Two points and a straight line
Manufacturer submittals give capacity and input power as a grid: entering water temperature down one axis, flow across the other, one block for heating and one for cooling. Between the printed rows the honest thing to do is interpolate, and over a twenty degree span a straight line is close enough that the error is smaller than the uncertainty in what the loop will actually deliver.
On the defaults, 3.1 at 30 degrees and 4.2 at 50 gives a slope of 0.055 COP per degree. At 38 degrees that is 3.54. Delivering 42,000 BTU/h at that COP takes 42,000 divided by 3,412 divided by 3.54, which is 3.48 kilowatts, and at 16 cents a kilowatt hour that is 56 cents an hour.
Push the line beyond the rating points and it stops being interpolation. Real performance curves bend, and they bend the unhelpful way: a straight line extrapolated to a colder entering water temperature is optimistic about what the machine will do there. The page says so when it happens rather than printing the number quietly.
COP is not EER, and the gap is 3.412
Both ratios measure heat moved against electricity used. COP does it in consistent units, watts over watts, so it is dimensionless. EER does it in mixed units, BTU per hour over watts, and since a watt is 3.412 BTU per hour, an EER is a COP multiplied by 3.412. A machine quoted at 17 EER in cooling has a COP just under 5.
The reason this matters beyond arithmetic is that cooling ratings on ground source equipment are commonly printed as EER while heating ratings on the same page are printed as COP. Reading down the column without noticing the header changes is easy, and the resulting error is a factor of three and a half, which is large enough to change a decision and small enough that both numbers still look like plausible performance figures.
Where the ground fits
In heating, the heat delivered indoors is the COP times the electricity in, and the compressor work is already part of that delivered heat. The ground supplies the rest: one minus one over COP. At 3.54 that is 71.8 percent, so out of 42,000 BTU/h delivered, 30,136 comes from the ground and 11,864 is the compressor, which is the 3.48 kilowatts arriving as heat.
In cooling it reverses. The ground takes the building heat and the compressor work together, so the ground-side load is the cooling capacity times one plus one over COP. The same machine at the same COP would push 53,864 BTU/h into the ground while removing 42,000 from the building. This is why loop lengths for cooling-dominated buildings come out longer, and it is worth carrying into any sizing conversation.
What the cost table is actually saying
Cost per million BTU delivered is one over the COP, converted. A million BTU is 293.1 kilowatt hours of pure electricity; at a COP of 3.54 it takes 82.8 kilowatt hours, and at 16 cents that is 13.25 dollars. Walk the entering water temperature down twelve degrees and the COP falls to 2.88, the same million BTU takes 101.8 kilowatt hours and costs 16.28 dollars. Twelve degrees of entering water is worth about 23 percent on the heating bill.
That is the argument for spending money on the loop rather than on the machine, and it is also the argument for caring whether the loop is drifting. Entering water temperature is not a fixed property of the installation. It falls through a heating season as the ground gives up heat, it recovers over the summer, and on a loop carrying an annual imbalance it can fall a little further each year without anybody noticing until the machine starts running out of capacity on cold nights.
The fuel comparison, and what it leaves out
Fuel cost per million BTU is the price per unit divided by the useful heat per unit, scaled up. Propane at 91,500 BTU a gallon burned at 92 percent gives 84,180 useful BTU a gallon, so a million BTU takes 11.88 gallons, and at 2.85 dollars that is 33.86 dollars. Against 13.25 dollars for the heat pump on the same defaults, the gap is 20.61 dollars a million BTU.
What that comparison does not carry is a standing charge, a delivery minimum, the tank rental, the circulator running alongside the compressor, the difference in maintenance, or the capital cost of getting from one system to the other. It also compares two prices on one day, and neither of them stays still. Use it to understand the shape of the difference rather than to settle an argument.
Questions people ask
How do I convert EER to COP?
Divide by 3.412. A watt is 3.412 BTU per hour, so an EER of 17 is a COP of 4.98. The conversion matters because ground source submittals often print cooling performance as EER and heating performance as COP on the same page, and reading straight down the column without noticing gives an answer wrong by a factor of three and a half. This page takes whichever scale the submittal used and converts internally.
Does a higher COP mean more heat comes from the ground?
Proportionally yes, but less than people expect, and the confusion comes from which side the compressor work sits on. Delivered heat is COP times the electricity in, and the compressor work is already inside the delivered figure. The ground supplies one minus one over COP: 67 percent at a COP of 3, 75 percent at 4, 80 percent at 5. Going from 3 to 5 moves the ground share by 13 points, while it cuts the electricity by 40 percent. The electricity saving is the point, not the ground share.
What entering water temperature should I assume?
It is an output of the loop design rather than something you pick, and this page will not supply one. It is what the ground and the loop hand the machine at the condition you care about, and it moves through the season, falls as the loop is worked, and drifts over years if the annual heat balance is one-sided. The table on this page walks it up and down twelve degrees around whatever you enter, so you can see what the uncertainty is worth before committing to a figure.
Why does the COP fall as entering water gets colder?
Because the compressor has to lift heat across a bigger temperature difference. In heating, the machine takes heat in at the loop temperature and delivers it at the distribution temperature, and the work needed grows with the gap between the two. Every degree the loop falls widens that gap. It is the same reason an air source unit loses performance on a cold night, except that a ground loop moves slowly and predictably instead of following the weather hour by hour.
Is a heat pump cheaper to run than my current fuel?
That is arithmetic on two prices you supply, and this page does it rather than answering it. Fuel cost per million BTU is the unit price divided by heat content times seasonal efficiency; heat pump cost per million BTU is 293.1 kilowatt hours divided by the COP, times your tariff. On the defaults here the gap is 20.61 dollars a million BTU in favour of the heat pump, but that is entirely a product of a 16 cent tariff and 2.85 dollar propane and it moves with both. Standing charges, tank rental, the circulator, maintenance and the capital cost of changing over are all outside it.