I have a theory that I am trying to develop empirical support for. I believe that the migration of people from the Rust Belt to the Sun Belt has caused an increase in overall electricity usage and carbon emissions due to increased air conditioning demand. This page from EIA indicates that about 13% of electricity usage is due to air conditioning, while only 9% is due to space heating. I suspect that much of this heating is actually used in the South (like my house in Virginia), because in colder climes houses are much more likely to be using oil or gas heaters (like the house where I grew up in Massachusetts).

Specifically, I want to examine the very high population growth in very hot states (Texas, Florida, Georgia, North Carolina) compared to the very high out-migration from Ohio, Michigan, New York, etc.

I need more evidence to test my theory. I am looking for a breakdown of electricity by usage and by state (or at least by region). If I know the average annual per-capita kW-h usage for air conditioning, space heating, water heating, and boiler fans in both Texas and Ohio, I can calculate the net carbon effect of moving from one state to another. Also, electricity doesn't cover everything; the farther north you go the more likely that heat will be provided by natural gas (or the older oil boilers).

The EIA has some information that I want, but does anyone know where to find detailed breakdowns by state (or region) of both electricity usage and fossil fuel heating?

  • 1
    Your use of the term "air conditioning" seems to imply that it is only used to cool homes. Where I live reverse cycle air conditioners are used to both cool and heat homes. Unless you compensate for this your numbers are likely to be way off.
    – Tim
    Commented Jul 7, 2018 at 21:10
  • @Tim Air conditioners cools homes. Heat pumps heat homes. The equipment is basically the same, but the name varies with the function.
    – kingledion
    Commented Jul 9, 2018 at 1:19
  • Really the more north you go the more likely to be on gas? TX is a net producer of natural gas and it is common. Where are you going to put all these people?
    – paparazzo
    Commented Jul 9, 2018 at 21:16

2 Answers 2


Original answer (which had a long explanation):

I proposed using degree days information to determine a baseload value for consumption of electricity and natural gas by region, and assume anything above this baseload was used for heating and cooling. This assumed that all seasonal changes in energy consumption were for heating and cooling only (so no change in lighting, cooking, water heating, etc).

New (short) answer:

@ChrisH made a comment on my original answer which sent me searching, and in the process I found some better resources. Here's the new process:

  1. Determine how much of each type of fuel (electricity, natural gas, propane, oil) is used for heating and air conditioning, by state.
  2. For electricity, determine the fuel mix by state.
  3. For each fuel, determine the carbon emissions.
  4. Put it all together to determine per capita energy use and carbon emissions for heating and cooling, by state.

New long explanation:

1. Fuel consumption by residential end use, by state

This information is available in the 2009 EIA Residential Energy Consumption Survey:

Excerpt from RECS data

(2015 RECS data is available as of early 2017.)

2. Fuel consumption for electricity generation, by state

Also from the EIA: consumption of fuels for electricity generation. To match the end use survey, take the data from 2009. For each state, add up all the fuels to determine what percent comes from each fuel type. Monthly data is available if you want to look at seasonal variation.

3. Emissions factors by fuel type

For electricity: The EIA provides used to provide a summary of carbon emissions per kWh of electricity produced, by fuel source. Now they provide a list of the data and conversion factors needed to make this determination per state. I did the math for the U.S. and the two states in the question title.

Metric tons of CO2 per MWh of electricity generated by region:
  US total
   - Coal          1.02 
   - Natural gas   0.45
   - Coal          1.05 
   - Natural gas   0.48
   - Coal          0.97 
   - Natural gas   0.45

The variations from state to state depend on the type of coal used, how plants are operated, and the age, efficiency, and emissions controls of the plants.

For heating fuels: The data in step 1 above gives consumption by BTU and physical units (kWh, cubic feet, gallons, etc). This summary gives the carbon dioxide factors by both measures.

  • This is really excellent information and some good strategies. Thank you.
    – kingledion
    Commented Feb 2, 2017 at 19:23
  • 2
    This is a great answer, but the "biggest assumption" in your penultimate paragraph is (IMO) too big. Lighting demand will change over a similar seasonal pattern (what % of total -- I don't know) and cooking habits do change as well (e.g. roasts and stews in winter, salads and barbecues in summer).
    – Chris H
    Commented Feb 3, 2017 at 9:19
  • Even better. You already had my vote but this makes a big difference
    – Chris H
    Commented Feb 3, 2017 at 16:21
  • 1
    @ChrisH Thanks for your help, both of you. With this new chart breaking down energy consumption by climate regions, the calculations became very easy. In short, my hypothesis was wrong. Comparing the Cold region to the Hot-Humid region, the cold region uses 68% more overall energy and 118% more energy for heating. However, since Natural Gas, Propane, and Fuel Oil are 2-3 times more efficient than electricity, the overall carbon impact in the Cold region is only 15% higher for heating and 7% overall. Still, its better for the Earth to live in Texas than Ohio. I guess. Thanks again for the help.
    – kingledion
    Commented Feb 3, 2017 at 18:49

This might be interesting in a theoretical sense of energy usage and locations to live, but isn't Texas having problems with negative energy prices? Texas has been quietly adding to their renewables portfolio, and they've only scratched the surface. Bloomberg.com: Power worth less than zero as green energy floods the grid

Of course, this is a problem only if you're trying to sell electricity. If you're trying to buy electricity, or encourage a growing economy, it's a "problem" we all want to have.

So as a theoretical exercise, sure. In terms of air conditioning and GHGs, Texas is actually moving the other way. The duck curve of solar energy production is actually a boon to a/c use since they tend to coincide in terms of time of day. (I know this sounds backwards and surreal - you've gotta love data collection. :)

  • It is an interestingly dynamic problem. The move to renewables favors electricity vs fossil fuel heating; increasing the carbon efficiency of air conditioning in Texas while not affecting the carbon cost of heating Ohio. On the other hand, future global warming, assuming it continues for another 1-2 C at a minimum, will strongly favor Ohio, since cooling demand will increase faster in Texas and heating demand will decrease faster in Ohio.
    – kingledion
    Commented Jan 10, 2019 at 21:36
  • I suspect that the reduction of colder temperatures in the winter for Ohio will matter more than hotter days in Texas (which supports your thought nonetheless). Winters warming faster than summers: vox.com/science-and-health/2018/12/20/18136006/… But overall Texas will be able to add much, much more in terms of renewables (they have sun and wind). This article has various maps of sun and wind in the US that can be tapped. vox.com/science-and-health/2019/1/7/18167351/…
    – FreeText
    Commented Jan 10, 2019 at 21:45
  • The OP is asking for data sources; this does not answer that.
    – user2451
    Commented Jan 11, 2019 at 9:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.