---
title: what lidar can see
canonical_url: https://ensurance.app/guide/what-lidar-can-see
markdown_url: https://ensurance.app/guide/what-lidar-can-see.md
subtitle: a laser times a return. height is that distance. a carbon number is a later step
category: ecosystem-services
---

# what lidar can see

*a laser times a return. height is that distance. a carbon number is a later step*

Lidar — light detection and ranging — transmits a brief pulse of laser light and times how long the return takes to come back. Time becomes a distance. Distance becomes an elevation. USGS describes the airborne version plainly: a laser scanner, a GPS receiver, and an inertial navigation system, typically flown on a small aircraft.

That is the instrument. Everything else lidar is said to "measure" — canopy height, biomass, carbon — is a step taken after the timing, by somebody, under assumptions you can write down.

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.

The ridge had its shape before the plane flew it. The cottonwoods were whatever height they were before anyone differenced two surfaces. The survey is a very good description of a thing that was already there.

## the only thing the laser measures is travel time

A pulse leaves the scanner. Part of it comes back from a roof, part from a branch, part from a leaf below that branch, part from the dirt. Each return arrives at a slightly different moment, and the clock is what turns those moments into geometry.

The clock sets the floor. NASA's GEDI records time in one-nanosecond intervals — a 15-centimeter range step, the round trip, not 15 centimeters of one-way travel. Where the aircraft was, and how the ground was classified, set the rest.

So when someone says lidar is accurate, ask accurate about what. A time is measured. An elevation is derived from that time plus the position and attitude of the aircraft. A height is the difference between two derived elevations. A tonne of carbon is several steps further out, and nothing in the laser knows about it.

## a point cloud is not a model yet

What a flight produces first is a **point cloud** — millions of returns from buildings, vegetation, and the ground, all mixed together in the same file. It is not a map of the land surface. It is a map of everything the pulse hit.

A **bare-earth model** is that cloud after structures and vegetation have been stripped away, leaving the ground. For the high-resolution ground elevation models it describes under the 3D Elevation Program, the USGS FAQ states a vertical accuracy of 10 centimeters — four inches. That is the figure USGS publishes for what it describes there, not a blanket promise about every lidar product anyone will ever hand you. A row that cites it should say which product it came from.

Then there is the part that surprises people who want to use a ground model for water. USGS notes that standard DEMs contain flattened water surfaces. Wide water gets flattened into a lid at a plausible elevation so the terrain makes hydrologic sense. A narrow creek usually gets thin returns and a filled gap. Where a pulse does hit water, that hit is one water-surface elevation on that morning's datum — not stage on a gauge, and not the stage today. If you want to know how high the water is, read the plate. What it can give you is the banks, the floodplain, and whatever part of the channel was dry on the day of the flight. The bed under the water takes a green bathymetric lidar, or someone wading it. That is a useful and completely different fact.

## canopy height is a subtraction

The second surface people want from lidar is the top of the vegetation — the first returns, the roofs and crowns. Subtract the bare-earth surface from that one and the difference is a **canopy height model**. The operation is ordinary: take a surface that includes the trees, take the surface of the ground, difference them. OpenTopography serves 3DEP point clouds and the surfaces derived from them.

That difference is a height, in meters — a real, checkable quantity, and one of the few things in ecological remote sensing where you can walk out with a clinometer and argue.

What it is not: a tape around one trunk, a species call, or a mass of carbon. Height relates to biomass through a model fit to field plots somewhere, for some species mix, under some assumptions. That model may be excellent. It is still a model, and the number it produces is a different kind of object from the height that went into it.

:::johnson
**a height is a measurement. a tonne is a conclusion.** Both can sit in the record. Only one of them came off the clock, and the row should say which is which.

[how a stack joins →](/guide/how-to-plug-a-sensor-stack-in?from=guide)
:::

## gedi listens to the whole return

Most airborne lidar records discrete returns — a handful of timed hits per pulse. GEDI, mounted on the International Space Station, is a **full-waveform** instrument: it records the shape of the entire returning signal, the whole curve of energy coming back through the canopy and off the ground.

The GEDI instrument overview is unusually direct about what this means. The waveform is the sole observable. The ground is identified within that waveform. Canopy height and vertical distribution are direct measurements, which is why they can be compared against field observations. Biomass is not the observable — it is derived.

Two other facts shape what GEDI can tell you about your place. From the space station, coverage runs between roughly 51.6° north and 51.6° south, so high latitudes are out. And the measurement is a footprint of about 25–30 meters, sampled along orbit tracks — GEDI is not a wall-to-wall map of forest height. It is a very large, well-characterized sample of specific spots, and a continuous surface built from it is a model someone fitted, with your parcel inside it.

That distinction decides what a GEDI-derived figure can support. A sampled footprint near your stand is evidence about that footprint. A smoothed national layer that happens to cover your parcel is a model output with your parcel inside it.

## what arrives, and what stays with the survey

The point cloud stays where it is. Nobody needs a copy of a few hundred million returns in a record about a place, and a waveform archive belongs with the mission that built it. A pointer is enough.

What can arrive is the conclusion, carrying the same five things any other reading carries:

| field | lidar example |
|---|---|
| **the measure, in the product's own name** | bare-earth ground elevation; canopy height; vertical distribution of returns |
| **result and unit** | 2,014.3 m; 18.4 m |
| **when it happened** | the acquisition date of the flight, or the footprint's date |
| **method** | an airborne lidar survey; a full-waveform spaceborne sample; a model, if the number was derived |
| **the place** | the parcel, reach, stand, or footprint the account already claims |

A carbon figure computed from canopy height is welcome and stays labeled as a model, with the model named if you know it. A result with no place attached can be kept as a note; it does not count as evidence.

When the 2019 flight and the 2024 flight disagree about the height of a stand, both readings stay. There is no blended score, no averaged surface that quietly replaces two real observations, and no confidence number we compute out of a survey. A laser does not write a price. A person who knows the stand decides what the difference means — growth, mortality, a different ground classification, a different vendor's algorithm.

### the honest part

There is no public place to upload a point cloud here, and no viewer that refreshes. The shape of a reading is decided — one dated result, with the device's own name, a unit, a method, and a place. The public intake is not switched on, and there is no URL to send a feed to. [The place keeps the record](/guide/the-place-keeps-the-record?from=guide) said the same thing to a corporate reader and that has not changed.

What exists today is the place: a named natural asset that can be held and can keep a record. The survey you already own has somewhere to belong when the public intake is on.

## lidar and radar answer different questions

Both instruments send a pulse and time a return, which is why they get grouped together, and the similarity ends there. Cloud stops the light, and a solid surface does too, but gaps let part of a pulse reach lower branches and the ground. That is the vertical profile. There is no view through cloud. Microwaves pass through cloud and do not resolve a branch. So a canopy height and a flood extent are not two versions of one thing — they are two readings about one place, and the record keeps both. [What radar can see](/guide/what-radar-can-see?from=guide) covers that instrument on its own terms.

## frequently asked questions

### what is lidar?

Lidar stands for light detection and ranging. It transmits a brief pulse of laser light and times the return, converting that time into a distance and then into an elevation. An airborne system pairs the laser scanner with GPS and inertial navigation, typically on a small aircraft, and produces a point cloud of millions of returns from buildings, vegetation, and the ground.

### can lidar measure tree height?

Yes, as a difference. A surface that includes the tops of the trees, minus the bare-earth surface, gives a canopy height. That is a genuine measurement in meters and it can be checked in the field. It is not a measurement of one trunk's diameter, and it is not a measurement of carbon — a carbon figure is derived from the height through a separate model that should stay labeled as one.

### can lidar measure water level?

No. A standard elevation model flattens wide water and fills gaps on a narrow creek. Where a pulse hits water, it is one water-surface elevation on that morning's datum, not stage on a gauge. Lidar can describe the banks, the floodplain, and the part of the channel that was dry. The bed under the water, and the height of the water through the day, belong to a gauge or to someone in the creek.

### what is the difference between lidar and radar?

Lidar uses pulses of laser light. Cloud stops them, and so does a solid surface, but gaps let part of a pulse reach lower branches and the ground — that is the vertical profile. They cannot see through cloud. Radar uses microwave pulses, which pass through cloud and image broad surface conditions but do not resolve individual branches. Different physics, different products, and neither one is a gauge in a channel.

## read next

**[how to plug a sensor stack in →](/guide/how-to-plug-a-sensor-stack-in?from=guide)** — the doors, the fields a note should carry, and what is not switched on yet.

**[what radar can see →](/guide/what-radar-can-see?from=guide)** — the other active instrument, on its own terms.

Or look at how a named place is held: [natural assets](/natural-assets?from=guide).

## sources

[USGS — 3D Elevation Program FAQs](https://www.usgs.gov/3d-elevation-program/science/faqs) — high-resolution ground elevation models, vertical accuracy of 10 centimeters (4 inches)

[USGS — what is lidar data and where can I download it?](https://www.usgs.gov/faqs/what-lidar-data-and-where-can-i-download-it) — laser scanner, GPS and inertial navigation on a small aircraft; brief pulses of light; point cloud and bare-earth model

[USGS — about 3DEP products and services](https://www.usgs.gov/3d-elevation-program/about-3dep-products-services) — standard DEMs contain flattened water surfaces

[GEDI — instrument overview](https://gedi.umd.edu/instrument/instrument-overview/) — full-waveform lidar; the waveform as sole observable; one-nanosecond (15 cm) timing; height and vertical structure as direct measurements

[NASA Earthdata — GEDI](https://www.earthdata.nasa.gov/data/instruments/gedi) — International Space Station platform, approximately 51.6° N to 51.6° S, footprints on the order of 25 meters

[OpenTopography](https://opentopography.org/) — access to 3DEP point clouds and derived surfaces, and instructional material on differencing a surface model against a terrain model

## the series

1. [what your sensors already measure](/guide/what-your-sensors-already-measure?from=guide) — five arrivals, one record
2. [a claim is not a reading](/guide/a-claim-is-not-a-reading?from=guide) — the sentence, the reading, the gap
3. [what mrv actually is](/guide/what-mrv-actually-is?from=guide) — three uses of one row
4. [the radio stays outside](/guide/the-radio-stays-outside?from=guide) — instrument, carrier, record
5. [what radar can see](/guide/what-radar-can-see?from=guide) — two radars, neither one a gauge
6. [how to plug a sensor stack in](/guide/how-to-plug-a-sensor-stack-in?from=guide) — the doors and the note to send
7. **what lidar can see** — this post