25 Reverse Osmosis Statistics

by

These reverse osmosis figures cover the measurement periods and populations identified in the available source material. Estimates, surveys and forecasts retain their reported scope.

Key Reverse Osmosis Statistics

The main findings are:

  • A typical point-of-use reverse osmosis system generates 5 gallons or more of reject water for every 1 gallon of treated water, while some inefficient units generate up to 10 gallons of reject water per gallon treated (EPA page accessed 2026; United States residential point-of-use systems).
  • Commercial and institutional RO units typically recover 50% to 75% of incoming feed water as permeate (March 2024 guidance; Commercial and institutional applications).
  • USGS reports that more than 300 million people globally receive water from desalination plants (USGS page accessed 2026; global figure reported by USGS from the International Desalination Association; Global desalination plants).
  • The DOE seawater-desalination bandwidth study used 2.7 kilowatt-hours electric per cubic meter as the state-of-the-art RO-unit intensity at the Carlsbad facility, with 35,000 ppm feed salinity and 50% recovery (2016 study baseline and 2017 report; Carlsbad, California, and U.S. seawater RO modeling).
  • In EPA’s Florida groundwater RO study, arsenic(V), calcium, cadmium, chromium(III), chromium(VI), copper, lead, molybdenum, sodium, radium, selenium(IV), selenium(VI), uranium, hardness, and TDS were classified as highly removed at above 95% (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study).
  • Replacing a typical point-of-use RO system with a WaterSense-labeled model reduces water use by more than 3,100 gallons per year on average (EPA page accessed 2026; annual estimate; United States residential point-of-use systems).
  • The 50%-75% commercial/institutional RO recovery range implies rejection of 25% to 50% of entering water (March 2024 guidance; Commercial and institutional applications).
  • The Claude Bud Lewis Carlsbad Desalination Plant produces 50 million gallons of fresh water per day for 110,000 customers in San Diego County (USGS page accessed 2026; current facility description; San Diego County, California, United States).
  • The DOE study used 4.00 kWh electric per cubic meter as a current-typical seawater RO intensity for 46% of U.S. potable seawater production, assuming 50% recovery and 35,000 ppm feed salinity (2016 study baseline and 2017 report; United States potable seawater desalination modeling).
  • The same EPA study classified fluoride, chloride, nitrate, and nitrite as moderately removed, at 85% to 94% removal (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study).
  • Lifetime WaterSense savings of 47,000 gallons when a typical point-of-use RO system is replaced with a WaterSense-labeled model (EPA page accessed 2026; system lifetime estimate; United States residential point-of-use systems).
  • Fully optimized commercial/institutional RO systems can achieve recoveries greater than 90% (March 2024 guidance; Commercial and institutional applications).
  • USGS reports that the Carlsbad plant serves a county where approximately 3.3 million people live (USGS page accessed 2026; current facility description; San Diego County, California, United States).
  • The DOE report cites a 1.58 to 9.00 kWh electric per cubic meter range for seawater reverse osmosis in the literature it reviewed (Literature reviewed in 2017 report; Seawater reverse osmosis literature).
  • The EPA study classified arsenic(III) and mercury(I) as poorly removed, below 85% (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study).

Contents

Residential point-of-use reverse osmosis water use

  • A typical point-of-use reverse osmosis system generates 5 gallons or more of reject water for every 1 gallon of treated water, while some inefficient units generate up to 10 gallons of reject water per gallon treated (EPA page accessed 2026; United States residential point-of-use systems). (Point-of-Use Reverse Osmosis Systems)
  • EPA estimates that replacing a typical point-of-use RO system with a WaterSense-labeled model reduces water use by more than 3,100 gallons per year on average (EPA page accessed 2026; annual estimate; United States residential point-of-use systems). (Point-of-Use Reverse Osmosis Systems)
  • EPA estimates lifetime WaterSense savings of 47,000 gallons when a typical point-of-use RO system is replaced with a WaterSense-labeled model (EPA page accessed 2026; system lifetime estimate; United States residential point-of-use systems). (Point-of-Use Reverse Osmosis Systems)
  • If every point-of-use RO system sold in the United States were WaterSense labeled, EPA estimates annual national savings of more than 3.1 billion gallons, equivalent to the annual household water needs of nearly 41,000 U.S. homes (EPA page accessed 2026; annual scenario). (Point-of-Use Reverse Osmosis Systems)
  • EPA’s WaterSense point-of-use RO performance criteria require the membrane to last at least 1 year before replacement (WaterSense specification described on EPA page, accessed 2026; United States WaterSense-labeled point-of-use RO products). (Point-of-Use Reverse Osmosis Systems)

EPA presents these as typical/example consumer-system figures, not a census of all installed systems.

Average modeled savings depend on EPA’s assumed typical-system usage and labeled-system performance.

This is an EPA lifetime estimate, not an observed follow-up of individual households.

Counterfactual national scenario based on EPA assumptions; it is not observed annual sales or installed-base performance.

Commercial and institutional reverse osmosis recovery

  • Commercial and institutional RO units typically recover 50% to 75% of incoming feed water as permeate (March 2024 guidance; Commercial and institutional applications). (WaterSense at Work Section 7.1:…)
  • The 50%-75% commercial/institutional RO recovery range implies rejection of 25% to 50% of entering water (March 2024 guidance; Commercial and institutional applications). (WaterSense at Work Section 7.1:…)
  • Fully optimized commercial/institutional RO systems can achieve recoveries greater than 90% (March 2024 guidance; Commercial and institutional applications). (WaterSense at Work Section 7.1:…)
  • EPA’s drinking-water technology overview states that RO and nanofiltration reject 15% to 30% of feed water entering the process (EPA overview accessed 2026; Drinking-water RO and nanofiltration processes). (Overview of Drinking Water Treatment…)

EPA describes a typical range drawn from cited best-management-practice references; it is not a universal equipment specification.

Direct complement of EPA’s stated recovery range; no additional measurement or arithmetic beyond that definition.

EPA describes an achievable optimized-system threshold, not typical performance across the market.

The range covers RO and NF together and is not limited to residential point-of-use systems.

Desalination scale and supply

  • USGS reports that more than 300 million people globally receive water from desalination plants (USGS page accessed 2026; global figure reported by USGS from the International Desalination Association; Global desalination plants). (Desalination)
  • The Claude Bud Lewis Carlsbad Desalination Plant produces 50 million gallons of fresh water per day for 110,000 customers in San Diego County (USGS page accessed 2026; current facility description; San Diego County, California, United States). (Desalination)
  • USGS reports that the Carlsbad plant serves a county where approximately 3.3 million people live (USGS page accessed 2026; current facility description; San Diego County, California, United States). (Desalination)
  • USGS reports 11 desalination plants in California and 10 additional plants proposed at the time of its page description (USGS page accessed 2026; status stated on page; California, United States). (Desalination)
  • USGS reports that the Jebel Ali Power Plant and Water Desalination facility in Dubai produces more than 588 million gallons of potable water per day (USGS page accessed 2026; facility description; Dubai, United Arab Emirates). (Desalination)
  • For 2015, reported saline surface-water withdrawals for public supply totaled 7.21 million gallons per day across the U.S. Virgin Islands, Texas, Florida, and Massachusetts. (Desalination)

USGS attributes the figure to the International Desalination Association and does not provide a plant-by-plant population census on the page.

Reverse osmosis energy intensity

  • The DOE seawater-desalination bandwidth study used 2.7 kilowatt-hours electric per cubic meter as the state-of-the-art RO-unit intensity at the Carlsbad facility, with 35,000 ppm feed salinity and 50% recovery (2016 study baseline and 2017 report; Carlsbad, California, and U.S. seawater RO modeling). (Bandwidth Study on Energy Use…)
  • The DOE study used 4.00 kWh electric per cubic meter as a current-typical seawater RO intensity for 46% of U.S. potable seawater production, assuming 50% recovery and 35,000 ppm feed salinity (2016 study baseline and 2017 report; United States potable seawater desalination modeling). (Bandwidth Study on Energy Use…)
  • The DOE report cites a 1.58 to 9.00 kWh electric per cubic meter range for seawater reverse osmosis in the literature it reviewed (Literature reviewed in 2017 report; Seawater reverse osmosis literature). (Bandwidth Study on Energy Use…)

DOE identifies this as a modeled state-of-the-art intensity and representative facility value, not a universal operating result.

This is DOE’s national modeling assumption for a production share, not a direct measurement from every U.S. plant.

The range aggregates literature values with differing plant conditions; it should not be treated as one standardized plant benchmark.

Contaminant removal measured in EPA studies

  • In EPA’s Florida groundwater RO study, arsenic(V), calcium, cadmium, chromium(III), chromium(VI), copper, lead, molybdenum, sodium, radium, selenium(IV), selenium(VI), uranium, hardness, and TDS were classified as highly removed at above 95% (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study). (Reverse Osmosis Treatment to Remove…)
  • The same EPA study classified fluoride, chloride, nitrate, and nitrite as moderately removed, at 85% to 94% removal (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study). (Reverse Osmosis Treatment to Remove…)
  • The EPA study classified arsenic(III) and mercury(I) as poorly removed, below 85% (Pilot study reported 1988; test runs varied by membrane and contaminant; Florida groundwater pilot study). (Reverse Osmosis Treatment to Remove…)
  • A short-term laboratory test cited by EPA found 97.8% atrazine removal with an NS-100 membrane versus 84.0% with a cellulose-acetate membrane (1975 laboratory test, as cited by EPA; Laboratory contaminated-water test). (Finalization of Guidance on Incorporation…)
  • EPA summarizes linuron removal from groundwater as 0% with cellulose acetate, 57% with polyamide, and 99% with thin-film-composite DSI membrane (Historical groundwater studies summarized by EPA; Groundwater samples). (Finalization of Guidance on Incorporation…)
  • EPA summarizes alachlor removal from surface water samples as 70% with cellulose acetate, 77% with polyamide, and 100% with thin-film-composite FT-30 membrane (Historical surface-water studies summarized by EPA; Surface-water samples). (Finalization of Guidance on Incorporation…)
  • EPA’s pesticide-class table reports thin-film-composite RO removal ranges of 80% to 100% for triazines, 98.5% to 100% for acetanilides, 100% for organochlorines, 98.5% to 100% for organophosphorus compounds, 99% to 100% for urea derivatives, and greater than 92.9% for carbamates (Historical studies summarized in EPA Table 11; RO membrane studies using contaminated water). (Finalization of Guidance on Incorporation…)

EPA used five membranes and non-identical operating conditions; this classification is study-specific and not a guarantee for every RO system.

The EPA summary notes wide variation for nitrate and limited nitrite testing; results depend on membrane, water chemistry, and conditions.

The classification is a study summary and EPA reports substantial variation for arsenic(III) and mercury.

Short-term laboratory result; it does not establish household or full-scale treatment performance.

Results compare membrane materials in study conditions and should not be generalized across current commercial systems.

Results compare membrane materials in study conditions and do not represent a market-wide product average.

These are ranges across cited studies and compounds; pesticide removal varies by membrane, compound, and water matrix.

Table of contents