What happens to an ecosystem when predators disappear?
Run the experiment on this page ↓
Short answer (general ecology)
Removing predators can change prey populations and how much vegetation is eaten, but the outcome depends on available food and other ecological interactions. There is no single result that applies to every ecosystem.
What you can do here
Compare two worlds that start from the same state. In one, predators remain; in the other, they are all removed at a set time. Follow the herbivore count, the grass, and how individuals die.
Limit
This is an abstract educational model. It does not forecast what would happen in any real ecosystem.
Run the experiment
- Predict
- Watch one run
- Compare 5 runs
- Read the result
After predators are removed in B, how do you think its herbivore peak will compare with A?
You can start without choosing. Your choice is not sent anywhere.
Watch one run
tick 0 / 0Preparing the world…
A: nothing changed
B: predators removed
| A | B | |
|---|---|---|
| Herbivores | ||
| Predators | ||
| Grass |
Keyboard, while the focus is in this panel: Space plays and pauses, → advances one step, R restarts, 1 and 2 show A and B.
Selected individual
Click or tap an individual in either world to see its values here. The button under each world selects one without a pointer.
What the shapes and colours mean
Shape = role
A herbivore has long ears and a round tail: a light body with a dark outline. A predator has pointed ears and a bushy tail: a dark body with a light outline. They are drawn to resemble a rabbit and a fox, but do not stand for any particular animal. The two can be told apart without colour.
Ground colour = amount of grass
The amount of grass is shown in three steps: the colour of soil where there is almost none, light green where there is grass, and deeper green where it is thick.
Dullness and thinness = reserves
As reserves run low, the body turns duller and thinner. An individual whose reserves run out dies.
Size = growth
Newborn individuals are drawn small and grow to adult size.
Compare repeated runs
One run can differ from the next by chance. The same comparison is repeated with different random seeds to see how consistent the difference is.
Try it: change a condition and compare again
Write down what you expect first. The random seeds stay the same, so a difference from the standard result comes from the setting you changed.
How much grass grows back on each tile at every tick, in both worlds. Standard: 8. Range: 1 to 40.
Saved results
Results are stored only inside this browser and are never sent to a server. What is stored is the settings, the random seeds and the summary numbers for each replicate; the tick-by-tick record is not stored, because the same settings always reproduce the same result. Up to 50 results are kept across all experiments and the free lab, and the oldest are removed beyond that. They do not carry over to another browser or device: use a share link to move one. Deleting cannot be undone.
Move saved items to another browser
Everything saved here (all experiments and the free lab) can be written to one file. Load that file in another browser or device and the same items are listed there. Loaded results use the numbers in the file as they are; run the settings again if you want to check them. The file stays on your device and is not sent anywhere.
Reference result (20 replicates)
These values were computed in advance with the same engine that runs in your browser. Running the same settings and seeds yourself gives exactly the same numbers.
A and B compared
| Measure | Result |
|---|---|
| Peak herbivore count (after the intervention) | B was higher in all 20 replicates (median A 786, B 1,264). |
| Herbivore count (mean over the observation window) | B was higher in all 20 replicates (median A 622, B 844). |
| Lowest grass level (after the intervention, as a share of the maximum) | B was lower in all 20 replicates (median A 38%, B 15%). |
| Grass (mean over the observation window, as a share of the maximum) | B was lower in all 20 replicates (median A 50%, B 24%). |
| Herbivores that starved (after the intervention) | B was higher in all 20 replicates (median A 966, B 4,051). |
Observed What was measured in these runs.Reading How the measurements can be read, given the rules of this model.
- ObservedIn 20 of 20 replicates, the herbivore peak after predators were removed was higher than in A (median A 786, B 1,264).
- ReadingIn 20 of 20 replicates, herbivores rose above A and grass then fell below A. This can be read as the loss of predators reaching the grass through the herbivores (an indirect effect).
- ObservedIn 20 of 20 replicates, more herbivores starved in B than in A (median A 966, B 4,051).
- ReadingIn B there are no deaths by predation after the intervention. This can be read as individuals dying from lack of food instead of being eaten.
Engine version 0.1.0, 20 replicate(s) with seed(s) 101–120. Observation window: ticks 1,400–2,400. These are observations from this model with these settings, not predictions about real ecosystems.
Charts
Show as a table
| Tick | A median | A range | B median | B range |
|---|---|---|---|---|
| 0 | 384 | 384–384 | 384 | 384–384 |
| 300 | 627 | 523–700 | 627 | 523–700 |
| 600 | 634 | 565–763 | 634 | 565–763 |
| 804 | 663 | 540–819 | 695 | 556–851 |
| 852 | 646 | 543–743 | 1,065 | 904–1,228 |
| 900 | 646 | 545–708 | 1,249 | 1,091–1,383 |
| 996 | 618 | 528–807 | 726 | 670–799 |
| 1,104 | 644 | 540–807 | 826 | 766–891 |
| 1,200 | 634 | 539–750 | 871 | 805–940 |
| 1,500 | 619 | 486–728 | 853 | 783–919 |
| 1,800 | 611 | 523–774 | 837 | 769–890 |
| 2,100 | 637 | 401–763 | 853 | 767–907 |
| 2,400 | 600 | 503–944 | 839 | 745–935 |
Show as a table
| Tick | A median | A range | B median | B range |
|---|---|---|---|---|
| 0 | 60% | 60%–60% | 60% | 60%–60% |
| 300 | 49% | 43%–58% | 49% | 43%–58% |
| 600 | 44% | 35%–52% | 44% | 35%–52% |
| 804 | 42% | 36%–49% | 42% | 35%–49% |
| 852 | 43% | 35%–51% | 29% | 21%–40% |
| 900 | 45% | 37%–53% | 16% | 14%–21% |
| 996 | 47% | 36%–56% | 21% | 18%–23% |
| 1,104 | 46% | 38%–54% | 26% | 23%–29% |
| 1,200 | 46% | 39%–57% | 23% | 21%–27% |
| 1,500 | 52% | 41%–68% | 25% | 22%–27% |
| 1,800 | 49% | 39%–58% | 24% | 21%–29% |
| 2,100 | 49% | 41%–64% | 24% | 21%–29% |
| 2,400 | 49% | 35%–64% | 24% | 21%–28% |
Show as a table
| Tick | A median | A range | B median | B range |
|---|---|---|---|---|
| 0 | 64 | 64–64 | 64 | 64–64 |
| 300 | 104 | 90–114 | 104 | 90–114 |
| 600 | 138 | 105–162 | 138 | 105–162 |
| 804 | 149 | 119–170 | 0 | 0–0 |
| 852 | 149 | 119–176 | 0 | 0–0 |
| 900 | 155 | 121–181 | 0 | 0–0 |
| 996 | 161 | 117–180 | 0 | 0–0 |
| 1,104 | 164 | 120–202 | 0 | 0–0 |
| 1,200 | 170 | 121–205 | 0 | 0–0 |
| 1,500 | 171 | 149–191 | 0 | 0–0 |
| 1,800 | 170 | 140–188 | 0 | 0–0 |
| 2,100 | 168 | 143–196 | 0 | 0–0 |
| 2,400 | 175 | 122–188 | 0 | 0–0 |
Show as a table
| Tick | A median | A range | B median | B range |
|---|---|---|---|---|
| 0 | 0.0 | 0.0–0.0 | 0.0 | 0.0–0.0 |
| 300 | 0.3 | 0.0–1.0 | 0.3 | 0.0–1.0 |
| 600 | 0.5 | 0.0–2.7 | 0.5 | 0.0–2.7 |
| 804 | 0.8 | 0.1–3.0 | 0.6 | 0.0–2.3 |
| 852 | 0.8 | 0.0–2.5 | 0.0 | 0.0–0.6 |
| 900 | 0.6 | 0.0–2.3 | 3.8 | 0.3–12.1 |
| 996 | 0.7 | 0.0–2.0 | 2.5 | 1.3–4.3 |
| 1,104 | 0.5 | 0.0–4.5 | 1.3 | 0.3–2.5 |
| 1,200 | 0.5 | 0.0–2.6 | 2.5 | 1.2–4.9 |
| 1,500 | 0.4 | 0.0–1.7 | 2.3 | 1.0–4.3 |
| 1,800 | 0.4 | 0.0–2.6 | 3.0 | 0.8–4.9 |
| 2,100 | 0.4 | 0.0–1.5 | 2.1 | 1.2–4.6 |
| 2,400 | 0.4 | 0.0–3.5 | 2.6 | 1.2–4.2 |
Show as a table
| Tick | A median | A range | B median | B range |
|---|---|---|---|---|
| 0 | 0.0 | 0.0–0.0 | 0.0 | 0.0–0.0 |
| 300 | 2.9 | 2.2–4.1 | 2.9 | 2.2–4.1 |
| 600 | 4.0 | 2.4–4.9 | 4.0 | 2.4–4.9 |
| 804 | 4.9 | 3.6–6.2 | 3.1 | 2.2–4.3 |
| 852 | 4.0 | 3.2–7.4 | 0.0 | 0.0–0.0 |
| 900 | 4.5 | 3.2–6.5 | 0.0 | 0.0–0.0 |
| 996 | 4.5 | 3.3–6.5 | 0.0 | 0.0–0.0 |
| 1,104 | 4.5 | 3.0–6.8 | 0.0 | 0.0–0.0 |
| 1,200 | 5.0 | 3.8–5.9 | 0.0 | 0.0–0.0 |
| 1,500 | 4.6 | 3.3–6.5 | 0.0 | 0.0–0.0 |
| 1,800 | 4.9 | 3.6–7.4 | 0.0 | 0.0–0.0 |
| 2,100 | 4.8 | 2.9–6.8 | 0.0 | 0.0–0.0 |
| 2,400 | 5.0 | 3.4–5.8 | 0.0 | 0.0–0.0 |
Printed from https://ecosimarium.com/experiments/predator-removal/