Whenever I sing the praises of solar PV as a means to hugely reduce U.S. carbon emissions while strengthening the security of the grid, I get people complaining that I'm not disclosing the carbon impacts of solar panel production.

So let's get this straight.. solar panels are at least 20x better on the climate than coal, kWh for kWh. I say at least because the most often cited report is by Danish utility Vattenfall from 1999. It looks at only 3 countries -- Japan, Sweden and Finland -- all of which are fairly dark and dreary, and it does not account for recent advances in PV production (new solar panels are significantly more efficient).

Based on that study, solar PV works out to about 50g of CO2 per kWh compared to coal's 975g of CO2 per kWh, or about 20x "cleaner."

Coal power plants in the U.S. are considerably less regulated (and therefore less efficient) than their European counterparts, making them more carbon-intensive. And the U.S. has much greater solar access than Japan or Sweden, making solar PV less carbon-intensive. 

Here's a comparison of several scientific studies on the carbon life-cycle costs of solar (both PV and thin film). The gist is an average 3 year payback... meaning after 3 years the solar panels create as much energy as was consumed in their initial production and installation. In the remaining 22-27 years of their life span they would be "carbon negative."

Now depending on where you live, this payback period could vary significantly. Here's a chart of "peak sun hours" in the U.S.  (a numerical equivalent for the amount of time that the sun averages 1 kWh per m2 of intensity):

Say you bought a standard 224 watt solar panel and stuck it on your roof in an optimal location with no shadows. The panel is rated to produce 224 watts of power in 1 hour of peak solar intensity. So if you live in Santa Fe, New Mexico you would generate 6 x 224 watts = 1.33 kWh in one day with that one panel. You have to shave off about 10% for efficiency losses (inverter and wiring) so in Santa Fe you're looking at 1.2 kWh of production per panel per day, or 438 kWh's per year.

In Cooperstown, New York you have roughly half the solar intensity, approximately 3 hours per day of peak, so there you would expect to produce 0.6 kWh with the same panel each day, or 219 kWh's per year. The national average is somewhere in the middle (about 4.5 hours per day) so on average in the U.S. a 224 watt panel creates 0.9 kWh per day or 328 kWh's per year. Over a typical 25 year lifespan that panel would produce about 8,200 kWh's.

Stockholm, Sweden where the Vattenfall study was performed, gets 3 peak sun hours just like Upstate New York, producing about 5,500 kWh's over its typical 25 year life span. According to the study, a solar panel in Sweden generates 50g of CO2 per kWh so that is a total of 274 kg of CO2, or 600 lbs of CO2.

Though there are some efficiency losses for solar panels in hotter climates, ostensibly the same 600 lbs of CO2 could get you twice the amount of energy in Santa Fe, New Mexico making it 40x less carbon-intensive than coal power and the national average would be about 30x less carbon intensity for domestic solar PV.

If you want to get really exact efficiencies for your area check out the awesome PVwatts calculator.