A review in Biological Diversity argues that repeated sampling of environmental DNA and RNA can trace the direction and speed of ecological change, turning a snapshot tool into a way of following trajectories.

Scoop a litre of water from a river and you can now find out, from the genetic fragments floating in it, much of what lives upstream. Environmental DNA has become one of the most useful tools in ecological surveying, capable of detecting a rare fish or an incoming invasive species without anyone having to catch or see it.

What such a sample gives you, though, is a single moment. A review published in Biological Diversity by researchers at the Chinese Academy of Sciences argues that the field is ready for something more demanding: using these molecules to track not the state of an ecosystem but its trajectory.

The paper was written by Wei Xiong, an associate professor at the CAS Research Center for Eco-Environmental Sciences and the first author, with Aibin Zhan, a professor at the same centre, as corresponding author.

What is actually in the sample

Environmental nucleic acids come in two forms. Environmental DNA, or eDNA, is genetic material shed into water, soil or air as skin cells, mucus, faeces and other debris. Environmental RNA, or eRNA, is the more fragile counterpart, and because RNA degrades faster it tends to indicate recent activity rather than mere presence.

The review stresses that what a sample contains is an integrated signal rather than a direct census. Several processes shape it: how much genetic material organisms release, the molecular properties that determine how quickly it breaks down, how it is transported and where it is retained, and the laboratory methods used to extract and sequence it.

That last point matters more than it might seem. Two teams sampling the same river with different filters, preservation methods or genetic markers can produce different species lists, which has been a persistent difficulty in the field.

Persistence is the key variable for the review’s argument. How long genetic material survives in the environment determines how far back in time a sample reaches. A molecule that degrades in hours reports on the present; one that persists for weeks carries a blurred record of the recent past.

In practice, eDNA in open water is generally thought to persist from hours to a few weeks depending on temperature, sunlight and microbial activity, while material bound to sediment can last far longer. That spread is part of what complicates interpretation: a detection may reflect an animal that passed through yesterday or genetic material stirred up from the bottom.

From a list of species to a trajectory

The authors argue that with ecological validation, these signals can describe more than which species are present. They can speak to demographic structure, physiological condition and the functions an ecosystem is performing, since the abundance and type of genetic material relate to how many organisms there are and what they are doing.

Sampling repeatedly over time, the review proposes, allows a set of features that no single sample can provide: the direction of change, its magnitude, its rate, its timing, the intermediate states a system passes through, and whether it is progressing, stabilising or recovering.

The difference is practical. A one-off survey can tell a manager that an invasive species is present. A sequence of surveys can tell them whether its population is expanding, holding steady or collapsing, which is the information that determines what to do next.

The interpretation problem

The review is candid about why this is hard. The same molecular signal can arise from different causes. A rise in detected genetic material might mean more individuals, or the same number of individuals shedding more because they are spawning, or a change in water chemistry that slowed degradation, or simply a shift in currents that carried material from elsewhere.

Statisticians call this an identifiability problem: when multiple underlying states produce indistinguishable observations, data alone cannot settle which one is real.

The authors recommend three responses. Collect complementary data, so that physical and biological measurements can constrain the possibilities. Use mechanistic or state-space models, which represent the processes generating the signal explicitly rather than treating the measurement as the thing itself. And treat identifiability and uncertainty as matters to be stated openly rather than assumed away.

Decisions that would change

The practical case rests on management. Trajectory-based monitoring, the review argues, could inform decisions about invasive species control, responses to disturbance, conservation priorities and restoration.

The release describes the shift plainly: instead of relying on isolated measurements, managers can judge whether to maintain surveillance, launch intervention or change course.

Restoration is a case where the distinction bites. A restored wetland or reforested slope is usually assessed against a target condition, and progress is slow enough that a single survey some years in says little about whether the project is working. A trajectory, by contrast, indicates whether a system is moving toward the target, has stalled short of it, or is drifting somewhere else entirely.

What the field needs next

The review sets out priorities. Better-calibrated temporal response functions would describe how a molecular signal changes over time in response to a biological event, which is what converts a measurement into a statement about timing. Non-redundant multi-marker combinations would mean choosing genetic markers that each add information rather than duplicating one another. And dynamic reference conditions would replace fixed baselines with ones that account for natural variation.

That last item addresses a real difficulty in environmental assessment. Ecosystems fluctuate seasonally and between years without any human cause, so judging a change against a single historical snapshot can mistake normal variation for degradation, or the reverse.

The long-term goal the authors set out is forecasting and adaptive management: not merely describing where an ecosystem has been, but anticipating where it is going and adjusting interventions as the picture develops.

The review contains no new measurements, and the authors are proposing a research agenda rather than reporting that it has been achieved. Whether the molecular signals prove reliable enough to support that agenda is the question the field will have to answer sample by sample.

The work was supported by the National Science Technology Major Project of China’s Ministry of Ecology and Environment, through the Jing-Jin-Ji regional integrated environmental improvement programme.

---

Notes

1. Xiong, W., Zhan, A. “From Ecological States to Trajectories: Revealing Ecological Dynamics Through Environmental Nucleic Acids.” Biological Diversity, 2026. DOI: 10.1002/bod2.70040.
2. South China Botanical Garden, Chinese Academy of Sciences. “From snapshots to trajectories: Environmental nucleic acids unlock dynamic ecological monitoring.” EurekAlert!, 22 October 2026. https://www.eurekalert.org/news-releases/1148640.