The Real Water Story: Evaporative Cooling Economics

The Misleading Metric

Evaporative cooling carries a reputation for high water consumption. The objection is intuitive: evaporative systems consume water on-site, while direct expansion (DX) systems do not. This framing is incomplete.

A full-system resource accounting must include the water consumed upstream at the power plant to generate the electricity that DX systems require. When this upstream consumption is included, the resource comparison shifts dramatically—and the economic comparison shifts further still.

Power Plant Water Consumption

NREL’s analysis of water consumption for U.S. power production (Technical Report NREL/TP-6A20-50900) quantifies the upstream impact.¹ Thermoelectric power plants—coal, natural gas, and nuclear facilities that generate the majority of U.S. electricity—evaporate approximately 0.47 gallons of fresh water per kWh of electricity delivered to the end user.¹

The consumption varies by plant type and cooling configuration:

• Coal plants with closed-loop cooling: 500-600 gallons per MWh
• Nuclear plants with closed-loop cooling: 700-1,100 gallons per MWh
• Hydroelectric plants: 18 gallons per kWh average (reservoir evaporation)

The national weighted average across thermoelectric and hydroelectric generation reaches 2.0 gallons per kWh.¹

The Energy Efficiency Gap

The comparison between evaporative and DX cooling is not close. Evaporative cooling can use up to 80% less energy than direct expansion air conditioning.² The efficiency gap is expressed most clearly in Coefficient of Performance (COP):

System Type	COP Range
Direct evaporative coolers	4.9 - 23.3
Coolerado hybrid evaporative	12.1
Air-cooled DX rooftop (80-100 ton)	~2.8 (1.18 kW/ton)
Water-cooled chiller + tower	~4.7 (0.75 kW/ton)

Source: PNNL Building America Solution Center³, NREL Technical Reports⁴

A COP of 12 means the system delivers 12 units of cooling for every unit of electrical input. A COP of 2.8 means only 2.8 units of cooling per unit of electricity. The evaporative system is moving heat using the latent energy of water evaporation rather than the mechanical compression of refrigerant.

Net Resource Accounting

Consider a 100-ton cooling load in an arid climate (ASHRAE Zone 5B, Denver):

Evaporative system:

• Water consumption: approximately 2.5 gallons per ton-hour⁴
• 100 tons × 8 hours × 2.5 gal = 2,000 gallons per day on-site
• Electricity: approximately 0.1 kW/ton = 10 kW
• Upstream water (at 0.47 gal/kWh): 10 kW × 8 hours × 0.47 = 37.6 gallons

DX system:

• Water consumption: 0 gallons on-site
• Electricity: approximately 1.2 kW/ton = 120 kW
• Upstream water: 120 kW × 8 hours × 0.47 = 451.2 gallons

The evaporative system uses more water on-site (2,000 vs. 0 gallons) but consumes far less electricity (80 kWh vs. 960 kWh) and therefore far less upstream water (38 vs. 451 gallons). When water has a cost—whether utility charges, treatment requirements, or regulatory constraints—the economic comparison must include the embedded water in electricity.

Real Facility Performance

The most efficient data centers in operation rely heavily on evaporative and free-air cooling. Their measured Power Usage Effectiveness (PUE) demonstrates what the physics predicts:

Facility	PUE	Cooling Strategy
Meta (fleet average, 2024)	1.08	Evaporative, free-air
Google (fleet TTM, 2024)	1.09	Mixed; best sites use evap/free-air
Google Central Ohio (Lancaster)	1.04	Free-air economizer
Google The Dalles, OR	1.08	Evaporative
NREL ESIF Data Center	1.06	Evaporative, waste heat recovery
Industry average	1.56	Predominantly DX

Sources: Meta Sustainability Reports⁵, Google Data Centers⁶, DOE Better Buildings⁷, Uptime Institute⁸

Google’s documentation states directly: “Compared with the industry average of 1.56, Google data centers used about 84% less overhead energy for every unit of IT equipment energy.”⁶

Meta’s Prineville, Oregon facility operates year-round without mechanical cooling, even when summer temperatures reach 110°F—a testament to what evaporative systems can achieve in suitable climates.⁵

Climate Zone Suitability

Evaporative cooling effectiveness depends on wet-bulb temperature. The practical limits:

• Direct evaporative: Effective below approximately 70°F wet-bulb. Above this threshold, the cooling delivered diminishes rapidly.
• Indirect evaporative: Functional to approximately 75°F wet-bulb. Because indirect systems do not add moisture to the supply air, they maintain comfort in moderately humid conditions where direct evaporative would cause indoor humidity problems.
• Two-stage (indirect-direct): Can deliver supply air 5-10°F below outdoor wet-bulb by pre-cooling the incoming air before the direct evaporative stage.

ASHRAE climate zone suitability:

Zone	Example Cities	Recommended System
1A (hot-humid)	Miami, Houston	Supplemental evap only; DX primary
2A (hot-humid)	Atlanta, Dallas	Hybrid with DX trim
2B (hot-dry)	Phoenix, Tucson	Primary evaporative
3B (warm-dry)	Las Vegas, Los Angeles	Primary evaporative
4B (mixed-dry)	Albuquerque	Primary evaporative
5B (cool-dry)	Denver, Boise	Primary evaporative
5A (cool-humid)	Chicago, Boston	Hybrid; economizer hours valuable

The Western Cooling Efficiency Center’s Western Cooling Challenge established a threshold: qualifying systems must be at least 40% more efficient than DOE 2010 standards with outdoor ventilation of 120 cfm per nominal ton.⁹ The Coolerado H-80, the first certified winner, uses less than half the energy of conventional units while achieving 90-120% wet-bulb effectiveness.⁹

The Refrigerant Liability

Beyond energy economics, evaporative systems avoid an increasingly problematic asset class: distributed refrigerant.

EPA GreenChill data from the supermarket sector—which operates thousands of distributed refrigeration systems—reveals the leak rate reality: the industry average is 25% of refrigerant charge lost annually.¹⁰ GreenChill Partnership members achieve 13%, and best-in-class operators achieve below 10%.¹⁰ GreenChill partners emit at least 65% less refrigerant than the average supermarket.¹⁰

R-410A, the dominant HVAC refrigerant, has a Global Warming Potential of 2,088.¹¹ The EPA’s 2025 GWP limit for air conditioning equipment is 700.¹² R-410A exceeds this by a factor of three.

The AIM Act mandates an 85% reduction in HFC production and consumption by 2036.¹³ The phasedown is already affecting pricing and availability. Building owners specifying new systems today must consider a 20-year lifecycle in which refrigerant costs, availability, and regulatory compliance are uncertain at best.

Evaporative systems use water. Water is not regulated as an ozone-depleting substance or a greenhouse gas. There is no “water phasedown.”

Total Cost of Ownership

The capital cost advantage favors evaporative systems, though the differential varies by configuration and capacity. The operating cost advantage, in suitable climates, is substantial.

NREL’s Research Support Facility data center achieved a PUE of 1.2 after optimization—down from 2.28 in its legacy configuration—with 84% reduction in energy overhead.⁷ The 1,900-square-foot data center had accounted for 33% of the 222,000-square-foot facility’s annual electricity consumption before the upgrade.⁷

The Energy Systems Integration Facility at NREL achieves a PUE of 1.06 and recovers 97% of its supercomputer waste heat for building heating.⁷ This represents the upper bound of what integrated design can deliver.

Implications

The case for evaporative cooling in suitable climates rests on three arguments:

First, the energy efficiency gap is not marginal—it is a factor of 4-8x in COP terms. This translates directly to operating cost.

Second, when upstream water consumption is included, evaporative systems do not necessarily use more total water than DX systems. They use water in a different place.

Third, evaporative systems avoid the regulatory and economic uncertainty of distributed refrigerant. As HFC phasedowns accelerate and refrigerant costs rise, this advantage compounds.

For facility operators in ASHRAE zones 2B through 5B, the question is not whether evaporative cooling can work. The question is why, given the measured performance of the world’s most efficient facilities, any other approach would be specified.

---

Works Cited

1. A Review of Operational Water Consumption and Withdrawal Factors for Electricity Generating Technologies, NREL/TP-6A20-50900, accessed February 2025, https://docs.nrel.gov/docs/fy11osti/50900.pdf
2. Evaporative Cooling vs Direct Expansion, Seeley International, accessed February 2025, https://www.seeleyinternational.com/us/evaporative-cooling-vs-direct-expansion/
3. Evaporative Cooling Systems Resource Guide, PNNL Building America Solution Center, accessed February 2025, https://basc.pnnl.gov/resource-guides/evaporative-cooling-systems
4. Coolerado Cooler Performance Analysis, NREL Technical Report, accessed February 2025, https://docs.nrel.gov/docs/fy13osti/56256-1.pdf
5. Meta Data Centers Sustainability, Meta, accessed February 2025, https://sustainability.atmeta.com/data-centers/
6. Google Data Centers Efficiency, Google, accessed February 2025, https://datacenters.google/efficiency/
7. Data Center Optimization at NREL Research Support Facility, DOE Better Buildings, accessed February 2025, https://betterbuildingssolutioncenter.energy.gov/showcase-projects/data-center-optimization-nrel-research-support-facility
8. Global PUEs: Are They Going Anywhere?, Uptime Institute, accessed February 2025, https://journal.uptimeinstitute.com/global-pues-are-they-going-anywhere/
9. Western Cooling Challenge, NREL News, accessed February 2025, https://www.nrel.gov/news/program/2009/731.html
10. EPA Applauds Smart Refrigerant Management in Supermarkets, EPA Newsroom, accessed February 2025, https://www.epa.gov/newsreleases/epa-applauds-smart-refrigerant-management-supermarkets-across-america
11. Technology Transitions GWP Reference Table, EPA, accessed February 2025, https://www.epa.gov/climate-hfcs-reduction/technology-transitions-gwp-reference-table
12. Global Warming Potential Explained, Trane, accessed February 2025, https://www.trane.com/residential/en/resources/glossary/gwp-meaning/
13. Frequent Questions on Phasedown of Hydrofluorocarbons, EPA, accessed February 2025, https://www.epa.gov/climate-hfcs-reduction/frequent-questions-phasedown-hydrofluorocarbons