In the US, the effort to slash the CO₂ emissions of our electrical grid has actually gotten an early bump from the low price of natural gas. New natural gas plants can produce electricity with about half the emissions of the older, dirtier (and now more expensive) coal plants they’re replacing. But there’s been a lot of debate over a drawback that eats into that gain: some of the natural gas leaks into the atmosphere during production and transportation. Precisely how much methane (a greenhouse gas) leaks is important—leak enough and natural gas isn’t actually better than coal.

Some estimates have suggested that leaks were a serious problem, while others produced much lower numbers. Now, a new study figures out why the different analyses produced a range of numbers:  it depends on the time of day people were looking for leaks.

Top down or bottom up?

Studies estimating real-world leakage have come in two primary flavors. The first is referred to as a “bottom-up” approach. This involves walking around gas wells and pipeline equipment while measuring leakage. With estimates for each process or type of equipment, you scale up to the big picture based on an inventory of equipment and records of the amount of natural gas produced.