Water emissions 2: electric boogaloo

I worked on this article until late april 2021, never quite fully finished it, but just changed some stuff today and released it now(october 2023) as I feel like it is a nice shitpost, enjoy!


So, a while ago, I thought, there's a lot of CO2 being emitted. But then I thought, wait, does that also mean there is a lot of water being emitted?

Take the following reaction formula, of methane combustion:

CH4 + 2O2 ⟶ CO2 + 2H2O

As you can see, there is CO2 being emitted, but also H2O: water!

Then I had a goal: To calculate a rough estimate of how much water there is being emitted. This is doable, I have made some assumptions to make it easier:

  1. Assume that every hydrogen atom in fuel will oxidize into H2O and assume that every carbon atom burns into CO2. Very wrong for obvious reasons. Such as not everything of oil is burned(ex. asphalt, plastic), coal leaves ash. Obviously all of this emits other substances as well.
  2. Assume that the burning reaction is 100% clean.
  3. All other atoms in the fuels ignored.
  4. Natural gas, oil, and coal are the only fuels being calculated. Flaring and such is ignored.

All these assumptions are very wrong, but again, the goal is calculating a rough estimate. Otherwise it would be tough for me to calculate. I am thoroughly impressed by how scientists calculate CO2 emissions.

Natural Gas ☁️

Natural gas is mainly comprised of methane. Methane is the simplest alkane, and it contains the highest hydrogen/carbon ratio, therefore I expect a lot of water from this.

Tbh natural gas is natural, but that doesn't mean it's good for nature, it's a huge source of CO2 emissions which causes dangerous climate change which destroys nature.

First start with the CO2 emissions from natural gas in 2018. I use the IEA source as this is slightly lower than the Our world in data source, leading to a more conservative estimate.

$$M_{CO_{2}} = 7.104 \space Gt = 7.104 *10^{15} g$$

Atomic masses(rounded, in the underlying code more precise atom masses are used):

H1 u
C12 u
O16 u

Now we need to determine the mass of the carbon. The ratio of carbon in carbon dioxide is: $${\frac {C} {C+2O}}={\frac {12} {12+16*2}}={\frac {12} {44}}$$

$$M_{C} = \frac{12}{44}*7.104*10^{15} \approx 1.938 *10^{15} g$$

That is enough to make a pencil, maybe even 5 of them. Next is the tricky part. It is figuring out how much hydrogen is per carbon atom.

This has several problems, as illustrated above. Also, it is really hard to find a good source. Please forgive me. I use the following source: Chemical Composition of Natural Gas, by union gas

I have decided to take methane, ethane, propane, and carbon dioxide in this equation. For simplicity I ignore the rest.

Now do the opposite for the maximum water scenario:

Now we have a minimum and maximum scenario. Here you can adjust it further:


The difference between the min and max scenario are not large, it seems.

Now calculate the hydrogen mass:

$$M_{H} = \frac{M_{C}}{12 u}*hcratio*1u$$

Next, create water:

There we have it!

Oil 🛢️

Next is oil. Here is the source of the chemical composition

Composition of oil, mass

This is by mass. Convert it by dividing the percentage by the atomic mass(ex. 83 / 12 = ~6.9%)

Composition of oil, mole

Start with the CO2 mass again, convert it to C mass

Calculate the minimum and maximum water scenario. X is not required as there are only two components:

Next, create water:

Coal ⛏️

Coal is not going to be calculated by min-max values, instead I take lignite as a minimum scenario and bituminous coal as a maximum scenario. These sources could definitely be improved. Previous sources had a little bit of higher difference between them, but not much(+/- 10%)

Composition of lignite & bituminous coal, mass
Lignite(source)Bituminous(source, page 6)

Apply the same trick:

Composition of lignite & bituminous coal, mole

As you can see the H/C ratios are not really that different, and yet lignite is a worse coal. How come? Well, in lignite there's much more other stuff(such as moist), which reduces the heat content of the fuel(i.e. you need more lignite to get the same amount of heat as it contains a lot of moist).EIA So you burn more of it. Also worth noting is that coal really doesn't contain all that much hydrogen.

Now water again:

Sum them all up, and now you have the total amount of water:

Comparisons(hover over the number to see the calculation):

That either sounds a lot to you, or little. But think about it this way: this water previously didn't exist. Once emitted, you can't easily undo it. Keep in mind this is just for the year 2018 only, accumulated over time these numbers get much bigger. As for sea level rise, the effects of climate change are eclipsed those from water emissions with several orders of magnitude(IPCC AR6).


I am not a scientist. In fact, I unfortunately didn't even finish the equivalent of high school. This increases the chance of errors in this calculation, primarily in assumptions, methodology and source selection. The core problem is figuring out the H/C ratio for each fuel.

This is a fermi problem and it's all math glued together. Also the assumptions are quite wrong. I am impressed how they managed to calculate CO2 emissions.

However, I am confident to say that, this at least gives an idea.

Feel free to output criticism which could make this calculation better.


Water emissions per fuel

It should be noted that this is the minimum and maximum using these sources and this methodology, not the actual minimum and maximum.

Relative share of CO2 emissions(left) compared to share of H2O emissions(right)

As you can see, gas accounts for the least CO2 emissions but has the most water emissions. Coal seems to do the opposite.

Difference between minimum and maximum H/C ratio, by fuel

Using this method and these sources, oil is by far the most uncertain.

Other notes

I haven't found anything related to sea level rise and water emissions, except later I found this infographic from a climate blog.

This is a second, improved version of an earlier article of mine, which is now defunct. I decided to rewrite the article to make an more educated guess and to make it more clear how I calculated it, as the earlier article was a bit crappy in that regard.