1.worldbank.org/curated/en/385761468330326484/pdf/394050Reducing1e0water0WSS81PUBLIC1.pdf · government ^[26] IDP Water Sector Input Report — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/City%20strategies,%20plans%20and%20frameworks/Water%20Services%20Development%20Plan.pdf · government ^[27] Policy Research Working Paper 5384 — https://documents1.worldbank.org/curated/en/476671468210593814/pdf/WPS5384.pdf · government ^[28] 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 ^[29] 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 ^[30] Debates to address in Barcelona’s water transition | Barcelona Metròpolis | Barcelona City Council — https://www.barcelona.cat/metropolis/en/contents/debates-address-barcelonas-water-transition · government ^[31] New renewably-powered desalination plant planned for Alkimos — https://www.watercorporation.com.au/about-us/media-releases/2022/june-2022/new-renewably-powered-desalination-plant-planned-for-alkimos · government ^[32] CAPE TOWN WATER OUTLOOK — https://www.capetown.gov.za/_documents/resource.capetown.gov.za/documentcentre/Documents/City%20research%20reports%20and%20review/Water-Outlook-June-2025.pdf · government ^[33] Strategies to tackle water shortages | Barcelona Metròpolis | Barcelona City Council — https://www.barcelona.cat/metropolis/en/contents/strategies-tackle-water-shortages · government ^[34] WATER EFFICIENCY POTENTIAL STUDY FOR WISCONSIN — https://psc.wi.gov/documents/water/WaterEffStudy2011.pdf · government ^[35] greywater — https://www.aquapublica.eu/sites/default/files/event/file/2024-06/Maria%20Jos%C3%A9%20Chesa%20PPT.pdf · professional ^[36] Imposition of a state of emergency - Spain facing long-term — https://wodnesprawy.pl/en/implementing-a-state-of-emergency-spain-facing-a-prolonged-drought/ · professional ^[38] Drought, Adaptation, and Inequality in Cape Town — https://www.nber.org/system/files/working_papers/w33468/w33468.pdf · professional ^[39] Spain’s water is getting more expensive: the challenge of cost recovery — https://smartwatermagazine.com/news/smart-water-magazine/spains-water-getting-more-expensive-challenge-cost-recovery · professional ^[40] CAPE TOWN'S NEW WATER PROGRAMME — https://greencape.co.za/assets/Uploads/Annexure-to-Water-Outlook-New-Water-Program-23-March-2019.pdf · professional ^[41] Spain's water desalination dilemma — https://newint.org/water/2025/spains-water-desalination-dilemma · professional ^[42] 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 ^[45] Smart Metering for Water Efficiency — https://www.gihub.org/infrastructure-technology-use-cases/case-studies/smart-metering-for-water-efficiency/ · professional ^[47] Seawater Desalination Power Consumption — https://watereuse.org/wp-content/uploads/2015/10/Power_consumption_white_paper.pdf · professional ^[48] NANCIAL AND ECONOMIC APPRAISAL OF WATER ... — https://cri-world.com/publications/qed_dp_4614.pdf · professional ^[49] 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 ^[50] 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 ^[51] Investment Tax Credit for Industrial Reuse | WateReuse Association — https://watereuse.org/advocacy/federal-priorities/investment-tax-credit-for-industrial-reuse/ · professional ^[52] The Cost of Alternative Water Supply and Efficiency Options in ... — https://voiceofoc.org/wp-content/uploads/2016/10/Pacific-Institute.pdf · professional ^[53] Catalonian desalination expanding with ERI technology — https://www.waterworld.com/home/article/16213430/catalonian-desalination-expanding-with-eri-technology · professional ^[54] WATER MARKET INTELLIGENCE REPORT — https://greencape.co.za/wp-content/uploads/2022/10/WATER_MIR_2021_31_3_21-3.pdf · professional ^[55] An Analysis of the Energy Intensity of Water in California — https://www.aceee.org/wp-content/uploads/proceedings-1980-2020/2006/SS06_Panel12_Paper14.pdf · professional ^[56] How much energy does desalinisation use? Is it “absurdly cheap”? — https://hannahritchie.substack.com/p/how-much-energy-does-desalinisation · general ^[57] Cape Town water crisis — https://en.wikipedia.org/wiki/Cape_Town_water_crisis · general ^[58] What permits are required for a commercial desalination installation? - Elemental Water Makers — https://www.elementalwatermakers.com/knowledge-base/desalination/what-permits-are-required-for-a-commercial-desalination-installation/ · general ^[59] How much does desalination cost per cubic meter in 2026? - Elemental Water Makers — https://www.elementalwatermakers.com/knowledge-base/desalination/how-much-does-desalination-cost-per-cubic-meter-in-2026/ · general ^[61] Cape Town desalination, explained — https://capetowndamlevels.co.za/articles/cape-town-desalination/ · general

Source quality: 3 academic, 29 government, 17 professional, 5 general.

Verification

• Existing versus funded versus proposed status is not credibly separated: Alkimos Stage 1 about 55 GL/year as approved, contracted and 2028 first water ^[6][9] includes prospective tunnelling to intake May 2026 and outfall late August 2026 and barge departure 28 July, Cape Town Faure 70 Ml/d Phase 1 ultimately 100 Ml/d first water March/mid-2031 and Paarden Eiland 50-70 Ml/d first water ~2030 full end-2031 ^{[32][24][61][5]} remain at Section 78, PPP registration October 2024, RFQ/RFP second half 2027 and procurement begun 25 September 2026 with EIA underway, Barcelona Tordera extension plus Foix 20 Hm3/year to 140 Hm3/year within five years ^[49] rest on budgets and credits not levelized yield.
• Conflicting projections are noted but not reconciled: Barcelona 33% in 2023 ^[41] versus 55% of total regional consumption ^[36] versus 58% desalination or reuse ^[33] with reservoirs below 16% ^[30][36]; Cape Town allocation 576 Mm3/year ^[19] versus ~590 Mm3/year against 547 Mm3/year revised yield ^[54], Faure as 70 Ml/d, 70-100 Ml/d and 100 Ml/d ultimate and desalination as 50, 50-70 and 70 Ml/d ^{[19][5][32][54][42][24]}, Berg June 2028 per City versus June 2029 per DWS ^[32][42], Faure July 2025 ^[54] versus August 2029 ^[42] versus March/mid-2031 ^[32][24]; Perth residual need 75-125, up to 125 and more than 110 billion litres/year ^[10][9][29] versus audit reliable to at least 2035 ^[28].
• Delivered-water cost comparison violates its own compatibility rule: Cape Town 2018-Rand R5/kl reuse P1, R9/kl desalination Phase 1, R3-5/kl Berg ^[19] versus GreenCape R5.7/kl Faure, R9.0/kl permanent desalination, R2.2/kl Groenlandberg, R4.62/kl Berg-Voelvlei ^[54] versus 2025-Rand URV R30/m3 Faure, R50/m3 desalination, R7.2/m3 Berg-Voelvlei ^[32] versus R5 billion in 2023 prices for 70 Ml/d Paarden Eiland ^[61] are shown together without inflation, scope, lifetime, discount rate or inclusion of distribution, recharge, brine or grid upgrades.
• Perth Alkimos $2.8 billion capital with $51.9 million/year at $1.00/kL benchmark explicitly excludes major renewals and membrane replacement and may understate lifecycle cost ^[6] with no Do Minimum base case for incremental comparison ^[6]; $1.4 billion 2021 Budget down payment ^[9][31] is not a full funding proof for 2035 operation.
• Barcelona has no scheme-levelized EUR/m3: €250 million Tordera extension plus hundreds of millions Foix ^[49], €467 million May 2024 credit ^[41] later rendered as $590 million, €2.4 billion Catalan Water Agency programme to 2027 ^[33], tariff 1.97 EUR/m3 ^[39][4], 33% ATL increase ^[39] and end-2023 >33% rise ^[41] cannot support delivered-water ranking or gap-closure costing.
• Energy intensity claims mix fenceline membrane with whole-system boundaries: average seawater ~4 kWh/m3 down from 8 to 2.3 optimised ^[49], modern RO 2.5-3.5 ^[56], Cape Town ~2 kWh per thousand litres reuse versus 3.5-4 desalination ^[23][40] and reclaimed one-third desalination ^[41] versus generic SWRO 6.8-8.2 kWh/kgal, indirect potable reuse total 7.0-11.5 kWh/kgal, Pacific seawater 10.0-14.0 kWh/kgal ^[47] and California import 9.8-12.1 kWh/kgal ^[47][55]; 1 kWh/kgal = 325.8 kWh/AF conversion ^[47] does not make rows comparable.
• Lead-time credibility to 2035 relies on non-transferable generics: desalination permitting typically three months to two years but longer near sensitive habitats ^[58] from Elemental Water Makers ^[58][59], Texas brackish planning 2001 to completion 2007 ^[11], standard South African owner design-build four to five years versus two to three years internationally ^[23][40], and emergency short-term licences in days or weeks ^[58] do not establish that Faure must start 2028/2029 for 70 Ml/d by 2031 without six-year delay if City-funded ^[24] will hold.
• Dependable drought yield overstates aquifer and reuse assurance: desalination, reuse and aquifers as almost 100% assurance ^[19] and desalination as totally independent of rainfall ^[23][40] and 100% assurance ^[19] are contradicted within the same evidence by uncertain climate resilience and environmental impacts of aquifers ^[19], rainfall-dependent recharge ^[38], metre-scale declines from largely unregulated private drilling exempt from restrictions with ~1 in 10 boreholes licensed ^[38], and reuse dependence on wastewater availability, 25% blend limit ^[24] or ~20% reuse to 80% dam water ^[61], SANS 241 700+ contaminants ^[32] and river and wetland flow trade-offs ^[23][40][49].
• Yield arithmetic conflates peak Ml/d, annual Mm3/year, licensed abstraction and sustainable recharge without capacity factors: Cape Flats licence 20 Mm3/year equals about 55 MLD year-round with ~80 MLD peak infrastructure ^[23], committed programme 347 Ml/d or 128 Mm3/year new supply and 417 Ml/d or 154 Mm3/year including demand management ^[19] versus updated alien clearing 30 Ml/d peak and 11 Mm3/year ^[32] versus 55 Ml/d or 20 Mm3/year ^[19], and demand management 70 Ml/d or 26 Mm3/year counted as supply.
• Environmental constraints are listed but not quantified as hard gates: aliens current 24 Mm3/year growing to 85 Mm3/year ^[42], Theewaterskloof 30 million m3/year loss ^[57] and DWS funding mechanism dependence ^[32]; Perth 40% RO recovery with diffuser ^[9], Barcelona brine oxygen and ferric chloride concerns ^[41], generic impingement, entrainment, salinity and temperature impacts ^[48] and September 2025 warning full impacts unknown with few operating rules despite EIA ^[41]; Barcelona 490 Hm3/year to sea bounding reuse ^[33] versus rivers stop functioning normally if all recycled ^[49].