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

Cold Air Is Dry Air

Don't be fooled by the relative humidity

The psychrometric chart on Henry Gifford's chest

Relative humidity (RH) is what everyone likes to talk about. It gets our attention but it can be a bit confusing, especially when the temperatures drop. For example, at one point yesterday, we had 97% RH. Seems humid, eh?

It’s not really. Not in terms of how much water vapor is actually in the air, that is. The psychrometric chart below shows how this works. The two points I’ve highlighted on the chart are:

  • Point A: 32°F, 100% RH
  • Point B: 70°F, 20% RH

They’re connected by the arrow, indicating that when that cold, seemingly humid, outdoor air leaks into a home by infiltration, it warms up. Let’s assume that the mass of air leaking in doesn’t gain or lose any water vapor molecules along the way.

The psychrometric chart shows how air with 100% relative humidity can be called dry air
The psychrometric chart shows how air with 100% relative humidity can be called dry air

The psychrometric chart rocks!

An unchanging number of water vapor molecules means the movement is purely horizontal on the psychrometric chart. The temperature changes but the concentration of water vapor does not. But as that mass of air warms up, the relative humidity does change. In simple terms, higher temperatures mean more moisture can be in the vapor state. So, rather than being saturated, as it was outdoors at 100% RH, that same mixture of air and water vapor can hold a lot more moisture when it moves inside and warms up.

The three main variables on the psychrometric chart are:

  • Temperature – Dry bulb, what we normally mean when we say the word; it’s shown along the horizontal axis.
  • Humidity ratio – A measure of how much actual water vapor is in the air, it’s defined as the ratio of the mass of water vapor to the mass of dry air. It’s usually measured in grains of water vapor per pound of dry air, where 1 grain = 1/7000 of a pound (or 7000 grains = 1 pound). As with so many things, it’s often shortened, sometimes to grains per pound (OK by me) or just grains (not OK by me).
  • Relative Humidity – What most people usually mean they say the word “humidity.” Technically, it’s the partial pressure of water vapor divided by the vapor pressure when at saturation. By definition, saturation means 100% relative humidity.

As you can see, there are only about 22 grains/lb. of water vapor when the temperature is 32°F and the RH is 100%. That temperature has a special name: dew point. It turns out that 22 grains/lb. isn’t a lot of water vapor.

On a nice spring day when the temperature has risen to 70°F and the RH is 50%, the absolute humidity is about 50 grains/lb. Can you find that on the chart above? Can you find what the dew point is for that condition? (Answer below)

One of the worst kind of days we have here in the Atlanta area is when it’s about 80°F an 80% RH. Ughhh. The absolutely humidity is about 122 grains/lb, and the dew point is 73°F. And that’s not the worst I’ve experienced in the Southeast. Atlanta has nice, fairly mild summers compared to where I was born: Houston, Texas.

Back to the main point, though, if we think of air by its absolute humidity, it’s easy to see that cold air is dry air. One hundred percent RH at the freezing point has only 22 grains per pound. Eighty percent RH at 80°F has 122 grains per pound. If you had 80°F air with only 22 grains/lb., the relative humidity would be less than 15%. You’d be in a desert.

Psychrometrics on your smart phone

If you have a smart phone, you can find apps that do psychrometric calculations for you. I have two on my iPhone. One is the Ultra-Aire Psychrometric Calculator made by Thermastor. They make great dehumidifiers, and their app can help you find how much water removal you’ll get based on the input and output conditions of the air. You can also do basic psychrometric calculations with relative humidity, absolute humidity, dry bulb temperature, and dew point. I also have one called PsychroApp, which does the basic calculations and allows you to adjust for altitude. Both of these are free. You can get more advanced apps that cost a few bucks.

Answer to Question: At 70°F and 50% RH, the dew point is ~50°F. You can see that by finding the point for 70°F and 50% RH and then going horizontally to the left until you hit the 100% RH curve.


Allison Bailes of Atlanta, Georgia, is a speaker, writer, building science consultant, and founder of Energy Vanguard. He is also the author of the Energy Vanguard Blog and writing a book. You can follow him on Twitter at @EnergyVanguard.


  1. user-723121 | | #1


    I would like to know more about HRV's vs ERV's for a cold climate. HRV ventilation air can really dry out a house in the winter when meeting ventilation requirements.

    1. GBA Editor
      Allison A. Bailes III, PhD | | #8

      Doug, yes, HRVs don't exchange moisture so they bring in the cold, dry air air and add heat to it without adding moisture. Since moisture from the house leaves with the exhaust air, the net effect of running an HRV is to dry out the indoor air. In a Passive House, that may be OK with the right amount of ventilation, but many houses will get too dry with an HRV.

      Twenty years ago, before I knew better, I built a tight house (1.4 ACH50) and put in an HRV. It was rated at ~100 cfm and at first, I ran it continuously. In winter, the indoor RH would drop to the bottom end of my hygrometer, which was 15%, I believe. Wood started popping. Gaps showed up in the hard wood floors. It dried out the house tremendously, and this house was in Georgia.

      If you've got an HRV in a cold climate, you can adjust the rate or add a humidifier to keep the indoor humidity at a tolerable level. If the house isn't airtight, of course, it's a good idea to do some air sealing, too, but that's really just another way of adjusting the rate of outdoor air.

      ERVs exchange moisture in addition to heat so they don't dry out the indoor air as much. They don't exchange all the moisture from the outgoing airstream, though, so they do have a net drying effect.

      Your question was very general, Doug, and you've been commenting here at GBA for a long time, so I'm wondering if you had something more specific in mind.

      1. user-723121 | | #12

        Thanks, Allison

        You have covered it. I am hearing you say, even when using an ERV there may be a need for some winter humidification.

        1. GBA Editor
          Allison A. Bailes III, PhD | | #14

          Doug, I'm saying it's possible that a house ventilated with an ERV could be too dry in winter. It could also be too humid. I have friends in Asheville, North Carolina who need an HRV in winter to keep the humidity low enough because they have a very tight house. They have two cores for their system. In summer they use the ERV core, in winter, the HRV core.

          The key variables are the ventilation rate, airtightness of the house, and the density of people in the house. Put two people in a 500 sf airtight condo, and it's likely to be too humid with an ERV. Put those same two people in a 3,000 sf house with the same level of airtightness and increased ventilation rate, and it might be too dry. It just depends on the situation.

      2. qofmiwok | | #24

        I live in a 6B very dry climate and for my new PH I'm planning a Zehnder ERV. Is that the better choice than HRV and controlled humidification? Not sure how I would distribute humidification though because I don't have any ducting.

        1. GBA Editor
          Allison A. Bailes III, PhD | | #25

          qofmiwok, if you're building a single-family detached home, an ERV is probably the way to go. HRVs can dry out the indoor air a lot in winter, and since you live in a very dry climate, it would dry it out in summer, too. An ERV keeps the moisture in the house. Sometimes, though, airtight houses get too humid in winter with an ERV. I have friends in Asheville, North Carolina who use an ERV core in summer and switch to an HRV core for winter. Another option would be to keep the ERV core in winter and increase the ventilation rate. Since you don't get 100% moisture transfer, every cubic foot or liter of air you bring in will result in drier indoor air when the outdoor air is drier than the indoor air.

  2. jameshowison | | #2

    I have some tags from which say that they measure temp and humidity. I set their app to show temp and dewpoint and I have one outside (on our back patio).

    Here's the thing, though, for the outside one the two curves basically follow each other, showing clear daily cycles (with the dewpoint lower than temp, obviously :). The temperature rises in the morning, but so does the dewpoint. I've tried to attach a screenshot to this post to show what I mean.

    I've wondered for a while if this is some sort of bug in the software, or if there is some daily cycle in humidity which is unrelated to temp (ie both temp and dewpoint are moving because of day and night?).

    My expectation, based off your explanation above, is that temp and dewpoint are independent of each other (or rather grains per pound of moisture determines dewpoint), so one would expect them to change together over long periods, as weather fronts come through, but not overnight or track so closely.

    I see this effect in our shed as well, which is unsulated and unheated (but fairly air-tight and has a vapor vent). I would expect to see relatively constant moisture present, perhaps with jumps when the door was opened or wet things inside were drying off, and that leads me to expect relatively constant dewpoints. Instead I see the same variance with temperature (but note that it's not exact, it's a different time series, not just some temp - 20° function). I've also attached the screenshot of the "Shed".

    Bug or am I'm missing something in this situation :)

    1. HFF | | #3

      Dew point and temperature are absolutely NOT independent of each other. In fact at constant RH they are very nearly linearly related. Your observation that the relationship is "not exact" is indicative of RH differences over time. I might add that over modest excursions of RH (say 10 to 20% or so), that the slopes of the lines are quite similar.

      1. jameshowison | | #4

        Thanks Fred. So perhaps what I'm mis-understanding here is whether dew point is a measure of absolute humidity akin to grains per pound. If one knows the dew point does therefore know the grains per pound?

        1. maine_tyler | | #5

          Dew point tells you 'absolute humidity.' That is correct. They are both the horizontal line.

          But dewpoint may also rise when temperature rises because vapor pressure (not 'partial' vapor pressure) is proportional to temperature (and only temperature). Meaning if temperature rises, vapor pressure rises, and therefore any H2O that is present is 'more likely' to enter vapor state. In turn, dewpoint is likely to rise.

          When cold 'dry air' enters a home and get's heated, there is typically not enough H2O present to fill the 'gap.' I.e. the relative humidity goes down but the dewpoint could very well rise if/when H2O sources are introduced.

          It is interesting that it apparently happens on the short cycle in the way you observe with your test equipment. I know nothing about that test equipment so can't say whether what you observe is accurate or not.

        2. HFF | | #11

          At a given pressure there is a unique relationship between water content and and dew point. However it is not at all linear, and in order to calculate it you need more information. Approaching the psychrometric chart can be daunting. There appear to be six independent variables in this graphical representation of the equations of state. All six must be satisfied based on thermodynamic considerations. If you know only dew point temperature, you don't have enough information to calculate water content. However, with knowledge of any two of the six variables, you can determine the values of all others. (Keep in mind that a specific chart is valid for only one pressure, so you should say that three variables are required, one being pressure.)

          I find it helpful to recall that for water or any liquid, at a given pressure, there is a well-defined relationship between temperature and partial pressure. From partial pressure, you can calculate RH from knowledge of the saturation pressure at the given temperature. From a mental picture perspective, I therefore like to start with temperature and RH as my "independent" variables. These are readily grasped and routinely measured. From this perspective, dew point temperature is a derived value that can be calculated from temperature (dry bulb) and RH. If you have access to a graphics program, or even Excel, you can play around with these relationships. Google "Magnus formula". This is a good approximation that allows the calculation of dew point from known temperature and RH. There are a couple of constants that first need to be calculated based on empirical relationships, but it's not too complicated. I find it difficult to create a mental picture of how this equation looks when graphed, so an actual set of graphs is quite beneficial to grasp the relationships.

          1. maine_tyler | | #13

            "If you know only dew point temperature, you don't have enough information to calculate water content."

            How do you mean? True about the pressure, but for a given chart with a given ambient pressure, if you know dewpoint you know the vertical location of the horizontal line. No?
            There are quite a few values that can be displayed on the y-axis. (I like this particular chart: Some depend on ambient pressure (such as humidity ratio) and some don't (such as partial pressure). But to keep it simple, all these y-axis values are essentially the 'amount of water vapor content' for a given system. Unless I'm missing something.

            Partial pressure (and also EVP) depends only on temperature. It's given in inches of mercury (for example) as is the pressure of the system. The vapor partial pressure contributes to overall system pressure rather than depends on it.

            "There appear to be six independent variables in this graphical representation of the equations of state."

            Maybe. Are they really independent variables, or maybe more like descriptors? Some charts show a lot more data, but it doesn't change the fundamentals of the system:

            Certainly the question of how many variables depends on what equation we're solving for.
            Hopefully this isn't too far off, I welcome clarification and correction.

          2. Deleted | | #16


          3. maine_tyler | | #18


            It sounds to me like you are confusing wet-bulb temp lines with dew-point. Dew-point is indeed the horizontal component. This certainly explains the discrepancy between how we are talking about this.
   (only a couple seconds of viewing needed)
   (only a couple seconds of viewing needed)

            If another 6 minutes and a cup of tea is to spare, the following video explains the saturation, dew-point, and vapor pressure (partial) quite well.

          4. HFF | | #20

            to # 18
            Tyler, you are correct of course. I don't know what I was thinking and I deleted my comment.

          5. HFF | | #21

            I've since come to my senses and acknowledge that dew point and water content are not independent of each other. For any given vapor concentration, there is a unique dew point. It is simply a physical property of water vapor and holds at a given barometric pressure. I apologize for the misleading information.

          6. maine_tyler | | #23

            No worries, we all get wires crossed on occasion, especially on subjects like this. When I say something inaccurate, I always appreciate and expect correction. Cheers.

        3. GBA Editor
          Allison A. Bailes III, PhD | | #15

          James, both Fred and Tyler make good points above (although Tyler raises some valid questions about Fred's statements). Let me go back to your primary question, though:

          " If one knows the dew point does therefore know the grains per pound?"

          The short answer to your question is yes, but only as long as we're dealing with constant pressure. As Fred said, the psychrometric chart has pressure as a hidden variable. If the pressure changes, you need a new chart calculated for the new pressure, and that changes the relationship between dew point and humidity ratio.

          NB: Sometimes humidity ratio is called absolute humidity, but that term is defined differently. Absolute humidity is actually mass of water vapor divided by the volume, whereas humidity ratio (plotted on the vertical axis on the psychrometric chart) is the mass of water vapor divided by the mass of dry air.

      2. maine_tyler | | #6

        Note that the dew point curve is simply a relative humidity curve (the one on top).

    2. GBA Editor
      Allison A. Bailes III, PhD | | #9

      James, I believe the thing you're missing is changes in water vapor concentration. As temperature rises throughout the day, more water vapor can enter the air by evaporation. As it cools down at night, water vapor condenses again. That accounts for the rising and falling you see in the dew point.

      I found the same thing in a spray foam encapsulated attic and wrote about it in an article titled, "Humidity in a Spray Foam Attic." Here's the link:

      At night, the water vapor goes through open-cell SPF and hangs out in the sheathing. When the sun blasts the roof the next door, it goes back into the air. See the attached graph.

      In the article I wrote here about cold air being dry air, I was talking about what happens when all you do is change the temperature of the air but not the concentration of water vapor. If you do that, the dew point stays the same while the relative humidity changes.

      Going back to your example of rising and falling dew points, if you watch long enough, you'll see some days where the dew point stays the same or maybe even falls as the temperature rises. That's because outdoor air moves. When a front moves through, you've got a whole new air mass coming in behind it and it may bring dry air.

      1. jameshowison | | #17

        Thanks Alison, yup that's it.

        Ok, so what I was missing is that there are water reservoirs everywhere and vapor migration is constantly happening. Outside that's going to be the ground and inside that's going to be (porous) surfaces of all kinds, especially the sheathing in my uninsulated/non-drywalled shed.

        And that water vapor will constantly move from (relatively) wet to (relatively) dry. So as temperature goes up, the moisture will migrate from the sheathing to the air (increasing absolute humidity in the air, driving up grains per pound, increasing dew point). As temperature falls (but stays above dew point) vapor will still be taken back in the sheathing/ground.

        I realize that I had been thinking of moisture only moving at the dew point, by becoming liquid and sitting on surfaces. But it's moving all the time (it's either adsorption or absorption, right, I know you've written about that :)

        This is why vapor vents work, even when the inside of an attic (on average) has less water in it than outside, because the vents are at the ridge, moisture migrates there, creating strata where the humidity on the inside of the vapor vent is higher than outside, further drying the inside (on average). (I now understand all over again :) And because the attic is hotter than the house, particularly the ridge of the attic, its ability to absorb water increases, prompting the moisture in the house to migrate to the roof, where its concentration in a small body of air raises humidity above the outside, prompting vapor to transfer through the vapor vent, and so on.

        So does this also explain why the uninsulated shed (which has vapor impermeable wrap with a vapor permeable strip at the ridge) is almost always noticeably colder than outside, even for long periods when the sun is out (but not late in the day). (btw, the screenshots I provided were not of the same time period, the shed has clearly lower dewpoints, and almost always lower temperatures, than the outside). The migrating water is latent heat and its leaving reduces the sensible temp measured by the sensor? Hmmm, perhaps I'm not 100% clear on that.

        1. maine_tyler | | #19

          I think your answer to your own question is frankly the most comprehensive.
          It is a salient point that the vapor content of a system can be reduced even when dewpoint temperatures aren't reached (via hygroscopic materials).

          In that sense, I like to think of it in terms of 'likelihoods,' i.e. warmer temps lead to higher likelihoods of water being in the vapor state, and visa-versa (even if the dewpoint threshold is not explicitly reached).
          It is also possible dewpoint temps ARE reached on certain surfaces due to radiative cooling, but either way it is not required to reduce the vapor content of the system.

          It does make one wonder though: if you had a shed with an interior made of well-sealed non hygroscopic materials on every surface (say mirrors), would you see greater swings in RH with temperature swings and smaller swings in dewpoint?

          On your last point. My guess is that a shed would be cooler inside by the mere fact of being insulated, so long as there isn't too much solar gain. An insulated structure, even unconditioned, will buffer the diurnal swings and bring the indoor temp closer to the average of the high and lows (which means warmer than ambient at night and cooler during the day).

          While vapor flow out of your assembly is certainly a form of energy loss, I don't believe it is substantial. (I have no proof in numbers for this, but I recall Dana talking about it semi-recently. It's mainly that not very large quantities of vapor (and therefore energy) is transported via diffusion, compared to other energy gains and losses in a typical system).

          1. jameshowison | | #22

            Thanks Tyler (and all!). My wife just looked at me and said, (duh) yeah, it’s diffusion and a concentration gradient.

            btw, Shed is not insulated. I’m going to check the data on it again but perceptually it’s often much colder than expected. Might still be a delay from nighttime temps, even without insulation. Also a metal roof so maybe nighttime radiative cooling? Anyway, the “solar dehumidifier” of the vapor vent works really well to keep it dry inside.

  3. maine_tyler | | #7

    If vapor wanted to throw a sick dance party, they'd need two things*.
    1) music (temperature)
    2) people who can hear that music (available H2O)

    The presence of people (H2O) set's the stage. The music (temperature) is the driver.
    If the people in the back can't hear the music, they won't dance (for the purposes of this analogy). If you crank the music so they can hear, more of them will dance (statistically speaking). Increasing music volume=increasing temperature. Increasing volume increases the 'dance driver'=increasing temperature increases 'vapor pressure'.

    You can also increase the chances of dancing (of those who can hear the music) by liquoring them up, i.e. break the tension and give them more space to dance. (Mist the H2O for example, or present larger surfaces of water to the air). Whether the percentage of dancers to non-dancers remains the same depends on the availability of people to hear the music and how ready they are to dance.

    Allison's crusade against relative humidity (I joke) is that people often talk about how sick a party vapor threw last night without specifying the sound system, "70% of people who could hear the music were dancing!"
    Well how loud was the music? If 10 people could hear it, 70% would be 7 people.

    The RH beef is really a beef with not having enough information [or at least less pertinent information in many cases]. We only know 'which curve.' If we knew temp, however, we'd then have plenty of info.

    Specifying dew-point temperature specifies a curve and a point on that curve [edit: actually, it really just specifies the horizontal line], which gives us the horizontal line that Allison is after. One could as easily do this with any of the RH curves. Let's call the 50% RH curve the '50-point' and do the same thing.
    What's the absolute humidity at a '50-point' of 85F? It's the same as a 'dew-point' of around 64F. A horizontal line is defined in both cases.

    If we have a horizontal line, then knowing either the temperature OR the RH will give us a point on the chart.

    *Actually they need a third thing: one of those freakin t-shirts! Heck yes.

    1. GBA Editor
      Allison A. Bailes III, PhD | | #10

      I like the dance party analogy, Tyler! And yes, you've explained my problem with RH well.

  4. Buzz_Burger | | #26

    I think I know the answer to this but it's bugged me for a long time and would greatly appreciate a clarification/confirmation from the expert. I understand that in addition to "recovering" heat an ERV, as opposed to an HRV, can partially transfer moisture from the wetter air to the drier air. But how do we define the wetter air? If it were relative humidity, then the transfer could be in the wrong direction. For a specific example, suppose an outside temperature of 30 degrees with a relative humidity of 40% and an inside temp of 68 degrees and 30% RH. So even though the RH is higher outside, the inside air contains more moisture, So, please confirm for me that my ERV is helping to retain inside humidity in this scenario, not transferring it outside.


    1. maine_tyler | | #27

      I'm not sure if Allison would approve of this source ;), but here's a good little write-up on it:

      and another:
      which states that, "Like temperature, moisture in the air will move from an area of
      high concentration to an area of low concentration. "

    2. HFF | | #31

      The driving force for vapor diffusion is the gradient of vapor pressure. The driving force for adsorption/desorption (for example to an ERV membrane within the core) is the relative humidity.

      1. maine_tyler | | #32

        you are right, however so as not to cause confusion, I think it's worth pointing out a few things:
        –It's a dynamic system with a temperature gradient across the core
        –The temperature of the membrane itself determines the adsorption rate, not the RH of the 'air' or 'space.' I.e. boundary conditions not average conditions. This is essentially how dehumidifiers work.
        –As temperature exchange occurs, the RH of the air streams change. Ultimately, initial RH conditions cannot strong-arm against vapor diffusion as predicted by partial vapor pressure.
        –The parameter of import (in the context of Buzz's question) is the the actual vapor content of the air streams (what we're calling absolute).

  5. Buzz_Burger | | #28

    My question is basically this: If one side has a higher RH but a lower absolute humidity, which way does the moisture move?

    1. maine_tyler | | #29

      Its the 'absolute' not the RH that dictates.
      High concentration implies absolute, not RH.

      1. Buzz_Burger | | #30

        Got it, thanks.

  6. SierraWayfarer | | #33

    Mr. Bailes,

    Thanks for teaching me how to read a psychrometric chart! Useful....

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |