eDNA and Ecosystem Condition: The Right Tools for Nature Reporting

Send a survey team onto a thousand-hectare estate with a budget for ten soil samples and the first question is the hard one: where do they dig? Get the placement right and ten samples tell a rich story. Get it wrong and you have ten precise answers to a question nobody asked. The sampling is superb either way. The aiming is the problem.

That is roughly where biodiversity measurement stands. Environmental DNA has handed ecology one of its great instruments, and the reflex is to point it at the ground and pull the trigger. This piece is about aiming it, and about why the right tools for nature-condition reporting are two tools, not one.

What eDNA does that nothing else can

Start by giving eDNA its due, because everything here depends on it. For the question it answers, there is nothing better.

From a few grams of soil or a flask of water, eDNA metabarcoding reads the genetic signature of a whole community: the species actually present, named, across the tree of life. It catches the rare, the cryptic and the nocturnal that no walked survey would find. It picks up invasives early. It speaks to genetic diversity inside a population. Most of all it gives hard, organism-level proof that a species is, or recently was, there.

EcoIntel does none of this and never will. No satellite will name the springtails in a spadeful of earth. When you need the species, you need eDNA or boots and a hand lens. Nothing here argues otherwise. The argument is about getting more from each sample you can afford.

The shape of a sample

That power comes with three facts of life, none of them a fault. They are simply what point sampling is, and naming them shows where a second, continuous layer earns its place.

A sample is one spot on one day. Communities shift across a few metres and across the seasons, so a result is a pinhole: sharp where you put it and dark everywhere else.

A sample has no past. eDNA begins the day you collect it. From one visit you cannot know what the ground was five years ago, whether it is recovering or sliding, or what hit it in between. That history was never recorded, and you cannot go back and collect it now.

And a sample costs money each time. In 2026 a single soil eDNA sample runs to several hundred dollars or more once collection, sequencing and the bioinformatics across several assays are counted, and the figure moves with provider, assay and country. The cost climbs more or less in step with the area you cover, how often you return, and how many sites are in the portfolio. Cover a whole property properly, repeatedly, across a supply chain, and the bill stacks up quickly.

The second tool: continuous condition

This is where the other tool comes in. EcoIntel is a satellite read of ecosystem condition, of whether land is functioning, for every parcel of a property, every year, back to 2018, with nobody on site. It scores condition one to nine, the Land Health Score, against the healthy reference state for the land’s own ecoregion, from four processes: the Water Cycle, Energy Flow, Mineral Cycle and Community Dynamics (the WCI, EFI, MCI and CDI indices).

The honest line, and the one ecologists rightly press on, is that condition is not biodiversity. EcoIntel reads ecosystem function, the land’s capacity to support life, not the life itself. Its Community Dynamics index is a structural-diversity proxy, never a species count. That is exactly why the two tools fit together rather than compete. One reads the stage; the other identifies the cast. Nature-condition reporting needs both, and here is what the condition tool does for the species one.

It tells you where to dig

Wall-to-wall condition turns a blind or gridded plan into a targeted one. EcoIntel shows which parts of the property are thriving and which are failing, which process is the limiting one, where the anomalies sit, and which patches are on the mend or sliding. You place scarce samples to span that range on purpose, catching the best ground and the worst, the steady and the moving.

It will not tell you which species you will find, and does not pretend to. It targets on habitat and condition, which is the variable a sound sampling design wants to stratify on anyway. Think of the MRI that tells a surgeon where to take the biopsy: the scan does not diagnose the tissue, it finds the tissue worth sampling.

It gives each result a context

A species list on its own is just a list, and the same list means very different things in thriving and failing ground. EcoIntel does not read or score your list. It describes the habitat the sample came from: its condition now, its eight-year trajectory, and how both compare with the ecoregion reference. That gives every finding a denominator. The same community means one thing in a high-condition patch that is improving and quite another where condition is low and falling.

It hands the sample a history

Because the condition record runs back to 2018, a one-off result can be set on a trajectory you never had to collect. Was this ground always like this, or is the snapshot the tail of a long decline, or the early part of a recovery? The same archive flags likely drivers: a disturbance, a drought signal, a change in land use. Two honest caveats. This is the habitat’s history, not the species’ history. And the drivers are candidates for interpretation, not proven cause. Even so, a snapshot with a plausible backstory is worth far more than a snapshot alone.

It keeps watch between visits

eDNA happens when the budget allows, once, or once every few years. Condition re-runs every year at almost no extra cost, so the habitat stays watched in the long gaps. Steady condition is reassurance that your last species picture probably still holds. A sharp move is your cue to go back. It can also show which unsampled areas resemble your sampled points in condition and habitat, which is a pointer to where to look next, not a claim about which species are there.

It triages a whole portfolio

This is the one that changes the budget. A fund or food company with hundreds of supplier farms cannot sample them all, certainly not every year. Condition-map every one of them cheaply, flag the handful where condition is poor, odd or moving fast, and spend the eDNA where it will pay. Triage is what turns a portfolio biodiversity programme from a wish into a plan. This is also the order TNFD itself recommends: a broad geospatial screen first, to find the sites that warrant a closer look, then a detailed, location-specific assessment where the risk is concentrated. Condition mapping is that first screen, sharpened from extent and context to function, and the eDNA goes where the screen says it will pay.

Why the changing places are the ones to sample

There is a deeper reason targeting on condition works, worth stating carefully because it is easy to overclaim. Function tends to turn before the species list does. A degrading habitat can still hold species that are present but no longer viable, an extinction debt not yet paid. A recovering one can be ready well before new arrivals show up, a colonisation credit not yet drawn. So the places where condition is moving fastest are, on average, where a fresh survey is most likely to turn up something the last one missed.

Read it the right way. This says where to spend a sample, not what the sample will contain. EcoIntel is not forecasting which species appear at a site. It is pointing at the sites where the species picture is most likely out of date. For more on that logic, see our piece on leading versus lagging indicators.

The economics that make it real

Set the two cost curves next to each other. eDNA, at several hundred dollars or more a sample in 2026, scales per sample and per visit. A credible single-site programme, enough samples to span the habitats and repeated often enough to see change, soon runs into the thousands, and a portfolio or a time-series multiplies it again.

EcoIntel’s Land Health System for one property is, as advertised in 2026, £600 a year standard, or £300 on the founding-member rate, or £55 a month, for the whole property, every year, with eight years of history included, ex-VAT for business users. Adding another site, or another year, costs almost nothing. eDNA’s next sample always costs the same as the last.

Put plainly, the price of a single eDNA sample buys roughly a year of continuous, whole-farm, retrospective condition. One tool is precious and priced that way; the other is cheap, repeatable and everywhere. That is why they belong in the same kit. By taking the cost, the time and the effort out of understanding the land at all, EcoIntel is what lets an organisation afford eDNA where it counts instead of nowhere.

EcoIntel outputs are indicative decision-support, not third-party-verified, and none of this is legal or financial advice. Confirm current eDNA and EcoIntel pricing before quoting either.

Where this goes next

Everything above works today, with no new science: targeting, context, history, monitoring, triage. There is a further prize, and honesty means labelling it as future rather than present. As eDNA and condition build up on the same sites, a two-way loop opens. The species results can calibrate the condition read, sharpening what the satellite signal means on the ground. And where condition and species turn out to be reliably linked, condition could begin to point the next survey with real statistical weight. That is a roadmap, not a current claim, and it runs the right way round: the ground truth teaching the satellite, not the satellite standing in for it.

The reporting payoff

This is where the two tools become the right tools for reporting. As of 2026, TNFD, the CSRD’s ESRS E4 and GRI’s biodiversity standard all expect a state of nature that covers both condition and species. Condition without species is half the account; species without condition is a snapshot with no context. A credible species layer across a whole portfolio is, on its own, more than most organisations can afford. By targeting the sampling, framing each result, and watching the habitat in between, EcoIntel is what makes a defensible portfolio biodiversity programme add up.

That is good news for eDNA providers, not a threat. A targeted, well-contextualised, continuously watched programme spends the same money on more useful samples, tells a cleaner story to an auditor, and, by lowering the cost of doing any of this, widens the market for the species evidence only they can supply. The satellite does not compete with the survey. It books more surveys, in better places.

The map and the expedition

The satellite is the map. eDNA is the expedition that tells you which species actually live there. The map does not replace the expedition, and never will. What it does is mean you stop sending it out blind: you know where to go, what the ground has been doing, and, across a portfolio, where the next survey will earn its keep. For nature-condition disclosure and reporting, that is the right kit, condition and species together, each making the other worth more.

---

Related reading

• Leading vs Lagging Indicators: why function turns before the species list does
• EcoIntel’s Land Health Score as Nature-Condition Metric: the full account of the condition half each eDNA result is set against
• Ecosystem Condition: Nature Disclosure’s Missing Metric: the short version of the state-of-nature gap
• Community dynamics: the structural-diversity process, a habitat-structure proxy and not a species count
• Land Health Score: the condition read each eDNA result is set against