The news is full of the surge in deforestation and fires in the Amazon rainforest, and and I have been fielding various media enquiries about what is causing this rise and what it means for our atmosphere. The increases fires have major consequences for regional climate, the rich Amazonian biodiversity, air quality and human health, and some consequence for global carbon emissions (though still small compared to the amount being emitted by fossil fuel combustion in industrialised parts of the world). One thing I am often asked on is to comment on the statement that "the Amazon provides 20% of our oxygen", a statement now being used by, among others, the President of France and the Secretary General of the UN. This statement is basically incorrect and based on a partial understanding of how ecosystems function. There are lots of reasons to be concerned about the Amazon and the current fires, including regional climate, human health affects of pollution, loss of the most biodiversity rich area of the planet and global carbon emissions. But running out of oxygen isn't one of them.
Below I lay out the science of where this number comes from, and why it is incorrect when you have a whole-ecosystem view of the Amazon.
The 20% figure comes from a partial understanding of the global oxygen cycle.
The tropical forests account for about a 34% of global land surface photosynthesis. This is shown in the figure below (Beer et al. 2010, Science). The figure shows the global land distribution of photosynthesis - the rainforests are the big red patches and the Amazon accounts for about one half of the world's rainforests. Tropical rainforests photosynthesise so much because they have a year-long growing season not constrained by winter or drought. The units in the figure at g of carbon per square metre, and in the summary Table below the units are Pg (petagrams) of carbon. Pg of carbon taken up by photosynthesis but can be converted to Pg of oxygen released by multiplying by 2.67. One petagram is 10^15 g, or a thousand million million grams, it is also often called a gigatonne, a thousand million tonnes).
Photosynthesis takes up carbon dioxide from the air but produces oxygen, as in the famous school textbook equation, which belie a fantastically complex and still incompletely understood marvel of nature: 6CO2 + 6H2O —> C6H12O6 + 6O2 The 2.67 multiplying factor comes from the molecular weight of O2 (32) divided by that of carbon (12).
Table 1 from the same paper shows the total photosynthesis (also known as Gross Primary Productivity or GPP) of each major land biome. We need to multiplying by 2.67 to convert to total oxygen production. Hence total oxygen production by photosynthesis on land is around 330 Pg of oxygen per year. The Amazon (just under half of the tropical forests) is around 16% of this, around 54 Pg of oxygen per year. Rounded up, this is where the 20% figure comes from. 16% of the oxygen being produced on land today is from photosynthesis in the Amazon.
But, rather like the Buddhist parable of the blind monks who each can only feel part of the elephant and therefore disagree on what the animal is, there are (at least) two important additional bits of information needed for a full picture.
First, the phytoplankton in the oceans also photosynthesise, generating around 240 Pg of oxygen per year. So total global photosynthesis (land and sea) produces about 570 Pg of oxygen per year. Therefore in terms of TOTAL global photosynthesis, photosynthesis in the Amazon contributes around 9%. This is smaller, but still substantial.
Second, a bigger point that is often missed is that the Amazon consumes about as much oxygen as it produces. This is shown in the diagram below. Plants produce oxygen through photosynthesis (green arrow). However, the the same plants consume the equivalent of over half the oxygen they produce in their own respiration (blue arrows: my own team's research suggests this is more like 60%). Plants metabolise just as animals do, just at a slower rate, and at night when there is no photosynthesis forests are net absorbers of oxygen. The remaining 40% of the Amazon oxygen budget is consumed mainly by microbes breaking down the dead leaves and wood of the rainforest, a natural process called heterotrophic respiration (dark blue arrows). These process of plant and heterotrophic respiration are effectively the reverse of the photosynthesis equation above. So, in all practical terms, the net contribution of the Amazon ECOSYSTEM (not just the plants alone) to the world's oxygen is effectively zero. The same is pretty much true of any ecosystem on Earth, at least on the timescales that are relevant to humans (less than millions of years).
The oxygen levels in the atmosphere are set on million year timescales by the subtle balance of geological, chemical and biological processes. They are not set by the short term (short term equals anything less than hundreds of thousands of years) activities or existence of current biomes.
A final point to make is that the atmosphere is awash with oxygen, at 20.95% or 209,500 ppm (parts per million). Carbon dioxide, by comparison, is around 405 ppm, over 500 times less than oxygen, and rising by around 2-3 ppm per year. Human activity (around 90% of which being fossil fuel combustion) has caused this oxygen concentration to drop by around 0.005% since 1990, a trivial amount. In parallel, the same activities have caused carbon dioxide concentrations to rise by by 37 ppm since 1990, or 10%. This is a much more substantial percentage because there is so little carbon dioxide in the atmosphere to begin with, so human activities that emit or absorb carbon dioxide can make a major difference. This is why we need to worry about the increase of carbon dioxide in the atmosphere (and its resulting impact on climate), and why we don't need to worry about running out of oxygen.
Since writing this piece I have received a number of questions about whether the "effectively net zero" contribution of the Amazon to our oxygen supply means that trees have no net contribution to the climate and atmosphere. To be clear, when a tree is planted up on deforested land, it absorbs carbon dioxide from the atmosphere and locks it away biomass. Following the same old balance of photosynthesis, an equivalent amount of oxygen is released to the atmosphere. It we were to plant up an area the size of the Amazon rainforest (something unrealistically ambitious), around 90 Pg C would be removed form the atmosphere, and atmospheric carbon dioxide would be lowered by around 40 ppm, or 10%. Conversely, if the entire Amazon rainforest were to go up in flames (something also unrealistic, I think), atmospheric carbon dioxide concentrations would rise by 10% and it would be almost impossible to keep global climate warming within safe boundaries such as 1.5 degrees Celsius. Using the same argument applied above, burning up the whole Amazon rainforest would use up around 240 Pg C of oxygen, causing atmospheric oxygen concentrations to drop by around 0.02%, an almost negligible amount because there is so much oxygen in the atmosphere. Keeping the Amazon rainforest largely intact matters for carbon dioxide and climate reasons (as well as many other reasons), it does not matter for oxygen reasons.
Addendum 2 (Aug 27) Yesterday Prof Scott Denning from Colorado State University published an article in the Conversation making basically the same point, but elaborating more on the long-term oxygen cycle and the role of the oceans. This can be found here: https://theconversation.com/amazon-fires-are-destructive-but-they-arent-depleting-earths-oxygen-supply-122369