Ecosystems like the african savanna have lots of megafauna which to some extent compete with each other for the same prey/vegetation, but still coexist just fine.

So how do I construct similarly diverse terrestrial ecosystems, while ensuring I don't have too many large carnivores, or any megafauna species which shouldn't be able to coexist?

I'm only concerned with scenarios which are extremely similar to earth and are similarly dominated by mammals and birds including many real modern or near-modern animals. So one needn't consider the sorts of dynamics you get when ectotherm/mesotherm organisms dominate such as they did in the age of dinosaurs.
I'm only really concerned with large animals for this question, because they have high caloric needs and are the part of the ecosystem I'm changing. Smaller animals and and plants are assumed to be similar or the same as those which exist on Earth.

Given those constraints how does one construct plausible ecosystems where all species are getting the calories they need and not being outcompeted by other species?

For the purposes of this question I'm using the definition of megafauna which refers to animals that reach or exceed 90 lbs/40 kg. So when I refer to birds in this question I'm talking about flightless birds which meet that mass. There are however giant flying bats in these ecosystems which do reach that mass limit.
Assume that within this setting human environmental influences are negligible.

  • 1
    $\begingroup$ Sounds like you want to get pretty detailed here. Are you just looking for rough ideas of how to balance things or suggestions on how to actually simulate such ecosystems? $\endgroup$
    – Joe Bloggs
    Mar 9, 2020 at 16:31
  • 1
    $\begingroup$ ReasonI ask is because the simple answer is very short. The detailed one is decidedly not short. $\endgroup$
    – Joe Bloggs
    Mar 9, 2020 at 16:41
  • $\begingroup$ I guess I want the long answer, since I'm not sure if the short answer would let one predict how ecosystems like the savannah can sustain so many different large herbivores and carnivores which compete to some extent. Though I deliberately limited the scope of the question to only deal with large mammals and birds, and I'm not creating ecosystems from scratch, but altering existing terrestrial earth ecosystems. $\endgroup$ Mar 9, 2020 at 16:58
  • $\begingroup$ The number of species on Earth is not even known ( see Global Biodiversity on Wikipedia ) but is at least 2 million-ish. The large creatures cannot exist in isolation from the small ones - it doesn't work like that. All the plankton die ? So will we (probably). All the grasses fail - we're dead. Balance has to include the smallest and largest. $\endgroup$ Mar 9, 2020 at 23:33
  • $\begingroup$ @StephenG You misunderstand, I'm starting with earth ecosystems and simply altering the composition of large animals. $\endgroup$ Mar 10, 2020 at 3:32

4 Answers 4


Short answer?

Stick a finger in the air and guess. Look at the relationship size/food consumption in existing animals (this paper has some nice data) and then extrapolate up the scale. Once you have a rough guess how much food they'll eat per day/month/year look at how much of that food is available in your suggested biosphere and boom: That's how many mega predators you can support.

Long answer?

You need to simulate it, or there's every chance that you're actually describing an unstable state where the mega predators will eat too many of the normal herbivores, then the megaherbivores will overpopulate and eat all the vegetation, leading to the mega predators starving to death and going extinct. Ecosystems are not simple things.

There are a variety of methods for this kind of simulation (from pen and paper all the way up to supercomputer), but I'd go with writing your own program in something with good statistical tools and random number generation libraries (R or Python might be good bets). That way the depth and complexity of the simulation is entirely up to you and can be upgraded if you want.

Some things are required for a good simulation.

  1. A good estimate for your starting state. By looking at existing earth biospheres you can get vital statistics on populations, the rough amounts of food herbivores and predators need, how often the species breed, and what kind of ranges they can be expected to require before they start killing each other. Also define some vegetation types and have a rough side of how many tonnes of food are being produced that your herbivores/small prey eaters can consume, then you can add in as many or few megafauna as you desire. The more detailed you want the more detailed you need to be. The key point here is that these stats don't need to be exact straight away. You may wish to pick a fixed area to begin with that you can estimate the populations of each animal for, that makes a lot of things simpler (including not having to worry about actual locations, migration etc).
  2. A good idea of how things change over time. You need functions to work out how much vegetation any given species will eat, and how much of any given species any other species will eat. The simplest way to do this is just by defining 'a Blort will eat 2.1 units of GlumbleWeed a day, a Snargle will eat 0.2 Blorts a day', then multiplying that rate by how many Blots/Snargles are available. A simulation that simple can be done with pencil and paper, if you're willing to spend the time doing it! Much more complex ways involve using random number generation to 'sample' random amounts of food each from the appropriate distributions (You will need to look up the concept of probability distributions for that though, so maybe just keep it steady to start with). Much more complex would be simulating each animal separately based on hunger, time since mating etc. You also need to define other functions for how many new creatures will be born, and how many will die of old age.
  3. A way to track what happens over time. You'll need to take your starting state, then apply all the functions for changes of state, then record the new state. Then do it again, this time putting the new state into the functions and recording the next day (or whatever time period you want to use. Weeks or months may also be good choices). Record the populations/vegetation amounts for each time step. If you're doing location based stuff (more complex), record that too. This is where computers are great: You can run millions of time steps and record the results over multiple species very, very quickly.
  4. A way to analyse what you've just done. Plots are good for this! Plot populations against time and you'll very quickly see some kind of equilibrium be established (or not, if your ecosystem is out of balance!). By looking at the plots you can then go back to your starting parameters and tweak them until you get a nice looking set of lines. Looks like the Blorts breed too quickly? Dial it back. Megapredators annihilating all their prey in the first few time periods? Start with fewer mega predators. Play about! If you're recording everything about how the simulation proceeds it should be quite simple to spot moments where odd things happen, and if you've got good repeatable steps for the simulation (easy on computer) then you can tweak and run and tweak and run...

With any luck your starting guesses won't be too far out. The complexity of your simulation will either be a boon or a burden here. Small changes in complex simulations can lead to unexpected results, but similarly complex simulations may reach equilibrium situations more easily since they have more freedom of movement. Once you've got some idea of how long it takes your simulation so settle down into equilibrium positions (either stable populations or populations that go up and down around a point), you can discard all the time periods before that and use statistics (simple means, or an idea of mean/standard deviation) to get a good grasp on your population densities.

Now: The biggest issue with simulations is that the real world is more complex. No matter how complicated you think it is: In reality it's more complicated than that. Depending on your simplifications/assumptions you might end up with numbers that are wildly different from reality. If you really, really care about getting that right then you need to only use real data, not tweak the starting parameters and instead increase the accuracy of the model to the point where your results start mimicking actual observed data. Then you can publish the model and probably get a PHD out of it.

Or you can make a simple model, aim for somewhere between guessing and accurate simulation, and blame any really big changes on the introduction of your megafauna.

Either way: If the people reading about your megafauna care more about the accuracy of the population statistics than they do about the giant animals then it's really their problem...

  • $\begingroup$ I think I can calculate a lot of what you described by hand, by considering the mass and metabolic rate of my megafauna and making comparisons to known figures. As well as by making some other assumptions based on existing animals. However how do I figure out how much food is available to my herbivores? This is especially tricky because different herbivores eat different stuff, and large xenarthrans in my setting (such as glyptodonts and giant sloths) are also able to sustain themselves on diets that wouldn't sustain many other herbivores, due to their lower metabolic rate. $\endgroup$ Mar 9, 2020 at 18:07
  • $\begingroup$ Also any advice on running some simulations given little to no coding knowledge? I suspect I could manage some of it with excel, but I'm not sure exactly how I'd go about doing that (such as which equations I'd need to plug in). $\endgroup$ Mar 9, 2020 at 18:09
  • 1
    $\begingroup$ One way to work out available food is to look at growing rates of various plants your herbivores might eat. Another would be to look at current small herbivore populations and how much they eat then work it backwards. $\endgroup$
    – Joe Bloggs
    Mar 9, 2020 at 18:19
  • 1
    $\begingroup$ @VakusDrake as for programming: the best idea given what I assume is your level of programming knowledge would be to write it all down on paper to start with. One you’ve got a better idea of the kinds of maths you do in between each step make a spreadsheet where you can do all the step-by-step maths on a single line. Break it up into lots of columns so you only ever have to do simple steps (like the simple steps you did on paper). $\endgroup$
    – Joe Bloggs
    Mar 9, 2020 at 18:22

The Problem

This would be surprisingly easy to accomplish based solely on what we know about ecology. the 10% energy rule of ecology is what applies here. Only 10% of each trophic level's energy is available to the next. So, if plants are the basis and absorb 100 units of energy from the sun, the hebivores only have 10 units of that energy available to themselves. Carnivores, in turn, will only have 1 unit of that original 100 available to themselves. Apex carnivores could have less than 0.1 units to themselves, depending on the number of trophic levels. This principle is based off of the fact that each organism is going to be using up most of that energy in biological processes, not just putting on mass for the next carnivore to eat. Example of trophic levels

The Solution

So, for our megafauna to exist in large numbers, we'll need our planet to have abundant resources and an easy time using them.

Abundant Oxygen

One huge limitation on size is oxygen availability. Take earth's history as an example. Bugs today can't grow large because of the need to get oxygen throughout their bodies. In prehistoric times, when oxygen was extremely abundant, you had massive insects like Meganeura. High amounts of oxygen takes away one limiting factor of your megafauna.

Carbon Dioxide

In addition to helping the fauna, we need to help their base as well. It needs to be hot, with abundant carbon for plant life. The hotter your planet is, the higher water capacity of your atmosphere becomes, and the more storms and rain you're going to get on average. Both of these things are very good for plant life, and will cover most of your planet in lush forests. (Note: high heat does not create deserts, geographical features do)

Low Gravity

Another huge limiter on your megafauna will be gravity. You want big bad animals? Make them as efficient as possible. Think how much an elephant exerts itself just to move. Imagine if it were able to move around in even slightly lower gravity. This would greatly increase the viability of large build animals, and especially those who wish to fly or glide.

Dense Atmosphere

Last but not least, make that atmosphere soupy. This will increase the availability of those vital atmospheric gasses for your flora basis and your megafauna, dramatically increasing the efficiency of plant stomata and lungs (more molecules of oxygen/CO2 per volume). This also has the added benefit of making flying easier again, leading to increased viability for your flying bats, and making flying animals be able to be much larger as well.

  • $\begingroup$ Given this setting is "extremely similar to earth" and shares most of the same species unfortunately many of your points don't really apply. $\endgroup$ Mar 10, 2020 at 17:07
  • 1
    $\begingroup$ I would still like to note a few things regarding your points: Making conditions more favorable to autotrophs in order to increase the available food for herbivores may backfire. Since for instance grasslands generally exist in areas unable to support many bigger plants, but have more available food for herbivores than forests do. Increasing co2 is also likely to either not work (since photosynthesizers tend to sequester it, thus why we no longer have >20% co2) or backfire since high co2 levels make aerobic respiration extremely inefficient or impossible. $\endgroup$ Mar 10, 2020 at 17:19
  • 1
    $\begingroup$ @VakusDrake Yeah, the last two points were extra tack-ons. You said similar, not the same, so I assumed a bit of wiggle room. As for the CO2, I used direct evidence of earth's history of megafauna. The Mesozoic era was characterized by higher oxygen, co2, and temperature levels with more plant mass on the earth as a whole, and certainly more megafauna. This is especially true in the Cretaceous, where temps were 10C higher and north-south temp gradient was almost non-existant. That said- sorry I misunderstood your qustion. $\endgroup$
    – Skyler
    Mar 10, 2020 at 17:45
  • 1
    $\begingroup$ I would recommend reading: en.wikipedia.org/wiki/Geological_history_of_oxygen as I found it very enlightnening, such as the fact oxygen levels were actually lower for part of the mesozoic (but this doesn't seem to have impacted archosaurs, though this could have been helped by their avian respiratory system). Also while high oxygen levels allowed for larger arthropods it's thought that the extinction of many of them was due at least in part to competition from vertebrates (modern giant arthropods are pretty cumbersome compared to vertebrates, especially endotherms like pterosaurs). $\endgroup$ Mar 10, 2020 at 18:11

As long as it is grassland of savannah you are fine, you really don't need to do much to justify it.

The African Savannah is the norm, the rest of the world is weird. And it basically comes down to humans. Most of the worlds grassland was like Africa with lots and lots of competing megafauna, but everywhere humans went they drastically reduced the local megafauna, the exception was Africa where said megafauna evolved alongside human and could deal with them. As long as the land is fertile grassland or savannah, you can have plenty of mega fauna.

Past that it all comes down to environment how much megafauna it can support is going to be different from cold taiga, to desert, to thick forest. And any answer that includes them all is going to be way out of scope.


Looks like there are just 2 conditions for megafauna to proliferate:

  1. Abundant grasslands (savanna, steppe, tundra);

  2. Lack of predators threatening megafauna (including humans).

During last Ice Age, megafauna was thriving throughout the Northern hemisphere. Woolly mammoths and rhinos were ubiquitous up to the polar circle. Those animals had enough food to eat, had freedom of movement and not so many predators to worry about. That was the case until the Ice Ice had come to close. No temperate or northern megafauna species could withstand a double impact of climate change and human hunting.

  • $\begingroup$ I've edited my question to make it clear which definition of megafauna I'm using, sorry for the confusion. One should also assume for this question that human hunting and other activities are not a major factor in this setting. $\endgroup$ Mar 9, 2020 at 17:48
  • $\begingroup$ There were plenty of predators eating those megafuana, dire wolves, cave lions, shortfaced bears, giant wolverines, saber tooth cats, ect. $\endgroup$
    – John
    Mar 9, 2020 at 19:17
  • $\begingroup$ @John all of those predators would have a difficult time taking on an adult mammoth, even more so if mammoths are in a pack. $\endgroup$
    – Alexander
    Mar 9, 2020 at 19:41
  • $\begingroup$ Not really true. As mentioned, there were many predators on megafauna in Pleistocene North America (and elsewere, presumably). Using the OP's 40 kg definition, we have such relationships today, and not just in grasslands. E.g. mountain lions and mule deer in the mostly forested US west. There are also moose, elk, pronghorn antelopes, and bears, which would qualify as megafauna, plus the re-introduced wild horse - none of which exclusively inhabit a savanna-like grassland. $\endgroup$
    – jamesqf
    Mar 9, 2020 at 19:49
  • $\begingroup$ @jamesqf with 40kg definition, different species need to adapt differently. In general, I don't see any natural threats that can cause species like deer or bison to go extinct. $\endgroup$
    – Alexander
    Mar 9, 2020 at 20:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .