Skip to content
NatureHQ

Ecologyphenomenon

Resilience and regime shifts

Suppressing small disturbances can reduce the size of shock a system can survive.

How much disturbance a system can absorb before reorganising into something else. Not the same as recovering quickly — a system can do that reliably and still collapse under one large shock.

Two properties get bundled under one word. Stability is how fast a system returns after a small knock. Resilience is how large a knock it can take before it stops returning at all and settles into a different configuration. They are independent, and the difference has an uncomfortable management implication: reducing variability — suppressing fires, smoothing flows, holding populations steady — can make a system return quickly from small perturbations while reducing the size of shock it can absorb. The reorganised state is the interesting part. A shallow lake can sit clear, with rooted plants holding sediment and taking up nutrients; loaded with enough nutrient it flips to turbid, with algae shading out the plants whose absence keeps sediment suspended, and the new arrangement holds itself in place. Crucially, reducing nutrients back to the level where the lake was clear does not usually make it clear again — the feedbacks maintaining turbidity have to be broken, often by removing fish or replanting, and that is what hysteresis means in practice. Similar arrangements are reported for reefs shifting to algal dominance, kelp to urchin barrens, and grassland to shrub. This is the vocabulary most in need of discipline. "Tipping point" is used about systems where no threshold has been located and no alternative state demonstrated, and the phrase implies both. The evidence is strong in shallow lakes, mixed elsewhere, and the honest position is that alternative stable states are real, that showing a particular system has them is difficult, and that thresholds are usually identified after they have been crossed.

Early coverage · 32% complete · reviewed 2026-09-04

What this page covers

Alternative states are best evidenced in shallow lakes, and reported with varying support in coral reefs, kelp systems, rangelands and woodlands.

Often confused with: Bouncing back quickly, which is stability rather than resilience; Tipping points, used confidently about systems whose thresholds have never been located

Quick facts

Two different properties
Stability is returning fast; resilience is how much it can absorb
They can trade off
A system managed for steadiness may take a large shock badly
Hysteresis
Reversing the driver does not usually reverse the shift
Best evidence
Shallow lakes; other systems are reported with weaker support

Bouncing back, and not falling over

Independent properties, routinely merged.

Returning quickly from a small knock is stability. Absorbing a large one without reorganising is resilience. They are independent — and suppressing small disturbances can reduce the size of shock a system can take.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

Stability denotes rapid return to equilibrium following small perturbation; resilience denotes the magnitude of disturbance absorbable before reorganisation into an alternative configuration. The two are independent properties, and management reducing variability may reduce resilience.

Who this applies to
A conceptual distinction applied across ecological systems.
Studied in
Animalia, Plantae
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

A settled conceptual distinction, though measuring resilience in a real system remains difficult.

How far it can be extended

The distinction is standard and illustrated across many system types.

Caveats

  • Measuring resilience requires knowing where a threshold lies, which is usually only apparent once one has been crossed.
  • The word has spread far beyond ecology and is often used without this precision.

Still unanswered

  • Whether early-warning indicators can reliably detect an approaching threshold in a real ecosystem.

Last reviewed 2026-09-04

The evidence (2 studies)
Four combinations, all of which occur
Under a big shockReturns fast from small knocksReturns slowly
Survives a big shockStable and resilientVariable, but hard to push over
Reorganises under a big shockLooks healthy until it does notFragile on both measures

The top-left cell is what people mean by a healthy ecosystem, and the bottom-left is the one worth worrying about: a system that recovers reliably from everything ordinary, and is closer than it appears to a threshold. Nothing in its ordinary behaviour distinguishes it from the top-left.

What a regime shift actually involves

A threshold, a feedback, and a return path that is not the way in.

The shallow lake is the clearest case. In the clear state, rooted plants hold the sediment down and take up nutrients, and the water stays clear enough for those plants to grow. Add nutrients gradually and little happens — until enough algae grow to shade the plants out. Without the plants, sediment is resuspended and nutrients are released, keeping the water turbid, which keeps the plants out. Each state maintains itself.

Diagram

The way out is not the way in

A shallow lake, and why reversing the cause does not reverse the effect.

The way out is not the way inClear lakePlants hold sediment downTurbid lakeAlgae shade the plants outAdd nutrientsRemove them again — usually not enoughEach state maintains itself, which is what makes the return path different.Restoring the clear state generally needs a separate intervention — removing fish,replanting — to break the loop keeping the water turbid. That is hysteresis, and itis why “we can undo this later” is a weaker plan than it sounds.
The same explanation in words

Two states face each other. A clear lake, where plants hold sediment down, and a turbid lake, marked out as the shifted state, where algae shade the plants out. A solid arrow runs from clear to turbid labelled "add nutrients". A dashed arrow runs back the other way labelled "remove them again — usually not enough". Notes beneath explain that each state maintains itself, that restoring the clear state generally needs a separate intervention such as removing fish or replanting to break the loop, and that this is hysteresis — which is why "we can undo this later" is a weaker plan than it sounds.

The consequence people find least intuitive is that reversing the cause does not reverse the effect. Bringing nutrient loading back to where the lake was clear generally leaves it turbid, because the feedback maintaining turbidity is not the nutrient loading. Restoring the clear state usually needs a separate intervention — removing fish, replanting — to break that loop. That is hysteresis, and it is why "we can undo this later" is a weaker plan than it sounds.

Related

Two words to use carefully

The concept is real; most of its uses are not evidenced.

Alternative stable states and thresholds are well evidenced in shallow lakes, reported with mixed support in reefs, kelp systems and rangelands, and asserted routinely about systems where neither has been demonstrated. Saying a system is approaching a tipping point implies two things: that it has an alternative state, and that a threshold has been located. Both are strong claims, and the second is usually established only by crossing it.

  • Can an approaching threshold be detected before it is crossed?

    Why it matters: The entire practical value of the concept depends on it. Statistical early-warning indicators — rising variance, slower recovery from small perturbations — have been proposed and work in models and some experimental systems; whether they are reliable in the field, at useful notice, is not established.

    What would settle it: Prospective prediction: identifying systems as near-threshold in advance and being right often enough to matter. Most published cases are retrospective.

The research behind this page

6 studies, newest first. Each one has a page explaining what it found and what it could not show.

2001Nature

Catastrophic shifts in ecosystems

Shallow lakes, coral reefs, rangelands and woodlands show evidence of alternative states with abrupt transitions and hysteresis, so that restoring prior conditions is insufficient to restore the prior state.

2000Annual Review of Ecology and Systematics

Mechanisms of maintenance of species diversity

Stable coexistence requires that intraspecific competition exceed interspecific competition.

1973Annual Review of Ecology and Systematics

Resilience and stability of ecological systems

The two properties are independent, and a system that returns quickly from small disturbances may reorganise readily under a large one, while a variable system may persist through much greater shocks.

1961The American Naturalist

The paradox of the plankton

Observed plankton diversity substantially exceeds the number of limiting resources, so the conditions under which competitive exclusion operates are evidently not met in open water.

1926Bulletin of the Torrey Botanical Club

The individualistic concept of the plant association

Species distributions along environmental gradients are individual and largely continuous, so communities have indistinct boundaries and are not repeatable units.

1916Carnegie Institution of Washington

Plant Succession: An Analysis of the Development of Vegetation

Vegetation change after disturbance follows recognisable sequences, and sites within a climatic region tend to converge on similar mature vegetation.

Where to go from here

Each of these follows from something on this page — a relationship in the evidence, a claim people ask about, or the next mechanism along.

How complete this page is, and what it is still missing

NatureHQ publishes its own gaps. This page is at 32% completeness against what we would call a finished subject, and was last reviewed on 2026-09-04. It carries 3 claims and answers 0 mapped search questions.

  • no research from the last few years is attached — check for newer work
  • more experiments could be explained in plain English
  • no popular claim about this subject has been checked yet
  • Early-warning indicator research is raised as an open question rather than covered.
  • Social-ecological resilience, where much of the modern literature sits, is out of scope here.
  • Coral reef regime shifts are referenced but covered on the coral pages.