That works out to 53,000 kilograms per tree. But how much of that is carbon? Trees are made of many different elements, like hydrogen and nitrogen, but let’s say it’s about half carbon. At least that’s an estimate that agrees with Wikipedia. So the mass of carbon would be 0.5 times the mass of the tree, or 26,500 kg. Simple!

### Counting Up the Atoms

But what I really need is the number of carbon dioxide molecules. Since each carbon atom would come from one molecule of CO_{2}, I need to convert from mass to numbers. This is where Avogadro’s number comes into play, with a value of around 6.022 x 10^{23}particles per mole. And one mole of carbon has a mass of about 12 grams. That gives us the number of carbon atoms (*n*) per tree:

Then, since everybody plants a tree, and assuming they’re all the same, the total amount of captured carbon atoms (*N*) would just be that number times 7.5 billion, the population of Earth.

We’re not done yet. We still need to find out how this changes the total concentration of CO_{2}in the air. For that, we need to estimate the total mass of Earth’s atmosphere …. well, that’s kind of daunting. What do physicists do in such situations? We Google it. I get a value of 5 x 10^{18}kilograms (from Wikipedia again).

So, to find the concentration in ppm, I need the molar mass of air. Air is 99 percent nitrogen and oxygen; a weighted average of their masses gives an air molar mass of 28.97 grams per mole. With that, I can calculate the number of air molecules. This uses the same formula as above for*n*, so I just built it into my computation code.

### The Grand Result

OK, let’s crank this sucker out! I’m attaching the code here, so if you want to change my assumptions—perhaps, in keeping with the tropical theme of Earth’s future, you’re envisioning palm trees instead of pine trees— you can click the pencil icon to edit it. Click Play to run the calculation.

Damn. That sucks. Even with 7.5 BILLION trees, it makes only a tiny dent in the carbon dioxide level. Yes, we made a lot of assumptions, and some of them are obviously wrong—but they’re not crazy-wrong. For example, we simplified by saying the trees are all the same. But allowing them to be different wouldn’t change the result if our generic tree is a good middle-of-the-pack average. The real question is whether our model is biased in one direction or the other.