GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Guest Blogs

The Methane Math for Tankless Gas Water Heaters

Methane emissions raise questions about the impact of these appliances on climate

Tankless gas water heaters hold a narrow efficiency edge over storage tank models, but they release unburned methane every time they cycle on and off. Photo courtesy Rinnai.

Once promoted as a more energy efficient option for providing hot water in homes and businesses than conventional tank-type water heaters, gas tankless water heaters may not be that much better for the climate when accounting for methane emissions—the primary component of gas and a climate-warming super pollutant—finds a study by Stanford researchers.

These tankless water heaters emit twice as much unburnt methane because they fire up their gas burners, and release a puff of methane, every time hot water is used. This is true even when only a small amount of water is used, like for hand washing or rinsing a cup.

In fact, unburnt methane from gas tankless water heaters has a much higher global warming potential (GWP) than carbon dioxide—86 times as much over a 20-year horizon. Small amounts of unburnt methane escaping to the atmosphere via the water heater flue add up quickly, offsetting much of the energy-saving benefits of tankless water heaters compared to tank-type gas heaters (also known as gas storage water heaters).

Water-heating is the largest use of gas in California homes, and second-largest in homes nationally after space heating. And the vast majority of new water heaters sold in the state today use gas. The tankless heaters save energy because they use electronic ignition instead of pilot lights, and they also don’t need to store water in a tank, where it gradually loses heat until it gets reheated. However, this conventional math only accounts for energy use, and carbon dioxide (CO2) emissions, the primary by-product of gas combustion. The Stanford study finds that gas tankless water heaters leak more than twice the amount of methane (CH4) as gas tank types.

Unbeknownst to the user, every time a tankless water heater fires up its burners, or switches them off, it emits a puff of methane, largely due to incomplete combustion:

Typical methane emissions when a tankless water heater switches on and off, showing the “On pulse” and “Off pulse” at the beginning and end of each hot water draw. Image courtesy Quantifying Methane Emissions from Natural Gas Water Heaters.

Tankless heaters also emit more than twice as much methane while the burners are on than tank heaters, adding up to 0.93% of unburnt methane in tankless vs. 0.39% in tank heaters.

This appears small but as noted earlier, methane is 86 times more powerful than CO2 from a heat-trapping perspective over a 20-year horizon. A small amount of leakage adds up to a large impact. In fact, while tankless water heaters are theoretically 22% more energy efficient than tank heaters, this advantage is in part offset by the higher methane emissions. NRDC estimates that when factoring in methane emissions rates per the study’s findings, tankless water heaters only reduce emissions by 14% vs. gas tank heaters, instead of the 22% conventional estimate.

This slight climate advantage of gas tankless over tank heaters is dwarfed by the 50% to 70% emissions reduction provided by electric heat-pump water heaters, and by the ultimate drop to zero emission as the electric grid becomes 100% clean, such as in the 13 states with 100% clean electricity goals. Heat pump models have another key benefit: they help clean up the air, eliminating air pollution from gas combustion that hurts public health.

Impact on climate and clean air goals

Gas tankless water heaters currently have a small market share, just around 2% of gas water heater sales in the United States. While the Stanford study has a relatively modest sample size, 35 water heaters, and the emissions numbers will be refined by larger studies, the direction is clear: methane emissions are a key piece of the climate equation.

Accounting for methane reinforces what other studies have already found: we cannot get anywhere close to our clean energy goals in buildings through energy efficiency alone, such as by moving from gas tank to tankless water heaters. Instead we must transition our homes and buildings off fossil fuels to clean and efficient electric space and water heating technology in the form of heat pumps, which are three to five times more efficient than the most efficient gas technology and can also help move to a 100% clean energy grid.

Starting with new construction

The first place to start is in new construction: we need to build homes right from the start to save future homeowners the much higher expense of upgrading later. This must include using clean and highly efficient heat pumps for space and water heating instead of fossil-fueled furnaces and water heaters.

California has a unique opportunity to lead in this area, building off the 34 cities and counties that have already adopted local building codes that require or strongly encourage electric construction, by shifting new construction from gas tankless to heat-pump water heaters.

It is time to stop building new homes with the gas tankless water heaters currently favored by the California energy code: they are only marginally more efficient than gas tank water heaters, and they lock in homeowners into higher emissions, pollution, and utility costs.

Less than 1% of all homes get built every year, but new construction has an outsize influence on the market (every new home needs new equipment, whereas water heaters only get replaced every 10 to 15 years in existing buildings). New construction is also the most cost-effective time to go electric.

Plan now for a water heater upgrade

Water heaters last for 10 to 15 years and can be expensive to replace. When your water heater gets close to its end-of-life, consider upgrading to an electric heat-pump water heater. Even if you currently use gas for water heating, an electric heat pump water heater is so much more efficient than even the best gas models, it will save energy and money over its life. Make sure to research available options and when possible, go ahead with the replacement before your water heater fails, as it is easier to upgrade when you can plan for it than in an emergency situation.

If you’re thinking of going for an electric tankless model, consider this: while they don’t have the methane problem, they use three times as much energy as heat-pump water heaters, costing you dearly, and they are arguably no better for the climate than gas water heaters. This is because they draw electricity, a lot of it, exactly when users need hot water, which tends to coincide with times of higher power demand on the grid and emissions from power plants. This makes it more difficult to transition the grid to 100% clean energy to clean our air and stave off the worst of the climate crisis. We need efficient and flexible heat-pump water heaters, not inefficient and inflexible electric tankless models.

Tips for minimizing methane emissions

If your water heater still has significant life in it, the Stanford study has an interesting tip on how to minimize those methane puffs every time your tankless water heater turns on and off.

You may be inadvertently causing the water heater to fire up and emit these methane puffs if you lift your sink’s or bathroom’s single faucet handle straight up, even when you don’t need hot water or don’t even wait for it. The researchers measured the angle required to trigger the water heater on six single-handle kitchen faucets, and found that all six triggered the tankless water heater when the handle was right in the middle.

Image courtesy Quantifying Methane Emissions from Natural Gas Water Heaters.

If you don’t need hot water, make sure to move your sink handle fully to the right so that it doesn’t inadvertently call for hot water.

This problem exists for all water heater types and eventually needs to be fixed by energy efficiency standards for faucets. In the meantime, you can greatly reduce these impacts by moving the handle to the right, and when the time comes, by upgrading to a clean energy and super-efficient water heater.

Methodological Notes:

Comparing minimum efficiency devices with 0.81 and 0.63 uniform energy factor respectively, which are the most-commonly sold models.

The net greenhouse gas impact of 14% differs from the study’s 18% because NRDC uses the more common minimum efficiency models instead of the average efficiency of the sample in the Stanford study.

Pierre Delforge is senior scientist, building decarbonization, Climate & Clean Energy Program, at the Natural Resources Defense Council. This article originally appeared at the NRDC Expert Blog and is republished here with permission.


  1. Jon_R | | #1

    > methane emissions when a tankless water heater switches on and off

    Don't let "minor tweaks" distract from the really important issues.

    > It is time to stop building new homes with the gas tankless water heaters

    It's time to eliminate gas (and propane and oil) from homes entirely. Go all-electric.

  2. Expert Member
    Peter Engle | | #2

    Interesting article. I've never been a big fan of tankless water heaters except in some specific applications. This new information just gives me more ammunition. All electric just keeps looking like a better way to go.

    1. JC72 | | #3

      The problem with all electric is reliability and reparability of electric ranges/ovens. On one spectrum gas ranges are the cheapest to acquire and repair and on the other spectrum you have induction ranges which are very expensive. The cheapest induction range is usually 3x the cost of the cheapest gas range.

      Electric resistance ranges fall in the middle. Simply the worst of both worlds.

  3. BirchwoodBill | | #4

    Heat pumps with some buffer tanks and solar PV panels appears to be the most viable approach. Would still have natural gas fired generator for emergencies. Natural gas has its place in a balanced energy portfolio, needs to be carefully applied.

    1. Jon_R | | #6

      It will be interesting to see if "heat pump + water storage" progresses beyond domestic hot water and gets used for space heating. If not, even more expensive grid upgrades will be needed. The 7pm load in a cool climate in January will substantially exceed the current Summer peaks.

      1. _Stephen_ | | #12

        Can you expand on this Jon?

        ACs are a lot less efficient than equivalent heat pumps would be, due to the waste heat produced by the compressor being pumped into the house instead of being waste, so I would expect that to help. Peak heating load will be in the middle of the night when it's coldest, not at 7 pm, unless you're worried about the superposition of heating demand and normal peaks?

        1. Jon_R | | #13

          > ACs are a lot less efficient than equivalent heat pumps

          In a cool climate, say the AC (Summer cooling) has a max delta-T of 25F (inside to outside) and a heat pump heating has a max delta-T of 70F. The latter will use more peak power. Unless time shifted, this means grid upgrades.

          > the superposition of heating demand and normal peaks
          Yes, talking about peak usage of the entire grid. Can't see that ever being middle of the night.

          1. _Stephen_ | | #14

            Ah, thanks!

      2. Expert Member
        Peter Engle | | #15

        Onsite storage, whether in the form of heat or electricity (or something else), won't really take off until residential utilities adopt smart meter and time-of-day billing. If they charge 2x as much for electricity at 7PM as they do at 7 AM, people will invest in technology to shift their costs. Until then, not very likely.

        1. Jon_R | | #16

          I agree. Unfortunately, politicians are afraid of irritating the utilities and fossil fuel suppliers. So we don't get enough dynamic pricing, coal plant shutdowns, utility scale solar, wind, national scale smart-grids, fossil fuel prices that reflect their true cost, etc. Fix these and the right details (like less water heater gas leakage) will follow.

        2. TimTuckerCom | | #19

          Time of day billing as a solution isn't without its drawbacks.

          It assumes that the best times to shift from / to remain relatively constant - or that people will accept / adapt to changes in the schedules for what constitutes a "peak" time.

          In colder climates, HVAC costs are already shifted to later in the evenings since it get colder at night.

          Widespread use of electric vehicles can also shift load to night.

          More people working from home with less fixed hours due to COVID has also shifted load in both time and location in ways that would have been hard to predict even a year ago.

          It can be particularly hard on low income residents in older homes with older/inefficient HVAC equipment and appliances that making up a higher percentage of their usage. For them, there's a lot less that they can do to shift usage to a different time of day other than just turning their HVAC systems down / off and trying to do without.

          It's also pretty common for them to be renters - meaning they can't make efficiency updates even if they had the money to invest. From the perspective of landlords, as long as the majority of the housing stock in the area is similar, there's little to no incentive for them to make upgrades as long as they're not the ones paying the utility bills.

          1. Kurt99 | | #21

            A billing model where the retail rate is the wholesale rate that the utilities buy the power at plus a markup. Not a straight time base. The utilities would be required to have a machine friendly website with projected electric rates hour by hour for the next day or maybe the next week. The algorithm would presumably use time of day, weather forecasts, day of week, holidays, etc. to predict the rates. Most appliances have microprocessors already so new ones could be upgraded to get rate information and choose the optimal time to run with little additional cost.

            When you set up your dishwasher, you make your default to be done by 5 pm so they will be clean by the time you come home from work. The machine determines the best time to start, whether it is immediately, 1 a.m., 11 a.m. or some other time. All that is important to you is that it is done by the time you get home to fix dinner. Most of the large energy consumers in the house can have there operating time adjusted to move some or all their power consumption to lower rate times of day. I can't help be believe this would be much cheaper than the massive energy storage that would be required to power our current grid with all renewable power sources and our few nuclear plants.

          2. TimTuckerCom | | #22

            Kurt -- The heart of what you're describing is referred to in industry as Demand Response, which is one of many elements that utilities are incorporating into "smart grids" as they upgrade their infrastructure:

            Note that having the devices figure out on their own when to shift load turns into a variant of the Prisoner's Dilemma from Game Theory -- if the predictions say that 2pm will be the cheapest rate, without coordination between the devices the most natural thing for them to do with that information would be to shift their load to 2pm.

            Worst case you have something like 5,000,000 AC motors timed to start operation at exactly the same time and the high startup current draw causes a brownout.

            The general solution to that is to introduce some 2-way communication between devices and something that coordinates telling them what to do (i.e: utility tells Google it wants to shift a target of 5 MW from 2-3pm, then Google randomly chooses 50% of people with a Nest in the each zipcode to send a signal to reduce their setpoint by 2 degrees starting at a random time between 1:45-2 and ending at 3).

            There are utilities running pilots with EV companies and thermostat manufacturers to test out the general idea, but the infrastructure to support making those types of predictions and decisions all the way down to the circuit level is still being built out in most places.

            From a more human standpoint, even basic questions like "What percent of customers will opt in to a demand response program?", "How often will customers choose to override the signals to shift use?", and "Do customers trust utilities or tech companies more to control their devices?" are still being researched.

            Note that you also still have the issue I mentioned above -- the gap between customers who can afford newer appliances with more advanced electronics and ones who either can't afford to upgrade their appliances or can only afford to rent apartments / homes with older appliances provided. Is it good economic policy to penalize someone for being unable to buy a more efficient dishwasher?

        3. joshdurston | | #20

          We need to see the true costs and savings at the consumption level to warrant investment. Here in Ontario we've literally paid to export electricity and peak rates can exceed $1kw occasionally but consumers don't see that. It would really help to justify storage by allowing the savings (and expenses) to trickle thru the system.
          Personally I thinks its crazy that every house gets a 100-200a service. I wish I could get a 15-20a service and use a local battery buffer. In exchange for guaranteeing my peak won't exceed 20amps, I should get a really good price. All the infrastructure is designed to carry the peak load back to the generation source, which means, more wires, bigger conductors, bigger transformers, switch gear etc.
          I have a whole home power meter and it's rare that I exceed 2000w so I would happily take a battery buffered 3000w service at say $0.05kw fixed rate. Internally I would have the infrastructure to surge to maybe 7000w.

  4. guilfordvtbuilder | | #5

    Yet another blow to living off grid...

    I have a new “energy efficient” Tankless condensing propane Water heater. Are they better from a environmental stand point? Do they give off less methane or more?

    1. Expert Member
      Peter Engle | | #7

      They don't give off methane. They give off propane, in roughly equal amounts. It's the puff of unburnt gas at the beginning and end of each cycle that the article discusses.

    2. JC72 | | #8

      The GWP of propane is only 3x that of CO2. I wouldn't sweat it. Just keep up with the required maintenance of the unit.

      1. Expert Member
        Michael Maines | | #9

        Because CO2 is the standard unit of measurement for greenhouse gasses, people seem to think that parity (or close to it) is nothing to be concerned about. Propane has 3.3 times the GWP of CO2, while methane (aka natural gas) has about 25 times the GWP of CO2. In all cases, GWP is not the best measurement to use, because it is based on a 100-year scale. We have a decade, maybe two at most, to get carbon emissions under control, not 100 years. On shorter time scales, greenhouse gas impact is magnified. I wish there was a GWP normalized for 10- and 20-year scales but I don't know of any.

        Living off grid CAN be a lower-impact way to live, but if you are trying to maintain a conventional American lifestyle, you will do less damage by being part of a community with shared infrastructure.

        1. b_coplin | | #10

          "In all cases, GWP is not the best measurement to use, because it is based on a 100-year scale."

          The usefulness of the time scale depends on what one is trying to communicate. I agree the 100 year time scale is not the best measurement to use in terms of communicating the severity of the problem of climate change or the paucity of time to solve it.

          If the goal is make a comparison of various fuels, GWP100 is a well understood and longstanding measurement. Using a GWP20 scale increases complexity and decreases transparency of the metrics, and would require reaching a consensus on methodology, as discussed in this paper:

          I'm not trying to split hairs, just point out the GWP100 metric is useful, if imperfect.

          1. Expert Member
            Michael Maines | | #11

            Bryan, thank you, that is a clear and more comprehensive response than mine. I agree that GWP100 is long-standing and useful for some things, and that GWP20 (or 10) is a different beast. But in the face of climate change, using a 100-year scale to measure how much our emissions are impacting the climate just doesn't make sense. We are essentially measuring the wrong thing and using the answers to make big decisions.

  5. CraigRo | | #17

    Interesting to note, from the refinery to the power plant to my house, to heat with a heat pump: roughly 1.7x becomes the efficiency of the HP when taking all factors (transmission losses, power plant efficiency) into account instead of closer to 3 COP.

    Conversely, my boiler is roughly 85% at converting NG to BTUs. So technically the HP is more efficient, "stealing" half my needed btus from the atmosphere. BUT: electrical rates are 20¢/KWh, and NG is $1.05/therm. Math says that's 5x difference in $/BTUs, which is not an economically smart model.

    I really want HP and electric systems to be viable, but lack of renewable energy, and energy source cost disparity, we're decades away from AK converting.

  6. AndrisSkulte | | #18

    I know this is nit-picky, but this statement isn't really true: "Heat pump models have another key benefit: they help clean up the air, eliminating air pollution from gas combustion that hurts public health."

    HP hot water heaters aren't cleaning the air - they just don't make it dirtier. That's like my four year old saying his messy room got cleaned because he didn't make it worse today. :)

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |