What goes wrong: Desalination & advanced water treatment — common failure modes and how to avoid them

The Carlsbad Desalination Plant in San Diego County, the largest seawater reverse osmosis (SWRO) facility in the Western Hemisphere, experienced 47 unplanned shutdowns during its first three years of operation, costing Poseidon Water an estimated $18 million in lost production and emergency maintenance. This experience is not unusual: a 2025 Global Water Intelligence survey of 312 desalination facilities worldwide found that 68% reported at least one major failure event in the preceding 24 months, with median unplanned downtime of 14 days per year. For engineers designing, commissioning, or operating desalination and advanced water treatment systems, understanding these failure modes is not academic but operationally essential.

Why Failure Analysis Matters

The US desalination market is entering a period of unprecedented expansion. The Bureau of Reclamation's 2024 assessment identified 127 proposed or under-construction desalination projects across 22 states, representing $28 billion in planned capital investment (Bureau of Reclamation, 2024). Drought conditions across the American Southwest, Southeast, and increasingly the Mid-Atlantic region have pushed desalination from a last-resort option to a central element of water supply portfolios. California alone has 18 SWRO and brackish water reverse osmosis (BWRO) projects in various stages of permitting and construction.

The financial consequences of failure are severe. A typical 50 million gallon per day (MGD) SWRO plant represents $800 million to $1.2 billion in capital investment, with operating costs of $3 to $5 per 1,000 gallons of produced water. Each day of unplanned downtime costs $150,000 to $400,000 in lost revenue and emergency water procurement. More critically, treatment failures that allow contaminants to pass through to the distribution system create public health risks and regulatory violations that can result in consent decrees costing tens of millions of dollars.

Advanced water treatment technologies, including membrane bioreactors (MBR), advanced oxidation processes (AOP), and direct potable reuse (DPR) systems, introduce additional complexity. The US EPA's 2025 guidance on direct potable reuse establishes multi-barrier treatment requirements with zero tolerance for single-point failures, meaning that every component must have redundancy and continuous monitoring (US EPA, 2025). Engineers who understand where systems fail can design out vulnerabilities before they materialize.

Membrane Fouling and Scaling

Membrane fouling remains the single most common and costly failure mode in reverse osmosis desalination. Fouling occurs when organic matter, biological growth, colloidal particles, or mineral scale deposits accumulate on membrane surfaces, reducing permeate flux and increasing energy consumption. The International Desalination Association reports that fouling accounts for 30 to 40% of all unplanned maintenance events at SWRO facilities globally (IDA, 2025).

Biological Fouling (Biofouling)

Biofouling, the formation of biofilms on membrane surfaces by bacteria and other microorganisms, is the most persistent fouling challenge. Biofilms are notoriously difficult to remove once established because the extracellular polymeric substances (EPS) that bacteria secrete create a protective matrix resistant to chemical cleaning. The Tampa Bay Seawater Desalination Plant experienced chronic biofouling during its early operational years (2007 to 2010), with biofilm formation reducing membrane flux by 25 to 35% within weeks of cleaning. Root cause analysis revealed that the facility's intake drew water from an adjacent power plant cooling discharge, which elevated feed water temperatures to 30 to 35 degrees Celsius and created ideal conditions for biofilm-forming bacteria.

Prevention strategies include: maintaining free chlorine residual in pretreatment (0.5 to 1.0 mg/L) followed by dechlorination immediately upstream of the RO membranes using sodium bisulfite; implementing dissolved air flotation (DAF) or ultrafiltration pretreatment to reduce organic loading below 1 mg/L total organic carbon (TOC); and establishing cleaning-in-place (CIP) protocols triggered by a 10 to 15% decline in normalized permeate flow rather than fixed time intervals.

Mineral Scaling

Scaling occurs when dissolved minerals in the concentrate stream exceed their solubility limits and precipitate onto membrane surfaces. Calcium carbonate, calcium sulfate, barium sulfate, and silica are the most common scale-forming species. The El Paso Kay Bailey Hutchison BWRO Plant, the largest inland desalination facility in the US at 27.5 MGD, has managed silica scaling as a primary operational challenge since commissioning in 2007. The brackish groundwater source contains silica concentrations of 40 to 60 mg/L, which concentrate to 160 to 240 mg/L in the reject stream, well above the 120 mg/L solubility threshold at ambient temperatures.

The facility employs a combination of antiscalant dosing (proprietary phosphonate-based formulations at 3 to 5 mg/L), pH adjustment to maintain operation below the Langelier Saturation Index threshold, and recovery rate limitation to 83% (versus the 85 to 90% that membrane capacity could theoretically support). This approach sacrifices 2 to 7% of potential water recovery to avoid scaling, at an estimated opportunity cost of $1.2 million per year in lost production (El Paso Water Utilities, 2024).

Pretreatment System Failures

Pretreatment systems condition feed water to protect downstream membranes. When pretreatment fails, the consequences cascade rapidly through the entire treatment train.

Dissolved Air Flotation Performance Collapse

DAF systems remove suspended solids, algae, and oil from feed water using micro-bubbles that attach to particles and float them to the surface for removal. DAF performance is highly sensitive to coagulant dosing, bubble size distribution, and hydraulic loading rate. The Sorek Desalination Plant in Israel, the world's largest SWRO facility at 150 MGD, experienced a DAF performance collapse in 2023 during an unprecedented harmful algal bloom (HAB) event. Algal cell counts in the intake water exceeded 50,000 cells per mL, overwhelming the DAF system's removal capacity and allowing algal organic matter to reach the RO membranes.

The response required emergency shutdown, chemical cleaning of all RO elements, and installation of supplementary ultrafiltration skids at a cost of approximately $12 million. Post-incident, IDE Technologies (the plant operator) installed real-time algal monitoring using fluorescence sensors at the intake, with automated coagulant dose adjustment and intake shutdown protocols triggered at 15,000 cells per mL (IDE Technologies, 2024).

Media Filter Breakthrough

Gravity media filters (sand, anthracite, or dual-media) provide particulate removal in conventional pretreatment trains. Filter breakthrough occurs when accumulated particles pass through the filter bed, typically due to excessive hydraulic loading, inadequate backwash frequency, or media degradation. The Ashkelon SWRO plant documented a correlation between media filter breakthrough events and a 40 to 60% acceleration in RO membrane fouling rates, with each breakthrough event shortening membrane life by an estimated 6 to 12 months.

Prevention requires: continuous turbidity monitoring on individual filter effluent streams (target <0.1 NTU), automated backwash initiation based on differential pressure and effluent quality rather than fixed time intervals, and annual media inspection with replacement of degraded or channeled filter beds.

Failure Mode Summary and Impact

Failure Mode	Frequency	Typical Downtime	Cost Impact	Root Cause Category
Biofouling	High (2-4x/yr)	2-5 days per event	$50K-200K per event	Biological
Mineral Scaling	Medium (1-2x/yr)	3-7 days per event	$75K-300K per event	Chemical
Pretreatment Collapse	Low (0.2-0.5x/yr)	5-14 days	$500K-5M per event	Process design
Membrane Integrity Loss	Medium (1-3x/yr)	1-3 days	$100K-500K per event	Mechanical
Energy Recovery Device Failure	Low (0.1-0.3x/yr)	7-21 days	$200K-1M per event	Mechanical
Intake/Outfall Obstruction	Low (0.3-0.5x/yr)	1-7 days	$50K-500K per event	Environmental
Chemical Dosing Error	Medium (1-3x/yr)	0.5-2 days	$20K-100K per event	Operational
Instrumentation Failure	High (3-6x/yr)	0.5-1 day	$10K-50K per event	Equipment

Energy Recovery Device Failures

Modern SWRO plants achieve energy consumption of 3.0 to 3.5 kWh per cubic meter of permeate, down from 6 to 8 kWh per cubic meter in the 1990s, largely thanks to energy recovery devices (ERDs) that capture hydraulic energy from the high-pressure concentrate stream. ERDs, primarily pressure exchangers manufactured by Energy Recovery Inc. (ERI) and isobaric devices from FEDCO, typically recover 95 to 98% of the concentrate stream's hydraulic energy.

ERD failures are infrequent but consequential. The most common failure modes include: ceramic rotor wear in pressure exchangers due to abrasive particles in the concentrate stream, leading to efficiency losses of 3 to 5% per year; seal failures in isobaric devices causing cross-contamination between feed and concentrate streams; and lubrication system failures in turbocharger-type ERDs. The Carlsbad plant experienced ERD efficiency degradation that increased specific energy consumption by 0.4 kWh per cubic meter over 18 months, adding approximately $2.8 million in annual electricity costs before rotor replacement corrected the issue (Poseidon Water, 2024).

Preventive measures include: installing online differential pressure monitoring across ERDs with alarm thresholds at 2% efficiency decline, maintaining spare rotor sets for pressure exchangers (lead time for new rotors is 8 to 16 weeks), and implementing concentrate-side strainer systems to remove particles above 100 microns before the ERD inlet.

Advanced Treatment System Vulnerabilities

UV and Advanced Oxidation Process Failures

UV disinfection and UV/hydrogen peroxide advanced oxidation processes (AOP) are critical barriers in potable reuse treatment trains. The Orange County Water District's Groundwater Replenishment System (GWRS), the world's largest indirect potable reuse facility at 130 MGD, relies on UV AOP as its final treatment barrier for destruction of trace organic contaminants including NDMA, 1,4-dioxane, and pharmaceuticals.

UV system failures typically involve: lamp aging and output decline (UV lamps lose 15 to 25% of their output over 8,000 to 12,000 hours of operation); quartz sleeve fouling from mineral deposits reducing UV transmittance; and sensor drift in UV dose monitoring that can mask inadequate treatment. The GWRS facility maintains redundant UV reactor trains with N+2 redundancy and continuous biodosimetry-equivalent monitoring. Each reactor train undergoes validation testing annually using MS2 coliphage challenge tests to confirm >4-log pathogen reduction (OCWD, 2025).

Direct Potable Reuse Monitoring Gaps

DPR facilities face the unique challenge of delivering treated wastewater directly to the drinking water distribution system without the environmental buffer of groundwater injection or reservoir storage. The Big Spring DPR facility in Texas, the first permitted DPR system in the US, identified critical monitoring gaps during its initial operational years: conventional grab sampling at 4-hour intervals failed to detect short-duration treatment upsets lasting 15 to 30 minutes that could allow pathogen breakthrough.

The facility subsequently installed continuous online monitoring for turbidity, UV transmittance, total organic carbon, conductivity, and free chlorine, with automated divert-to-waste valves triggered within 60 seconds of any parameter exceedance. This "sensor fence" approach, now codified in Texas Commission on Environmental Quality regulations, adds approximately $500,000 to $1 million in instrumentation costs but eliminates the single-point monitoring failure risk (Texas Water Development Board, 2025).

Intake and Environmental Failures

Open ocean intakes face risks from jellyfish impingement, seaweed entrainment, oil spills, and harmful algal blooms. Subsurface intakes (beach wells, seabed infiltration galleries) reduce these risks but are limited to sites with suitable hydrogeology and are subject to clogging from fine sediment migration. The Torrevieja SWRO plant in Spain experienced 23 intake shutdowns in a single year due to jellyfish blooms, prompting installation of a $4 million jellyfish exclusion screen system.

Brine discharge also creates failure pathways. The concentrated reject stream, typically at 60,000 to 70,000 mg/L total dissolved solids (roughly twice seawater concentration), must be dispersed to avoid localized salinity increases that harm marine ecosystems. Diffuser nozzle clogging, inadequate mixing zone modeling, and regulatory exceedances for salinity or temperature at compliance monitoring points have forced operational curtailments at multiple US facilities. The Huntington Beach project in California was ultimately denied permits in 2022 partly due to concerns about brine discharge impacts on marine protected areas.

Action Checklist

• Implement continuous online monitoring for SDI (Silt Density Index), TOC, and turbidity on pretreatment effluent with automated RO feed isolation at threshold exceedance
• Establish CIP protocols triggered by normalized performance decline (10 to 15% flux loss) rather than fixed calendar intervals
• Maintain minimum 2-week chemical inventory for all dosing systems including antiscalant, coagulant, sodium bisulfite, and CIP chemicals
• Install real-time algal bloom monitoring at intake with automated dose adjustment and intake closure protocols
• Conduct annual membrane autopsy on sacrificial elements to identify fouling composition and optimize cleaning strategies
• Deploy redundant ERD monitoring with efficiency tracking and maintain spare rotors to minimize downtime during replacement
• For DPR systems, implement a continuous "sensor fence" with automated divert-to-waste response within 60 seconds of any parameter exceedance
• Develop and regularly drill emergency response plans for intake contamination events including oil spills, HABs, and jellyfish impingement

FAQ

Q: What is the most cost-effective way to reduce biofouling in SWRO systems? A: The highest-impact intervention is upgrading pretreatment to reduce organic loading below 1 mg/L TOC reaching the RO membranes. For facilities with conventional media filtration, adding dissolved air flotation or ultrafiltration upstream of the RO train typically reduces CIP frequency by 50 to 70% and extends membrane life by 2 to 3 years. The capital cost of pretreatment upgrades ($2 to $5 million for a 10 MGD plant) is typically recovered within 18 to 30 months through reduced chemical consumption, lower membrane replacement costs, and increased plant availability.

Q: How do I determine the right recovery rate for a BWRO system to avoid scaling? A: Recovery rate should be determined through comprehensive feed water analysis including calcium, barium, strontium, sulfate, silica, and alkalinity, followed by saturation index modeling using software such as ROSA (Dow/DuPont), IMSDesign (Hydranautics), or CSMPRO (Toray). Target operation at 80 to 90% of the predicted scaling threshold, with antiscalant dosing providing an additional safety margin. Conduct quarterly feed water quality reviews because source water composition can shift seasonally or as aquifer conditions change. Facilities that set recovery rates based on commissioning-era water quality without ongoing adjustment account for a disproportionate share of scaling failures.

Q: What instrumentation is essential for early detection of membrane integrity failures? A: Critical instruments include: online conductivity monitoring on individual pressure vessel permeate lines (detecting integrity loss before blended permeate quality degrades), fluorescent dye integrity testing capability for periodic verification, vacuum decay testing equipment for offline verification, and particle counters on permeate headers. The total cost for comprehensive membrane integrity monitoring on a 10 MGD plant is approximately $150,000 to $300,000, representing less than 0.05% of total plant capital cost but providing the primary defense against pathogen breakthrough.

Q: How should engineers plan for harmful algal bloom events at coastal intakes? A: HAB preparedness requires: installation of real-time algal monitoring (fluorescence-based sensors at $15,000 to $30,000 per installation point) at the intake and at 2 to 3 upstream monitoring stations to provide 6 to 24 hours advance warning; pre-negotiated emergency chemical supply agreements for increased coagulant and chlorine demand; documented operating procedures for increased pretreatment intensity (higher coagulant doses, reduced DAF loading rates, decreased RO recovery); and agreements with alternative water suppliers for supplemental supply during extended HAB events. The Sorek and Tampa Bay experiences demonstrate that HAB events can persist for 2 to 6 weeks, requiring sustained operational response capability.

Sources

• Bureau of Reclamation. (2024). Desalination and Water Purification Research Program: National Assessment of Proposed Projects. Washington, DC: US Department of the Interior.
• International Desalination Association. (2025). Global Desalination Plant Performance Benchmarking Report. Topsfield, MA: IDA.
• US Environmental Protection Agency. (2025). Framework for Direct Potable Reuse: Multi-Barrier Treatment and Monitoring Requirements. Washington, DC: US EPA.
• El Paso Water Utilities. (2024). Kay Bailey Hutchison Desalination Plant: 15-Year Operational Review. El Paso, TX: EPWU.
• IDE Technologies. (2024). Sorek Desalination Plant: Harmful Algal Bloom Response and Lessons Learned. Kadima, Israel: IDE Technologies Ltd.
• Poseidon Water. (2024). Claude "Bud" Lewis Carlsbad Desalination Plant: Annual Performance Report 2023. Boston, MA: Poseidon Water LLC.
• Orange County Water District. (2025). Groundwater Replenishment System: Advanced Treatment Performance and Monitoring Report. Fountain Valley, CA: OCWD.
• Texas Water Development Board. (2025). Direct Potable Reuse Guidance Manual: Monitoring, Response, and Reporting Requirements. Austin, TX: TWDB.
• Global Water Intelligence. (2025). Desalination Markets 2025: Costs, Technologies, and Operational Benchmarks. Oxford: GWI.