Climate Zones and Design Temperatures in Plain Terms

Why this matters

A heat pump that cannot hold the house at the design temperature generates a no-heat call every cold snap, and a furnace sized to the coldest night anyone remembers short-cycles all winter and dies early. Both mistakes come from mixing up two different numbers about the place: the climate zone, which the energy code uses to set what the envelope needs, and the design temperatures, which a load calculation uses to set how big the equipment is. They come from different tables, answer different questions, and one does not give you the other.

Two maps, two jobs

The climate zone is a code compliance label. The International Energy Conservation Code (IECC) assigns every US county a zone number and, for most zones, a moisture letter. The zone decides insulation levels, window performance, duct insulation and sealing requirements, and, through the residential code, which vapor retarder class goes on a frame wall. It does not size anything.

The design temperatures are sizing inputs. A Manual J load calculation (ACCA's residential load method) needs an outdoor heating temperature, an outdoor cooling temperature with its coincident humidity, and indoor setpoints. The heating and cooling loads come from those, and Manual S then selects equipment against the loads. The IRC's equipment sizing section (M1401.3, as adopted by your jurisdiction) requires that chain: equipment sized per Manual S from loads calculated per Manual J or another approved method.

The trap is assuming one implies the other. Zone 5 covers places whose 99 percent heating temperatures sit many degrees apart, because the zone boundary is drawn on yearly heating degree days, not on the cold extreme. Two towns in the same zone can need different equipment.

The zones, in plain terms

The IECC zones run from 0 (hottest) to 8 (coldest), assigned by county in the code's climate zone table (Table R301.1 for residential in recent editions, as adopted by your jurisdiction). Zone 0 was added in the 2021 edition for the hottest locations and is rare in the US. The thermal part of the label comes from heating degree days base 65 F and cooling degree days base 50 F; zone 5, for example, spans roughly 5,400 to 7,200 heating degree days. See related: Degree-Days and Season Length, for what those units are and how to use them on a bill.

The letter is the moisture regime:

• A, moist. Enough rainfall and humidity that drying is slow. Most of the eastern half of the country.
• B, dry. Evaporation outruns precipitation. Much of the interior West.
• C, marine. Mild winters, cool summers, and a dry summer with a wet winter. The Pacific coastal strip is the main US example.

The letters matter mostly through moisture control: dehumidification, vapor retarders, and where condensation risk sits in a wall. See related: Humid, Dry and Marine Regimes, which owns what each regime changes about equipment and envelope choices.

What the design percentiles mean

A design temperature is not an average and not an extreme. It is a percentile of hours.

• The 99 percent heating value is the temperature the outdoor air stays above for 99 percent of the hours in a typical year. The other 1 percent, about 88 of 8,760 hours, is colder. Those hours cluster in a few cold snaps, mostly overnight.
• The 1 percent cooling value is the temperature exceeded in about 88 hours a year. Cooling design also carries a coincident wet-bulb temperature or moisture content, because the latent (moisture) part of the cooling load comes from outdoor humidity, not outdoor temperature.
• The 99.6 and 0.4 percent values are the more extreme columns (about 35 hours a year beyond them) in ASHRAE's climatic design data. Manual J uses the 99 and 1 percent values. Mixing a 99.6 percent heating value into a Manual J calculation oversizes the heating side, and a reviewer comparing your inputs against the table will catch it.

Indoor design conditions are the other half. Manual J's standard indoor conditions are 70 F heating and 75 F cooling, and the IECC (R302.1 in recent editions, as adopted) caps the heating indoor design at 72 F and floors the cooling indoor design at 75 F. A load calculation run at a 76 F heating setpoint because the customer likes it warm is outside what the energy code allows for sizing, wherever that section is adopted.

Where the numbers come from

In order of authority for a given job:

1. The jurisdiction's own design table. The IRC's climatic and geographic design criteria table (in Section R301.2, numbered R301.2(1) in the 2018 edition) has a winter design temperature column that each adopting jurisdiction fills in. Some jurisdictions and a few states publish mandated design conditions for load calculations. Where they exist, a plan reviewer expects those, not yours.
2. Manual J's outdoor design tables for the nearest listed location, which is what most load software pulls.
3. ASHRAE climatic design data, updated with each Handbook of Fundamentals edition. Values move a few degrees between editions as the underlying weather years change, so write down which edition your software uses.

When the nearest listed station is in a different setting from the house (an airport on a plain versus a house in a valley that pools cold air, or a station at a much lower elevation), the station value can be wrong for the site by several degrees. See related: Microclimates on One Lot, and Altitude Beyond Combustion.

Why sizing to the record day is the error

The coldest night on record is one event, often decades old. Sizing heating equipment to it does three things at once: the equipment runs short cycles on nearly every day of the season, efficiency and comfort both suffer, and on a furnace the extra cycling wears the ignition and heat exchanger. The 99 percent value already includes all but the worst 88 hours, and during those hours the house's own thermal mass and the slow decline of indoor temperature carry most homes through with a degree or two of droop.

Cooling oversizing is worse for comfort than heating oversizing. An oversized air conditioner satisfies the thermostat before the coil has run long enough to pull moisture out, so a humid-climate house sits cool and clammy. See related: Design Temperature Versus the Coldest Day Ever, for the customer conversation when the design night comes and the house falls a degree.

A worked example

A shop is replacing a furnace and air conditioner with a heat pump. All values below are illustrative, chosen to show the method rather than describe a real town.

Inputs. The county table gives a winter design temperature of 9 F. The Manual J table for the nearest station agrees at 9 F (99 percent) and gives 91 F (1 percent) for cooling. The customer says it got to -18 F one January about twenty years ago, and the cooling record is 104 F. Indoor design is 70 F heating, 75 F cooling.

Heating load. The Manual J run gives a design heat loss of 36,600 Btu/h at a 61 F difference (70 minus 9). That is 600 Btu/h per degree of indoor-outdoor difference, which holds reasonably well for heating, where conduction and infiltration dominate.

• At the record low, the difference is 88 F (70 minus -18). 600 times 88 is 52,800 Btu/h.
• 52,800 divided by 36,600 is 1.44, so sizing to the record night would buy 44 percent more heating capacity than design requires.

Cooling load. Do not scale it the same way. Cooling load is solar gain through windows, internal gains from people and appliances, and latent load from humidity, as much as it is conduction from the 16 F difference (91 minus 75). The record 104 F is a 29 F difference, but the cooling load does not rise in the 29-to-16 ratio, because the solar and internal parts do not change with outdoor temperature. Rerun Manual J if you want to see a hot-day load. Do not multiply.

Selecting the heat pump. The candidate unit is rated 36,000 Btu/h at 47 F, the standard rating point. Its extended performance data (from the manufacturer's engineering data, not the brochure) shows 24,000 Btu/h at 9 F, also illustrative.

• At design, the house needs 36,600 and the unit delivers 24,000. 36,600 minus 24,000 leaves a 12,600 Btu/h shortfall.
• 12,600 divided by 3,412 Btu/h per kW is 3.7 kW of backup heat needed at design. A 5 kW strip covers it with margin.
• The alternative is a cold-climate unit whose data shows full design capacity at 9 F, which avoids running resistance heat on the design night.

The comparison the customer needs is not "36,000 versus 36,000". It is capacity at 9 F, the condition the house was calculated at.

What changes the answer. If the jurisdiction mandated a design temperature of 4 F, the difference would be 66 F, the design loss 600 times 66 is 39,600 Btu/h, and the unit's capacity must be read at 4 F, where it is lower again. If the house sits in a valley several degrees colder than the station, use a site-adjusted value and write down why. If the house is getting new windows and air sealing in the same job, run the load on the house as it will be, not as it is.

How to check you used the right numbers

• The heating and cooling outdoor values come from the same percentile family (99 and 1, not 99.6 and 1) and the same source, written on the load report.
• The indoor setpoints are inside the energy code's caps where that section is adopted.
• Equipment capacity is read at the design temperature, from extended performance data, and the backup heat calculation shows its subtraction.
• If the job needs a permit, the design values match what the jurisdiction's table or reviewer expects. A mismatch is a plan review comment at best and a failed final at worst.

References

• International Energy Conservation Code, climate zone table and interior design conditions (R302.1), in the edition your jurisdiction has adopted
• International Residential Code, M1401.3 equipment sizing and R301.2 climatic and geographic design criteria, as adopted
• ACCA Manual J (residential load calculation) and Manual S (equipment selection)
• ASHRAE Handbook of Fundamentals, climatic design conditions chapter, current edition
• See related: Reading the Place Before the Quote; Design Temperature Versus the Coldest Day Ever; Degree-Days and Season Length