Canopy signal

(Rv)

Ground signal

(Rg)

Lidar Waveform Total LAI

GORT

model

C: clumping index

G: projection coefficient

ρv: leaf reflectance

ρg: ground reflectance

Vertical

structure

Lidar Remote Sensing of Vertical Foliage Profile (VFP) and Leaf Area Index (LAI) Hao Tang & Ralph Dubayah, Department of Geographical Sciences, University of Maryland, College Park, MD, 20742. Contact: htang@umd.edu

Introduction Framework

Implications

Lidar Data

Forest structure parameters, such as LAI and VFP can provide

potentially important information that can be used not only to better

quantify biomass and its changes, and to refine input variables of

ecosystem models and help improve carbon flux estimates.

The overall goal of this study is to derive and validate continental

scale LAI and VFP products using full waveform lidar data acquired

by the Geoscience laser Altimeter System (GLAS) onboard ICESat

(Ice, Cloud, and land Elevation Satellite).

In order to achieve this goal the following objectives will be met:

1. Develop and test a physically-based model to derive LAI and VFP from

airborne waveform lidar data.

2. Develop a robust Lidar-LAI model and validation framework for

deriving LAI and VFP from spaceborne lidar.

3. Derive and validate Lidar-derived LAI and VFP products at different

scales across major forest types.

Results

The potential LAI products resulting from this study will be used

to provide the first significant attempt to obtain large-scale estimates

of LAI from lidar remote sensing. Such products will improve

current climate and ecological models through more accurate

parameterization.

Acknowledgements: A. Strahler (Boston Univ.), C. Schaaf (Univ. MASS), F.

Zhao (UMD), S. Ganguly & G. Zhang (NASA AMES)

Derive LAI from lidar waveform using a GORT model

Methods

Table 1 Data Acquisition Time

Individual LAI Measurement

Plot-scale LAI Measurement

Regional-scale LAI Mapping

Large-scale LAI Measurement

Average

Validate

Calibrate

Validate

Calibrate

Spaceborne

Lidar

Airborne

Lidar

Terrestrial

Lidar

Echidna

LAI-2000

Hemi-photo

Plot LAI

LVIS LAI

ICESat/GLAS LAI

Site location LVIS Acquisition Field Acquisition

La Selva, Costa Rica 2005 2003-2005

Sierra Forest, CA 2003, 2008 2008

Howland Forest, ME 2003, 2009 2007

Harvard Forest, NH 2003, 2009 2007

Bartlett Forest, NH 2003, 2009 2007

Field LAI Measurement

LVIS at La

Selva GORT model

LAI&VFP

at La Selva Destructive

sampling of

LAI&VFP

at La Selva Model

sensitivity

analysis

LVIS

LAI&VFP

map (~25 m)

Continental -

scale GLAS

Modeling

Filtering

GLAS

LAI&VFP

map

Additional

Field LAI/

MODIS LAI

Echidna/Field

LAI

Refined Lidar-

LAI Model

Regional

GLAS

Regional LVIS GLAS

LAI&VFP

(~65 m )

Error analysis

(~ slope +

forest type)

Obj 1

Obj 2

Obj 3 Input Process Result Validate

General framework for validating GLAS LAI products

Destructive

Sampling

Hemispherical

Photo LAI-2000 Echidna®

GLAS

~70 m

footprint

60

0 k

m

LVIS

10

km

~25 m

footprint Full-Waveform Lidar

Laser Vegetation Imaging Sensor (LVIS)

Geoscience Laser Altimeter System (GLAS)

1064 nm laser

Polar-orbit Satellite

Airborne

a) b)

0 2 4 6 8 10 12

02

46

81

01

2

Cumulative LAI from tower

Cu

mu

lative

LA

I fr

om

LV

IS

N = 185

R2 = 0.63

Bias =0.00

RMSE = 1.36

a) Examples of cumulative LAI profiles from destructive sampling and derived from LVIS.

b) Cumulative LAI estimated from LVIS against cumulative LAI for towers with a distance

less than 5 m from the laser pointing center.

RMSE and R2 calculated upon comparison of all LVIS LAI and GLAS LAI values as a

function of slope threshold (x). For a given slope α, the y-axis represents the RMSE (or r2)

of GLAS shots with slope < α.

Slope threshold

R2

Comparison between GLAS LAI and Landsat LAI over the entire state of California

Histogram of LAI difference Landsat LAI

GL

AS

LA

I