Design Temperature Versus the Coldest Day Ever

Why this matters

During a record cold snap the phones light up with the same call: the furnace runs nonstop and the house is sitting at 65. Some of those are real failures. Many are systems working exactly as designed on a day colder than anyone designs for. A tech who cannot tell the two apart either replaces parts on a healthy system or tells a customer with a cracked heat exchanger that "it's just the cold." And a shop that cannot explain the difference to a customer ends up selling an oversized replacement that short-cycles for the other 8,670 or so hours of the year.

How design temperatures are published for your area belongs to the climate-zones card. This card is about what the design number means, what happens below it, and how to talk about it.

What the 99 percent design temperature actually says

Heating design temperatures come from long records of hourly weather observations, summarized by ASHRAE and carried into the ACCA Manual J tables residential load calculations use. The common residential heating value is the 99 percent dry-bulb temperature: the outdoor temperature that the hourly readings are at or above for 99 percent of the hours in a year.

So it is not the coldest temperature on record, and it is not the average winter low. It is the temperature the location is colder than for about 1 percent of the year's 8,760 hours, which is roughly 88 hours a year. Those hours mostly come in a handful of cold nights and early mornings, not spread evenly. ASHRAE also publishes a 99.6 percent value, colder, exceeded about 35 hours a year; Manual J uses the 99 percent figure.

Cooling works the same way from the other side: the common cooling design value is the 1 percent temperature, exceeded about 88 hours a year.

Why not design for the record? Because a system sized to carry the coldest day in fifty years is much larger than the house needs on every other day. It cycles on and off in short bursts, wears its ignition and controls faster, swings temperatures, and on the cooling side removes less moisture. In most of the country the record low sits well below the 99 percent value. Sizing to it trades a few uncomfortable hours every decade for poor performance every day.

The code layer: IRC M1401.3, as adopted by your jurisdiction, requires heating and cooling equipment to be sized by ACCA Manual S from loads calculated by ACCA Manual J or an equivalent method, and IRC Table R301.2(1) is where many jurisdictions publish their own winter and summer design temperatures for that calculation. Where the jurisdiction states a design temperature, that is the one the load calculation uses.

What happens below the design temperature

A house's heat loss rises roughly in proportion to the indoor-outdoor temperature difference. A heating system has a fixed maximum output. Below some outdoor temperature the house loses heat faster than the system can supply it, the system runs continuously, and the indoor temperature drifts down until the house's heat loss at the lower indoor temperature matches what the system delivers.

That is not a failure. It is the predictable shape of any properly sized system on a day past its design point.

Worked example: a properly sized furnace on a record morning

A house in a location with a 99 percent heating design temperature of 5 F. Indoor design 70 F. The Manual J heating load is 40,000 Btu/h at design. The installed furnace delivers 48,000 Btu/h output, which is 120 percent of the load and inside the range Manual S allows for furnaces.

Heat loss per degree. The design difference is 70 minus 5, 65 F. 40,000 divided by 65 is about 615 Btu/h per degree of difference. This treats the loss as linear and ignores internal gains (people, appliances, sun), which makes it slightly pessimistic about the indoor temperature that results.

On a design day (5 F). Load 40,000 Btu/h against 48,000 available. The furnace runs about 40,000 divided by 48,000, roughly 83 percent of the time, and holds 70 F with room to spare.

On the record morning (minus 12 F). The difference to hold 70 indoors is 70 minus (minus 12), 82 F. 82 times 615 is about 50,400 Btu/h. The furnace can supply 48,000, short by about 2,400 Btu/h. It runs 100 percent of the time, and the indoor temperature falls until the loss matches 48,000. 48,000 divided by 615 is about 78 F of difference, so the house settles near minus 12 plus 78, about 66 F.

With wind. Real cold snaps usually bring wind, which raises infiltration above what the Manual J estimate assumed. If the effective loss rose 10 percent, to about 677 Btu/h per degree, the same furnace holds 48,000 divided by 677, about 71 F of difference, and the house settles near 59 F. That is the range of what customers report on the worst mornings, from a system with nothing wrong with it.

What would make it worse and actually be a fault. A furnace delivering less than its rated output because of a dirty filter, a closed or blocked return, low gas pressure or an underfired burner, a duct disconnected in the attic, or a cracked heat exchanger forcing a lockout. Those show up as the house falling below the expected settling temperature, or as the furnace not running continuously when it should.

The heat pump case. A heat pump's output falls as the outdoor temperature falls, and its load rises, so its own balance point (the outdoor temperature where capacity equals load) is often above the design temperature. Below that point auxiliary or backup heat carries the difference. On a record morning, a heat pump house that is cold usually comes down to whether the backup heat is working and properly sized, not the heat pump itself.

Tell design limit from failure before you say anything

Before telling a customer it is the weather, confirm the system is delivering what it should:

• It is running continuously. A system cycling off on a 60 F house is not at its limit; something is stopping it, from a thermostat to a limit switch to a flame-sensing fault.
• Temperature rise is within the furnace nameplate's stated range, measured supply minus return, with a clean filter. A rise below the range suggests underfiring; above it, restricted airflow.
• For a heat pump, the auxiliary heat is energizing when called for, and the defrost cycle is completing.
• The supply registers in the coldest rooms are delivering air comparable to the rest of the house, which catches a disconnected or crushed duct.
• Gas pressure, or the fuel supply, is adequate. During a long cold snap a propane tank can drop low, and a low tank delivers less vapor in extreme cold; ask when it was last filled.

If all of that checks and the house is settling within a few degrees of what the arithmetic predicts, you are looking at a design limit.

The conversation with the customer

The customer is cold, and probably has not slept well. They need three things: a clear answer whether anything is broken, what to expect, and what they can do. A version that works:

CUSTOMER: It's 64 in here and it hasn't shut off all
          night. Is it broken?
TECH:     I checked it. It's putting out full heat and
          the airflow and gas are right. Your system is
          built for about 5 degrees outside, which covers
          all but a few nights a year. It's 12 below
          right now, so it's running flat out and the
          house is losing a little faster than it can
          keep up. When it warms past 0 or so it'll catch
          up and start cycling again.
CUSTOMER: So should we get a bigger one?
TECH:     I'd say no. A bigger one would be too big the
          other 360 days and run worse. Leave the
          thermostat where it is, no setback tonight, and
          keep the shades down on the north side.

What the conversation does: confirms the check happened, gives the design number and today's number in the customer's terms, says when it ends, and turns the upsell question into honest advice. Write the readings you took on the ticket, so if the customer calls again next week the next tech sees the system was verified.

Practical advice worth giving:

• No thermostat setback during a cold snap. A system already running continuously cannot recover from a setback, and the house will spend the day colder than if it had been left alone.
• No unvented combustion for heat. Never heat with a gas range or oven, a charcoal or gas grill indoors, or a generator in a garage, basement or near a window or door. Each one can produce lethal carbon monoxide. A working CO alarm on every level is worth confirming on the same visit.
• Portable electric heaters plugged directly into a wall receptacle, not an extension cord or power strip, kept at least 3 ft from anything that burns, and not left running unattended or while sleeping.
• Protect the plumbing. If indoor temperatures are falling toward the 50s, cabinets on exterior walls opened and a slow drip on vulnerable fixtures help prevent frozen pipes.

When the answer changes

• The house was altered after the system was sized. A finished basement, an addition, or removed insulation means the system is undersized for the current house, and a new load calculation is the honest next step, not a bigger furnace by guess.
• The house settles well below the predicted temperature with everything checked. The load calculation may have been wrong, or the house leaks far more than assumed. A blower door test and a new Manual J find out which.
• The design temperature itself is old. Updated ASHRAE climate data has moved some locations' design values, generally warmer. Use the value your jurisdiction specifies; where it does not specify one, use the current Manual J table.
• The customer has a medical or vulnerability reason that makes 60 F unacceptable. That is a reason to discuss supplemental heat or a backup source sized for extremes, not a reason to oversize the primary system.

References

• ACCA Manual J (residential load calculation) and Manual S (equipment selection), as required by IRC M1401.3 in the edition your jurisdiction has adopted.
• ASHRAE Handbook, Fundamentals, climatic design conditions (99 percent and 99.6 percent heating values, 1 percent cooling values).
• US CDC and US Consumer Product Safety Commission guidance on carbon monoxide and portable heater safety.
• See related: Climate Zones and Design Temperatures in Plain Terms; Degree-Days and Season Length; The Local Design Criteria Table.