Yes, I've made mostly Volt Amp (VA) measures on the input and assume the power factor is close to 1... in any case the power in is measured to include that to the loop pumps, the compressor and the circulating fan. Using the VA I always have an overstatement of the power being used. VA is always greater than or equal to Watts, so the ture COP is in fact a little higher In any case, the resistive back-up and supplemental is fused separately as those units use a lot more power (amps) I have a two stage resistive backup which isn't suppose to run unless the outside is zero and I'm holding the inside at 70. Neither is the case here... it may have been zero a night or two in the last 20 years but not more than that.
The best measure I have of the BTU output is to measure the air temperature at a duct well removed from the plenum and then switch in a 5KW (17 KBTU/Hr) resistive heating element and measure the delta temperature... with the circulating fan running at the same speed. This allows me to estimate the number of BTU/Hr it takes to give the temperature rise without the resistive supplement. This number comes out very close to the factory specified heating BTU. Using this and the VA (more than Watts) input to get those BTUs I get a COP approaching 4 on low (compressor and fan, and only one loop pump) and 3.5 on high. This will depend on the ground loop temperature, but the factory rates the units with the loop at 32 degrees best I can recall, colder than it ever is I believe. The loop has antifreeze in it just in case. My loop is a twin 250' vertical loop which hits the water table at about 30 feet, so it has very good heat conduction with the earth.
Given my electric usage, I can check but think it is about 1,000 KWH per month average (12K KWH per year). Said another way with an all electric, including hot water, cooking, washer/dryer and heat/cooling my annual electric cost is about $1,800 ($150 per month) for a 2000 sq ft (well insulated) house. I have neighbor, with a 3,000+ sq ft house who uses over $1,000 in oil in one winter month. They spend more than my total energy cost just to heat during the 3 month winter high heat period.
The high cost of oil is the big problem in the above, when I had my Geo installed I estimated it would cost me about the same as oil - oil must have been about $1 gallon then and electricity about 10 Cents/KWH. Now it is $3.50 a gallon verses 15 Cents/KWH in NJ, one of the higher cost electricity states. The 350% increase in heating oil and a 50% increase in electric makes the big difference. Happily Obama didn't get his Cap and Trade legislation through, but he's working on those goals via the regulatory end and I may see the electric rates rise to make it as expensive as oil when heating with an electric HP. In NJ electricity is generated with about 50% nuclear and with a large component from coal, not oil or natural gas.
Hope I didn't ramble too much, or make too many typos... all the numbers above are from memory I haven't consulted any notes.
My HP is approaching 20 years old, the newest has slightly higher COP and EER, but mine works on. I had no maintenance in the first 15 years other than changing the forced air filters, no service calls or maintenance. In the last 4+ years I have had to spend about $2,000 on repairs, including loop pumps and circulating fan. Still, averaged out the maintenance cost have been only $100 a year, that's less than an oil furnace tune-up I will guess.
As for the subject water heater using outside air, I assumed a closed loop, not just exhausting - of course that's not good for heating water if that makes the input air at 20 degrees or less.
I think this at least lends credibility to the 3.9 COP estimate from my installer. His other calculations, such as winter design temp, have been right on the money (actually a little conservative).
