Foundation species and the myth of functional redundancy

A persistent claim in popular ecology is that biodiversity provides "functional redundancy" — multiple species perform the same role, so the loss of any one species can be absorbed without consequence. This claim is comforting. It is also largely wrong.

The empirical record from long-term ecosystem studies suggests something closer to the opposite: ecological function is unevenly distributed across species, and a small number of foundation species disproportionately determine the structure of the systems they inhabit. Lose them and the ecosystem reorganizes — not because no species can take their place, but because no species is doing what they did at the scale they did it.

What foundation species actually are

Foundation species are organisms that physically structure their environment in ways that other species depend on. The classic examples are large, abundant, and slow-growing: kelp in a kelp forest, corals in a reef, oaks in an oak savanna, Sphagnum mosses in a peat bog, beavers in a riparian network.

The shared property is not size or abundance per se — it is that the foundation species creates the physical and chemical conditions under which much of the rest of the community exists. Removing them does not subtract a single role from the system; it removes the conditions that make most of the other roles possible.

This is what makes the redundancy framing misleading. There is no other species in a kelp forest that performs the function of "being a kelp forest" — the forest is the function, and the function is the species.

How redundancy survived as long as it did

Functional redundancy is not a baseless idea. It comes from a real observation: in well-functioning systems, perturbing or removing many species does not measurably change ecosystem-level properties like productivity or nutrient cycling. The community absorbs the loss.

What this masks is which species the experiments removed. Most early redundancy experiments worked with non-foundation species — common but interchangeable members of the community. When experiments specifically targeted foundation species, the redundancy effect collapsed. The ecosystem reorganized; productivity dropped; new community states established.

The lesson is not that redundancy doesn't exist. It does, for most of the species in most communities. The lesson is that the species redundancy doesn't apply to are the ones whose loss actually matters. Treating "biodiversity" as a single number obscures the asymmetry.

Why this matters for conservation

Conservation prioritization frameworks often fall back on species counts (alpha diversity), with secondary attention to rarity. Foundation species are frequently neither rare nor charismatic. Kelp is not rare. Sphagnum moss is not charismatic. Yet the loss of either reorganizes entire systems.

A foundation-species-aware framework would prioritize:

• Identifying the foundation species in a given system, which is empirical work that has been done for fewer ecosystems than people assume
• Monitoring their condition with the same intensity we monitor flagship species
• Treating their decline — even if absolute populations remain large — as an early-warning signal for ecosystem reorganization

This is not how most conservation budgeting works. It probably should be.

The functional-diversity correction

Some of the redundancy literature has corrected itself by moving from species diversity to functional diversity. Functional diversity counts the number of distinct roles in a community — fast vs. slow growers, deep vs. shallow rooters, generalist vs. specialist consumers — rather than just species. By that metric, communities are much less redundant than species counts suggest.

Even this is incomplete. Functional diversity treats roles as discrete units, but foundation species often perform roles that other species in the system cannot replicate at any scale. They are not just one functional unit among many — they are the substrate on which other roles operate.

What survives the correction

Two pieces of the older biodiversity-as-insurance argument do survive scrutiny.

First, in the long run, the existence of redundant species does buffer ecosystems against perturbation, because over decades to centuries, redundant species can shift roles or evolve to fill gaps. The redundancy framing fails as a description of short-term dynamics; it succeeds as a description of long-term evolutionary buffering.

Second, intact biodiversity is correlated with the persistence of foundation species. We do not always know which species are foundational in a system, and protecting overall diversity is a hedge against guessing wrong. This is a weaker argument than the one usually made, but it is the strongest one the evidence supports.

The distinction matters because the weaker argument leads to different policy than the stronger one. If you believe biodiversity is fungible, you can substitute restoration credits across systems. If you believe it isn't, you can't.