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
Energy Solutions

How Much Insulation Is Enough?

Recommended residential insulation levels by climate zone

Image 1 of 2
Most leading energy experts today recommend installing a lot more insulation than is common practice. This wall for a Passive House in Seattle will hold about a foot of insulation.
Image Credit: Dan Whitmore
Most leading energy experts today recommend installing a lot more insulation than is common practice. This wall for a Passive House in Seattle will hold about a foot of insulation.
Image Credit: Dan Whitmore
The United States is divided into seven climate zones by the U.S. Department of Energy and the International Energy Conservation Code.
Image Credit: Building Energy Codes Resource Center

I’m often asked the question, “How much insulation should I install in my house”? It’s a great question. Let me offer some recommendations:

First of all… it depends. It depends to a significant extent on where you live. And it depends on whether we’re talking about a new house or trying to squeeze insulation into an existing house.

To simplify the discussion, let’s assume, for the time being, that we’re talking about new construction

As for location, I’ll provide recommendations for three different climates, based on the U.S. Department of Energy (DOE) and International Energy Conservation Code (IECC) climate zones. These DOE climate zones range from Zone 1 at the extreme southern tip of Florida, to Zone 7, which covers the tip of Maine, stretches across the northern reaches of Michigan, Wisconsin, Minnesota, and North Dakota, and includes a few high-elevation places in the Rockies (see map). In my recommendations, I group these into three larger zones for simplicity.

Cold climates: Zones 5-7

Zones 5-7 cover much of the northern half of the U.S., from roughly the Mason-Dixon Line at the East Coast across the northern third of Missouri and the northern edge of eastern Kansas, then dipping south in the higher-elevation Plains States through northern New Mexico, northern Arizona, nearly all of Nevada (except the Las Vegas area), and the northeastern corner of California and the eastern three-quarters of Oregon and Washington.

For these locations, I follow the widely quoted recommendations from Building Science Corporation and aim for the 5-10-20-40-60 rule. These numbers refer to the R-value recommendations for windows, foundation slabs, foundation walls, above-ground walls, and attics (or roofs), respectively.

These recommendations are for “true” R-values, not the nominal values listed on insulation packaging. For example, if you install R-19 fiberglass batts in 2×6 frame walls, with the studs 16 inches on-center, double top-plates, and other elements of “standard” framing, the actual R-value of the whole wall with the R-19 insulation will be about R-15. The whole-wall R-value is lower because of thermal bridging through the wood framing.

To achieve R-40 in the walls requires a lot of insulation — far more than is found in standard construction. This level of insulation, if combined with strategies for minimizing air leakage, will result in a house that will be affordable to heat even if energy prices double or triple. And if combined with some passive solar heating will result in a house that should never come close to freezing in winter, even if the heat is turned off.

With window R-values, the recommendation refers to the “unit R-value,” a measure that averages the center-of-glass R-value and the R-value at the window edges — where the heat loss is greater (at least with high-performance windows). These unit R-values are the inverse of the U-factors listed on NFRC (National Fenestration Rating Council) labels found on most new windows: R = 1/U.

Hot climates: Zones 1-2

Zones 1-2 include the hottest areas in the U.S., covering most of Florida and a band west to central Texas, as well as southern Arizona and the Imperial Valley of extreme southeastern California.

Here, I recommend a 3-5-10-20-60 rule: R-3 windows, R-5 under slabs and for any below-grade foundation walls, R-10 for above-grade foundation walls and slab perimeter (full foundations are rare in these climates), R-20 for above-ground walls, and R-60 for attics. These recommendations come from an informal conversation with John Straube of Building Science Corporation. Again, these are true R-values (unit values for windows).

It will surprise some to see the recommendation for attics to be the same as in cold climates. This is because of the difference in temperature (delta-T) between the living space and the attic on a hot summer day can be as high as wintertime delta-T in a cold-climate between indoors and outdoors. With windows, I further recommend a solar heat gain coefficient (SHGC) of 0.3 or lower to minimize unwanted solar gain.

Moderate climates: Zones 3-4

Zones 3-4 include much of the southern half of the country, with the boundary between Zones 4 and 5 dipping south across the center of New Mexico and Arizona. This moderate region excludes Florida and the Gulf Coast, but includes most of California and the western edge of Oregon and Washington.

For these locations, I recommend intermediate insulation values between those for cold climates and hot climates. I suggest a 4-5-10-30-60 rule: R-4 windows, R-5 under slabs, R-10 foundation walls or slab perimeter, R-30 above-grade walls, and R-60 in the attic or roof.

What about existing houses?

In new construction, the incremental cost of increasing insulation levels are relatively modest. With existing houses, retrofit insulation costs are usually much higher, so it is usually difficult to justify such high insulation levels. The exception is attics, where adding lots of additional insulation is usually quite affordable.

So, in existing homes, determining reasonable insulation levels is project-specific. In a full gut-rehab (where the house is taken down to the structure, or the frame is opened up on either the interior or exterior), achieving close to the recommended insulation levels for new construction may be possible (though higher costs for extending window and door jambs and, sometimes, roof overhangs also need to be considered).

And with windows, whether to replace or improve existing windows is a key question. Look for recommendations in future blogs.

Alex is founder of BuildingGreen, Inc. and executive editor of Environmental Building News. He also wrote Insulation: The BuildingGreen Guide to Insulation Products and Practices, which provides in-depth guidance on the selection of insulation materials. To keep up with his latest articles and musings, you can sign up for his Twitter feed.


  1. user-757117 | | #1

    Thumbs up.
    Those sound like "pretty good" recommendations to me.

  2. user-723121 | | #2

    I Agree

    It is good you pointed out true R-values, this is important especially for single walls. What about some ACH50 targets for the zones you have listed.

  3. Alex Wilson | | #3

    Air tightness targets
    Here's what I proposed in our just-posted EBN feature article, "Resilient Design: Smarter Building for a Turbulent Future":

    Aim for 1.5 air changes per hour at 50 pascals or 0.15 cfm per square foot of building shell at 50 pascals in Climate Zones 5–7, as measured with a blower door. In milder climates, 2 ACH (0.2 cfm/sf of shell) at 50 pascals is adequate.

    I'd be interested in input on airtightness from others.

    That article is available here, though it requires a log-in:

  4. user-1047602 | | #4

    R Value in Cold Weather Climates

    In cold weather climates 5 - 7. Could you speak to fibrous insulation r values not taking into consideration convection. As in we can't compare cellulose/fiberglass r value levels to foam r value levels since foam do more to stop convection. At least this what i am coming to understand. Please advise!

  5. Alex Wilson | | #5

    Convection in fiber insulation
    From my understanding, this is really a non-issue today with most fiber insulation--except for batt insulation, where poor installation may leave corners and gaps where convection occurs. With older lose-fill fiberglass insulation in attics, convection in very cold weather significantly degraded the R-value, but that is far less of a problem today, even with loose-fill fiberglass in attics. It should make almost no difference with damp-spray or dense-pack cellulose.

  6. jklingel | | #6

    too light
    I am no expert, but from all I can gather reading and looking at my own rough calcs, the 5-10-20-40-60 rule is awfully light on the dirt end. Ground is pretty chilly 24 x (most of) 365 in colder zones, and, especially if one has slab heat, then it "feels like" the dirt numbers ought to be maybe twice as high.

  7. Gregory La Vardera | | #7

    not enough below slab
    I think your recommendations for below slab insulation are way too low, certainly out of proportion with your other locations.

  8. gusfhb | | #8

    I know John is in Alaska, where the frost line is closer to the earth's core than the floor, but I think a lot of people are putting pretty crazy levels of insulation under slabs.
    There was a link to a guy in maybe Ohio a few months back that I wanted to read more about who built a pretty well 'super' insulated house with no sub slab insulation. I will try to find it. I do not know how that worked out. I kn w the PH guys go crazy, but in most places with a max delta T of less than 20, how much insulation do you need?

  9. jklingel | | #10

    still waiting
    Keith: as for "...but in most places with a max delta T of less than 20, how much insulation do you need?
    ..." I'm still waiting for The Guru to develop an exacting method to calc floor heat losses. In the meantime we take our best-guess model and run the numbers. Calc the heat loss given that that delta T is over 12 months (+ or -) instead of 7 or so, and you may be surprised. Couple that with the fact that IF you discover later that your R20 is light, what are you going to do about it but cry? It is a gamble, sure, to add "a tad extra" and everyone has to be comfortable with their decisions. PassivHaus using 12" may seem obscene, but 4" in zone 7 may be just as bad in the other direction. I wish I knew the answer for everyone, but I think most anyone can pay for a couple extra inches over the life of a house. But, until The Guru does his/her thing, we gamble a little bit.

  10. jklingel | | #11

    our frostline is about a foot closer to the surface than it is to the core of the earth.... on average.

  11. gusfhb | | #12

    re John
    Oh, you must live in the warm spot................

    I had wanted to dig thru the guys website more, but haven't really had the time. He was however predicting PH like levels of total energy usage[~39 kwh/sqm], and as I remembered, with no sub slab foam

  12. LukeInClimateZone7 | | #13

    Perhaps this is too nerdy of a building physics question--
    Alex, I think you make some really great recommendations here, however if I might take an opportunity to comment such that we reframe the question and attempt to improve the understanding of the questioner via it's, errr, reframing (pardon the pun):
    the question of "What R-value do I put in my wall?" strikes me as an all too specific question... almost like asking about a car's efficiency by asking about what kind of compression ratio we should look for in an engine. All other things being equal, such things as compression ratios are quite pertinent, but as you know, they are many many steps removed from the ultimate efficiency, and service that vehicles provide us.

    And so, I make a similar claim about R-values. Ultimately, I know you wish you were asked about a better metric for energy efficiency in a home (and often times, you are asked!). And so, all of what follows is example and contextual evidence for the following question:
    Can't we start talking about (UA) values instead of R-values in this forum, especially when we're talking about the overall performance of these projects, solar decathalon homes, and everyday recommendations?

    Here's my real-world building-specific inspiration:
    Some time ago, I was asked by a colleague what the prescriptive U-value and floor-to-area limitations were on for a prospective project on Lake Michigan that we were simply 'advising' on. After reading and re-reading the 2009 IECC, as well as checking with your fellow Vermonter Martin (via this forum-- thank you by the way!), I confirmed that while there was a prescriptive U-value for glazing in the IECC, there was no limitation on how much of the wall it could take up. The project in question was actually a glass house (something like 60% glazing-to-floor area) with some very toothsome views of the Lake. So, we said that as long as it was greater than the prescriptive value, it was fine.
    My colleague then asked me what R-value we might suggest they put in the walls and roof to make it energy efficient. I was a bit flustered-- "What walls?" I asked sarcastically. No matter if the walls and roof were essentially infinitely resistant, the house was destined to be an energy hog. Oh yeah, it was in the middle of a forest, so no luck on counting on even passive solar design with some good direct gain mass areas. The elephant in the room was the glazing area, which was like a bull in the china shop of shattering records for heat loss in a home (how's that for a mixed metaphor!).

    For my projects, big or small, the metric that I find comes in most handy (and is also referenced by the IECC for modeling) is a comparison of (UA) values. Perhaps it's simply too much to expect our average homeowners to be fluent in such metrics, but is it too aggressive to ask our building professionals to know this metric? It implicitly takes into account all our discomforts we have about floor area, home size, glazing area ratios, air leakage, etc. And, it's before the potentially confusing, and perhaps arbitrary layer of expected monthly energy bills.

    I'm not sure if some sort of combined (UA) value is the perfect metric; a perfect metric is one that is well understood, but cannot be all-encompassing anyway.
    But at least it's something to budget around and talk about in forums such as these. And perhaps you could even provide (UA) recommendations for each climate zone, as tricky as that might sound.

    Speaking of (UA), since my house here in Montana is historic and mostly uninsulated (and furnace is off), I've redefined the relevant (UA) in terms of clo, and the clo of my duvet is looking pretty good about now. I just thought I'd throw out another reframing of the possible solution space. :)

  13. Alex Wilson | | #14

    Maybe it should be energy consumption per person
    You make some great points. I agree that insulation targets are only one part of the story. But instead of UA, isn't the real issue how much energy we should each be using? What about a per-person or per-bedroom target for consumption? Let's consider two houses. House A is a 6,000 sf starter castle that's insulated to state-of-the-art levels (R-5-10-20-40-60, say) and 1.5 ACH50. House B is a 600 sf house built of insulated 2x4 walls, no foundation insulation, R-20 roof, and cheap double-glazed windows with no low-e (R-2-1-1-10-20). Both houses are for retired couples. Which is better? I didn't do the calculations, but my guess is that the House B would use less energy per person per year. Isn't that better? And if we build that house in a place where the occupants can walk to the store, that's even better....

    In an article I wrote years ago, "Small is Beautiful: House Size, Resource Use, and the Environment," ( — requires log-in), we ran a more reasonable comparison and concluded that the smaller house with less insulation was indeed better. Some recent discussion about Passive House standards for North America make similar arguments--that we should remove the benefits realized by large houses (having a metric that is per-square-foot or per-square-meter).

    Anyway, thanks for your observations Lucas. Great food for thought--even with the mixed metaphors!

  14. Mica239 | | #15

    Floor insulation
    What R value would you recommend for floor insulation over an open crawl space in Zone 7?

  15. Alex Wilson | | #16

    Floor insulation
    I'd recommend R-40 minimum--and a very good air barrier.

  16. LukeInClimateZone7 | | #17

    Thanks for the

    Thanks for the excellent response. I've thought about a per person metric as well. I am working in my current practice among various projects with rather disparate performances according to this metric-- an 8000 sq. ft. home for 2 people, and a 5000 sq. ft. home for 12. There's nothing to say that the former or latter's occupancy couldn't change dramatically-- partition walls are fairly inexpensive to construct/deconstruct. In fact, I think of another nice comment in response to mine previously-- a reference to "A Pattern Language" (which I have yet to get my hands on).

    I suppose the challenge I see is not technical, but political. Setting a standard for energy use per person is within the fears of creeping socialism that many in this country would object to. We know that some embrace inequality as a god given right, and energy is included. After all, energy use is well correlated with lifestyle, and energy use is yet another way to distinguish haves from have-nots.

    Of course, many of us in this forum know that energy use does not actually mean energy service. But our national dialogue unfortunately does not reflect that level of understanding.

    So, to summarize-- I still see upgrading the standards for our shells of buildings to be the most expedient path forward. How we use those shells, and their associated occupancies is a political/social issue with some design ramifications. So long as we follow your recommendation of adaptable, resilient structures, then we'll have an easier time with adaptation ourselves in response to many evitable (and increasingly inevitable) pressures we face in the future: global weirding, peak oil, etc.

    ug... again, pardon my prolixity... :)

  17. user-757117 | | #18


    So long as we follow your recommendation of adaptable, resilient structures, then we'll have an easier time with adaptation ourselves in response to many evitable (and increasingly inevitable) pressures we face in the future: global weirding, peak oil, etc.

    I couldn't agree more.
    Resiliency should be foremost in the mind of all designers in this day and age.

  18. ecdunn | | #19

    North/south side insulation
    Also, we should pay attention to the idea of putting more insulation on the north side of the house than on the south (north of equator, of course)

  19. user-705006 | | #20

    Large vs. small - new vs. rehab
    Love the comparison of the large home vs. the small home. I think that there are a lot of those existing small homes around (particular in our towns and cities), where the owners are interested in further reducing the operational costs (energy consumption) of the building.

    We are not helping those owners by tip toeing around the complexity of (for instance) how much to insulate in an existing home. I know there is no one answer. But couldn't GBA start featuring precedents and examples, so much so that most owners of existing buildings could find something that would be similar to their conditions?

    We are sitting on a huge existing building stock, and I believe we need to address that. I would love to see GBA help. Thx.

  20. escalante | | #21

    re-purposed "junk" house
    I really appreciate comments regarding evaluation of housing in terms of energy use per person. I think we have all seen way to many extravagant "green" house projects. My specific insulation questions is a best guess for floor insulation /vapor barrier? on my somewhat unique gut-rehab project. Our rural town (Southern Utah) is full of what I call "junk" housing. Modulars and "double-wides" that are deteriorated to the point where most everyone considers them only as tear downs. We also have a solid waste disposal problem here so these eyesores dot the landscape as in much of the west. I inherited a doublewide on a partial basement and daunted by the cost and effort involved in demo, removal, fill, regrading and moving utilities decided to attempt re-building into a "pretty good house" The underfloor area is to be semi-conditioned. 20% of the he ERV capacity will ventilate the crawlspace/basement area and the perimeter "super skirting" will insulate the wall assembly to the frostline (R20). The partial basement/crawlspace makes any kind of comprehensive basement insulation scheme very complicated so instead I want to also insulate the subfloor which I know is not usually recommended for unvented crawlspaces. We are in climate zone 5 and the walls will be R35 and attic R60. I am taking structure down to the deck and will remove the exist subfloor so can insulate from above, floor framing is 2X6. I appreciate any suggestions or comments.

  21. watercop | | #22

    Insulate under Florida slabs? (R5 recommended in zones 1-2)
    Why recommend any insulation under a slab that is on top of soil at 70*F?

    My ICF house near Jacksonville has no slab insulation - basement / garage feels warm in winter and cool in summer...what's not to like?

  22. p9BuEvB79F | | #23

    I am not understanding why
    I am not understanding why the R-value recommendations are higher for moderate zones than hot zones, specifically windows and above grade walls? Also, Im surprised R-60 is constant even for moderate zones.

  23. Alex Wilson | | #24

    R-value recommendations
    I'll be interested to hear how others respond on this, but for me it's mostly about expected delta-T values (difference in temperature across the insulation). The ceiling R-value recommendations are as high in hot climates as in cold climates because the delta-T from a hot attic in summer can be as great as the delta-T in a cold climate in winter. An unvented dark roof in Zone 1 could reach 160°F, which would be an 85°F delta-T if the indoor temperature is 75°F. This is similar to Burlington, Vermont in the winter if it's -10°F outdoors. But for walls, the wintertime delta-T is greater than the summertime delta-T in all but the hottest climates (because walls don't heat up as much at roofs). So, for walls, I think it makes sense to have the most insulation in the coldest climates, less as you move to warmer climates, and the least in hot climates.

    Relative to Curt's suggestion, I'll concede that the R-5 for sub-slab insulation in a hot climate is hard to justify from a thermal standpoint, but the separation between ground and slab that is ensured by the insulation might still be justification for that insulation.

    I'd love to hear other opinions on this.

  24. user-1109117 | | #25

    Zone 8 recommendations
    Are your recommendations any different for zone 8?

  25. Alex Wilson | | #26

    Zone 8
    I would boost the foundation and floor slab insulation levels in Zone 8 (and perhaps even Zone 7): R-20 under slabs and R-30 for below-grade foundation walls. I guess I'd try for R-50 in walls as well.

  26. andyfrog | | #27

    What's the ballpark on nominal R-values look like to achieve the cold weather true 5-10-20-40-60 R values?

    1. Expert Member
      Michael Maines | | #28

      R-5 windows are U-0.20. NFRC values are for an average-sized window and include the frame, so unless your windows are mostly very large or very small, the listed U-factor should be close to accurate.

      R-10 sub-slab insulation is continuous so there is no reduction for thermal bridging. XPS is a common selection and although sold at R-5/in, it will eventually degrade to about R-4.2/in. If you use EPS, a more environmentally responsible option, it starts out at about R-4/in and maintains R-4/in. Either way, 2.5" is enough to get R-20.

      R-20 foundation walls are insulated in different ways so the answer would depend on your assembly.

      R-40 above-grade walls are done in different ways so the answer would depend on your assembly.

      R-60 roofs are often done with loose-blown insulation on the attic floor, in which case what you order is what you get. When roofs are compact--i.e., cathedral ceilings--if the assembly is open-web trusses or I-joists, there is little thermal bridging, maybe a 10% reduction in listed R-values, so you would want R-66 or R-70 insulation. There are several other ways to build roofs so the answer would depend on your assembly.

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |