What is an ecosystem? A working scientist's definition

An ecosystem is the simultaneous flow of energy, matter, and information through a community of organisms and their physical surroundings (UNEP: Ecosystems). The word "ecosystem" was coined by Arthur Tansley in 1935 specifically to displace the older, more romantic notion of "community" — Tansley wanted a term that emphasized exchange rather than membership.

Almost a century later, that emphasis is more relevant than ever.

A flow, not a place

The popular conception of an ecosystem is a place — a forest, a wetland, a coral reef. The scientific definition is closer to a process. The forest is the substrate; the ecosystem is what the forest does. Photosynthesis converts solar energy into biomass. Decomposition releases mineral nutrients back into circulation. Herbivores transfer that energy upward; predators regulate the herbivores. Water moves through soil, root, leaf, atmosphere, and back. Nitrogen cycles through fixers, plants, animals, and decomposers. Information — pheromones, alarm calls, mycorrhizal signaling — coordinates the whole.

Take the substrate away and the ecosystem stops. Take the flow away and the substrate is dead. This distinction matters when you start asking what conservation actually conserves.

Three properties that define an ecosystem

The substantive content of "ecosystem" reduces to three measurable properties.

Energy throughput. The rate at which a system fixes solar energy as primary production, and the efficiency with which that energy moves up trophic levels. This is the metric that makes a tropical rainforest and a tundra recognizably the same kind of object — they differ in magnitude, not in kind.

Material cycling. The rate at which nutrients (nitrogen, phosphorus, carbon, water) move through biological and physical compartments. Healthy ecosystems are characterized by tight cycles where nutrients are quickly reabsorbed; degraded ecosystems leak.

Information flow. The signaling — chemical, behavioral, ecological — that coordinates organisms across space and time. This is the property we measure least well and arguably the one that breaks first under stress.

Any system that has all three is an ecosystem. A bare rock with lichen has all three (at small scale). A swimming pool does not. The boundary is fuzzy at the edges, which is fine — most useful definitions in biology are fuzzy at the edges.

Why scale is hidden in the definition

An ecosystem is whatever level of organization captures meaningful flux. A drop of pond water is an ecosystem. A continental biome is an ecosystem. The Earth's biosphere is an ecosystem. The choice of scale is the researcher's, and it should be motivated by what flow you are trying to study.

This is liberating. It means you can talk about a "soil microbiome ecosystem" and a "coral microbiome ecosystem" without abusing the word — both have energy throughput, material cycling, and information flow at the relevant scale.

It also means that "the ecosystem" without a scale qualifier is rarely a meaningful object. When public discourse refers to "saving an ecosystem", the operative question is always which one, at what scale, measured by which flux. Without that, the claim is doing rhetorical work but not scientific work.

What this lets us do

Defining ecosystems by flow rather than place leads to several useful reframings.

Restoration is restarting flux. A restored ecosystem is one in which the characteristic flows have resumed at a rate comparable to the reference state. Replanting a forest without restoring its hydrology, soil community, or trophic structure produces a stand of trees, not an ecosystem.

Resilience is buffer capacity. A resilient ecosystem is one whose flows persist under perturbation. Two systems can look identical and behave very differently under stress because they have different buffers — different species redundancy, different soil carbon, different functional diversity.

Tipping points are flow reorganizations. When an ecosystem "flips" — savanna to desert, kelp forest to urchin barren, coral reef to algal mat — what flips is the dominant energy and material pathway. The same place persists; the ecosystem does not.

Ecosystem services are quantified flows. When we value an ecosystem for carbon sequestration, water purification, or pollination, we are valuing a specific flux. The accounting only works if we treat the flux, not the place, as the unit. The IPBES Global Assessment makes this framing the operational basis for its biodiversity and ecosystem-services accounting.

What this rules out

The flow-based definition also rules out some popular but unhelpful framings.

It rules out the idea that ecosystems have a "natural state" they should be returned to. Ecosystems are dynamic by definition. There is no static reference; there is only a chosen baseline.

It rules out the strict separation of "wild" and "managed" ecosystems. A heavily managed agricultural field still cycles energy, matter, and information — it is an ecosystem, just a low-diversity one with anthropogenic energy subsidies. The interesting question is not whether it is "natural" but how its flows compare to what came before.

It also rules out the idea that you can preserve an ecosystem by drawing a line around it. Ecosystems exchange flows across their boundaries — water, nutrients, propagules, atmospheric gases. A line on a map preserves the substrate; whether it preserves the ecosystem depends on what crosses the line.

The takeaway

When a working ecologist says "ecosystem", the meaningful content is the flux. That framing is what makes the rest of the field — disturbance ecology, restoration, ecosystem services, conservation biology — coherent.

Once you internalize it, much of public ecological discourse becomes easier to evaluate. Claims about "destroying ecosystems" are claims about disrupting specific flows; claims about "restoring ecosystems" are claims about restarting them. Without the flow framing, both kinds of claim float free of the evidence they would need to be true.