---
title: what mrv actually is
canonical_url: https://ensurance.app/guide/what-mrv-actually-is
markdown_url: https://ensurance.app/guide/what-mrv-actually-is.md
subtitle: "measurement, reporting, and verification are three uses of one reading. they are not three products to buy"
category: natural-capital
---

# what mrv actually is

*measurement, reporting, and verification are three uses of one reading. they are not three products to buy*

MRV is measurement, reporting, and verification. Measurement writes down what a place did. Reporting shows that reading to the people who rely on the claim. Verification is an independent look at the method and the data against stated criteria. A second instrument can corroborate a reading. A person who did not take the result can stand behind it. Those are different jobs, and both are part of the word.

A creek, a soil profile, a canopy, and a storm already produce a signal whether or not a network stores it. A probe, a lab assay, a camera, a radar pass, and an indicator sheet are readings of that living system. [**Ensurance**](/?from=guide) keeps the reading beside what the place claims, so funding can cite the condition. It does not replace anyone's sensors, and it is not the living system.

Most people meet these three letters from the opposite direction. Someone hands them a methodology, a spreadsheet of variables, and a quote from a verification body, and they conclude MRV is a category you procure. That is how MRV gets procured. It is one use of the record, not the whole of it.

:::johnson
**mrv is one reading, used three times.**

Measurement is the row. Reporting is who gets to see it. Verification is the second source. The row is the same row in all three. The reading your gauge already takes is the input.

[the same three words, in the onchain register →](/guide/ecological-claims-mrv?from=guide)
:::

## three uses, one row

A hydrologist who reads a staff plate at the same bend every Tuesday is doing measurement. When that number goes into the quarterly packet, it is reporting. When a second instrument on the same reach agrees, that corroborates the reading. When an independent party assesses the method and the data and stands behind the result, that is verification. The plate did not change. The use of the plate changed.

Each use has its own failure mode.

| use | what it does | failure mode |
| --- | --- | --- |
| **measurement** | writes down what the place did — a result, a unit, a time, a method, a place | a number nobody stored; the season ends and the notebook is the only copy |
| **reporting** | puts that reading in front of the people who rely on the claim | a report with no place; an average across a program, with no way back to the reach |
| **verification** | a second look — another source, or a person who stands behind a result they did not take | a single source briefed as confirmed; one probe, restated in the summary as fact |

The third is the expensive one. A single instrument can be right, and often is. It is still a single instrument. Calling it verified because it appeared twice in one document is formatting dressed up as a check.

## the row carries six things

The shape of a reading is not mysterious. Whatever the instrument, the useful row carries the same six things:

1. **the name the device already uses** for what it measures — not a name we assign it
2. **the result**
3. **the unit**
4. **the time it happened**
5. **the method** — a person, a lab, an instrument, or a model
6. **the place** — the parcel, the reach, or the station the result belongs to

A number with no method, no unit, or no place can be kept as a note. It does not count as evidence. That is not a technicality. "Soil moisture improved" with no depth, no plot, and no date is a sentence about a feeling, and a reviewer cannot do anything with it except believe you.

Two things follow. First, two sources that disagree both stay. A sonde and a staff plate on the same reach can differ, and the honest record shows both. There is no blended score, no automatic one-to-ten rating from a probe. Second, a sensor does not write a price. A reading describes condition; what anyone pays to protect that condition is a decision people make, and the probe is not in that room.

## verification is a person or a second instrument

Verification has no magic in it. There are only two honest versions.

A **second source** is another instrument, another method, or another season looking at the same ground. A radar retrieval and a soil probe on the same field. A lab assay against a hand-held meter. Last year's transect against this year's.

A **person who signs** is someone who stands behind a result they did not take themselves, and who can be asked why. That is what a verification body sells, and it is a real service. It is also the part that cannot be automated, because what is bought is accountability, not arithmetic.

One rule holds across both. A satellite retrieval, a device-side species call, or any model output stays labeled as a model. Not demoted — labeled. A canopy figure from imagery is useful, and it is a model of canopy. An acoustic unit that concludes *something living was detected* is useful, and that conclusion belongs to a model. The label travels with the number, into the field notes and into the report. A model that loses its label on the way to a summary gets read as a site visit, and the people relying on the claim are then relying on something nobody offered them.

## digital mrv is a packaging standard

If you searched for MRV because of a carbon methodology, you have met digital MRV. The IWA and GBBC **dMRV Framework version 3.0** shows what that work is for. Its headline change is **extension sets** — groupings of templates covering entities, messages, formulae, and variables tied to the specific quality standard or methodology a project is bound to, organized into shared libraries that other participants can reuse. The spec says plainly what it is aimed at: unifying the manufacturing and origination of ecological and environmental products, such as voluntary carbon credits. Version 3.0 circulated as a public draft through 2025, and the spec's own version history lists a released 3.0.0 dated January 8, 2026.

That is a good standard for its job. If two projects run the same methodology, they should not each rebuild the same formulae by hand.

It is also not a creek gauge. A methodology formula computes a product — a credit, with a vintage and a buyer. A staff plate reports a water level. Both are MRV, and they are not interchangeable. We do not implement the dMRV token schema, and we are not a credit registry. We can keep a number a project already computed, labeled for what it is: a model output or a registry result, attached to the place it describes.

## carbon markets built the words. the words are bigger than the market.

MRV grew up inside the UN climate process and the carbon markets that followed it, and dismissing that is cheap. The people who argued about additionality, baselines, and leakage for twenty years are the reason anyone now expects a nature claim to carry a method. The vocabulary came out of that work.

And the same three words name something larger. A ranch that watches soil temperature because the grass tells it to has measurement. A conservancy that sends a lab sheet to its board has reporting. A watershed council with two gauges on one creek has verification. None of them is manufacturing a credit. They are keeping the record of a living place, which is what MRV described before it became a procurement category.

One adjacent case, named and left alone: a contract that pays out when a fast reading crosses a line can later read the same row. The trigger design is not this post's job. The row does not need rebuilding for it.

## what is decided, and what is not switched on

The shape above is settled: one dated result, with the device's own name, a unit, a method, and a place. That is the design.

The public writer is not on. There is no webhook URL to point a gateway at, no observation feed on ensurance.app, no observation log on a public page, and no continuous refresh behind a login. If you came looking for a dashboard, there is not one. [The place keeps the record](/guide/the-place-keeps-the-record?from=guide) already told corporate readers the same thing, and we would rather repeat it than let a design read as a product.

We also do not sell MRV as a feed. The [corporations page](/solutions/corporations?from=guide&topic=sensor-plug-in) lists **MRV & monitoring** among the work a company can fund on a place it depends on. That is field work and record-keeping on named ground, paid for by someone who needs the condition to hold. It is not a data subscription, and it is not a login.

What is live today is the place. [Browse the natural assets](/natural-assets?from=guide) and find the watershed, the stand, or the floodplain your own readings already describe. The record belongs with that place, which is the whole argument.

## frequently asked questions

### what is mrv?

MRV is measurement, reporting, and verification. Measurement writes down what a place did, reporting puts that reading in front of the people who rely on the claim, and verification is a second look — another source, or a person who stands behind a result they did not take. It is three uses of one reading, not three products.

### what is digital mrv?

Digital MRV is MRV expressed as structured, machine-readable data instead of a document. The IWA and GBBC dMRV framework version 3.0 is the reference example: it packages methodology-specific entities, messages, formulae, and variables into reusable extension sets, aimed at products such as voluntary carbon credits. It standardizes credit manufacture. It does not replace a gauge in a creek.

### how is mrv different from a carbon credit?

A carbon credit is a manufactured product with a methodology, a vintage, and a buyer. MRV is the record behind any claim about a place, credit or no credit. MRV need not support a credit at all. A water level, a soil temperature, or a lab assay on one field is already the record. The credit is one use of it.

### who verifies an ecological claim?

A second source corroborates a reading: another instrument or method on the same ground, not merely a later season. Verification is an independent party assessing the method and the data against stated criteria, then standing behind the result. A signature without that assessment is not the check. Neither is produced automatically by the instrument that took the first reading.

## read next

Much of today's measurement arrives from hardware in a field talking over a radio. That is the next post: [the radio stays outside](/guide/the-radio-stays-outside?from=guide) — the probe measures, the gateway speaks, and the record is still one row about the place.

## sources

[IWA dMRV Framework v3.0 technical specification](https://interworkalliance.github.io/TokenTaxonomyFramework/dmrv/spec/) — InterWork Alliance and Global Blockchain Business Council: extension sets, version history, and stated scope

## the series

- [what your sensors already measure](/guide/what-your-sensors-already-measure?from=guide)
- [a claim is not a reading](/guide/a-claim-is-not-a-reading?from=guide)
- [what mrv actually is](/guide/what-mrv-actually-is?from=guide)
- [the radio stays outside](/guide/the-radio-stays-outside?from=guide)
- [what radar can see](/guide/what-radar-can-see?from=guide)
- [how to plug a sensor stack in](/guide/how-to-plug-a-sensor-stack-in?from=guide)
7. [what lidar can see](/guide/what-lidar-can-see?from=guide) — a laser times a return; height is that distance