Wednesday 21 April 2021

"People have shaped most of terrestrial nature for at least 12,000 years". Really?

This week, Ellis et al. published an important paper in PNAS (https://www.pnas.org/content/118/17/e2023483118). Their main message is that people have shaped most of terrestrial nature for at least 12,000 years. The paper comes with several maps that show how, already 12,000 years ago, places like the Amazon were mostly "Cultured anthromes" with only small parts of it classified as "Wildlands". These are defined as follows: "Wildlands are characterized by the complete absence of human populations and intensive land uses. Cultured anthromes are less than 20% covered by intensive land uses".

You can explore the maps here https://anthroecology.org/anthromes/12kdggv1/maps/ge/

One week before the paper was published, I was sent an embargoed version by a journalist working for an important scientific magazine to comment on the paper. However, none of my comments made it to the final version of the journalist's article. So I decided to post here the questions I was asked and my answers.

It could well be that I didn't understand important aspects of the paper, nevertheless, I read it and I formed an opinion that, even if wrong, might lead to some useful discussion.

Here it goes:

Do I call you an environmental scientists? If not, then what?

I am a geographer

What impresses your about this paper and why?

The most impressive thing is the scale of the analysis: the whole world during 12,000 years! It is a very ambitious project.

What new ideas come from it?

Well, in terms of ideas, it is pretty much in line with what some of the authors, and of course Ellis, have being saying for a while now. They published a paper in Science less than two years ago where they made the same point based on a collaborative assessment of many archaeologists. This time, it is based on HYDE 3.2, a model of past land use and past population density, but the main message is basically the same.

What surprises you about the paper and why?

There is almost no change in the amount of “wildland” across the last 12.000 years. This is quite surprising because I would say that 12.000 years ago America was almost empty, far less populated than it was by 1000 years ago.

What relevant work are you doing?

I study the evolution of the landscape in the Bolivian Amazon, looking at both natural and human causes of the change.

What is the paper’s broad significance?

This paper shows that it is very hard to establish baselines for what “natural” or “wild” actually mean. I think most of the academic debate is just about setting that baseline. For example, can I say that a forest is “natural” even if people where there and planted trees?

Do you expect it to be controversial? If so, why?

Yes. The fact that the amount of what they consider Wildland is almost constant during the last 12.000 years will be hard to accept. Of course, it depends on how they define the anthromes and the methodology. Nevertheless, the figures in the paper will raise a few eyebrows. Take for example Amazonia. They produced their maps by dividing the space into 96 km2 hexagons, meaning that Amazonia, which is around 7M km2, is made of 73,000 hexagons. However, based on a recent review (https://www.nature.com/articles/s41598-019-43086-w), in Amazonia there is only 1 site older than 11,000 BP (uncalibrated), 9 older than 10,000 BP and 46 older than 8000 BP. This means that by 8000 BP, only 0.06% of the hexagons that make Amazonia contain at least 1 archaeological site. How is it possible that, by 12000 BP, almost all Amazonia is classified as Cultured instead of Wildland? And that there is almost no change in the size of anthromes between 12.000 and 8.000 years BP when the number of archaeological sites grow from 1 to 46?

What are the paper’s shortcomings and strengths?

As I said, the most important aspect of this paper is the scale of the study. It is very important to make these reconstructions. Even if the uncertainties are very large, someone has to start. However, I think that there is an important issue with the definitions of the anthromes they use. They say that Wildlands are characterized by complete absence of humans, while Cultured anthromes are characterized by less than 20% of intensive use. But, there should be something in the middle! There is a threshold below which the presence of people should not count as causing any permanent change. If we consider a tropical forest, a few groups of hunter gatherers do not change the environment in any significant way, but still Ellis et al include that forest into the Cultured anthrome. It is possible that for most of the world, and for most of the time, Cultured anthromes were characterized by 0.5% or less of intensive use and people didn’t have any permanent impact. I think that what they call Wildlands should have included the presence of people as long as they did not cause any significant and permanent change. Then, we would have probably observed much bigger changes in the amount of Wildlands across the Holocene that would have likely been more in line with the archaeological and paleoecological evidence. 

P.S.

By zooming on the actual map that is online (https://anthroecology.org/anthromes/12kdggv1/maps/ge/), I see that I was wrong in saying that "almost all the Amazon is classified as Cultured" because it is roughly half of it. Anyway, my point is that it is a lot and doesn't change until 3000 BP. For large parts of the world, there are huge margins of error in their reconstruction. In my opinion, these are too big to justify the title of the paper and many of the comments I read in the press and on twitter.