Why Salvador changes the design baseline in 2026
A 20–80 MW data center campus in Salvador, Bahia faces a hydrological and regulatory envelope that does not match Brasília's Cerrado baseline or a US Southwest reference. The site sits on a tropical-Atlantic coast with year-round high wet-bulb temperatures and high ambient humidity, which keeps evaporative cooling demand and makeup volume elevated relative to a Cerrado or semi-arid site. For a campus at PUE 1.3–1.5, plan makeup of 200–800 m³/day, as the Open Engineering reference of roughly 2 million L/day for a 100 MW facility sets the upper benchmark (Genesis Water Technologies, 2026).
CONAMA Resolution 430/2011 sets the national discharge floor of pH 5–9, BOD ≤ 120 mg/L to sewer, O&G ≤ 50 mg/L, plus explicit toxicity caps for biocides (HydropureWater Brasília guide, 2026). Those toxicity caps are what hit cooling blowdown first, not the sanitary BOD and TSS numbers — biocide and scale-inhibitor residuals carry the load that triggers permit review. INEMA's Bahia effluent overlay and EMBASA's industrial allocation posture are tightening in dry-season stress, so for a 2026 procurement manager the discharge-permit conversation must be raised in front-end engineering, not deferred to commissioning.
All Saints Bay as receiving water raises the bar on TDS, chloride, sulfate, and biocide residual load limits, as documented in the February 2026 TNFD case study on data-center water-quality impacts (Water Utility Report, 2026-04-14). The practical effect is that polishing blowdown to cooling-makeup quality converts a discharge-permit problem into a wet-weather overflow problem — a fundamentally different regulatory posture than a dry inland site with a deep aquifer dilution buffer.
The two effluent streams a Salvador campus must treat
A hyperscale Salvador campus produces two distinct effluent streams that converge at the discharge point but require separate treatment trains. Stream 1 is cooling tower blowdown — the bleed cycle that holds cycles of concentration (COC) at 4–6 to prevent scale. At 4–6 COC, expect 1,500–2,500 mg/L TDS, 400–800 mg/L CaCO₃ total hardness, 5–50 mg/L free chlorine or bromine residual, and 5–30 mg/L TSS, with ranges shifting on scale-inhibitor chemistry and ambient dust load (HydropureWater Brasília guide, 2026). Stream 2 is domestic sanitary wastewater from staff, cafeteria, and restrooms, typically 50–100 L/person/day at a hyperscale campus, characterized by BOD 150–300 mg/L, TSS 150–250 mg/L, and NH₃-N 20–40 mg/L.
Blowdown volume is governed by the relationship Blowdown = Makeup / (COC − 1); at 4–6 COC the bleed runs 0.3–0.8% of makeup volume (HydropureWater Brasília guide, 2026). For a 40 MW campus with PUE 1.4 and an adiabatic/economizer hybrid cooling loop, that translates to roughly 60–240 m³/day of blowdown on top of 50–80 m³/day of sanitary flow. Both streams normally discharge to the EMBASA sewer under CONAMA 430 unless on-site reuse is designed in; the permit path differs by stream because biocide and scale-inhibitor residuals trigger toxicity caps that sanitary BOD and TSS do not.
| Parameter | Cooling tower blowdown (4–6 COC) | Domestic sanitary wastewater |
|---|---|---|
| Flow at 40 MW reference (m³/day) | 60–240 | 50–80 |
| TDS (mg/L) | 1,500–2,500 | Low (not a governing parameter) |
| Total hardness as CaCO₃ (mg/L) | 400–800 | Low |
| Free chlorine / bromine (mg/L) | 5–50 | Negligible |
| TSS (mg/L) | 5–30 | 150–250 |
| BOD (mg/L) | Low | 150–300 |
| NH₃-N (mg/L) | Low | 20–40 |
| Biocides (isothiazolinone, mg/L) | Variable, drives toxicity cap | Not present |
The 2026 five-unit-operation train for cooling blowdown

The 2026 treatment train for a Salvador cooling-blowdown stream is five unit operations, designed to hit either CONAMA 430 discharge or cooling-makeup reuse without changing the upstream sequence. Each step targets a specific parameter in the table above; together they convert a biocide-laden, scale-prone bleed into either a compliant discharge or a high-quality makeup stream that lifts the site's COC and reduces EMBASA allocation pressure.
- Hold-and-decay equalization sized at 24–48 h of blowdown volume, paired with sodium bisulfite dosing to drop free chlorine residual to ≤ 0.5 mg/L before the next stage (HydropureWater Brasília guide, 2026). Isothiazolinone destruct requires longer holding time or activated-carbon polishing.
- DAF or lamella clarifier when TSS > 20 mg/L. The high-efficiency sedimentation tank cuts chemical consumption by up to 30% versus conventional trains and shrinks mechanical-room footprint. A DAF system sized to peak blowdown flow handles the bulk-removal step before media filtration.
- Multi-media filter targeting turbidity and iron to protect downstream membranes; feed to RO should be filtered to less than 10–15 microns, chemically conditioned against scaling, and pH-adjusted (Genesis Water Technologies, 2026). A multi-media filter in this position drops TSS below the threshold that would otherwise foul RO elements.
- Side-stream RO on the cooling loop, or RO on blowdown at 150–400 psi, with hybrid antiscalant. RO permeate of 10–50 mg/L TDS is suitable for direct return to the cooling tower as high-quality makeup or for blending with standard makeup to raise overall COC (Genesis Water Technologies, 2026). A staged industrial RO system sized to 50,000 GPD is the standard block in this position.
- ClO₂ polishing on the reuse loop for residual biological control, dosed through an automatic chemical dosing system with an on-site ClO₂ generator sized to the polishing flow.
| Step | Unit operation | Target parameter | Typical sizing at 40 MW |
|---|---|---|---|
| 1 | Hold-and-decay + sodium bisulfite | Free Cl₂/Br₂ ≤ 0.5 mg/L | 24–48 h blowdown volume |
| 2 | DAF or lamella clarifier | TSS < 20 mg/L downstream | Peak blowdown flow |
| 3 | Multi-media filter | Turbidity, iron, TSS to < 10–15 µm | RO feed flow |
| 4 | Side-stream RO at 150–400 psi | TDS 10–50 mg/L permeate | 50,000 GPD reference |
| 5 | ClO₂ polishing | Residual biological control | Reuse loop flow |
Sanitary wastewater: MBR for reuse, packaged STP for discharge-only
For discharge-only sites below roughly 80 m³/d of sanitary flow, a WSZ underground integrated sewage treatment plant in the 1–80 m³/h envelope handles the load fully buried with no on-site operator and a single annual sludge pump-out (HydropureWater Brasília guide, 2026). That envelope covers the sanitary envelope of most Salvador campus sub-sites that do not intend to reuse treated sewage.
For a 20–80 MW hyperscale site that wants to reuse treated sewage for cooling-tower makeup, irrigation, or toilet flush, specify an MBR with submerged PVDF at < 1 µm pore size — roughly 60% smaller footprint than conventional activated sludge, with effluent BOD < 5 mg/L and TSS < 1 mg/L (HydropureWater Brasília guide, 2026). DF-series replaceable flat-sheet elements cover 10–2,000 m³/day campus WWTPs and clean in place with standard CIP chemistry, so the reuse train scales with campus headcount without a step-change in footprint or operator skill. The full sizing methodology for this block is covered in the containerized MBR sizing guide, which applies the same flux and CIP logic to a tropical coastal site.
Three CAPEX/OPEX tiers a Salvador buyer can actually procure

Three tiers, each tied to COC and reuse targets rather than a fixed process flow, let a procurement manager pick a scope by campus size and water-stewardship ambition instead of forcing a one-size-fits-all tender. The JY integrated water purification system and the high-efficiency sedimentation tank (lamella clarifier) fit Tier 2 and Tier 3 polishing loops; both cut chemical consumption by up to 30% versus conventional trains and reduce footprint in the mechanical room (HydropureWater Brasília guide, 2026).
- Tier 1 (discharge-only, ~4 COC, no RO): DAF only if TSS > 20 mg/L; no RO, no softening; suitable where EMBASA has confirmed allocation capacity and All Saints Bay discharge is permitted (HydropureWater Brasília guide, 2026).
- Tier 2 (reuse, 6–8 COC): DAF + industrial water softener system + side-stream RO + ClO₂ on the reuse loop; a JY integrated water purification system fits this polishing envelope.
- Tier 3 (reuse at 7–10 COC with optional ZLD): full pretreatment-RO train, with brine concentrator (MVC at 15–25 kWh/kgal distillate) and forced-circulation crystallizer added only when basin TDS exceeds 1,500 mg/L or discharge is effectively prohibited (Genesis Water Technologies, 2026; HydropureWater Brasília guide, 2026).
| Tier | Scope | Target COC | RO CAPEX (50,000 GPD) | OPEX | Notes |
|---|---|---|---|---|---|
| 1 | DAF only, discharge to EMBASA | ~4 | None | Discharge fees only | Requires confirmed EMBASA allocation |
| 2 | DAF + softener + side-stream RO + ClO₂ | 6–8 | $250,000–500,000 | $1.50–3.00/kgal | Standard hyperscale reuse scope |
| 3 | Full pretreatment-RO + optional MVC/crystallizer | 7–10 | $250,000–500,000 RO + $1–3 M MVC | $1.50–3.00/kgal RO; $5–15/kgal ZLD | ZLD only when basin TDS > 1,500 mg/L or discharge prohibited |
RO CAPEX for a 50,000 GPD blowdown unit runs $250,000–500,000 installed, with OPEX of $1.50–3.00 per thousand gallons treated including energy, chemicals, membrane replacement, and maintenance (Genesis Water Technologies, 2026). Full ZLD CAPEX runs $3–8 M with OPEX of $5–15 per thousand gallons at 95–99% overall recovery (Genesis Water Technologies, 2026; HydropureWater Brasília guide, 2026). The 60–80% potable-draw reduction from reusing treated sanitary effluent as cooling makeup compounds the savings when EMBASA scarcity is declared, because the industrial allocation cap stops being binding once on-site reuse is the baseline.
When ZLD becomes defensible on a Salvador site
ZLD becomes defensible only when basin TDS exceeds 1,500 mg/L, EMBASA industrial allocation is binding, and discharge is effectively prohibited — at that point add MVC and a forced-circulation crystallizer to the standard five-step train (Genesis Water Technologies, 2026; HydropureWater Brasília guide, 2026). Outside that envelope, ZLD is a $3–8 M outlay chasing a problem the standard train already solves.
A partial-ZLD configuration that concentrates blowdown to reduce discharge volume by 80–90% captures most of the water-recovery benefit at a fraction of full-ZLD cost; the remaining concentrated brine may then qualify for hauling to an approved disposal facility (Genesis Water Technologies, 2026). Without side-stream RO, target 4–6 COC to keep blowdown TDS manageable for discharge under CONAMA 430 and the INEMA Bahia overlay; with side-stream RO and softening, push to 7–10 COC and recover more than 75% of blowdown as reuse water (HydropureWater Brasília guide, 2026).
Beyond direct savings, discharge fees in water-stressed regions of $5–15 per thousand gallons (Genesis Water Technologies, 2026) add a second economic line to the reuse case and accelerate RO payback once EMBASA industrial tariffs cross the threshold the Brasília analog demonstrates is already in range for Brazilian utilities. The Brasília data center treatment guide applies the same decision logic to the Cerrado basin, and the Córdoba data center treatment guide covers the arid-temperate analog where ZLD is the default rather than the exception.
Frequently Asked Questions
What is the order-of-magnitude BRL CAPEX for a Tier 2 reuse train on a 40 MW Salvador campus?
A Tier 2 reuse train (DAF + softener + side-stream RO + ClO₂) on a 50,000 GPD blowdown unit installs for $250,000–500,000 USD per Genesis Water Technologies (2026); the BRL equivalent moves with the dollar-quotation rate of the procurement month, so request a quotation pegged to the closing dollar of the tender date. OPEX runs $1.50–3.00 per thousand gallons treated including energy, chemicals, membrane replacement, and maintenance. For a Tier 3 ZLD scope, CAPEX runs $3–8 M USD and OPEX $5–15 per thousand gallons at 95–99% overall recovery (Genesis Water Technologies, 2026; HydropureWater Brasília guide, 2026) — confirm the dollar-to-BRL conversion against your treasury's hedge rate before submitting the budget baseline to the CFO.
How do I pick a supplier for a 50,000 GPD RO + MBR skid for a Salvador site, and what lead time should I budget?
Qualify suppliers on three inputs: documented PVDF flat-sheet MBR performance at < 1 µm pore size with effluent BOD < 5 mg/L and TSS < 1 mg/L (HydropureWater Brasília guide, 2026), RO operating-pressure envelope of 150–400 psi with hybrid antiscalant, and a ClO₂ polishing block sized to the reuse loop rather than a generic potable-water generator. For lead time, request a written schedule tied to the equipment-list review and FAT milestones — Brazilian coastal EPCs typically run 16–24 weeks for a skid of this scope, but the actual number depends on membrane stocking at the supplier and the INEMA submission window. Ask each bidder for a recent Brazilian hyperscale reference with the same influent envelope before scoring the technical proposal.
What CONAMA 430/2011 and INEMA Bahia limits will the permit reviewer focus on first?
CONAMA Resolution 430/2011 sets the national floor: pH 5–9 for discharge to receiving waters, BOD ≤ 120 mg/L to sewer, O&G ≤ 50 mg/L, and explicit toxicity caps for biocides (HydropureWater Brasília guide, 2026). The INEMA Bahia overlay layers tighter limits on TDS, chloride, and sulfate, especially in sub-basins that feed All Saints Bay — the specific numeric cap is a site-and-license input to request from INEMA during the front-end engineering phase rather than a value to assume. For the biocide residual specifically, the standard permit approach is a 24–48 h hold-and-decay tank plus sodium bisulfite to drop free chlorine to ≤ 0.5 mg/L at the point of discharge.
When is ZLD actually justified for a hyperscale campus in Salvador?
ZLD becomes defensible only when basin TDS exceeds 1,500 mg/L, EMBASA industrial allocation is binding, and discharge is effectively prohibited — at that point add MVC and a forced-circulation crystallizer to the standard five-step train (Genesis Water Technologies, 2026; HydropureWater Brasília guide, 2026). For most 2026 Salvador campuses with confirmed EMBASA allocation and All Saints Bay discharge permitted, Tier 2 reuse with side-stream RO is the economic optimum. The intermediate option is partial-ZLD, which concentrates blowdown to reduce discharge volume by 80–90% and captures most of the water-recovery benefit at a fraction of full-ZLD cost (Genesis Water Technologies, 2026) — request a partial-ZLD bid alongside the full-ZLD line item when the basin TDS reading approaches the 1,500 mg/L trigger.