tl;dr -- For the west coast of the U.S., a 20% decrease in heating demand and 18% increase in cooling demand results in a 5% reduction in carbon emissions from electricity and natural gas usage
Here's the approach I came up with to try and answer this question with some hard data.
1. Find a region of the U.S. which had a cold winter and mild summer, followed by a mild winter and a hot summer.
This is the sort of change we'd expect to see with global warming. If we look at the data for consecutive years, we minimize the effects of changes to the grid and/or population.
The U.S. Energy Information Agency publishes monthly regional data for heating and cooling degree days as part of the short term energy outlook.
Looking at data for all regions, I determined that 2013 to 2014 for the Pacific region best meets the criteria. The area includes Alaska, Washington, Oregon, California, and Hawaii. This area stretches pretty far from north to south, so makes a decent sample that we could use to extrapolate. From 2013 to 2014 this region experienced:
- 20% increase in cooling degree days (= more energy needed for cooling)
- 17% decrease in heating degree days (= less energy needed for heating)
The population in this region grew by slightly more than 1% over that time period, so likely not a huge factor in any changes in energy usage:
2013 2014 % change
Alaska 735,132 736,732
California 38,332,521 38,802,500
Hawaii 1,404,054 1,419,561
Oregon 3,930,065 3,970,239
Washington 6,971,406 7,061,530
51,373,178 51,990,562 1.2%
2. Determine total usage of electricity and natural gas for heating over that time period
EIA provides monthly natural gas consumption by state. This is broken down by end use, so I looked at residential and commercial uses, which would predominantly cover heating. I left out use by vehicles and industry. Some industry uses would include heating, but for all five states this usage was fairly consistent month-to-month, and less than either residential or commercial usage for most months.
EIA also provides monthly generation by fuel source by state (scroll to "Generation" then "State-level generation and fuel consumption data" then "Monthly (back to 2001)") .
Here's a chart showing degree days (left axis) and energy for heating and electricity (right axis) using the above data sets:
3. Compare changes in heating, electric generation, and associated CO2 emissions
Here's a chart comparing 2013 and 2014 usage for the largest sources:
And here's the data summarized in a table with emissions data and percent change calculated:
2013 2014 % change
Heating degree days 3,365 2,776 -18%
Cooling degree days 890 1,069 20%
Heating natural gas (MWh) 291,012,539 256,759,857 -12%
CO2 emissions (metric tons) 54,613,680 48,185,555
Coal generation (MWh) 26,701,933 25,562,047 -4%
CO2 emissions (metric tons) 27,235,971 26,073,288
Natural gas generation (MWh) 297,462,431 289,771,833 -3%
CO2 emissions (metric tons) 133,858,094 130,397,325
Hydroelectric generation (MWh) 273,042,255 263,321,494 -4%
Nuclear generation (MWh) 52,745,666 52,966,598 0%
Solar generation (MWh) 7,708,892 19,940,312 159%
Wind generation (MWh) 55,861,453 58,717,774 5%
Other generation (MWh) 68,296,211 67,585,528 -1%
Total electric (MWh) 781,818,840 777,865,587 -1%
Total CO2 (metric tons) 215,707,745 204,656,168 -5%
This is a pretty limited (and possibly spurious) analysis, but it supports your hypothesis: a warmer climate reduces energy needs in the winter by a greater factor than the summertime increase.
The really interesting thing, especially in a U.S. context, is that a huge proportion of natural gas is used for heating in the winter. But in the summer, electricity from cooling comes from a more diverse set of sources. So any reduction in heating demand necessarily reduces CO2 emissions, but as the grid adds more wind and solar, increases in cooling demand don't necessarily increase CO2 emissions.
Some caveats to this:
- Much of the population of California is on the southern coast, where heating and cooling demand is already limited due to the mild climate
- None of this accounts for changes in population, economic activity, policy, or precipitation
- Electric usage is for all uses (not just heating/cooling) so doesn't account for any economic factors
- California added A LOT of solar, and some wind, during this time period, which contributed to the overall reduction in CO2 emissions from the electric sector