Intermediate Disturbance Hypothesis: Why Peak Diversity Sits in the Middle

What the hypothesis claims and where it applies

The Intermediate Disturbance Hypothesis (IDH) predicts that local (alpha) species richness follows a humped, unimodal curve along a disturbance gradient, peaking at intermediate levels rather than at either extreme. "Disturbance" here means any discrete event that removes biomass and frees resources or space — a storm, fire, flood, landslide, treefall, grazing pulse, or a wave flipping a boulder.

Crucially, IDH describes disturbance along three axes that can each be tuned:

• Frequency — how often events recur (e.g., fire every 2 vs 200 years)
• Intensity / severity — how much biomass each event removes
• Time since the last disturbance — where a patch sits in succession

It applies most cleanly to sessile, space-limited organisms whose competition plays out on a fixed substrate: corals, intertidal barnacles and macroalgae, forest trees, and grassland plants. Joseph Connell framed it in 1978 explicitly to explain the paradox of extraordinarily high diversity in coral reefs and tropical rainforests — systems once assumed to be stable climax communities but in fact repeatedly hammered by cyclones and treefalls.

The mechanism, step by step

IDH is a non-equilibrium coexistence mechanism. Left undisturbed, communities march toward competitive exclusion: the single best competitor for the limiting resource (light on a reef, space on a rock) eventually displaces everyone else. Disturbance interrupts that march before it finishes.

Walk the gradient:

• Low disturbance: succession runs to completion. The competitive dominant (a slow-growing, shade-tolerant K-strategist) monopolizes the substrate. Richness collapses to one or a few species.
• Intermediate disturbance: events reopen patches often enough that the dominant never completes exclusion, yet rarely enough that slower competitors can mature and reproduce. Early-successional colonizers and late-successional competitors coexist in a shifting mosaic of patches at different recovery stages. Richness peaks.
• High disturbance: mortality outpaces recovery. Only fast-dispersing, fast-maturing pioneers (r-strategists) or stress-tolerant species persist. Richness falls again.

The engine is a competition–colonization trade-off: no single species is both the best competitor and the best colonizer, so different disturbance regimes favor different life-history strategies. The hump emerges because the two strategies overlap only in the middle of the gradient.

Key players, life-history strategies, and characteristic numbers

IDH has no molecules — its "key players" are life-history strategies and the traits that define them, drawn from r/K selection (MacArthur & Wilson 1967) and Grime's C-S-R plant scheme (1974–1977):

• Colonizers (r-strategists / ruderals): high dispersal, rapid growth, early reproduction, short lifespan. Reef example — fast-growing branching corals (Acropora) and turf/green algae (Ulva), which can recruit within weeks.
• Competitors (K-strategists): slow growth, large size, competitive dominance, long lifespan. Reef example — massive Porites colonies growing only ~1–2 cm per year, or in intertidal zones the mussel Mytilus californianus.

Characteristic scales: Connell's Great Barrier Reef and Queensland rainforest plots were tracked for over 30 years. Intertidal disturbance cycles operate over weeks to years; forest treefall gaps reopen on decadal scales; fire regimes span years to centuries. In Sousa's (1979) boulder-field experiment, small (frequently overturned) boulders held ~1–2 algal species, large (stable) boulders were monopolized by the red alga Gigartina canaliculata, and intermediate boulders supported the highest richness (~3–4+ species) — a clean field demonstration of the hump.

How IDH is studied and tested

Because the prediction is a shape (a hump), testing IDH means plotting richness against a measured disturbance gradient and checking for unimodality — statistically, a significant negative quadratic term, or a formal test such as Mitchell-Olds & Shaw for the peak.

• Natural gradients: Connell compared reef zones by wave exposure and cyclone history; ecologists compare fire-frequency zones, flood regimes, or grazing intensity.
• Manipulative field experiments: Sousa (1979) used natural boulder size as a proxy for overturning frequency; others physically clear or scour plots at set intervals and intensities.
• Microcosms: Buckling et al. (2000) grew Pseudomonas fluorescens in vials, imposing disturbance by periodic mixing/bottlenecking; diversity of morphotypes peaked at intermediate disturbance — one of the few controlled, replicated confirmations.

A vital caveat: IDH concerns local/alpha diversity, and the disturbance gradient must be controlled independently of confounds like productivity, patch size, or spatial scale. Many apparent failures trace to conflating frequency with intensity, or to measuring the wrong diversity scale. Modern analyses (e.g., Fox 2013) also insist you specify the underlying competitive model before claiming IDH support.

How IDH relates to neighboring theories

IDH is one member of a family of non-equilibrium coexistence ideas and is easily confused with its cousins:

• Competitive exclusion principle (Gause): the baseline IDH pushes against — without disturbance, one competitor wins. IDH is the mechanism that suspends exclusion.
• Huston's dynamic-equilibrium model (1979/1994): a richer framework where diversity depends on the balance of disturbance frequency and productivity/growth rate — the hump shifts with resource supply. Often considered a generalization that supersedes plain IDH.
• Grime's humped productivity–diversity curve (1973): a parallel unimodal pattern along a fertility gradient rather than a disturbance gradient.
• Storage effect & lottery models (Chesson): coexistence via temporal environmental variation and buffered recruitment — a more mechanistically explicit modern account of variable environments.
• Janzen–Connell hypothesis: a different Connell idea (density-dependent seedling mortality near parents) that also maintains tropical diversity — related name, distinct mechanism.

IDH is best seen as a verbal, pattern-level hypothesis; the others supply the equations and additional axes it originally lacked.

Significance, controversy, and open questions

For three decades IDH was a textbook staple and one of ecology's most-cited ideas, shaping conservation and management: prescribed burns to keep fire-adapted systems in the diversity-maximizing middle, environmental flow releases below dams, and rotational grazing all invoke IDH logic. Empirically the hump does appear in many surveys.

But since around 2010 the hypothesis has faced sharp criticism:

• Mackey & Currie (2001) meta-analyzed ~200 studies and found the humped pattern in only ~16–19% of cases; flat, U-shaped, and monotonic relationships were just as common.
• Fox (2013), in Trends in Ecology & Evolution ("The intermediate disturbance hypothesis should be abandoned"), argued the verbal logic is internally inconsistent and that the competition–colonization trade-off does not reliably generate a hump under standard population models. A vigorous rebuttal debate followed (Sheil & Burslem 2013).

Open questions: Under exactly which competitive and dispersal conditions is a hump predicted versus a flat or reversed pattern? How do the three disturbance axes interact? And should IDH be retained as a useful heuristic or replaced entirely by explicit coexistence theory (modern coexistence / storage-effect frameworks)? The pattern is real where it appears — the debate is over whether IDH is the right explanation.

Community outcome across the disturbance gradient (frequency / intensity / time since disturbance)

Disturbance level	Dominant process	Who wins	Species richness
Low (rare, mild, long recovery)	Competitive exclusion reaches equilibrium	Best competitor (K-strategist) monopolizes	Low
Intermediate (moderate frequency/intensity)	Non-equilibrium coexistence; competition reset before exclusion completes	Competitors + colonizers together	High (peak)
High (frequent, severe, short recovery)	Repeated mortality; only pioneers establish	Fast r-strategist colonizers / tolerant species	Low
Example — Connell reef	Cyclone-scoured vs sheltered coral	Mid wave-exposure reefs richest	Humped curve observed