csiro marine & atmospheric research (cmar) & ensis 1 the csiro canopy lidar initiative, its...

CSIRO Marine & Atmospheric Research (CMAR) & ENSIS 1

The CSIRO Canopy Lidar Initiative, its ECHIDNA® and an EVI

David LB Jupp1, Darius Culvenor2, Jenny Lovell1 & Glenn Newnham2 1 CSIRO Marine & Atmospheric Research (CMAR); 2 CSIRO Forestry and Forest Products (ENSIS)

David LB Jupp1, Darius Culvenor2, Jenny Lovell1 & Glenn Newnham2 1 CSIRO Marine & Atmospheric Research (CMAR); 2 CSIRO Forestry and Forest Products (ENSIS)

Presented at the IWMMM-4 Meeting in Sydney, Australia, March 20-24 2006Presented at the IWMMM-4 Meeting in Sydney, Australia, March 20-24 2006


Canopy Structure

Forest structure is complex – very complex

Canopy, trunks and stems are rarely measured as a total

Every method for measuring LAI gets a different answer

The best methods are laborious and time consuming – ie expensive

Foresters only see the trunks, environmental people see the leaves

The most significant aspects of canopy structure remain unmeasurable at all but a few sites

Forest structure is complex – very complex

Canopy, trunks and stems are rarely measured as a total

Every method for measuring LAI gets a different answer

The best methods are laborious and time consuming – ie expensive

Foresters only see the trunks, environmental people see the leaves

The most significant aspects of canopy structure remain unmeasurable at all but a few sites


The simplest of illustrations (1)Same Cover/DAI higher for Clumped

Constant size Lognormal size Clumped

DAI is mean area of disks per unit area

Cover is mean area covered by disk per unit area


The simplest of illustrations (2)Same DAI but Cover changes for Clumped

CF=5.3% CF=28.2% CF=71.5%


How should we measure the things that are the same and the things that are different?

The histograms are the same if the cover is the same.

In one case the DAI was the same in the other the Cover so what is different?

The spatial statistics change: length distributions, spatial density, the way histograms change with scale, variograms and local variance change

The difference is in the morphology therefore to measure the differences (and the similarities) you need an instrument that measures morphology

The histograms are the same if the cover is the same.

In one case the DAI was the same in the other the Cover so what is different?

The spatial statistics change: length distributions, spatial density, the way histograms change with scale, variograms and local variance change

The difference is in the morphology therefore to measure the differences (and the similarities) you need an instrument that measures morphology


Echidna® – A Ground Based Lidar

CSIRO EOC canopy Lidar Initiative (CLI) arose to promote innovative R&D, applications and commercial opportunities for airborne and ground based Lidar

ECHIDNA® is a ground based lidar technology identified by CSIRO as a potential tool for forest and vegetation structural measurement

The ECHIDNA® and its research prototype – the ECHIDNA® Validation Instrument (or “EVI”) have key differences from scanning rangefinders

Digitise the full ‘waveform’

Have variable beam divergence

Use full hemispherical scanning

Have linear response and calibration

CSIRO EOC canopy Lidar Initiative (CLI) arose to promote innovative R&D, applications and commercial opportunities for airborne and ground based Lidar

ECHIDNA® is a ground based lidar technology identified by CSIRO as a potential tool for forest and vegetation structural measurement

The ECHIDNA® and its research prototype – the ECHIDNA® Validation Instrument (or “EVI”) have key differences from scanning rangefinders

Digitise the full ‘waveform’

Have variable beam divergence

Use full hemispherical scanning

Have linear response and calibration


Ground Based Lidar (ECHIDNA®)


EVI (The ECHIDNA® Validation Instrument)


Principles of Lidar Ranging

0 2

2( ) t

p

RI t t t

c R


Hard & Soft Returns in EVI Data

Tree Trunk Foliage


Styles of product and processing

There are three basic ways that the EVI data are being analysed Stand based information from foliage profiles and stem returns (eg cover and LAI

with height, layering, stand & bole height variation, mean DBH, density and basal area);

Stand based counting information from stems and trees (eg basal area, stem density, size and stand and bole heights);

Tree based estimation of stem and foliage factors (eg leaf to stem ratios, crown size, form factor (taper), multi-stems and defect);

Each one can be made easier by projecting and re-formatting the data in different ways

There are three basic ways that the EVI data are being analysed Stand based information from foliage profiles and stem returns (eg cover and LAI

with height, layering, stand & bole height variation, mean DBH, density and basal area);

Stand based counting information from stems and trees (eg basal area, stem density, size and stand and bole heights);

Tree based estimation of stem and foliage factors (eg leaf to stem ratios, crown size, form factor (taper), multi-stems and defect);

Each one can be made easier by projecting and re-formatting the data in different ways


ECHIDNA® Data Projections

Hemispherical

Plate Carre (simple cylindrical)

Horizontal & Radial Slices


Hemisphere Data – Generalising Hemispherical Photography

EVI Data – Mean over range Hemispherical Photograph


However, there is a lot more information about the trees in EVI than in Hemispherical photography

EVI data provides strong separation between foliage profile (LAI), green height and stem profile (BA) – they are now analysed separately

Larundal Biomass Site - Holbrook

Crowns

Trunks


EVI can provide Pgap as a function of Range


Gap to range - animation


Pgap Model for EVI data

( , ) ( , ) / cos ( , )( , ) LG F r A rgapP r e


Mean Waveforms

2

( )( , ) * ( ) ( , )shot Hit

K rI r CI p P r

r


Model for Lidar Returns

( , ) ( cos ) / cos ( , )

2

( ) ( ) ( , )( , ) * ( , ) ( ) ( ) LG F r A r

shot L L t t

K r F z A rI r CI p g G p e

r z r


The data can also be “sliced” by radial distance providing tree silhouettes

Range Moments 10, 12 & 14 (Near Range)

Range Slice 15-17 m away from and above EVI for branching, defect and shape of stems

H

eight

Zenith


The data can be “sliced” by height providing stem (trunk) plots and horizontal canopy slices

Range Moments 18, 20 & 22 (Far Range comparison)

Height Slices 0.25, 1.75 & 3.75 m above EVI provide stem information

Z

enith

Radius


Field Data Stem Plot & EVI Stem Plot

R

adius Field Data

EVI


ECHIDNA® Products – height, LAI & Stem location, size distribution and density


Current applications of ECHIDNA®

Primary Information Foliage profile & LAI Stocking, Basal Area & DBH distribution (C) Stem maps and identification (C) Tree silhouettes (C) Bole height & branching (C)

In Progress Stem form factor, taper and sweep (by size class) (C) Separating branches and foliage Allometry from ground to airborne data

The potentials in forestry & ecology are almost unlimited

Primary Information Foliage profile & LAI Stocking, Basal Area & DBH distribution (C) Stem maps and identification (C) Tree silhouettes (C) Bole height & branching (C)

In Progress Stem form factor, taper and sweep (by size class) (C) Separating branches and foliage Allometry from ground to airborne data

The potentials in forestry & ecology are almost unlimited


Mapping the Canopy Structure – the “Star”

The eye normally see the “trees” rather than the “gap”

Watch the light areas and not the black!

The “Star” is the radial extent of the Laser illumination and displays the structure of the “gap” from the EVI position and where the laser “illuminates” the forest and where it does not


The radial average “Star” as an ECHIDNA®

The structure of the Star has the same information as the Trees.

The first and second order properties give us cover and BRDF by range


A “Real” (not ®) Echidna – in the forest


Current & Future work program

Commercialisation of a focused EVI/Echidna® for forest measurement

Test sites and a field mission in the US (spring of 2007?) where airborne canopy Lidar has been used (LVIS) for airborne/ground based allometry

Development of new methods for LAI, clumping, gap size distributions, BDRF functions, visibility and multi-component characterisation of forests as ecological systems

Applications to ecology and environment (using the “Star” and its structure) are major scientific goals

Commercialisation of a focused EVI/Echidna® for forest measurement

Test sites and a field mission in the US (spring of 2007?) where airborne canopy Lidar has been used (LVIS) for airborne/ground based allometry

Development of new methods for LAI, clumping, gap size distributions, BDRF functions, visibility and multi-component characterisation of forests as ecological systems

Applications to ecology and environment (using the “Star” and its structure) are major scientific goals

csiro marine & atmospheric research (cmar) & ensis 1 the csiro canopy lidar initiative, its...

Documents