Here's a test. Can you answer the following three questions about seasonal peak demand in the Southeast USA?

Is the Southeast, as a region,

1. Winter peaking,
2. Summer peaking, or
3. Dual peaking?

Is the Southeast's winter peak demand

1. Increasing,
2. Decreasing, or
3. Flat?

At a winter peaking utility in the Southeast, are the frequency of winter peak hours

1. Much more than,
2. About the same, or
3. Much fewer than

the frequency of summer peak hours?

If you answered #3 to all of the questions above, you are already pretty knowledgeable about the Southeast's peak electric load. But there still may be a few surprises in store for you.

This week, the Southern Alliance for Clean Energy released "Seasonal Electric Demand in the Southeastern United States." This in-depth look at two decades of hourly load data from 22 utility systems has many findings that will interest electric utility wonks. Here are a few highlights:

• The Southeast is now a dual peaking region due to declining summer peaks. Individual utility systems are evenly balanced between winter and summer peaking systems. Eight appear to be winter peaking, six appear to be dual peaking, and eight appear to be summer peaking. Winter peaks are more common recent years than in the earlier years of our 20-year analysis. Even so, winter peaks at all utilities are less frequent than summer peaks.
• Even though the Southeast is a dual peaking region, the vast majority of peak hours occur during the summer.
• Winter peak variability is higher than summer peak variability, but there is no evidence of an increase or decrease to seasonal peak variability.
• Summer peak events tend to be of similar length (on average, under 5 hours) and have a similar load shape, although a few utilities tend to have longer duration peak events of up to 12 hours.
• For most utilities, winter peaks are of shorter duration than summer peaks. However, the most strongly winter peaking systems also have occasional long duration peak events of 14 hours or more and occasionally more than 20 hours.

It's been an interesting project, and sets up a number of questions for future research that we weren't able to answer with the data sets we examined. I'd be happy to discuss our findings, or how we built this unique dataset, in the comments.