Every factual sentence carries assertions checked against the recorded evidence: references must resolve, figures and quotations must be in the cited passage, arithmetic must reproduce, and an independent reviewer must approve the wording. Automated validation says whether the narration is ready for human editorial review; it does not establish truth.
Open a sentence to see its assertions, their status and the passages behind them. Sentences without a marker make no factual claim.
verified s04.a An NH Patch headline said heat pumps are tech that falls flat in the cold.
supported (approved): no direct support, 6 contradicting, 1 source statement.
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s05.a Mass Save says of heat pumps: "many provide efficient heat at outdoor temperatures as low as -15 °F".
supported (approved): no direct support, 3 qualifying, 1 source statement.
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s08.a Electric resistance heat operates at a COP of 1, less efficient than a heat pump.
qualified (qualify): 1 supporting (1 independent, strongest 0.55), 1 qualifying.
warning needs-qualification: Edwards 2024, inequities in heat pump bill savings: Edwards assumes single-stage heat pumps fall to COP 1 (no better than resistance) below 5F, while cold-climate two-stage units keep a higher COP, so the claim depends on equipment type. Add the qualification: the evidence supports a narrower or weaker statement.
verified s09.a An evaluation metered delivered heat from dozens of ductless mini-splits over winter 2015 and winter 2016; in the 2016 winter their COPs averaged 1.5 at -10 to 0 °F.
qualified (qualify): 2 supporting (1 independent, strongest 0.6), 2 qualifying, 1 context only.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: Over 25,000 products on the NEEP cold-climate list keep a COP of 2 or more at full capacity at 5F, and modelled per-household GHG savings are generally higher in cold climates. [partial: a count of products rated at 5F on a certification list, not measured operation in homes] Edwards 2024, inequities in heat pump bill savings: Edwards assumes singl…
verified s10.a At a COP of 1.5 the mini-splits gave effective heating efficiencies 50% higher than electric resistance heating, which the evaluators called far superior.
qualified (qualify): 1 supporting (1 independent, strongest 0.6), 1 qualifying.
warning needs-qualification: Edwards 2024, inequities in heat pump bill savings: Edwards assumes single-stage heat pumps fall to COP 1 (no better than resistance) below 5F, while cold-climate two-stage units keep a higher COP, so the claim depends on equipment type. Add the qualification: the evidence supports a narrower or weaker statement.
verified s10.b The units operated less efficiently in winter 2015, which had prolonged deep snow and cold, with snow and ice around outdoor units, than in winter 2016.
qualified (qualify): 1 supporting (1 independent, strongest 0.6), 1 qualifying.
warning needs-qualification: Edwards 2024, inequities in heat pump bill savings: Edwards assumes single-stage heat pumps fall to COP 1 (no better than resistance) below 5F, while cold-climate two-stage units keep a higher COP, so the claim depends on equipment type. Add the qualification: the evidence supports a narrower or weaker statement.
verified s11.a In the evaluation, observed non-cold-climate models operated at outdoor temperatures below 0 °F, at lower efficiency than cold-climate models.
qualified (qualify): 1 supporting (1 independent, strongest 0.4), 3 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. [partial: the verdict note limits its own scope ("though")] Winkler 2023 (NREL), cold-climate heat pump field validation: Among …
verified s12.a Excluding periods with auxiliary heat or defrost, the prototype cold-climate heat pumps in the field had a median COP of 1.9 in the 0 to 5 °F bin.
qualified (qualify): 2 supporting (1 independent, strongest 0.55), 2 qualifying, 1 context only.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: Over 25,000 products on the NEEP cold-climate list keep a COP of 2 or more at full capacity at 5F, and modelled per-household GHG savings are generally higher in cold climates. [partial: a count of products rated at 5F on a certification list, not measured operation in homes] Edwards 2024, inequities in heat pump bill savings: Edwards assumes singl…
verified s13.a The prototypes' laboratory performance was significantly more efficient, but most units' observed field COP was above 1.75, the EPA Energy Star requirement for cold-climate heat pumps at 5 °F.
qualified (qualify): 1 supporting (1 independent, strongest 0.55), 1 qualifying, 1 context only.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: Over 25,000 products on the NEEP cold-climate list keep a COP of 2 or more at full capacity at 5F, and modelled per-household GHG savings are generally higher in cold climates. [partial: a count of products rated at 5F on a certification list, not measured operation in homes] Context only, RMI 2018, the economics of electrifying buildings (wrong): …
verified s15.a Technology Connections said that along with getting less efficient, heat pumps' actual heat output also drops.
supported (approved): no direct support, 1 source statement.
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s16.a One 1-ton ductless heat pump in Northern Minnesota delivered a consistent median capacity from 10 °F to -15 °F.
qualified (qualify): no direct support, 3 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. Add the qualification: the evidence supports a narrower or weaker statement.
verified s17.a The unit continued to operate from -15 °F to -25 °F at reduced capacity; between -20 and -25 °F its median delivered capacity was approximately 20% of its rated capacity at 5 °F.
supported (approved): 1 supporting (1 independent, strongest 0.5).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s18.a The researchers concluded the reduction in capacity shows these units should not be used without backup at extreme temperatures below -15 °F.
qualified (qualify): no direct support, 3 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. Add the qualification: the evidence supports a narrower or weaker statement.
verified s19.a In the Minnesota study, outdoor air temperature had the largest impact on COP; heat-pump-only heating events had COPs around 1.3 at 10 °F, rising to around 3.0 to 4.5 in the shoulder heating seasons (around 50 to 60 °F), so 1.3 is less than half of the mild-weather figure.
supported (approved): 1 supporting (1 independent, strongest 0.65).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s21.a In cold, humid conditions frost accumulates on the outdoor coil and cuts heat exchange; defrost reverses the refrigerant flow so the heat pump rejects heat from the conditioned space to melt the frost, which can drop the supply air temperature.
supported (approved): 1 supporting (1 independent, strongest 0.55).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s22.a Defrost was most energy intensive at 5 to 35 °F, where the median share of energy expended was between 5% and 8%, a small portion; one site spent 16% of its time in defrost under 40 °F against an overall average of 5%.
supported (approved): 1 supporting (1 independent, strongest 0.55).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s24.a Jordan Dearden said electric backup heat strips will absolutely destroy your electric bill.
supported (approved): no direct support, 4 qualifying, 1 source statement.
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s25.a The units primarily operated in heat-pump-only mode even at the lowest outdoor temperatures, and auxiliary-only operation was minimal above 0 °F.
qualified (qualify): 1 supporting (1 independent, strongest 0.45), 3 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. [partial: the verdict note limits its own scope ("though")] Winkler 2023 (NREL), cold-climate heat pump field validation: Among …
verified s26.a In the coldest outdoor temperature bins with sufficient data (-10 to 0 °F), auxiliary elements were engaged 25% of the time across sites; the two coldest bins had a sample of 8 and 11 of the 22 sites.
qualified (qualify): 1 supporting (1 independent, strongest 0.45), 4 qualifying.
warning needs-qualification: Winkler 2023 (NREL), cold-climate heat pump field validation: Auxiliary heat runtime generally rose in colder bins and heavy use cut system COP over 30% at two sites, but the heavy users were undersized or faulty while six sites used little auxiliary heat. CEE 2018, ducted and ductless cold-climate heat pumps in the field: In all-electric systems the booster resistance heat ran at COP 1 below abou…
verified s27.a Switchover to supplemental heat was observed around 30 to 35 °F; the researchers say a typical switchover temperature without homeowner setpoint interference, or with better setpoint controls, would likely be much lower.
qualified (qualify): 1 supporting (1 independent, strongest 0.45), 3 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. [partial: the verdict note limits its own scope ("though")] Winkler 2023 (NREL), cold-climate heat pump field validation: Among …
verified s28.a Two sites engaged higher stages of auxiliary heat during defrost; at least one did so in normal operation too, which the manufacturer team said was a controls problem and not intended operation.
supported (approved): 1 supporting (1 independent, strongest 0.55).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
warning number-dense (deterministic): 10 figures from s29 to s36: keep the ones the argument needs, chart the rest and say the point in words
verified s29.a In the all-electric system, events with electric resistance booster heat occurred below 30 °F, but because the heat pump was prioritized, system COPs were greater than 1.0 even at the coldest outdoor temperature, around -7 °F.
qualified (qualify): 1 supporting (1 independent, strongest 0.55), 5 qualifying.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: In all-electric systems the booster resistance heat ran at COP 1 below about 30F, but because the heat pump was prioritized the system COP stayed above 1.0 even at -7F, so backup use dented rather than erased efficiency. "all" is stronger than the evidence (strongest support 0.55). Add the qualification: the evidence supports a na…
verified s30.a At Site 10 the compressor did not run below 0 °F although its lockout temperature was disabled, likely due to an on-board control setting that cannot be adjusted in the field.
qualified (qualify): no direct support, 3 qualifying.
warning needs-qualification: Winkler 2023 (NREL), cold-climate heat pump field validation: Among these ducted variable-speed units, Site 10's compressor did not run below 0F despite a disabled lockout, and the dual-fuel unit did not run below about 4F. Add the qualification: the evidence supports a narrower or weaker statement.
verified s31.a In Minnesota, median bill savings were $182/year for high-efficiency ASHPs sized for the heating load, while median bills rose $96/year when they were sized for the cooling load, because of higher use of electric resistance backup.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: In Minnesota a high-efficiency heat pump sized for the heating load saves a median $182/yr, but sized for cooling it raises bills $96/yr through more resistance backup; reliance and the bill hit depend on sizing, not cold-climate units as such. Add the qualification: the evidence supports a narrower or weaker statement.
verified s33.a For Minnesota homes replacing propane furnaces or electric heating, performance maps, home heating loads and normalized Minnesota weather were used to estimate homeowner cost savings of 30% to 56%.
qualified (qualify): 1 supporting (1 independent, strongest 0.6), 1 qualifying, 1 context only.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: Over 25,000 products on the NEEP cold-climate list keep a COP of 2 or more at full capacity at 5F, and modelled per-household GHG savings are generally higher in cold climates. [partial: a count of products rated at 5F on a certification list, not measured operation in homes] Context only, RMI 2018, the economics of electrifying buildings (wrong): …
verified s34.a For natural gas, the cost breakpoint was above 70 °F in either winter, so the mini-splits only operated cost-effectively above 70 °F compared with an 80% efficient heating system, and the evaluators said they do not provide a viable direct replacement for a gas-fired system at those energy prices.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Cadmus 2016, ductless mini-split evaluation (MA and RI): Switching between heat pump and fuel at the crossover temperature is cheapest in principle, but at 2016 prices the gas crossover was above 70 F for the measured units, so pairing with gas saved nothing; only a modelled HSPF-13 unit (28-40 F crossover) would beat gas alone. Add the qualification: the evidence supports a narrower or weaker sta…
verified s35.a The balance point for propane was -15 °F in both winters, meaning the mini-split was always less expensive than propane.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Cadmus 2016, ductless mini-split evaluation (MA and RI): Switching between heat pump and fuel at the crossover temperature is cheapest in principle, but at 2016 prices the gas crossover was above 70 F for the measured units, so pairing with gas saved nothing; only a modelled HSPF-13 unit (28-40 F crossover) would beat gas alone. "always" is stronger than the evidence (strongest support 0). Add the…
verified s36.a For the measured units listed as cold climate units, gas balance points were above 58 °F for both winters; based on the rating curve of a current cold-climate unit with an HSPF of 13, the natural gas balance point dropped to 40 °F for 2015 and 28 °F for 2016.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Cadmus 2016, ductless mini-split evaluation (MA and RI): Switching between heat pump and fuel at the crossover temperature is cheapest in principle, but at 2016 prices the gas crossover was above 70 F for the measured units, so pairing with gas saved nothing; only a modelled HSPF-13 unit (28-40 F crossover) would beat gas alone. Add the qualification: the evidence supports a narrower or weaker sta…
verified s37.a In Toronto, standard heat pumps with gas backup are the lowest-cost option for nearly all households under mid-range assumptions; in Edmonton they are the lowest-cost option for single family homes and townhouses built in 1940 or 1980, and in other cases gas and air conditioning is the most cost-competitive option.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Canadian Climate Institute 2023, heat pumps pay off: Lifetime-cost model: in Toronto a heat pump with gas backup is cheapest for nearly all households, and in Edmonton for 1940 and 1980 detached and townhouses, but gas with AC wins other Edmonton cases; this is lifetime cost including equipment, not running cost alone. "all" is stronger than the evidence (strongest support 0). Add the qualificatio…
verified s38.a Edmonton's result was due to low gas prices and climate conditions; the cold climate makes heat pumps relatively less efficient and results in greater use of the backup heating system.
qualified (qualify): no direct support, 8 qualifying.
warning needs-qualification: Canadian Climate Institute 2023, heat pumps pay off: Lifetime-cost model: in Toronto a heat pump with gas backup is cheapest for nearly all households, and in Edmonton for 1940 and 1980 detached and townhouses, but gas with AC wins other Edmonton cases; this is lifetime cost including equipment, not running cost alone. Canadian Climate Institute 2023, heat pumps pay off: In Edmonton's cold climate…
verified s40.a Most metered mini-splits kept heating at -10 to 0 °F, and a ductless cold-climate unit kept operating down to -25 °F; not every unit did (one field site's compressor did not run below 0 °F).
qualified (qualify): 1 supporting (1 independent, strongest 0.55), 4 qualifying, 1 context only.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. Winkler 2023 (NREL), cold-climate heat pump field validation: Among these ducted variable-speed units, Site 10's compressor did …
verified s40.b Their efficiency and output fell as it got colder.
supported (approved): 2 supporting (1 independent, strongest 0.65).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s40.c Their COP stayed above that of electric resistance heat in the cold.
qualified (qualify): 2 supporting (2 independent, strongest 0.6), 1 qualifying.
warning needs-qualification: Edwards 2024, inequities in heat pump bill savings: Edwards assumes single-stage heat pumps fall to COP 1 (no better than resistance) below 5F, while cold-climate two-stage units keep a higher COP, so the claim depends on equipment type. Add the qualification: the evidence supports a narrower or weaker statement.
verified s41.a Auxiliary heater usage generally increased at colder temperature bins.
qualified (qualify): 1 supporting (1 independent, strongest 0.45), 4 qualifying.
warning needs-qualification: Winkler 2023 (NREL), cold-climate heat pump field validation: Auxiliary heat runtime generally rose in colder bins and heavy use cut system COP over 30% at two sites, but the heavy users were undersized or faulty while six sites used little auxiliary heat. Add the qualification: the evidence supports a narrower or weaker statement.
verified s41.b A heat pump sized for the cooling load used more electric resistance backup.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Wilson 2024 (NREL and LBNL), heat pumps for all?: In Minnesota a high-efficiency heat pump sized for the heating load saves a median $182/yr, but sized for cooling it raises bills $96/yr through more resistance backup; reliance and the bill hit depend on sizing, not cold-climate units as such. Add the qualification: the evidence supports a narrower or weaker statement.
verified s41.c A controls problem made at least one site engage higher stages of auxiliary heat.
supported (approved): 1 supporting (1 independent, strongest 0.55).
info wording: Wording differs from the canonical claim; the reviewer confirms the passages entail it.
verified s43.a Against propane furnaces or electric heating, the Minnesota study estimated cost savings, though for retrofits paybacks will typically be longer than 10 years unless equipment replacement is necessary; and the mini-splits were always cheaper than propane.
qualified (qualify): 1 supporting (1 independent, strongest 0.6), 5 qualifying, 1 context only.
warning needs-qualification: Cadmus 2016, ductless mini-split evaluation (MA and RI): Switching between heat pump and fuel at the crossover temperature is cheapest in principle, but at 2016 prices the gas crossover was above 70 F for the measured units, so pairing with gas saved nothing; only a modelled HSPF-13 unit (28-40 F crossover) would beat gas alone. "always" is stronger than the evidence (strongest support 0.6). Add t…
verified s43.b Against gas, the result varied with the unit (measured units vs a rated cold-climate unit), the climate and gas prices (Toronto vs Edmonton), and for the measured mini-splits gas was cheaper below 70 °F.
qualified (qualify): no direct support, 4 qualifying.
warning needs-qualification: Cadmus 2016, ductless mini-split evaluation (MA and RI): Switching between heat pump and fuel at the crossover temperature is cheapest in principle, but at 2016 prices the gas crossover was above 70 F for the measured units, so pairing with gas saved nothing; only a modelled HSPF-13 unit (28-40 F crossover) would beat gas alone. Canadian Climate Institute 2023, heat pumps pay off: Lifetime-cost mo…
verified s44.a The field tests followed limited samples: dozens of metered mini-splits, 22 prototype sites, and one ductless unit for the coldest readings.
qualified (qualify): 2 supporting (2 independent, strongest 0.55), 8 qualifying, 1 context only.
warning needs-qualification: CEE 2018, ducted and ductless cold-climate heat pumps in the field: A ductless cold-climate unit delivered a consistent median capacity from 10F to -15F and kept operating to -25F at reduced capacity, though the authors say it needs backup below -15F; one monitored unit. Add the qualification: the evidence supports a narrower or weaker statement.
33 scenes, 8:14 at 150 words per minute including pauses. Narration is copied from the script; every chart value and on-screen citation is checked against a passage.
1. sc01 · 0:00 to 0:08 typography
Picture the coldest night of a northern winter: frost on the windows, and the heat pump outside still humming.
Visual. Deep blue field; "The coldest night" settles in large type, frost-white, as the sentence begins.
On screen. The coldest night
2. sc02 · 0:08 to 0:17 typography
Is it still heating the house, or has it turned into a very expensive space heater? Meters, not marketing, can settle this.
Visual. The question types on, centred; "very expensive space heater" turns amber as it is spoken.
On screen. Still heating, or a very expensive space heater?
3. sc03 · 0:17 to 0:25 typography
An NH Patch headline put the doubt bluntly: "Taxpayers On The Hook For Tech That Falls Flat In The Cold".
Visual. Quote card, text only; "Falls Flat In The Cold" turns red as it is spoken.
On screen. "Taxpayers On The Hook For Tech That Falls Flat In The Cold"
Cite. NH Patch headline
4. sc04 · 0:25 to 0:34 typography
The Mass Save heat pump page answers that "many provide efficient heat at outdoor temperatures as low as -15 °F".
Visual. Quote card, text only, mirrored from the previous one; "-15 °F" turns green as it is spoken.
On screen. "many provide efficient heat at outdoor temperatures as low as -15 °F"
Cite. Mass Save heat pump page
5. sc05 · 0:34 to 0:42 typography
So does a heat pump give up in the cold, or carry on as if nothing happened?
Visual. The question types on; "give up" in red, "carry on" in green.
On screen. Give up in the cold, or carry on?
6. sc06 · 0:42 to 0:52 typography
Start with the bluntest question: does the heat keep coming at all? Engineers score heating efficiency as a COP, and plain electric resistance heat scores 1.
Visual. Two lines build one per sentence: the question first, then "Resistance heat: COP 1" with the 1 held large.
On screen. Does the heat keep coming? / Resistance heat: COP 1
7. sc07 · 0:52 to 1:21 data-chart
In one evaluation, engineers metered the heat that dozens of ductless mini-splits delivered over back-to-back winters, and in the later one the units averaged a COP of 1.5 between -10 and 0 °F. That is 50 percent more heat per unit of electricity than resistance heat, which the evaluators called far superior; in the earlier winter, with deep snow and ice around the outdoor units, the units ran less efficiently.
Visual. Three tiles: resistance heat COP 1, metered mini-splits COP 1.5, and 50% more heat per unit of electricity; each lands as it is spoken.
| Resistance heat COP | 1 | 1 | count | CEE 2018, ducted and ductless cold-climate heat pumps in the field, lines 425-L435 |
| Mini-splits, later winter, average COP | 1.5 | 1.5 | count | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 2926-L2935 |
| More heat per unit of electricity | 50% | 50 | percent | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 2926-L2935 |
8. sc08 · 1:21 to 1:29 typography
And not only the cold-climate models: the ordinary ones in that evaluation also ran below 0 °F, just less efficiently.
Visual. One line; "below 0 °F" holds in ice blue.
On screen. Ordinary ones ran below 0 °F too
9. sc09 · 1:29 to 1:55 data-chart
A newer field test followed prototype cold-climate heat pumps, and leaving out the periods when backup heat or defrost was running, their median COP between 0 and 5 °F was 1.9. Those prototypes tested better in the lab, but in the field most still beat 1.75, the COP the EPA's Energy Star label requires of a cold-climate heat pump at 5 °F.
Visual. Two tiles: median field COP 1.9 at 0 to 5 °F, then the Energy Star requirement 1.75 at 5 °F beside it.
| Median COP, 0 to 5 °F | 1.9 | 1.9 | count | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 180-L189 |
| Energy Star requirement at 5 °F | 1.75 | 1.75 | count | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 180-L189 |
10. sc10 · 1:55 to 2:01 transition
So the heat keeps coming. But efficiency is only half of it.
Visual. Split field: "Efficiency" on the left dims, "Output" on the right lights up.
On screen. Efficiency / Output
11. sc11 · 2:01 to 2:08 typography
As the channel Technology Connections puts it, "along with getting less efficient, their actual heat output also drops."
Visual. Quote card, text only; "heat output also drops" turns amber.
On screen. "along with getting less efficient, their actual heat output also drops."
Cite. Technology Connections, YouTube
12. sc12 · 2:08 to 2:31 data-chart
In northern Minnesota, researchers watched one small ductless unit hold a steady median output from 10 °F all the way down to -15 °F. Below that it kept running at reduced output, and at -20 °F and colder it delivered only about 20 percent of what it was rated for at 5 °F.
Visual. One large tile: about 20% of rated output at -20 °F and colder, landing on "20 percent".
| Output at -20 to -25 °F, share of its rating at 5 °F | 20% | about 20 | percent | CEE 2018, ducted and ductless cold-climate heat pumps in the field, lines 478-L489 |
13. sc13 · 2:31 to 2:38 document-excerpt
The researchers' verdict was that units like it should not be used without backup below -15 °F.
Visual. The report passage scrolls in; the line about not using these units without backup highlights.
14. sc14 · 2:38 to 2:53 data-chart
Efficiency follows the same slope: in the same Minnesota research, outdoor temperature had the biggest effect on COP, and heat-pump-only cycles managed only around 1.3 at 10 °F, less than half what they reached on mild days.
Visual. Two tiles: COP around 1.3 at 10 °F in cold blue, COP 3.0 to 4.5 on mild days in warm orange.
| At 10 °F | 1.3 | about 1.3 | count | CEE 2017, cold-climate heat pumps in Minnesota, lines 1143-L1157 |
| Mild days, 50 to 60 °F | 3.0 to 4.5 | about 4.5 | count | CEE 2017, cold-climate heat pumps in Minnesota, lines 1143-L1157 |
15. sc15 · 2:53 to 3:10 document-excerpt
Then there is the frost. In cold, humid air, frost builds up on the outdoor coil, and to melt it the heat pump runs in reverse, pulling heat out of the house, which can cool the air coming out of the vents.
Visual. The PNNL defrost passage scrolls in; the line about rejecting heat from the conditioned space highlights.
16. sc16 · 3:10 to 3:26 data-chart
In the prototype field test, defrost used a median of 5 to 8 percent of the energy at the temperatures where it cost most, a small share, though one site spent 16 percent of its time in defrost.
Visual. Two tiles: median defrost share of energy 5% to 8%, then the outlier site's 16% of time in defrost in amber.
| Median share of energy, 5 to 35 °F | 5% to 8% | 8 | percent | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 1249-L1310 |
| One site's time in defrost | 16% | 16 | percent | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 1249-L1310 |
17. sc17 · 3:26 to 3:29 typography
Which brings us to the real catch: backup heat.
Visual. One line punches in; "backup heat" in red.
On screen. The real catch: backup heat
18. sc18 · 3:29 to 3:37 typography
When a heat pump's electric backup strips kick in, Jordan Dearden warns, "They'll also absolutely destroy your electric bill."
Visual. Quote card, text only; "destroy your electric bill" turns red.
On screen. "They'll also absolutely destroy your electric bill."
Cite. Jordan Dearden, YouTube
19. sc19 · 3:37 to 4:01 data-chart
In the prototype test, the units ran mostly on the heat pump alone, even at the lowest temperatures, and backup-only operation was minimal above 0 °F. But in the coldest range with enough data, below 0 °F, backup elements were on 25 percent of the time, and only some of the 22 sites had enough data that cold.
Visual. Two tiles: 25% of the time with backup on below 0 °F, and the 22 sites in the study, the second tile smaller.
| Time with backup on, -10 to 0 °F | 25% | 25 | percent | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 1169-L1193 |
| Sites in the study | 22 | 22 | count | PNNL 2025, DOE Cold Climate Heat Pump Challenge field results, lines 1169-L1193 |
20. sc20 · 4:01 to 4:15 typography
Some backup use came from the thermostat: backup heat began at around 30 to 35 °F, and the researchers expect it would start much lower without homeowners changing setpoints, or with smarter thermostat controls.
Visual. One line; "30 to 35 °F" in amber.
On screen. Backup began at 30 to 35 °F
21. sc21 · 4:15 to 4:28 document-excerpt
Some came from faults: two sites fired up higher stages of backup heat during defrost, at least one did it in normal heating too, and the manufacturer called that a controls problem.
Visual. The PNNL passage on defrost strategies scrolls in; the manufacturer's controls-problem line highlights.
22. sc22 · 4:28 to 4:43 typography
In the Minnesota study's all-electric system, the resistance booster joined in below 30 °F, yet because the heat pump ran first, the whole system still beat plain resistance heat even at the coldest reading, around -7 °F.
Visual. One line; "-7 °F" in ice blue.
On screen. Still beat resistance heat at -7 °F
23. sc23 · 4:43 to 4:57 typography
In another field study, one site's compressor would not run below 0 °F at all, even with its lockout disabled, most likely because of an on-board control setting that cannot be adjusted in the field.
Visual. One line; "below 0 °F" in red.
On screen. One compressor would not run below 0 °F
24. sc24 · 4:57 to 5:18 data-chart
Sizing matters too: one study estimated that in Minnesota a high-efficiency heat pump sized for the full heating load cut median bills by $182 a year, while one sized for the cooling load raised them by $96, because it leaned harder on electric resistance backup.
Visual. Two tiles: $182 a year saved in green for heating-load sizing, $96 a year more in red for cooling-load sizing.
| Sized for heating: saved | $182/year | 182 | usd | Wilson 2024 (NREL and LBNL), heat pumps for all?, lines 942-L946 |
| Sized for cooling: raised | $96/year | 96 | usd | Wilson 2024 (NREL and LBNL), heat pumps for all?, lines 942-L946 |
25. sc25 · 5:18 to 5:34 data-chart
So what does all this do to the bill? For the Minnesota homes, which had heated with propane or electricity, the researchers used performance maps and typical Minnesota weather to estimate cost savings of 30 to 56 percent.
Visual. One tile: estimated savings of 30% to 56% against propane or electric heat.
| Estimated cost savings | 30% to 56% | 56 | percent | CEE 2018, ducted and ductless cold-climate heat pumps in the field, lines 495-L518 |
26. sc26 · 5:34 to 5:56 data-chart
Gas is harder to beat: at the prices of those winters, the mini-splits from the first evaluation cost less to run than an 80 percent efficient gas system only above 70 °F, which the evaluators said made them no direct replacement for gas. Against propane, they were cheaper at every temperature.
Visual. One tile: the break-even outdoor temperature against 80% efficient gas, above 70 °F, in red; it dims as s35 turns to propane.
| Against 80% efficient gas | 70°F | gt 70 | count | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 3436-L3459 |
27. sc27 · 5:56 to 6:10 data-chart
Even the measured units listed as cold-climate broke even with gas only above 58 °F; on paper, the rating curve of a then-current cold-climate model brought it down to between 28 and 40 °F.
Visual. Three tiles: measured cold-climate units above 58 °F, then the rated model at 40 °F and 28 °F, the paper tiles outlined rather than filled.
| Measured units | 58°F | gt 58 | count | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 3436-L3459 |
| Rating curve, 2015 winter | 40°F | 40 | count | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 3436-L3459 |
| Rating curve, 2016 winter | 28°F | 28 | count | Cadmus 2016, ductless mini-split evaluation (MA and RI), lines 3436-L3459 |
28. sc28 · 6:10 to 6:41 typography
A cost model for Toronto and Edmonton that counts equipment as well as energy points the same way: in Toronto, standard heat pumps with gas backup came out the lowest-cost option for nearly all households, but in Edmonton that held only for some house types, and gas with air conditioning won the rest. The model's reasons were cheap gas and the cold itself, which makes heat pumps less efficient and calls on the backup more.
Visual. Two lines build one per sentence: Toronto in green, Edmonton in amber.
On screen. Toronto: heat pumps with gas backup / Edmonton: cheap gas and the cold
29. sc29 · 6:41 to 6:46 typography
So do heat pumps still work when it gets really cold?
Visual. The question types on, centred.
On screen. Do heat pumps still work when it gets really cold?
30. sc30 · 6:46 to 7:09 typography
In these field tests most units did: they kept heating below zero, with less efficiency and less output, and still got more heat from each unit of electricity than resistance heat. The catch is backup heat, which ran more as it got colder, and more where a unit was sized for cooling or its controls misbehaved.
Visual. Two lines build one per sentence: the first in green, the second in amber.
On screen. Most kept heating below zero / The catch is backup heat
Cite. Cadmus 2016; CEE 2018; NREL 2023; PNNL 2025 field tests
31. sc31 · 7:09 to 7:33 typography
And do they still save money? Against propane or electric heat, the studies here say yes on running costs, though the Minnesota researchers put retrofit paybacks at over 10 years unless the old system needs replacing; against gas, it turns on the unit, the climate and local prices, and for the mini-splits measured in the first evaluation, gas won.
Visual. Two lines build one per sentence: "Save money?" then "Against gas: it turns on the unit, climate and prices".
On screen. Do they still save money? / Against gas: it turns on the unit, the climate and prices
Cite. CEE 2018; Cadmus 2016; Canadian Climate Institute 2023
32. sc32 · 7:33 to 7:44 typography
Each of these field tests followed a limited number of units, so they show what heat pumps can do in the cold, not what every installation will.
Visual. One line, smaller and grey.
On screen. A limited number of units
33. sc33 · 7:44 to 8:14 typography
If your winters bite, ask the installer whether the unit is a cold-climate model, and whether it is sized for your heating load or only for your cooling. Ask what temperature the backup is set to start at, then keep the thermostat steady and watch how often the backup comes on. And if you heat with gas, work out the break-even temperature at your own prices before you decide what the furnace is for.
Visual. Three lines build one per sentence, as a checklist.
On screen. Cold-climate model, sized for heating / When the backup starts / Break-even at your own prices