Amazonian Forest Islands: Accidental Products of Ancient Human Occupation

The Amazon basin isn’t all a lush tropical forest.  In large portions of western Amazonian wet seasonally flooded grasslands–not trees–are the norm.   The upper branches of the Amazon River wind themselves across massive flood plains like snakes writhing across a sandy surface.  One feature of these nearly featureless flat plains that has long puzzled visitors is the presence of hundreds of small forest islands. These 50 to 200-foot diameter forest islands found amongst seas of grass and marshes sit only one to three feet above the surrounding ground level.  That small elevation change is enough to reduce the water content of the soil allowing trees to persist on these mounds.

An example of forest islands on a flood plane in central Bolivia.  These islands had been thought to have formed from natural (non-human induced) processes but now some have been shown to be the result of human activity.
An example of forest islands on a flood plain in central Bolivia. These islands had been thought to have formed from natural (non-human induced) processes but now some have been shown to be the result of human activity.

What is the origin of these forest mounds?

Rocks are virtually unknown in this part of the world and indeed no stones are found in these mounds.  Could they be ancient termite mounds or the result of animal burrowing?  Or could they have been intentionally built by native South Americans much like some of the famous Native American mounds found in North America?  For some forest islands, the answer is apparently: none of the above!

Dr. Umberto Lombarto and colleagues also wondered about the origin of these strange landforms (Early and Middle Holocene Hunter-Gatherer Occupations in Western Amazonia: The Hidden Shell Middens).  They set about to investigate three of these island mounds near Trinidad, Bolivia and made some surprising discoveries.  

When they dug into these mounds they discovered they were composed of vast quantities of snail shells, bones of various mammals including human bones, and fire-blackened shell and bone material.  That humans had occupied this islands throughout their development, even without intent to create the mounds, was of little doubt.

Evidence of human influence can be seen in the millions of snail shells that are mostly missing their opercula which would have been present had they died naturally and deposited here.  Furthermore, bones of game animals and evidence of cooking hearths identify these sites as being locations where people brought prey and other food sources and prepared them to eat.

How might the mounds have formed?

Armed with the observation that the mounds contain massive numbers of snail shells, Lombardo and colleagues hypothesized that shells were deposited at this location over a considerable time.  As they accumulated the location would have risen slightly about the flat grassy plain. This in turn would have made it a more attractive site to spend time while hunting for game or eating more snails which would have accelerated the accumulation of mound material. Thus, a positive feedback loop developed where the “island” became a place to come in high water times and eventually because it provided the cover of trees which would have been able to grow once the locations reached a certain height above the water table of the region.  

The formation of these mounds is not thought to have originally been the product of of any intentional design, such as some form of landscape engineering that would be done thousands of years later in the same region where ditches and hills were built to control the wetlands by pre-Columbian dwellers of this same area. Rather the mounds came about by an accidental, yet human-induced, series of events.

Screenshot of Google maps showing the possible identification of forest island SM3 from Bolivia.
I took this screenshot of Google maps showing the possible identification of study site forest island SM3 from Bolivia.   Toward the left side near the road there is another smaller forest island. The remainder of the light and dark greens here are grasses and aquatic plants.  I could identify more than 20 similar looking mounds in my short search of the areas using Google maps/earth.

Forest islands suggest a long chronology of human habitation.

How old are these mounds of shells and bones?  Data collected from these mounds gives us some clues. Cores into the mounds were dated from top to bottom and it was found that the oldest material on the bottom dated to around 10,000 years ago. I’ve written a series of posts about how context is key to understanding chronology of human remains (See The frequently overlooked geological context of hominid fossils  and Implications of artifacts and bones on ancient human butchery practices and Geological context II: Neanderthals and the Italian supervolcano.  Context is no less important here either when one considers how old these mounds must be.

Claims of dates provided by radiometric dating are usually dismissed by young-earth creationists who insist that archeological sites such as these must be no more than about 4000 year old and have been developed over a very short period of time. However, we will see that an examination of the physical evidence from these forest islands reveal that these mounds tell a  story of a long history of occupation by humans in this portion of South America irrespective of reservations about radiometric dating methods.

Let’s look at some of that evidence more closely to see why.

I searched on Google maps and found the location of forest island site SM1 from the Lombardo et al. 2013 paper. This is a screen shot I took of that site from the Lombardo et al. 2013 paper describing the formation of these islands. On the left is a marsh/lake area and to the left are high grasses. The island itself only sits about 2 feet above the area around it but that is enough for trees to find a home there.

Figure 1 from Lombardo et al. 2013 PLosOne paper. This shows the forest island SM1 which I have take a screenshot from Google maps below.
Figure 1 from Lombardo et al. 2013 PLosOne paper. This shows the forest island SM1 which I have taken a screenshot from Google maps below.
Figure 3 Lombardo
Figure 3 from Lombardo et al. 2013. This presents a cross-section of the island (shell midden as they call it) that is pictured in the figures above.

Lombardo and colleagues dug a pit near the center of one of these forest islands, making close observations of what they found as they dug down until this hit groundwater level about six feet below the mound surface. They also took sediment cores starting near the center of the mound and then moving further out from the mound into the grassy plain.  In figure 3 from their publication you can see what they found.

This massive mound about 200 feet in diameter is composed of hundreds of millions of discarded freshwater shells that extend just over six feet below the current ground surface. The figure shows that this mound sits on a flat plane of silty soils that is three to four feet below the present ground surrounding the mount.  This reveals that at some point after the mound formed, local flooding deposited several feet of silt and clay around the mound. 

How does this series of events in the history of the mound and the new layers of soil impact our understanding of the chronology of human occupation?

Another insight from Lombardo’s work at this location is that in the top foot of material that makes up the mound surface, there are pieces of pottery and other cultural items that record the presence of more recent human use of these mounds. How recent?  The radiocarbon dating suggests this portion of the mound records the events from 4000 years to the present day. This includes many cultural artifacts that archeologists and anthropologists would consider evidence of very early developed societies in South America.  

Below these relatively young (4000 years or less) top layers of artifacts evidence of advanced culture including pottery skills is lacking.  Most of the mound contains only evidence indicative of primitive hunter-gatherer group-use.  This includes the use of these locations for seasonal use as hunting camps which included extensive use of snails as a food source. 

The size of the mounds and their composition of hundreds of millions of shells dictates a very long period of occupation and use (thousands of years).  The types of animal bones, plant fragments and shells found in the mound tells us that the environment was considerably drier than it is today.  Snails would only have been available seasonally rather than all-year round which also limits the rate at which these shell-mounds could have grown.  

At some point, probably after the hunters and gatherers had disappeared from this area, other people moved here and these mounds would have become natural places to gather and cook food.  The C14 dating, evidence of climate change in the mound material, and other archeological clues from the region reveal this was thousands of years ago. At this point local floods deposited many feet of new soil around the mounds and the region was converted into the marshy wetlands that we see today and the mounds became the only sanctuary where trees could live making the forest islands that we see today.  The current “modern” soil that envelops these mounds is likely many hundreds to thousands of years old itself.

Once again the facts and the context of this archeological site defy typical creationists explanations.

Creationist are committed to a timeline in which the peopling of South America could have taken place no more than 4200 years ago and likely less than 4000 years ago.  As a result they must compress all of these human occupation sites into a chronology that spans that short period of time.  

Explaining these developments of these mounds is a tremendous challenge to that chronology.  And it is made even more difficult by the fact that the ground that lies below these mounds is likely composed of hundreds and hundreds of feet of sediments from floods of earlier times that must also be compressed into a post-Noahic flood world. In the creationist’ chronology there was an ice age 400 years after the flood and it was only during that time that people first migrated from the Tower of Babel across the globe. They traveled across Alaska and then all the way down into South America. They then populated this region and ate millions and millions of snails forming these large mounds which then became partially buried by a flood of new sediments on which the “newer” civilizations of ancient South Americans then lived thousands of years to present.

For anthropologists these landscapes were thought to have been one of the later locations to have been occupied by human immigrants. With so many other environments more suitable for human occupation, one would not expect that the very first people to arrive in South America would immediately take up residence in this area.  There are many other locations in South America that have been identified as being inhabited much earlier than 10,000 years ago. The discovery of these mounds are of human-induced origin shows that there could have had a sizable population in this part of South America many thousands of years ago.

Summing up…

Observations of only three easily accessible forest islands have been reported thus far.  The authors of this work not that hundreds of such forest islands are found in this region of Bolivia. It is expected that many, if not most, of these forest islands share the human-influence origins that the three they investigated do. If so, the sheer number of shells alone would strongly point to thousands of years of development for each of these mounds just as the C14 dates suggest.  

As I have said many times before, when specific sites of human occupation are examined in detail the data tell a story that is consistent with long periods of occupation in which the occupants developed technologies slowly over time or were replaced by different people groups with new technologies.  Creationist’s explanations must compress the development of these occupation sites into a very short time frame and typically ignore the evidence of changing technology and other cultural advancements (art, writing, etc…).

Addendum: Snail Shell Accumulation Rates

I did some quick calculations of the number of shells in a single mound and what the minimum time for the mound formation to take place. 

Taking a medium sized mound the volume of such a mound would be about 100 feet x 100 feet x 5 feet which is 50,000 cubic feet. Most of this volume consists of discarded snail shells.  The snails found here average less than one inch in length.  To be conservative we will say 1 cubic inch. There are 1728 cubic inches in a square foot (12 x 12 x 12 inches).  In this single mount then there would be at least 86 million snail shells.  Given compaction (most are broken under foot) and I overestimated the original volume of a single shell this is a massive underestimate.  The number would likely be more than 200 million shells in one mound. But let’s stick with an even one million to make the calculation easier. 

Let’s say an individual at 15 snails every day of the year (very unrealistic since they would only have been available for parts of the year). They could have disposed of 5475 snail shells per year. Let’s further allow that 10 individuals used the same mound every day for a full year and all ate 15 snails (again, unrealistic for the same reason but also because the snails in the area would also have been completely eaten up and not able to reproduce fast enough to continue to be eaten at this pace).  That would be 54750 snails per year eaten by a group of 10 individuals.  At this pace how long would it have taken if they continuously disposed of all of their snail shells at this single location? 

To get that number we take the one hundred million snail shells and divide the number of shells disposed of in a single year (54,750) to get the number of years. That value is 1826 years. 

This value superficially fits the YEC timelines.  They could argue that people arrived in this region just after 4000 years ago and then the mounds were formed over the next 2000 years and then utilized by pre-Columbians as hunting camps.  But as I said, the real number of snail shells probably is in the range of 200,000 million which pushes our time to formation to 4000 years.  But realistically, these sites were almost certainly not used on a daily basis but were seasonally used thus reducing the number to harvest and dispose of snail shells to less than half of a year. That also doubles the time of mound formation. Also, just think about the population limitations.  If there are 20 mounds within a few square miles and hundreds of people all eating 15 snails each day how many snails would be needed to sustain that population? A hundred thousand a month? The snail population would be decimated and probably forced to extinction if more than a few dozen people are constantly eating them. 

When you put it all together the 6-8 thousand year timeline suggested by the C14 dating is the most biologically realistic timeline. It is common sense that these mounds are evidence of long-term ancient human occupation of this region.

I have also made a YouTube video based on this post:


Early and Middle Holocene Hunter-Gatherer Occupations in Western Amazonia: The Hidden Shell Middens.   Umberto Lombardo et al. PLOS One 2013

Humans explored ‘Treasure Island’ much earlier than thought.  By Charles Chol

Comments are closed.

Up ↑

%d bloggers like this: