A heat pump water heating plant that cuts a building's site energy by two thirds can leave its source energy nearly unchanged and its emissions somewhere in between, depending on which grid serves the building. That is three different metrics measuring three different things, and every building performance standard picks one.
Washington and Oregon set site EUI targets. New York City's Local Law 97, Colorado, Maryland and Boston set greenhouse gas limits. The ENERGY STAR score, which several programs reference, is built on source energy.
An owner planning hot water electrification needs to identify the building's scoring currency before the engineer runs the calculation. The same project moves each metric by a different amount.
In short, site EUI counts the energy that crosses the meter, so a heat pump wins by its COP. Source EUI multiplies electricity by 2.80 and natural gas by 1.05 in ENERGY STAR Portfolio Manager, so a heat pump replacing an 80 percent boiler breaks even only above a COP of about 2.13 and gains slowly beyond it.
Greenhouse gas intensity multiplies by the emissions factor of the building's eGRID subregion. Electricity ranges from 57.16 kgCO2e per MBtu on the Southwest Coast to 115.09 in New York City, against 53.11 for gas. The break-even COP therefore ranges from about 0.86 to 1.73.
On the worked example, converting 1,000 MMBtu of annual hot water from gas to a COP 2.5 heat pump cuts site energy by 68 percent, source energy by 15 percent and emissions by 31 to 66 percent depending on the grid. Thermal storage changes none of these annual figures directly; its value sits in the demand charge, the service size and the plant that fits the room.
In this article
The three currencies
Portfolio Manager defines site energy as the energy consumed at the building as shown on utility bills. Source energy is the total raw fuel required to operate the building, including losses in generation, transmission and distribution.
EPA's technical reference for source energy states that it is “most equitable to employ national-level source-site ratios,” so that no building is credited or penalized for the efficiency of its utility provider. It also states that geography is fixed and the interconnected grid prevents electricity from being traced to a specific plant.
The same document states that source energy and the ENERGY STAR score are focused on energy consumption, not on the supplier or resulting emissions. Green power purchases do not change source energy or the ENERGY STAR score.
Greenhouse gas intensity is the third currency. Portfolio Manager converts each fuel with an emissions factor. For electricity, it maps the building to an eGRID subregion by ZIP code or, where a ZIP spans subregions, by distribution utility.
Portfolio Manager calculates location-based emissions and added a market-based calculation in 2024. A city may use Portfolio Manager factors, fixed coefficients or a projected grid trajectory. The metric type is the same, but the numbers differ.
The building performance standards covered on this site show the split. Washington's Clean Buildings Performance Standard and Oregon's standard set site EUI targets built on ASHRAE Standard 100.
New York City's Local Law 97 sets emissions limits per square foot computed with fixed coefficients. Colorado and Maryland set emissions-based standards with alternative pathways.
The limits and dates are in each article. An owner with buildings in Seattle and Baltimore is being scored in two different currencies for the same project.
The conversion factors, and what they do to a heat pump
The factors determine how a heat pump appears under each metric. They are listed in full below.
The Portfolio Manager thermal conversions are 3.412 kBtu per kWh and 100 kBtu per therm.
- Site-to-source ratios, Portfolio Manager, United States:
- Electricity, grid purchase: 2.80
- Natural gas: 1.05
- District steam: 1.20
- District hot water: 1.20
- District chilled water: 0.91
- Fuel oil: 1.01
- Propane: 1.01
- Emissions factors, Portfolio Manager, kgCO2e per MBtu:
- Natural gas: 53.11
- District steam and district hot water: 66.40
- Fuel oil No. 2: 74.2
- Propane: 61.95
- Electricity, eGRID 2023 subregion NEWE, New England: 72.21
- Electricity, RFCE, Mid-Atlantic: 79.65
- Electricity, CAMX, Southwest Coast: 57.16
- Electricity, NYCW, New York City: 115.09
- Electricity, U.S. average: 102.48
Break-even COP by metric
The break-even COP is the point at which the heat pump plant scores the same as the boiler it replaces. For a boiler at efficiency η and a factor ratio F, defined as the electricity factor divided by the gas factor, the break-even COP is η × F.
At 80 percent boiler efficiency, the break-even values are:
- Site energy, F = 1.0: break-even COP 0.80. Any heat pump wins.
- Source energy, F = 2.80 / 1.05 = 2.67: break-even COP 2.13.
- Emissions, U.S. average, F = 102.48 / 53.11 = 1.93: break-even COP 1.54.
- Emissions, New York City, NYCW, F = 2.17: break-even COP 1.73.
- Emissions, Mid-Atlantic, RFCE, F = 1.50: break-even COP 1.20.
- Emissions, New England, NEWE, F = 1.36: break-even COP 1.09.
- Emissions, Southwest Coast, CAMX, F = 1.08: break-even COP 0.86.
Why source-energy performance can differ from site-energy performance
The source-energy break-even of 2.13 is the number that surprises owners. A central heat pump water heating plant that spends a winter running at a COP below 2 in a cold mechanical room, or that relies on resistance backup, shows as a source-energy increase in Portfolio Manager even while site EUI falls.
A condensing boiler at 95 percent raises every break-even by 95 / 80 = 1.19: source energy becomes 2.53 and U.S. average emissions become 1.83.
Measured annual COPs at central plants are the subject of the central heat pump water heater article. The metric sets the COP bar the plant has to clear.
Worked example: 1,000 MMBtu of hot water
Labeled assumptions: a building delivers 1,000 MMBtu per year of thermal heat to domestic hot water, measured at the water; the existing plant is a gas boiler at 80 percent; the replacement is a heat pump plant at an annual COP of 2.5, with 1.9 and 3.2 as sensitivity cases; and building area is 200,000 square feet for intensity figures. Thermal and electrical energy are labeled throughout.
- Gas baseline:
- Site gas: 1,000 / 0.80 = 1,250 MMBtu, or 12,500 therms; 6.25 kBtu per square foot of site EUI from hot water.
- Source: 1,250 × 1.05 = 1,312.5 MMBtu.
- Emissions: 1,250 × 53.11 = 66,388 kgCO2e, or 66.4 metric tons; 0.33 kgCO2e per square foot.
- Heat pump at COP 2.5:
- Site electricity: 1,000 / 2.5 = 400 MMBtu, which is 400,000 / 3.412 = 117,228 kWh electrical; 2.0 kBtu per square foot.
- Source: 400 × 2.80 = 1,120 MMBtu.
- Emissions, U.S. average: 400 × 102.48 = 40,992 kg, or 41.0 metric tons.
- Emissions, NEWE: 400 × 72.21 = 28,884 kg, or 28.9 metric tons.
- Emissions, RFCE: 400 × 79.65 = 31,860 kg, or 31.9 metric tons.
- Emissions, CAMX: 400 × 57.16 = 22,864 kg, or 22.9 metric tons.
- Emissions, NYCW: 400 × 115.09 = 46,036 kg, or 46.0 metric tons.
- Change from the gas baseline at COP 2.5:
- Site energy: 1,250 to 400 MMBtu, down 68 percent.
- Source energy: 1,312.5 to 1,120 MMBtu, down 15 percent.
- Emissions: down 38 percent on the U.S. average grid, 56 percent in New England, 52 percent in the Mid-Atlantic, 66 percent on the Southwest Coast and 31 percent in New York City.
- Sensitivity at COP 1.9:
- Site electricity: 526.3 MMBtu, or 154,250 kWh electrical, down 58 percent on site energy.
- Source: 526.3 × 2.80 = 1,473.7 MMBtu, up 12 percent against the boiler.
- Emissions, U.S. average: 526.3 × 102.48 = 53,935 kg, or 53.9 metric tons, down 19 percent; NEWE: 38.0 metric tons, down 43 percent.
- Sensitivity at COP 3.2:
- Site electricity: 312.5 MMBtu, or 91,588 kWh electrical, down 75 percent.
- Source: 875 MMBtu, down 33 percent.
- Emissions, U.S. average: 32.0 metric tons, down 52 percent; NEWE: 22.6 metric tons, down 66 percent.
Interpreting the sensitivity cases
The spread between COP 1.9 and 3.2 is a 12 percent source-energy increase at one end and a 33 percent decrease at the other, from the same tank, piping and building.
Under a site EUI standard, the project is a clear reduction at every COP. Under a source-energy metric, it is a reduction only if the plant is designed and operated to hold a COP above 2.13 across the year. Under an emissions standard, it is a reduction at any realistic COP on every listed grid, and a larger one each year a grid decarbonizes if the program updates its factors.
What storage does and does not change in these metrics
Thermal storage does not appear in any of the three formulas. It does not change delivered heat, and it changes COP only indirectly. Its effects are listed below in order of size.
- Demand and service, not energy. Charging storage overnight moves kW out of the billed window and lets the heat pump plant be sized to the daily average instead of the morning peak. That appears in the demand charge and electrical service calculation, covered in the demand charge article and NEC 220.87 article, not in the EUI or emissions line.
- COP, through operating conditions. A plant that runs long, steady charging cycles at night instead of short cycles against the peak tends toward its rated COP. A plant that stores at a higher temperature than needed pays for it in COP. The 140 F storage article sets out the Legionella trade-off, and the stratification article explains why a mixed tank delivers less usable heat for the same charge.
- Standby loss, small. The TEHQ 700-gallon module loses 2.4 F per 24 hours under free convection. At 700 gallons and 8.34 pounds per gallon, that is 700 × 8.34 × 2.4 = 14,011 Btu per day, or about 5.1 MMBtu per year per module, equal to 0.5 percent of the 1,000 MMBtu load in the example. The heat loss article has the full calculation.
- Hour of charging: not credited. Portfolio Manager's factors are annual. A plant that charges at 2 a.m. on a windy night and one that charges at 5 p.m. on a gas-peaking afternoon report the same kgCO2e because the eGRID factor is an annual average for the subregion. Programs that adopt hourly or marginal factors would change that; none of the programs covered on this site does today. Re-verify before claiming otherwise.
Owner sequence
Find the metric, compute the break-even COP, specify a plant that clears it with margin, then size the storage for the demand charge and service. The peak shaving article explains how the final step is decided.
Reading a Portfolio Manager report after electrification
Check three items during the first full year of data after a hot water conversion.
- Site EUI should fall by roughly the boiler's gas use times (1 − 0.80 / COP). If it fell less, the COP is lower than modeled or the gas meter still carries another load.
- Source EUI will fall far less than site EUI because of the 2.80 electricity ratio. A board shown a site EUI projection will see a smaller change in a source EUI report and should be told why in advance.
- Emissions depend on the eGRID edition. Portfolio Manager now uses eGRID 2023; a program with its own coefficients may lag or lead. Boston, for example, publishes a projected grid factor by year. Keep the factor and its vintage in the project file.
Frequently Asked Questions
What is the difference between site EUI and source EUI?
Site EUI is the energy delivered to the building per square foot as metered. Source EUI multiplies each fuel by a site-to-source ratio that adds energy lost in generation and delivery: 2.80 for grid electricity, 1.05 for natural gas, and 1.20 for district steam and hot water in Portfolio Manager's U.S. factors.
Does electrifying hot water lower a building's EUI?
It lowers site EUI whenever the heat pump's COP exceeds the boiler's efficiency, which is always for the 80 percent boiler assumption. It lowers source EUI only when the COP exceeds about 2.13 against an 80 percent boiler, or 2.53 against a 95 percent condensing boiler, because the 2.80 electricity ratio is 2.67 times the 1.05 gas ratio.
Which metric do building performance standards use?
Washington and Oregon use site EUI targets. New York City, Colorado, Maryland and Boston use greenhouse gas limits with their own coefficients or pathways. Programs that reference the ENERGY STAR score use a source-energy basis. Each TEHQ article on those programs carries the specific limits.
What emissions factor does Portfolio Manager use for electricity?
Portfolio Manager uses an eGRID 2023 subregion factor assigned by ZIP code or distribution utility. Examples are 72.21 kgCO2e per MBtu in New England, 79.65 in the Mid-Atlantic, 57.16 on the Southwest Coast, 115.09 in New York City and 102.48 for the U.S. average. Natural gas is 53.11 kgCO2e per MBtu.
Can a heat pump increase a building's emissions?
Only at a COP below the break-even for its grid: 1.54 on the U.S. average, 1.73 in New York City, 1.09 in New England and 0.86 on the Southwest Coast against an 80 percent boiler. Any plant operating near its rated COP is above those values on every listed grid.
Does thermal storage reduce EUI or emissions?
Not directly. Thermal storage does not change delivered heat, and annual factors do not credit the hour of charging. Its effects are on demand charges, electrical service size and operating COP, plus a small standby loss of about 5.1 MMBtu per year for a 700-gallon module at 2.4 F per 24 hours.
Why does Portfolio Manager use national source ratios?
EPA's technical reference states that national ratios are the most equitable approach because no building is credited or penalized for its utility's efficiency, and because electricity on an interconnected grid cannot be traced to a specific plant.
Does buying renewable energy certificates change source EUI?
No. Portfolio Manager states that green power purchases do not change source energy or the ENERGY STAR score, which address consumption. Market-based emissions, added in 2024, are where certificates register, and each program decides whether to accept them.
The bottom line
The same heat pump water heating plant is a 68 percent reduction, a 15 percent reduction or a 31 to 66 percent reduction depending on whether the standard counts site energy, source energy or emissions.
The factors are public: 2.80 against 1.05 for source energy, and 53.11 for gas against 57.16 to 115.09 for electricity by grid. Find the metric first, compute the break-even COP, and specify a plant that clears it with margin across the year.
Then size the storage, because that is where the demand charge, service size and mechanical room are decided. None of those variables appears directly on the compliance form.
The source-energy ratios are from the ENERGY STAR Portfolio Manager Technical Reference: Source Energy, August 2023, with ratios last revised in 2018. The emissions factors are from the Portfolio Manager Technical Reference: Greenhouse Gas Emissions, August 2025. Re-verify both documents and the Thermal Conversions reference on publish day; EPA updates the eGRID edition annually.


