Vegetation height estimation from Shuttle Radar Topography Mission and National Elevation Datasets

www.elsevier.com/locate/rse

Remote Sensing of Environm

Vegetation height estimation from Shuttle Radar Topography Mission and

National Elevation Datasets

Josef Kellndorfera,*, Wayne Walkera, Leland Piercea, Craig Dobsona, Jo Ann Fitesb,

Carolyn Hunsakerc, John Vonad, Michael Cluttere

aRadiation Laboratory, EECS Department, The University of Michigan, 1301 Beal Avenue, Ann Arbor, MI 48109, United StatesbUSDA Forest Service, Adaptive Management Services, Tahoe National Forest, Nevada City, CA, United States

cUSDA Forest Service, Forestry Sciences Laboratory, Pacific South West Research Station, Fresno, CA, United StatesdPlum Creek Timber Company, Watkinsville, GA, United States

eWarnell School of Forest Resources, The University of Georgia, Athens, GA, United States

Received 16 April 2004; received in revised form 26 July 2004; accepted 27 July 2004

Abstract

A study was conducted to determine the feasibility of obtaining estimates of vegetation canopy height from digital elevation data collected

during the 2000 Shuttle Radar Topography Mission (SRTM). The SRTM sensor mapped 80% of the Earth’s land mass with a C-band

Interferometric Synthetic Aperture Radar (InSAR) instrument, producing the most complete digital surface map of Earth. Due to the

relatively short wavelength (5.6 cm) of the SRTM instrument, the majority of incoming electromagnetic energy is reflected by scatterers

located within the vegetation canopy at heights well above the bbald-EarthQ surface. Interferometric SAR theory provides a basis for properly

identifying and accounting for the dependence of this scattering phase center height on both instrument and target characteristics, including

relative and absolute vertical error and vegetation structural attributes.

An investigation to quantify the magnitude of the vertical error component was conducted using SRTM data from two vegetation-free

areas in Iowa and North Dakota, revealing absolute errors of �4.0 and �1.1 m, respectively. It was also shown that the relative vertical error

due to phase noise can be reduced significantly through sample averaging. The relative error range for the Iowa site was reduced from 13 to 4

m and for the North Dakota site from 7 to 3 m after averaging of 50 samples. Following error reduction, it was demonstrated that the SRTM

elevation data can be successfully correlated via linear regression models with ground-measured canopy heights acquired during the general

mission timeframe from test sites located in Georgia and California. Prior to outlier removal and phase noise reduction, initial adjusted r2

values for the Georgia and California sites were 0.15 and 0.20, respectively. Following outlier analysis and averaging of at least 20 SRTM

pixels per observation, adjusted r2 values for the Georgia and California sites improved to 0.79 (rmse=1.1 m) and 0.75 (rmse=4.5 m),

respectively. An independent validation of a novel bin-based modeling strategy designed for reducing phase noise in sample plot data

confirmed both the robustness of the California model (adjusted r2=0.74) as well as the capacity of the binning strategy to produce stable

models suitable for inversion (validated rmse=4.1 m). The results suggest that a minimum mapping unit of approximately 1.8 ha is

appropriate for SRTM-based vegetation canopy height mapping.

D 2004 Elsevier Inc. All rights reserved.

Keywords: SRTM; InSAR; NED; Vegetation canopy height; Biomass; Carbon; Noise reduction

0034-4257/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.rse.2004.07.017

* Corresponding author. Tel.: +1 734 763 9442; fax: +1 734 647 2106.

E-mail address: [email protected] (J. Kellndorfer).

1. Introduction

Accurate estimates of aboveground terrestrial biomass

and carbon stocks are dependent on the availability of

biophysical measures that capture both the horizontal and

vertical structural character of vegetation. In general,

ent 93 (2004) 339–358

http:\\www2.jpl.nasa.gov\srtm\coverage.html

J. Kellndorfer et al. / Remote Sensing of Environment 93 (2004) 339–358340

optical remote sensing systems are well suited to the

acquisition of structural information in the horizontal

dimension (e.g., canopy cover, community type, etc.)

given their long-established sensitivity to variations in

pigment composition and surface biochemistry. However,

in the vertical dimension, passive optical systems such as

Landsat and SPOT suffer from an inability to penetrate

through layers of vegetation, making difficult the accurate

retrieval of vertical structural metrics such as canopy

height, particularly as canopy density increases. Only in

the last decade have significant advances in the develop-

ment of active sensor technologies made it possible to

obtain consistent and reliable estimates of vertical

structural metrics. These advances include sensors such

as Light Detection and Ranging (LIDAR) (Dubayah &

Drake, 2003; Lefsky et al., 1999) and Synthetic Aperture

Radar (SAR) (Bergen & Dobson, 1999; Dobson, 2000;

Dobson et al., 1995; Kasischke et al., 1997; Kellndorfer

& Ulaby, 2003; Kellndorfer et al., 1998, 2003; Ulaby et

al., 1995). Another technology for estimating forest

vertical structure based on an airborne microwave

scatterometer has been proposed by Hyyppa and Halli-

kainen (1996) and Martinez et al. (2000). While both

airborne LIDAR and microwave scatterometers can

achieve high accuracy and high spatial resolution esti-

mates of vegetation height, neither technology is currently

capable of providing regional-scale to global-scale data-

sets. On the other hand, to date, several spaceborne SAR

Fig. 1. Conceptual representation of a forest stand indicating the relative position

SRTM resolution cell (~30 m).

missions have generated an abundance of global-scale

data, which have proven useful in the estimation of forest

biophysical parameters.

In particular, Interferometric Synthetic Aperture Radar

(InSAR) has proven to be an invaluable tool in the

determination of vegetation canopy height and a number

of studies have had success retrieving canopy height from

InSAR measurements (Brown, 2003; Hagberg et al., 1995;

Kobayashi et al., 2000; Papathanassiou & Cloude, 2001;

Rosen et al., 2000; Sarabandi & Lin, 2000; Treuhaft &

Siqueira, 2000). Papathanassoiu and Cloude (2001) used a

repeat-pass (temporal baselinec10 min) polarimetric (using

both horizontal- and vertical-polarized fields) airborne

instrument at L-band (23 cm wavelength) to acquire

estimates of canopy height. Their results showed a standard

deviation between estimated and observed heights of

approximately 2.5 m. In another work, Hagberg et al.

(1995) used the spaceborne ERS-1 C-band (5.6 cm wave-

length with vertical polarization) platform as part of a

repeat-pass (temporal baselinecseveral weeks) processing

scheme over a boreal forest at Hfkmark in northern Sweden.

Due to the influence of temporal decorrelation, their results

were less accurate with rms errors on the order of 5 m. In a

third study conducted by Treuhaft at al. (1995), the

Topographic Synthetics Aperture Radar (TOPSAR) operat-

ing at C-band (vertical polarization) was flown over

multiple boreal forest stands within the Bonanza Creek

Experimental Forest, Alaska. The estimation error reported

s of mean canopy height and scattering phase center height within a single

J. Kellndorfer et al. / Remote Sensing of Environment 93 (2004) 339–358 341

in this case varied between 3 and 6 m, depending on the

forest stand and environmental conditions.

In February 2000, an unprecedented near-global eleva-

tion dataset based on single-pass InSAR technology was

acquired as part of the National Aeronautics and Space

Administration (NASA) Jet Propulsion Laboratory (JPL)

Shuttle Radar Topography Mission (SRTM), resulting in the

most complete digital topographic database of the Earth. In

accordance with a memorandum of understanding between

NASA and the primary mission sponsor, the National

Geospatial Intelligence Agency (NGA; formerly NIMA), a

raster digital elevation model (DEM) dataset was released at

a resolution of 1-arc sec (c30 m) within the United States

and 3 arc sec (c90 m) elsewhere. Given the relatively short

operating wavelength (C-band, 5.6 cm) of the SRTM sensor,

the interferometric height response over vegetated terrain is

expected to reflect the interaction of the InSAR signal with

various scatterers associated with leaves, branches, and

stems. As a result, the height surface [i.e., the scattering

phase center height (hspc)], retrieved via InSAR processing,

will be higher than the underlying bbald-EarthQ surface (Fig.1). Preliminary work by Brown (2003) investigated the use

of SRTM data for estimating the height of vegetation

canopies. Based on both simulated and real data from red

pine stands in Michigan, canopy height estimates were

achieved with an rms error of approximately 4 m.

The objective of this study is to determine the feasibility

of deriving vegetation canopy height from the SRTM digital

elevation data in conjunction with the National Elevation

Dataset (NED), which provides a reference surface for bald-

Earth topography. This effort focuses on two pilot study

areas: one in southeastern Georgia near Jesup and a second

in the northern Sierra Nevada of California near Quincy.

The Jesup site lies in the center of a heavily managed

landscape characterized by large homogeneous forest stands

and moderately undulating topography. Conversely, the

Fig. 2. SRTM global coverage map showing the number of data takes acquired ov

and North Dakota test sites are shown (modified from http://www2.jpl.nasa.gov/s

Sierra Nevada site represents less intensely managed, more

heterogeneous forests and highly variable terrain. Thus, the

biogeophysical characteristics of the two test sites provide a

unique opportunity to evaluate the SRTM data as a source

for canopy height estimates across a range of vegetation

densities and structural classes as well as a variety of

topographic conditions.

2. Shuttle radar topography data

2.1. Mission and instrument characteristics

The SRTM was flown on board the Space Shuttle

Endeavor during mission STS-99, which was in orbit from

February 11 to February 22 of 2000. SRTM was the first

spaceborne fixed-baseline InSAR and marked the largest

rigid structure to be deployed in space with a payload

weight of 13,660 kg. During the mission, a total of 12.3

Tbyte of data were collected. The orbit inclination was 578,which allowed for a targeted coverage of 80% of the total

Earth landmass lying between the latitudes of 608 north and

568 south (Fig. 2). Of the total targeted area, 99.97% was

mapped with at least one data take (i.e., one overpass),

corresponding to 119.51 million km2. Additionally, 94.59%

(113.10 million km2) of the targeted area was covered at

least twice, 49.25% (58.59 million km2) at least three times,

and 24.10% (28.81 million km2) at least four times. Six

locations, totaling 50,000 km2, were not imaged during the

mission, all of which are located in the conterminous United

States (USGS, 2003).

SRTM was flown at an altitude of 233 km where two

InSAR instruments were operated. These included the

United States C-band (5.6 cm, 5.3 GHz) sensor and the

German X-band (3.1 cm, 9.6 GHz) sensor. To enable the

collection of single-pass fixed-baseline interferometric data,

er land and water. Approximate locations of the California, Iowa, Georgia,

rtm/coverage.html).

http:\\www2.jpl.nasa.gov\srtm\coverage.html


a 60-m-long boom was extended from the shuttle cargo bay

with C-band- and X-band-receiving antennas attached to its

end. Dual-purpose transmit/receive antennas were operated

in the cargo bay. The 60-m fixed-baseline configuration of

the InSAR system, in conjunction with a ScanSAR mapping

mode, resulted in a swath width of 225 km for C-band. The

ScanSAR mode was comprised of four subswaths, which

alternated between horizontal and vertical polarizations. The

look angle ranged between 308 and 608 (Hensley et al.,

2000).

2.2. SRTM data characteristics

According to the mission objectives, SRTM data were

expected to have an absolute horizontal circular accuracy of

less than 20 m. Absolute and relative vertical accuracy was

anticipated to be less than 16 and 10 m, respectively.

Performance evaluations by NIMA, the USGS, and the

SRTM project team have shown the absolute vertical error

to be much smaller, with the most reliable estimates being

approximately 5 m (Curkendall et al., 2003; Rosen et al.,

2001a,b; Smith & Sandwell, 2003; Sun et al., 2003). Given

the single-pass interferometric configuration of the SRTM

sensor, the dataset was not subject to the temporal

decorrelation errors that are common in repeat-pass systems.

Following the mission, the SRTM dataset was interfero-

metrically processed by the JPL and made available to the

public. The processor included averaging where multiple

data takes were acquired and filtering of the interferogram to

reduce noise using either a boxcar lowpass filter or power

spectral filtering (Hensley et al., 2000). Within the Unites

States, data were released at a spacing of 1 arc sec (c30 m).

For all other regions, data are being released at a spacing of

3 arc sec (c90 m). The projection parameters are set to

geographic coordinates (unprojected) and the data are

horizontally referenced to the North American Datum

1983 (NAD83). The vertical reference datum is the North

American Vertical Datum 1988 (NAVD88). SRTM data for

the United States are accessible via the National Map

Seamless Data Distribution System provided by the United

States Geological Survey (http://seamless.usgs.gov).

3. NED

The NED is a compilation of various elevation data

sources including 7.5 min, 15 min, 2 arc sec, and 3 arc sec

DEMs dating back as far as 1978. In 1999, and for the

first time, the NED was assembled completely for the

continental United States from 7.5-min DEM source data

(10 and 30 m resolution) (Gesch et al., 2002). Develop-

ment of the NED required the merging of 57,000 different

DEM data files—54,000 within the conterminous United

States alone. The NED is a seamless dataset where

procedures were developed to maintain the database with

periodic updates to insure the integration of higher-

resolution elevation data as these data become available.

The NED is released in geographic coordinates at a

resolution of 1 arc sec. The horizontal and vertical data

are NAD83 and NAVD88, respectively. Thus, the NED is

available in the same resolution and projection parameters

as the SRTM dataset. Given the production history of the

NED, the accuracy varies spatially with the quality of the

individual source data. The USGS is assessing the quality

of the NED by comparing it to the High-Accuracy

Reference Network (HARN). Releases of the NED

currently contain accuracy statistics where available (Gesch

et al., 2002).

4. Study area datasets

4.1. Test site Jesup, Georgia

During a NASA-EOCAP SAR program, an intensive

survey of slash pine (Pinus elliotii) plantations owned by

the Plum Creek Timber was conducted for a test site near

Jesup (31875VN, 82800VW), southeast Georgia (Kellndorfer

et al., 2003). The field campaign was conducted during the

early months of 2000, which coincided well with the

timeframe of the SRTM mission. A total of 22 stands were

biometrically surveyed using a cluster plot design (Vona,

2001). Fifteen of the 22 stands were unthinned and a 0.02-

acre (0.008 ha) cluster-plot survey was performed. The

remaining seven stands were thinned and the cluster plot

size was 0.05 acre (0.02 ha). Each cluster contained a center

plot and four diagonally aligned subplots at a distance of

one chain (c20 m) from the center. Throughout each stand,

cluster plots were sampled on a regular grid of 3�6 chains

(c60�120 m). The tally consisted of recording all stems

greater than 1�1 in. (2.54 cm) diameter at breast height

(DBH) classes. Additionally, a minimum of 8 dominant and

16 codominant trees in each stand were selected at random,

and height measurements were made using a handheld

clinometer.

Published biometric models were used to calculate

several biophysical parameters including basal area, dom-

inant height, stem density, and volume from the plot survey

data. According to the standards for forest measurement in

the United States, the following equations reflect units of the

English system. Using the observed dominant height

measurements (hd), the heights of all remaining slash pine

stems were predicted (hp) after Pienaar et al. (1993) with:

hp ¼ 1:12hd

�1� 1:257e

�2:058�

dDq

��ð1Þ

where hd=height of observed dominants [ft]; d=DBH of

remaining stems [in.]; Dq=quadratic mean diameter of

remaining stems [in.].

For further analysis, the stand biometric parameters were

translated from English to SI units. The mean canopy height

http:\\seamless.usgs.gov


(hcan) of each stand was computed by averaging the heights

(ht) for all trees (NT) present in all plots:

h¯ can ¼1

NT

XNT

t¼1

ht ð2Þ

where hcan=mean canopy height of stand; NT=number of

observed (hd) and predicted (hp) stems; h t=height of

observed (hd) or predicted (hp) stems.

As a measure of vertical structural variability, the

standard deviation (rcan) of canopy height was computed

as well.

Because the Jesup survey was part of an experiment

designed around acquisition of data by the JPL Airborne

Synthetic Aperture Radar (AirSAR), all stands were located

Fig. 3. (a) NED and (b) SRTM digital elevation models of the Jesup, Georgia test s

Bright areas in both images represent areas of higher relative elevation. Image sc

within a 7-km-wide swath, which spanned the test site from

the southwest to the northeast. The original stand bounda-

ries were supplied by Plum Creek in the form of a vector

polygon layer, which was derived from uncorrected aerial

photography and, hence, many polygons contained minor

location errors. Although the stand boundaries were

accurate in terms of their general shape, their location and

extent were not completely reliable. In some cases,

polygons were moved to the correct location, which was

identified using the georeferenced AirSAR data.

Fig. 3a shows the NED DEM image of the Jesup test site

as well as the locations of the surveyed stands. The test site

is moderately hilly with elevations ranging from 9 to 46 m,

and a drainage network formed by several creeks and

streams is apparent. The test site lies in an intensively

ite. Boundary locations of 22 surveyed stands are shown as white polygons.

ale is approximately 33�24 km (790 km2).


managed timber-producing region and the majority of the

area consists of slash pine plantations. Lowland vegetation

is also common and forms narrow bands of hardwoods that

buffer the drainage network. Fig. 3b shows the SRTM DEM

image with the stand locations superimposed. The SRTM

image demonstrates the sensitivity of the sensor to topo-

graphic features like drainage networks and undulating

terrain that are also seen in the NED image. Additionally,

the SRTM image reveals the marked sensitivity of the

sensor to features relating to the spatial distribution of

Fig. 4. (a) SRTM minus NED difference image of the Jesup, Georgia, test site. Th

surface that reflects the height of the scattering phase center (hspc). (b) Transects

(upper curve), NED (center curve), and SRTM–NED (lower curve) difference (h

vegetation. For example, taller features like forest stands

appear brighter (higher elevation) relative to adjacent

features like clearcuts and fields, which are shorter and

appear darker (lower elevation). The sensitivity of the

SRTM sensor to features which extend vertically above the

bald-Earth surface is even more clearly observed when the

NED DEM elevations are subtracted from the SRTM DEM

(Fig. 4a). In this difference image, the gray-value range

corresponds to increasing feature height from dark to light.

Discussion of Fig. 4b is deferred to Section 5.3.

e differencing procedure removes the underlying topography, resulting in a

A–C in (a) were used to extract elevation profiles (A–C) from the SRTM

spc) images. The reference (flat) line represents zero elevation.


4.2. Test site Sierra Nevada, California

As part of NASA-funded remote sensing research

focused on the structure of forests in the Sierra Nevada

mountain range, a 10,000-ha test site was established 40 km

southwest of Quincy (39857VN, 120856VW) on the Plumas

National Forest, northern California. The test site is

dominated by mixed coniferous forests characterized by

varying amounts of Douglas-fir (Pseudotsuga menziesii),

red fir (Abies magnifica), white fir (Abies concolor), sugar

pine (Pinus lambertiana), ponderosa pine (Pinus ponder-

osa), and incense cedar (Calocedrus decurrens), with

elevations ranging from 1200 to 1850 m.

Fig. 5. (a) NED and (b) SRTM digital elevation models of the Sierra Nevada, Califo

shown as white circles. Bright areas in both images represent areas of higher rela

During 2000 and 2001, an intensive survey was

conducted to characterize the three-dimensional structure

of forest stands within the test site. A grid of 1.0-ha

circular (56.4 m radius) sample plots representing approx-

imately 3% of the test site was generated and super-

imposed on layers of transportation and hydrology. Roads

and rivers were each buffered by 10 m; plots intersecting

the buffered network were systematically shifted to the

west by 56.4 m. If the western shift was unsuitable, plots

were shifted east, then north, and, finally, south in

sequence. The final coordinates of each plot were then

calculated and used to identify plot locations in the field

via global positioning system (GPS).

rnia, test site. The locations of 227 56.4-m radius (1.0 ha) surveyed plots are

tive elevation. Image scale is approximately 24�18 km (430 km2).


Each sample plot consisted of two concentric annuli with

radii of 15.0 m (0.07 ha) and 56.4 m (1.0 ha). In the 15-m

inner annulus, all live trees z10 cm in DBH were targeted

for structural measurements. Structural variables measured

and recorded for live trees included: (1) species, (2) DBH,

(3) crown diameter, (4) crown form, (5) height to bottom of

partial crown, (6) partial crown wedge angle, (7) height to

bottom of full crown, (8) height to top of live crown, and (9)

total tree height. In the outer annulus (15–56.4 m), structural

measurements focused on large (late successional/old

growth) live trees (z76 cm DBH). Measurements of live

Fig. 6. SRTM minus NED difference image of the Sierra Nevada, California, test si

in a surface that reflects the height of the scattering phase center (hspc). Transects

curve), NED (center curve), and SRTM–NED (lower curve) difference (hspc) ima

trees included: (1) species, (2) DBH, and (3) and total tree

height. All heights were measured to the nearest 0.01 m with

an Impulse 200 LR Laser Rangefinder.

In total, 227 sample plots were established and measured

during the northern Sierra Nevada field campaign. For each

plot, the mean canopy height (hcan) was computed according

to Eq. (2), where NT is the number of measured stems in

each plot and stem height (ht) corresponds to the measured

height of each stem.

Fig. 5a shows the locations of all sample plots super-

imposed on the NED DEM of the test site. In general, the

te. The differencing procedure removes the underlying topography, resulting

A–C were used to extract elevation profiles (A–C) from the SRTM (upper

ges. The reference (flat) line represents zero elevation.


site is characterized by rolling to steeply sloping terrain. A

number of drainage networks are visible in the NED DEM

including the prominent middle fork of the Feather River in

the southeast corner. Bucks Lake, shown in black, is located

along the west-central edge of the image. The SRTM DEM

of the test site is shown in Fig. 5b. The marked sensitivity to

terrain features that was clearly visible in the Jesup SRTM

image is similarly demonstrated here. However, unlike the

Jesup SRTM image, the Sierra Nevada SRTM image

appears to be nearly identical to its NED complement, and

seemingly devoid of all vegetation and anthropogenic

features. Nevertheless, much of the northern Sierra Nevada

study site is, in fact, heavily forested. Whereas in the Jesup

SRTM image, individual forest stands and other features

stand out as brighter compared to their surroundings, in the

corresponding Sierra Nevada image, brightness differences

reflecting finer-scale vegetation and anthropogenic features

are almost completely masked by the larger-scale brightness

differences that result from the pronounced variation in

topographic elevation. However, with the influence of

topography being effectively removed in the SRTM–NED

difference image (Fig. 6a), tracts of taller trees can be

identified as they appear brighter and are interspersed with

occasional clearcuts, shrub fields, and otherwise unforested

land, which appears darker. Discussion of Fig. 6b is deferred

to Section 5.3.

5. Determination of the scattering phase center heights

5.1. Theory

Given the nature of the interferometric height response,

the SRTM dataset represents the scattering phase center

height (hspc), and only reflects bald-Earth elevations in

vegetation- and structure-free areas. The scattering phase

center height is dependent on target and sensor character-

istics and can be described according to:

hspc ¼ ftarget vs; vm; sr; smÞofsensor k; b; p; h; gp��

ð3Þ

where vs=vegetation structure; vm=vegetation moisture;

sr=soil roughness; sm=soil moisture; o=concatenation oper-

ator; k=wavelength; b=baseline length and attitude;

p=polarization; h=incidence angle; gp=phase noise.

For the SRTM C-band sensor configuration, wavelength

(k) and baseline length/attitude (b) are constant. Therefore,

the dependency of hspc on these parameters is unvarying

across the SRTM dataset. On the other hand, polarization

( p) and incidence angle (h) are variable within each

ScanSAR swath. Simulations conducted by Sarabandi and

Lin (2000) suggest that the C-band interferometric height

response for coniferous vegetation is lower for hh polar-

ization than for vv polarization by approximately 1–2 m. For

deciduous vegetation, the simulation results show no

significant difference in the hh and vv polarization height

response. Simulations of the incidence angle dependency in

the 30–608 range corresponding to the SRTM ScanSAR

swath indicate variations in the scattering phase center

height on the order of 1–2 m for coniferous vegetation

(based on an average stand height of 9 m) and no significant

difference for deciduous vegetation (based on an average

stand height of 17 m). While both simulations were based

on stands with closed canopies, it is understood that in open

canopies at steep incidence angles, the SAR signal will

penetrate deeper, thus lowering the center of the interfero-

metric height response. Nevertheless, in order to reduce the

dependency of the height response on p and h, integration(i.e., averaging of multiple data takes), ranging from 2 to as

many as 10 or more, was performed during the NASA-JPL

product generation phase (USGS, 2003). Since incidence

angle information from the individual data takes cannot be

retrieved from the integrated dataset, the influence of

incidence angle on the interferometric height response

cannot be determined. Thus, a p- and h-dependent residualerror in the estimation of hspc is to be assumed.

The influence of soil roughness and moisture on hspc is

dependent on k, h, and canopy density. Canopy density

directly affects the extent to which an electromagnetic wave

can penetrate through the canopy to the ground. For

moderate to dense canopies, the C-band signal return at

the incidence angle range of SRTM can be largely attributed

to volume scattering from within the upper canopy rather

than from the underlying surfaces (Sarabandi & Lin, 2000).

For low-density canopies, the C-band signal has a greater

likelihood of penetrating to the ground. Ignoring signal

dependencies on soil moisture and roughness for the time

being, as the amount of signal penetrating to the ground

increases, backscatter from the ground surface will increase.

As a result of the higher ground surface response, hspc will

be observed to decrease (i.e., the center of phase scattering

will be closer to the ground). Superimposed on this

observation, varying soil moisture and roughness character-

istics can influence hspc as well. Simulations by Sarabandi

and Lin (2000) suggest that the soil-related influence on the

height response for the SRTM incidence angle range and

wavelength is less than 0.5 m. Since soil roughness and

moisture measurements were not available for the Jesup and

Sierra Nevada test sites, and accounting for these parameters

would prove rather difficult in the context of the SRTM

processing scheme, an additional yet minor residual error

component remains in the estimation of hspc due to these

influences.

A more pronounced source of measurement uncertainty

as compared to the error sources described above relates to

phase noise. For single-pass interferometers like SRTM,

phase noise is mainly caused by thermal and quantization

noise of the radar receivers (Bamler, 1999). The magnitude

of the phase noise error is largely dependent on the number

of radar looks used in signal averaging. Hence, the error is

less where multiple data takes were acquired and multiple

samples are averaged. For Gaussian phase noise, the


uncertainty decreases with the square root of the number of

looks (e.g., averaging four looks would result in an error

reduction of 50%). For the SRTM mission, a phase noise

error of less than 10 m was targeted (Hensley et al., 2000).

To summarize the potential error sources described thus

far, a total relative error, which is decoupled from the

influence of vegetation on the estimation of hspc can be

defined as:

er ¼ esoil þ ep;h þ epn ð4Þ

where er=total relative error; esoil=soil error due to soil

roughness and moisture uncertainty; ep,h=error due to

polarization and incidence angle uncertainty; epn=error dueto phase noise uncertainty.

Separate from the existence of a relative error, an

absolute error in SRTM-derived elevation is also observed

and is related to (1) error in the attitude (roll) of the

interferometric baseline, and (2) error in the measurement

of the baseline length. Both of these errors result in large-

scale deviations in SRTM elevation from that of the true

surface, but generally, this deviation can be corrected

using ground control points (GCPs) (Bamler, 1999).

Depending on the source of the GCPs (e.g., the NED),

the reference elevation data may have inherent errors as

well. Thus, in order to compare SRTM elevations to

reference elevations, we define a spatial variable that

represents a vertical offset between the SRTM and a given

reference dataset according to:

dv x; yð Þ ¼ hSRTM�b x; yð Þ � href x; yð Þ ð5Þ

where dv=vertical offset; hSRTM-b=noise-free SRTM bald-

Earth elevation; href = reference bald-Earth elevation;

x,y=geographic location.

Given that the absolute error associated with the SRTM

and reference DEMs varies at a relatively broad spatial scale

(thousands of kilometers), dv is expected to result in a

constant (i.e., trendless) offset for relatively small regions

and a spatially variable offset or trend for larger regions.

Following the relative and absolute error formulations

provided in Eqs. (4) and (5), respectively, Eq. (3) can now

be reformulated as:

hspc ¼ fveg vs; vmÞ þ dv þ erð ð6Þ

It follows from Eq. (6) that in order to accurately model

the functional relationship between vegetation character-

istics (vs and vm) and hspc, it is desirable to identify and

remove dv and minimize, to the extent possible, the error

component er, which is primarily dependent on the

reduction of epn.

5.2. Investigation of SRTM phase noise error and vertical

offset

In an attempt to quantify the potential range of dv and epnvalues to be expected within the SRTM coverage of the

conterminous United States, an analysis was carried out

using SRTM and NED elevation data from two large (c1.5

km2) agricultural fields, one in Iowa (IA) and one in North

Dakota (ND) (Fig. 2). Agricultural fields were selected

because they tend to be flat, and during the February 2000

timeframe of the SRTM mission, they would have been

devoid of vegetation. The selected fields were equal in size,

covering an area of approximately 1600 SRTM pixels.

Based on inspection of the SRTM coverage map (Fig. 2),

Field IA (41855VN, 94827VW) was located in a region

where only one SRTM data take was obtained. Conversely,

Field ND (48851VN, 101800VW) was selected from a region

where at least four data takes were acquired. The averaging

of multiple data takes, ranging from 2 to as many as 10 or

more, was performed during the NASA-JPL product

generation phase (USGS, 2003). Given the timeframe of

the SRTM mission and the northerly latitude of Iowa and

North Dakota, both fields are assumed to have been snow-

covered when the SRTM data were acquired. At northerly

latitudes, snow is assumed to be essentially dry in February

and, hence, the presence of snow was not expected to affect

the estimation of dv (Rignot et al., 2001). Given that several

more data takes (N4) were available for Field ND compared

with Field IA (Fig. 1), less phase noise was expected in the

SRTM pixel values for Field ND.

The first step in the analysis was the elimination of

existing topographic variation within the fields by subtract-

ing the NED DEM from the SRTM DEM. Under the

assumption that noise-free SRTM and NED data should

provide for consistent bald-Earth elevation estimates, it

follows that: (1) the residuals of the SRTM–NED difference

image can be attributed to epn, and (2) the mean of the

residuals, should it differ from zero, is an estimate of the

vertical offset (dv) between the SRTM and NED datasets.

From the frequency distribution and summary statistics for

the two difference images, a number of inferences can be

made (Fig. 7, Table 1). First, the phase noise within both

fields is observed to be Gaussian in nature. Second, the

mean value calculated from the difference images (IA=�4.0

m, ND=�1.1 m) is other than zero, reflecting a vertical

offset between the SRTM and NED datasets, which differs

between sites by approximately 2.9 m. Third, the narrower

noise range associated with Field ND (7 m) compared to

that of Field IA (13 m) confirms the hypothesis that greater

noise reduction occurs in locations (i.e., North Dakota)

where multiple data takes have been averaged. In order to

quantify the relationship between sample (i.e., pixel)

averaging and subsequent noise reduction, samples from

each of the difference images were block-averaged with a

window size ranging from 3�3 to 25�25 (i.e., 9–625

samples per block). Fig. 8 shows the results for each field.

In both cases, the results are shown after the removal (via

subtraction) of the respective dv offset. As expected, the

noise reduction is reflected in a decrease in the noise range

with increasing window size. This reduction is more

pronounced up to ca. 7�7 samples averaged, with a sharper

Fig. 7. Histogram of sample values extracted from Fields IA (Iowa) and ND (North Dakota) illustrating the presence of Gaussian phase noise (epn).


decrease observed in Field IA. At 7�7 samples (~50 pixels)

averaged, the range for Field IA was reduced from 13 m to

approximately 4 m, and the range for Field ND was reduced

from 7 m to approximately 3 m. Beyond the 7�7 threshold,

a more gradual decrease is observed in both fields and the

maximum, minimum, and mean values begin to converge.

Table 1

Summary statistics for phase noise (epn) from Fields IA (Iowa) and ND

(North Dakota)

Field IA Field ND

41.918N, 94.458W 48.858N, 101.008W

SRTM Min 330.0 449.0

Mean 337.5 454.1

Max 346.0 457.0

Range 16.0 8.0

NED Min 337.7 451.4

Mean 341.5 455.1

Max 345.8 456.7

Range 8.1 5.3

SRTM–NED Min �10.1 �4.1

Mean �4.0 �1.1

Max 2.6 2.7

Range 12.7 6.8

At a threshold of 25�25 (=625) samples, the noise range is

reduced to less than 1 m.

Compared to the relatively flat, nonvegetated regions

studied here, the phase noise inherent to SRTM data from

forested terrain should be far less pronounced due to the

higher signal-to-noise ratio associated with InSAR back-

scatter from vegetation canopies. Hence, the noise curves

shown in Fig. 8 should represent worst-case scenarios with

respect to the relative error to be expected in the SRTM data,

particularly for Field IA where only one SRTM data take

was acquired.

5.3. Test site error reduction and vertical offset identification

An effort was made to apply the knowledge of SRTM

phase noise and vertical offsets gained above to the analysis

of the Jesup and Sierra Nevada test sites. The extent to

which these error sources were present in the datasets was

explored graphically via the assessment of elevation profiles

extracted from the SRTM, NED, and SRTM–NED differ-

ence images. A series of elevation profiles (A–C) was

extracted from the Jesup test site along selected transects

(Fig. 4b). In general, Profile A, representing a transect of

Fig. 8. Comparison of phase noise (epn) statistics (minimum, mean, and maximum) for Fields IA (Iowa) and ND (North Dakota) following a block averaging

procedure in which block size was increased linearly from 3�3 (nine samples) to 25�25 (625 samples). All values were adjusted by the mean absolute error

determined from each field.


approximately 15 km, demonstrates the close overall

correspondence between the SRTM and NED elevation

datasets. Further inspection of the profile reveals the

pronounced relative difference in elevation between the

two DEMs, which is best reflected in the SRTM–NED

difference (hspc) curve shown at the bottom of Profile A.

This difference in elevation reflects the presence of, and

variability associated with, vegetation and anthropogenic

features along the transect. Profiles B and C are approx-

imately 2–3 km in length and correspond to transects across

riparian (B) and upland (C) areas (Fig. 4b). In profile B, the

SRTM–NED difference curve reveals quite clearly the

presence of peaks associated with bands of riparian forest.

Similarly, profile C shows the sharp boundaries between

alternating plantations and clearcuts.

Elevation profiles (A–C) extracted from the Sierra

Nevada test site illustrate the greater topographic variability

and elevation range that distinguishes this site from Jesup

(Fig. 6b). Whereas Sierra Nevada Profile (SP) A exhibits an

elevation range of approximately 400 m, Jesup Profile (JP)

A has a range that is closer to 25 m (Fig. 4b). In general, SP

A also suggests the more continuous nature of the forest and

the presence of taller trees. Sierra Nevada Profiles B and C

are approximately 2–3 km in length and are intended to

complement the corresponding Jesup profiles. Profile B

reflects a transect crossing three more or less parallel

riparian zones (Fig. 6b). Again, the SRTM–NED difference

curve (bottom) captures the presence of taller vegetation in

these moister, more hospitable areas. The transect associated

with Profile C crosses a region of alternating plantations and

clearcuts. The SRTM–NED difference curve (bottom) is

consistent with that of JP C, revealing the sharp boundaries

between these features.

In general, the elevation profiles presented in Figs. 4b

and 6b illustrate the marked sensitivity of the SRTM sensor

to vegetation canopy height. At the same time, however, the

profiles also reveal the presence of phase noise error. As a

result, and in order to improve subsequent scattering phase

center height calculations, efforts were undertaken to reduce

the observed noise component. In both test sites, this was

accomplished using a sample (i.e., pixel) averaging

approach.

For the Jesup site, sample averaging was carried out

within the boundaries of each of the 22 stands described in

Section 4.1. The field data (i.e., cluster plots) were

aggregated within these boundaries as well. The interior of

stand boundaries was buffered by 30 m (i.e., one SRTM

pixel) to minimize edge effects resulting from geocoding

errors. Following this procedure, the number of samples per

stand ranged from 2 to 574.

For the Sierra Nevada site, sample averaging was

accomplished by extracting all pixels within a radius of 45

m around the center point of each plot. This zone of

inclusion was selected as a basis for averaging all pixels

having a majority of their area within the boundary of the

56.4-m (1.0 ha) sample plot annulus. Given the relatively

low pixel-to-plot area ratio, the extraction procedure

resulted in the averaging of just 6–10 pixels per plot. Based

on the results presented in Fig. 8, this sample size was

deemed much too small to effectively reduce the phase

noise component. Operating under the assumption that

larger sample plots (e.g., coincident with stand-level

boundaries) would increase the pixel-to-plot area ratio, a

novel strategy was devised based on a linear binning or

partitioning of the observed canopy height variable. The

goal of this strategy was to effectively simulate stand-level

statistical averaging using plot level data. To implement the

approach, the original 227-plot dataset was divided among

20 bins, with the number of plots per bin ranging from 3 to

29. With the exception of the lowermost and two uppermost


bins, the bin width was set to 1.5 m. Bin width was

determined based on a desire to obtain a minimum of 20

pixels per bin. Under this binning strategy, the number of

pixels per bsimulated standQ or bin ranged from 24 to 253—

a sample size much more suitable for reducing epn.Within each bin, the observed mean canopy height (hcan)

was computed according to Eq. (2), with NT representing

the sum of all stems in all plots within the respective bin,

and ht corresponding to the observed heights of these stems.

In general, the binning strategy can be equated to simulating

mean observed canopy and associated scattering phase

center heights within homogeneous (in terms of observed

height), contiguous stands, with the caveat that individual

sample plots are not co-located within contiguous stands but

are distributed across a larger forested region (see plot

locations in Fig. 5).

To determine if a vertical offset (dv) existed between the

SRTM and NED datasets for the test sites, regions

representing barren, nonvegetated areas were identified

from the National Land Cover Dataset (NLCD) (Vogelmann

et al., 2001) and recent Landsat ETM+ imagery. For the

Jesup site, the analysis revealed a small offset of �0.2 m

(i.e., bald-Earth SRTM elevations were, on average, 20 cm

lower than NED elevations). For the Sierra Nevada test site,

the offset was �1.3 m.

5.4. Test site scattering phase center height calculation

Following the procedures for error accounting described

above, the mean scattering phase center height (hspc) for a

vegetated region can be estimated within relatively narrow

error bounds if a large-enough sample population can be

averaged. Substituting the unknown functional dependency

on vegetation characteristics in Eq. (6) with the SRTM–

NED difference, the mean scattering phase center heights

for the Jesup and Sierra Nevada test sites were calculated

according to:

h¯ spc ¼1

Np

XNP

i¼1

hSRTM � hNEDÞ þ dvð ð7Þ

where hspc=mean scattering phase center height; hSRTM=

SRTM pixel elevation; hNED=NED pixel elevation; NP=

number of pixels in a stand, plot, or bin.

The remaining error component er associated with the

hspc of each stand (Jesup), plot, or bin (Sierra Nevada) is

primarily dependent on the number of samples (NP)

averaged and can be approximated from Fig. 8.

6. Regression models for canopy height determination

The primary objective of this study was to determine

whether a functional relationship exists between the STRM

scattering phase center height and the observed vegetation

canopy height for forested areas in the Jesup and Sierra

Nevada test sites. Given the significant relative error

inherent in the SRTM data, this relationship cannot be

established at the individual pixel scale. Hence, a formula-

tion of this relationship must necessarily be based on mean

estimates of observed canopy (Eq. (2)) and scattering phase

center height (Eq. (7)), and can be expressed in the most

general form as:

h¯ can ¼ f h¯ spc��

ð8Þ

Preliminary inspection of the data suggested that linear

models might be adequate for describing the functional

relationship in Eq. (8). Thus, the following model was tested

for both sites:

h¯ can ¼ b0 þ b1h¯spc ð9Þ

where b0=model intercept; b1=model slope.

Based on the relationship in Eq. (9), a least squares

regression analysis was performed on the data from each test

site to determine the model parameters and associated

coefficients of determination. To quantify the deviation

between observed and predicted mean canopy height values,

the root mean square error was computed as:

rmse ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXNi¼1

�h¯ obscan � h¯ predcan

�2

N

vuuuutð10Þ

where rmse=root mean square error; hcanobs=observed mean

canopy height; hcanpred=predicted mean canopy height; N=

number of observations.

7. Results

7.1. Test site Jesup, Georgia

A summary of biophysical parameters for the 22 Jesup

stands is included in Table 2. Included in this table are the

hspc and observed hcan values used in the regression

analyses. Fig. 9 shows the hspc values plotted against the

observed hcan values for the 22 stands. The linear regression

that included all 22 stands in the analysis resulted in an

adjusted r2 value of 0.15. The weak linear relationship

observed was largely due to the presence of several outlier

stands. Two possible explanations for the outliers can be

offered. First, several stands have a relatively small area and

therefore too few samples were available for averaging to

reduce the phase noise component present in the SRTM

data. Three of the outlier stands, indicated by open triangles

in Fig. 9, have fewer than 20 pixels (i.e., b1.8 ha) within

them and significantly fewer pixels than the remaining

stands (31–574). Removal of the three stands resulted in an

improved adjusted r2 value of 0.42 with an rmse of 1.8 m

(Table 4). Second, four of the remaining 19 stands exhibit

Table 2

Summary statistics for 22 stands surveyed within the test site Jesup, Georgia

Stand ID Age Strata Number of

surveyed

plots

Plot size

[m2]

Number

of SRTM

pixels

Mean canopy

height [m]

(hcan)

S.D.

height

[m]

Basal

area

[m2/ha]

Trees per

hectare

SRTM–NED

height [m]

(hspc)

S.D.

SRTM–NED

height [m]

195766 18 Unthinned 24 81 574 17.1 2.0 21.5 788 9.2 2.4

224800 15 Unthinned 29 81 447 16.4 1.5 19.1 1114 7.9 3.0

224305 20 Unthinned 24 81 377 20.1 2.1 21.5 592 10.1 2.4

224542 25 Thinned 25 202 327 15.6 3.2 15.4 458 8.5 2.9

224472 15 Unthinned 30 81 258 14.0 2.2 11.6 724 6.0 2.0

195835 32 Unthinned 24 81 247 21.3 2.8 31.0 766 12.8 2.6

224796 22 Thinned 11 202 236 12.7 2.6 8.9 464 5.8 2.3

224353 14 Unthinned 53 81 210 10.8 2.0 9.1 899 4.7 2.1

224368 18 Unthinned 24 81 187 18.3 1.7 24.0 1083 9.0 2.0

195816 18 Unthinned 22 81 127 18.9 1.5 22.4 675 9.1 2.9

224278 20 Unthinned 28 81 82 15.8 2.6 15.5 932 8.9 2.9

224303 20 Unthinned 27 81 72 18.0 4.4 25.5 692 10.8 2.0

224224 18 Unthinned 24 81 37 16.5 2.0 20.4 762 7.1 2.2

224799 18 Unthinned 24 81 36 16.9 1.8 24.4 997 10.4 2.2

224784 18 Unthinned 28 81 31 16.1 1.5 26.3 1217 8.3 2.7

224302 21 Thinned 31 202 165 19.8 5.0 15.1 333 7.5 2.3

224401 22 Unthinned 43 81 113 18.0 5.5 14.2 386 9.6 2.5

224550 20 Unthinned 36 81 59 16.7 6.0 30.3 700 13.0 2.6

224546 26 Thinned 41 202 128 15.7 7.5 19.9 426 10.6 2.4

224335 17 Unthinned 25 81 13 10.4 2.0 23.5 1034 8.3 1.9

224304 20 Unthinned 26 81 6 12.0 2.5 21.1 556 8.8 2.1

224547 26 Thinned 41 202 2 15.7 7.5 19.9 426 16.2 2.2


relatively large standard deviations in observed height (on

the order of 5 m and above), which points to the presence of

a dissimilar stand structure characterized by a more

heterogeneous canopy. Whereas this heterogeneity should

also be reflected in the corresponding standard deviations of

the SRTM hspc values, it is not (Table 2). This is due to the

fact that the variability in observed height within the four

stands exists at a scale less than the resolution of the SRTM

data. Visual inspection of the stands both on the ground and

Fig. 9. Plot of SRTM scattering phase center height (hspc) versus observed stand he

key differences in stand variables including stand size [e.g., triangles indicate sma

indicate stands with high height variability (S.D.N5)].

in high-resolution imagery confirms that these stands exhibit

a very heterogeneous structure including windfalls and

small clearings, which contribute to the large variance in

observed height. Thus, these stands were deemed not

representative of the remaining sample and would have to

be treated with an alternative approach (e.g., elimination of

the windfall/clear areas from both the ground and image

sample). Since the ground survey did not record exact plot

locations, this was not possible retrospectively. Removal of

ight (hcan) for all stands (22) in the Jesup, Georgia, test site. Symbols reflect

ll stands with few (b20) pixels per stand] and stand structure [e.g., squares

Fig. 10. Plot of SRTM scattering phase center height (hspc) versus observed stand height (hcan) for 15 stands in the Jesup, Georgia, test site.


the four stands resulted in a significant improvement in the

correlation between SRTM hspc and hcan. The adjusted r2

value of the model with seven outlier stands removed was

0.79 with an rmse of 1.1 m (Fig. 10).

Under the assumption that additional sample averaging

could further reduce the influence of phase noise on model

performance, a minimum threshold of 50 pixels per stand was

explored since (1) according to Fig. 8, the noise range is

reduced significantly for stands of this size, and (2) a sufficient

number of stands (i.e., 12) remained available with which to

develop the regression model (Table 2). The resulting adjusted

r2 value was 0.86 with an rmse of 1.0 m (Fig. 10).

Table 3

Summary statistics for 20 bsimulated standsQ (i.e., bins) based on data from 227

Bin

number

Bin range

[m]

Number

of plots

per bin

Plot

size

[m2]

Number

of SRTM

pixels per bin

Mean canop

height [m]

(hcan)

1 b5.00 3 707 26 3.5

2 5.00–6.49 3 707 25 5.9

3 6.50–7.99 3 707 26 7.4

4 8.00–9.49 11 707 85 8.8

5 9.50–10.99 29 707 253 10.2

6 11.00–12.49 24 707 214 11.8

7 12.50–13.99 20 707 173 13.3

8 14.00–15.49 22 707 191 14.7

9 15.50–16.99 19 707 158 16.1

10 17.00–18.49 21 707 184 17.8

11 18.50–19.99 10 707 87 19.2

12 20.00–21.49 16 707 139 20.9

13 21.50–22.99 7 707 56 22.0

14 23.00–24.49 13 707 114 23.8

15 24.50–25.99 5 707 44 24.9

16 26.00–27.49 4 707 35 27.1

17 27.50–28.99 6 707 47 28.5

18 29.00–30.49 4 707 35 29.4

19 30.50–33.99 4 707 36 33.0

20 34.00–38.00 3 707 24 37.3

The parameters of the linear regression models associated

with the 20- and 50-pixel thresholds indicate that the height

of the scattering phase center increases moderately

(slope=1.2–1.3) with increasing canopy height and lies

approximately 6+ m (intercept=5.8–6.6) below the canopy

surface.

7.2. Test site Sierra Nevada, California

An overview of biophysical parameters, including hspcand observed hcan values, corresponding to the 20 simulated

Sierra Nevada stands (i.e., bins) is provided in Table 3. Prior

plots surveyed within the test site Sierra Nevada, CA

y S.D.

height

[m]

Basal area

per bin

[m2/ha]

Number

of trees

per bin

SRTM–NED

Height [m]

(hspc)

S.D.

SRTM–NED

height [m]

0.9 0.2 17 3.3 3.4

2.8 2.6 91 3.7 2.2

4.0 3.7 122 8.7 1.8

3.9 10.3 220 7.5 1.8

5.0 47.2 990 13.0 2.9

6.3 57.0 805 12.2 2.8

7.5 55.5 663 14.1 2.6

7.5 68.2 688 16.1 2.6

9.3 63.0 483 12.8 2.7

8.1 80.2 684 18.1 2.9

10.5 46.0 289 18.3 1.9

10.3 60.5 331 19.5 3.0

10.4 25.8 172 20.8 2.4

11.1 59.7 321 21.3 3.4

11.0 18.1 78 19.7 4.0

8.2 20.2 100 18.6 2.4

10.7 33.7 139 19.9 3.0

12.4 20.0 69 13.5 3.0

12.8 18.3 52 22.1 2.9

10.1 15.1 38 26.3 2.8

Fig. 11. Plot of SRTM scattering phase center height (hspc) versus observed plot height (hcan) for all plots (227) in the Sierra Nevada, California, test site.


to binning, a simple linear regression analysis was con-

ducted with observed hcan regressed on hspc (Fig. 11). Data

from all 227 plots were included in model development. The

result was an adjusted r2 value of 0.20 with an rmse of 5.8

m (Table 4), indicating the presence of only a slight positive

correlation. This relatively weak relationship was attributed

to the presence of residual phase noise resulting from the

fact that only 6–10 pixels were averaged per plot under the

45-m buffering strategy. This number was observed to be

quite low in light of the findings presented in Fig. 8 as well

as in comparison to the Jesup site where the average stand

size was significantly larger and hence the number of pixels

extracted was markedly greater (i.e., N20 pixels following

outlier removal) . The binning approach described in

Section 5.3 was therefore undertaken in an effort to simulate

stand-level statistics using plot-level data. Under this

strategy, the number of pixels being averaged per observa-

tion was increased from a minimum of 6 to a minimum of

20 (Table 3). Fig. 12 shows the hspc values plotted against

observed hcan values for the resulting 20 height bins. The

regression analysis based on this strategy resulted in a

significantly improved adjusted r2 value of 0.75 with an

rmse of 4.5 m (Fig. 12; Table 4). Based on the Jesup results,

Table 4

Results of simple linear regression analyses performed on the test sites Jesup, Ge

Dataset Observed Adjusted r

Jesup test site: linear least squares regression

Number of pixels N20 19 0.42

Number of pixels N20 and hcan S.D. b5 m 15 0.79

Number of pixels N50 and hcan S.D. b5 m 12 0.86

Sierra Nevada test site: linear least squares regression

No pixels threshold (based on all plots) 227 0.20

Number of pixels N20 (based on all plots) 20 0.75

Number of pixels N50 (based on all plots) 11 0.86

Number of pixels N20 (based on 126 training plots) 20 0.74

an attempt was made to determine whether any remaining

phase noise could be removed. Again, a minimum threshold

of 50 pixels per bin was tested, which resulted in 11 height

bins (Table 3). The least squares regression based on these

data resulted in an adjusted r2 value of 0.86 with an rmse of

1.7 m (Fig. 12; Table 4). Inspection of Fig. 12 reveals the

influence that bins with fewer than 50 samples had on the

20-bin model prior to their removal. In the initial 20-bin

model, the slope and intercept were 1.336 and �1.892,

respectively. The regression based on the remaining eleven

bins resulted in slight decreases in both the slope (1.088)

and intercept values (�0.940).

7.3. Model inversion and validation

In the context of this research, validation efforts were

focused on the bin-based modeling approach developed and

applied as part of the Sierra Nevada pilot study. Validation

was limited to the Sierra Nevada site given (1) the limited

amount of field data available for the Jesup site (22

georeferenced stands) compared to the Sierra Nevada site

(227 georeferenced plots), and (2) a desire to conduct a

formal and independent assessment of the bin-based

orgia, and Sierra Nevada, California

2 F statistic p value Intercept (b0) Slope (b1) rmse [m]

14.30 0.00149 9.785 0.784 1.8

54.93 N0.0001 6.619 1.160 1.1

67.88 N0.0001 5.778 1.261 1.0

58.81 N0.0001 10.883 0.370 5.8

58.15 N0.0001 �1.892 1.336 4.5

63.72 N0.0001 �0.940 1.088 1.7

55.44 N0.0001 �1.775 1.351 4.6

Fig. 12. Plot of SRTM scattering phase center height (hspc) versus observed binned plot height (hcan) for 20 bsimulated standsQ (i.e., bins) in the Sierra Nevada,

California, test site.


modeling strategy employed as part of the Sierra Nevada

analysis.

A validation effort based on independent training and

testing datasets was carried out in order to test the robustness

of the binning strategy summarized in Table 3. Here, a 20-

pixel minimum threshold was established to pool the plot

data into 20 bins (height classes), thereby simulating 20

stands of relatively homogeneous height. To properly

validate the binning approach, sufficient data were needed

within each bin to allow for the 20-pixel minimum threshold

to be maintained in both training and testing datasets. Thus,

Table 5

Summary statistics for bsimulated standsQ (i.e., bins) used in regression model de

Bin

number

Bin range [m] Training (126 plots)

Number

of plots

per bin

Number

of SRTM

pixels per bin

Mean canopy

height [m]

(hcan)

SR

hei

(hs

1 b5.00 3 26 3.5 3.

2 5.00–6.49 3 25 5.9 3.

3 6.50–7.99 3 26 7.4 8.

4 8.00–9.49 6 47 8.8 8.

5 9.50–10.99 14 119 10.3 12.

6 11.00–12.49 12 108 11.9 11.

7 12.50–13.99 10 90 13.3 13.

8 14.00–15.49 11 99 14.6 16.

9 15.50–16.99 10 85 16.1 12.

10 17.00–18.49 10 88 17.8 18.

11 18.50–19.99 5 45 19.0 18.

12 20.00–21.49 8 68 20.8 20.

13 21.50–22.99 4 32 22.2 19.

14 23.00–24.49 6 51 23.8 20.

15 24.50–25.99 3 27 24.9 20.

16 26.00–27.49 4 35 27.1 18.

17 27.50–28.99 3 23 28.6 17.

18 29.00–30.49 4 35 29.4 13.

19 30.50–33.99 4 36 33.0 22.

20 34.00–38.00 3 24 37.3 26.

for all bins containing 40 (i.e., 2�20) pixels or more (see

Table 3), half of the sample plots (102 plots) and associated

pixels were randomly selected for model development (i.e.,

training), while the remaining half (101 plots) were reserved

for inversion and validation (i.e., testing) (Table 5). Bins

containing fewer than 40 pixels (24 additional plots) were

used exclusively for model development. Thus, whereas the

model was trained on the entire range of height classes (0–38

m) represented by the 20 bins (126 plots), validation was

conducted on height classes ranging from 8 to 26 m and from

27.5 to 29 m represented by 13 bins (101 plots) (Table 5).

velopment (20 training bins) and validation (13 testing bins)

Testing (101 plots)

TM–NED

ght [m]

pc)

Number

of plots

per bin

Number

of SRTM

pixels per bin

Mean canopy

height [m]

(hcan)

SRTM–NED

height [m]

(hspc)

3

7

7

3 5 38 8.7 6.5

0 15 134 10.1 13.9

4 12 106 11.8 13.1

5 10 83 13.3 14.8

2 11 92 14.7 16.0

6 9 73 16.1 12.9

2 11 96 17.8 17.9

6 5 42 19.3 17.9

0 8 71 21.0 19.0

9 3 24 21.9 21.9

1 7 63 23.8 22.3

2 2 17 24.8 18.9

6

2 3 24 28.3 22.5

5

1

3


The regression analysis applied to the training dataset

resulted in an adjusted r2 value of 0.74 with an rmse of 4.6 m

(Table 3). These results are consistent with the adjusted r2

and rmse values (0.75, 4.5 m) reported for the original 20-

bin model developed using data from all 227 plots.

Furthermore, the slope (b0) and intercept (b1) parameters

associated with the original 20-bin model (b0=1.336,

b1=�1.892) and the 20-bin training model (b0=1.351,

b1=�1.775) are virtually identical, indicating that the

original model parameters were appropriate. When the

training model is inverted to predict observed canopy height

values for the independent 13-bin testing dataset, an rmse of

4.1 m is achieved (Table 4), a result consistent with those

reported above for both the original 20-bin dataset (rmse=4.5

m) and the 20-bin validation dataset (rmse=4.6 m).

8. Discussion

In this study, two test sites representing a range of

vegetation structure, moisture, and topographic character-

istics were selected for use in determining the feasibility of

deriving vegetation canopy height from SRTM digital

elevation data. Whereas the Jesup test site is dominated

by heavily managed, relatively homogeneous plantations of

slash pine on moderately rolling terrain, the Sierra Nevada

test site is characterized by less heavily managed, relatively

heterogeneous tracts of coniferous forest of variable

composition and structure on highly variable terrain. All

else being equal, these structural-, topographic-, and

associated moisture-related differences have a direct and

differential influence on the height of the scattering phase

center (i.e., the vertical position within the canopy from

where the majority of backscattered electromagnetic energy

is being returned). As expected, different empirical models

are needed to describe the differing relationships that exist

between the two test sites with respect to observed and

remotely sensed estimates of vegetation canopy height.

Preliminary inspection of the data indicated that simple

univariate linear models would be sufficient for describing

these relationships. It follows that efforts to extend this

research to regional, continental, or global scales will

necessarily require development of a more comprehensive

region-specific and/or multivariate approach. To be most

effective, such an approach would be conducted in the

context of an ecoregional framework where regional

ecological drivers of vegetation composition and structure

including climate and topography can be appropriately

controlled. Within individual ecoregions, remotely sensed

descriptors of horizontal structure (e.g., canopy density,

cover type, etc.) can then be used together with SRTM

scattering phase center height information as part of a

multivariate and structurally hierarchical approach to

regression model development.

Whatever the spatial scale at which the SRTM data are

used, based on the research presented here, it is clear that

accurate estimates of vegetation canopy height cannot be

expected at the level of individual SRTM resolution cells

(c900 m2). Due to the relative error component resulting

largely from phase noise, sample averaging is a critical

prerequisite of any effort to derive canopy height informa-

tion from the SRTM data. The results from the Jesup and

Sierra Nevada test sites suggest that 20 pixels is perhaps the

minimum acceptable threshold for sample averaging (Figs.

10 and 12). The results of the Iowa and North Dakota field

studies support this finding but also indicate that a higher

averaging threshold may be warranted depending on what

level of residual relative error is deemed acceptable (Fig. 8).

Regardless of the final threshold selected, thought needs to

be given to the specific method used to aggregate pixels

prior to averaging. The stand-based approach applied in the

Jesup case is generally ideal assuming that (1) survey data

are collected on a per-stand basis; (2) stand boundaries are

available for use in the analysis, and (3) stands exhibit

relatively homogeneous structural characteristics. Since in

the general case one or more of these conditions are likely

not to be met, an alternative averaging strategy with more

universal application is desirable. Future work in this regard

will focus on the use of knowledge-based segmentation,

which provides for the automatic and optimal delineation of

local homogeneous regions reflecting a partitioning of the

landscape into units of similar height and overall structure.

This approach is superior to kernel-based methods (i.e., the

simple buffer employed in the Sierra Nevada case) for

several reasons. First, image segments can be highly

adaptive to the irregular shapes of homogeneous regions

(e.g., bands of riparian forest). Second, ancillary information

including other high-resolution remote sensing imagery

such as Landsat-derived vegetation indices or land cover

information can be incorporated in the retrieval of image

segments in order to refine the process of boundary

delineation. National-scale data of this sort will soon be

available in the United States through the ongoing release of

the National Land Cover database 2001 (Homer et al.,

2004).

Despite the presence of phase noise and other contrib-

utors to the overall relative and absolute height errors (see

Eqs. (3) and (4)), efforts undertaken as part of this research

to both understand and minimize errors sources have been

quite successful. In most cases, the errors themselves are

relatively minor. Some of these errors such as those related

to soil moisture and roughness are nearly impossible to

quantify given the complex nature of SRTM data acquisition

and processing (e.g., averaging of multiple data takes). As a

result, and given their minimal influence on the height

estimation process, no effort has been made to account for

them explicitly. On the other hand, more significant sources

of error such as those attributed to phase noise and vertical

offsets are reasonably well understood and can be mini-

mized, if not removed, using the techniques (e.g., sample

averaging) described here. Future work in the arena of error

characterization will focus on development of methods for


minimizing this suite of errors across broader geographic

extents. For example, regional-scale efforts to estimate

vegetation canopy height require techniques for spatially

modeling the absolute error associated with vertical offsets

or trends between the SRTM and NED datasets. Results

from the Iowa/North Dakota investigation of phase noise

indicated a difference in vertical offset amounting to almost

3 m. In the context of regional-scale analyses, spatially

varying vertical trends such as that observed in the Great

Plains can be identified using trend analysis together with

SRTM–NED difference data from vegetation-free areas

(e.g., snow-covered fields, pastures, meadows, etc.). The

trend can then be modeled using geostatistical techniques

such as kriging and accurately removed from the regional

estimates of scattering phase center height.

A number of additional research areas are currently being

explored in the context of efforts to extend SRTM-derived

canopy height estimation to larger scales. Perhaps most

important among these is the need to acquire accurate and

consistent field survey information for use in larger-scale

applications of the SRTM data. One potential source of such

data is the Forest Inventory and Analysis (FIA) program of

the USDA Forest Service (Gillespie, 1999). This federal

program represents the only nationwide source of timely

and reliable forest inventory and monitoring information.

The inventory consists of some 150,000 permanent plots

nationwide; 20% of these are surveyed annually. Generally

speaking, the FIA sampling design is quite similar to that

employed as part of the Sierra Nevada field survey. The

encouraging results obtained from the Sierra Nevada test

site suggest that FIA survey data may be useful for scaling

the SRTM-based height estimation process up to the

conterminous United States and beyond.

A second key area of research involves developing a

more complete physical understanding of the interferometric

scattering phase center height and determining what this

measure best reflects in the context of traditional field

metrics for describing vegetation canopy height. In this

study, the functional relationship between hspc and the

observed mean canopy height was explored exclusively.

However, it is possible that other alternative height metrics

(e.g., median, maximum, or dominant height) may in fact be

more highly correlated with the hspc. In areas like the Jesup

test site, dominated almost exclusively by even-aged pine

plantations, variability in canopy height is very low,

resulting in a situation where very little difference exists

between values of the mean, median, maximum, or

dominant height. In this case, the mean represents an

entirely appropriate measure of the observed height.

However, in regions like the Sierra Nevada where variability

in vertical structure tends to be significant, mean canopy

height is a very different measure from that of maximum or

dominant height. Thus, in the presence of heterogeneous

structural conditions, it may be that hspc is a much better

predictor of one of these alternative height metrics,

especially in stands of moderate to high density.

9. Conclusions

The research presented here demonstrates that elevation

data from the NASA JPL SRTM holds considerable potential

for developing statistically significant inversion models for

deriving estimates of vegetation canopy height. Various

relative and absolute SRTM error sources were identified as

part of this work and broadly applicable methods have been

proposed for removing or significantly reducing these errors.

Most important among these relates to the minimization of

the SRTM-inherent phase noise. Using elevation profiles and

outlier analysis together with an analysis of noise statistics in

vegetation-free areas, it was established that averaging a

minimum of 20 SRTM pixels yields stable estimates of the

mean scattering phase center height. This corresponds to a

minimum SRTM mapping unit for SRTM-based vegetation

canopy height estimation of approximately 1.8 ha.

The possibility of expanding this work to the scale of the

conterminous United States is now being explored. Com-

plementary field and ancillary data sources including the

national FIA database and the soon-to-be released NLCD

2001 represent unprecedented sources of information for

enabling the production of the first-ever nationwide map of

vegetation canopy height. Such a map would undoubtedly

prove invaluable for a wide range of applications including

biomass estimation and carbon accounting as well as timber

management and habitat characterization.

In summary, SRTM elevation data exhibit a significant

sensitivity to the vertical structure of vegetation. Although

part of the user community views vegetation as amajor source

of noise and nuisance in what would otherwise be highly

accurate SRTM-derived bald-Earth elevation data, it is shown

that for the purposes of vegetation height estimation, the

SRTM dataset is an invaluable resource of near-global scope.

Acknowledgements

The collection of ground data within the Jesup test site

was supported by the NASA EOCAP-SAR program and the

Plum Creek Timber. Analysis of the Jesup data was

supported by a NASA Solid Earth grant titled bVegetationCorrected SRTM-Derived Digital Elevation Models.Q Datacollection within the Sierra Nevada test site was supported

by a NASA Earth Science grant titled bForest Structure fromMultispectral Fusion.Q The analysis was supported, in part,

by the aforementioned grant as well as a NASA Earth

System Science Fellowship awarded to the second author.

References

Bamler, R. (1999). The SRTM Mission: A world-wide 30 m resolution

DEM from SAR Interferometry in 11 days. In D. Fritch, & R. Spiller

(Eds.), Photogrammetric week ’99 (pp. 145–154). Heidelberg7

Wichmann.


Bergen, K. M., & Dobson, M. C. (1999). Integration of remotely sensed

radar imagery in modelling and mapping of forest biomass and net

primary production. Ecological Modelling, 122, 257–274.

Brown, C. G. (2003). Tree height estimation using Shuttle Radar

Topography Mission and ancillary data. PhD Thesis, Department of

Electrical Engineering and Computer Science, The University of

Michigan, Ann Arbor, MI.

Curkendall, D., Fielding, E., Cheng, T., & Pohl, J. (2003). A computational-

grid based system for continental drainage network extraction using

SRTM digital elevation models. IEEE ICPP Workshops, 181–190.

Dobson, M. C. (2000). Forest information from Synthetic Aperture Radar.

Journal of Forestry, 98, 41–43.

Dobson, M. C., Ulaby, F. T., Pierce, L. E., Sharik, T. L., Bergen, K. M.,

Kellndorfer, J. M., et al. (1995). Estimation of forest biophysical

characteristics in Northern Michigan with SIR-C/X-SAR. IEEE Trans-

actions on Geoscience and Remote Sensing, 33, 877–895.

Dubayah, R. O., & Drake, J. B. (2003). Lidar remote sensing for forestry.

Journal of Forestry, 98, 44–46.

Gesch, D., Oimoen, M., Greenlee, S., Nelson, C., Steuck, M., & Tyler, D.

(2002). The National Elevation Dataset. Photogrammetric Engineering

and Remote Sensing, 68, 5–11.

Gillespie, A. (1999). Rationale for a National Annual Forest Inventory

Program. Journal of Forestry, 97, 16–20.

Hagberg, J. O., Ulander, L. M. H., & Askne, J. (1995). Repeat-pass SAR

interferometry over forested terrain. IEEE Transactions on Geoscience


Hensley, S., Rosen, P., & Gurrola, E. (2000). Topographic map generation

for the Shuttle Radar Topography Mission C-band SCANSAR

interferometry. Proceedings of SPIE, 4152, 179–189.

Homer, C., Huang, C., Yang, L., Wylie, B., & Coan, M. (2004).

Development of a 2001 National Landcover Database for the United

States. Photogrammetric Engineering and Remote Sensing, 70,

829–840.

Hyypp7, J., & Hallikainen, M. (1996). Applicability of airborne profiling

radar to forest inventory. Remote Sensing of Environment, 57, 39–57.

Kasischke, E. S., Melack, J. M., & Dobson, M. C. (1997). The use of

imaging radars for ecological applications—A review. Remote Sensing

of Environment, 59, 141–156.

Kellndorfer, J. M., Dobson, M. C., Pierce, L. E., & Ulaby, F. T. (1998).

Toward consistent regional- to-global-scale vegetation characterization

using orbital SAR systems. IEEE Transactions on Geoscience and

Remote Sensing, 36, 1396–1411.

Kellndorfer, J.M, Dobson, M. C., Vona, J., & Clutter, M. (2003). Towards

precision forestry: Plot-level parameter retrieval for slash pine

plantations with JPL AIRSAR. IEEE Transactions on Geoscience and

Remote Sensing, 41, 1571–1582.

Kellndorfer, J. M., & Ulaby, F. T. (2003). Forest biomass inversion from

SAR using object oriented image analysis techniques. IEEE Geoscience

and Remote Sensing Symposium, Toulouse, 6, 3465–3467.

Kobayashi, Y., Sarabandi, K., & Dobson, M. C. (2000). An evaluation of

the JPL TOPSAR for extracting tree heights. IEEE Transactions on

Geoscience and Remote Sensing, 38, 2446–2454.

Lefsky, M. A., Harding, D., Cohen, W. B., Parker, G., & Shugart, H. H.

(1999). Surface Lidar remote sensing of basal area and biomass in

deciduous forests of Eastern Maryland, USA. Remote Sensing of

Environment, 67, 83–98.

Martinez, J. M., Floury, N., Le Toan, T., Beaudoin, A., Hallikainen, M., &

Makynen, M. (2000). Measurements and modeling of vertical back-

scatter distribution in forest canopy. IEEE Transactions on Geoscience


Papathanassiou, K. P., & Cloude, S. K. (2001). Single-baseline polarimetric

SAR interferometry. IEEE Transactions on Geoscience and Remote

Sensing, 39, 2352–2363.

Pienaar, L. V., Shiver, B. D., & Rheney, J. W. (1993). Yield prediction for

mechanically site-prepared slash pine plantations in the southeastern

Coastal Plain. Southern Journal of Applied Forestry, 17, 163–173.

Rignot, E., Echelmeyer, K., & Krabill, W. B. (2001). Penetration depth of

Interferometric Synthetic Aperture Radar signals in snow and ice.

Geophysical Research Letters, 28, 3501–3504.

Rosen, P., Eineder, M., Rabus, B., Gurrola, E., Hensley, S., Knoepfle, W.,

et al. (2001a). SRTM mission cross comparison of X and C band data

properties. IEEE Geoscience and Remote Sensing Symposium, 2,

751–753.

Rosen, P., Hensley, S., Gurrola, E., Rogez, F., Chan, S., Martin, J., et al.

(2001b). SRTM C-band topographic data: Quality assessments and

calibration activities. IEEE Geoscience and Remote Sensing Sympo-

sium, 2, 739–741.

Rosen, P. A., Hensley, S., Joughin, I. R., Li, F. K., Madsen, S. N.,

Rodriguez, E., et al. (2000). Synthetic Aperture Radar interferometry.

Proceedings of the IEEE, 88, 333–382.

Sarabandi, K., & Lin, Y. C. (2000). Simulation of interferometric SAR

response for characterizing the scattering phase center statistics of forest

canopies. IEEE Transactions on Geoscience and Remote Sensing, 39,

115–125.

Smith, B., & Sandwell, D. (2003). Accuracy and resolution of Shuttle

Radar Topography Mission data. Geophysical Research Letters, 30,

1467–1470.

Sun, G., Ranson, K. J., Kharuk, V. I., & Kovacs, K. (2003). Validation of

surface height from Shuttle Radar Topography Mission using shuttle

laser altimeter. Remote Sensing Environment, 88, 401–411.

Treuhaft, R. N., Madsen, S. N., Moghaddam, M., & van Zyl, J. J. (1995).

Vegetation characteristics and underlying topography from interfero-

metric radar. Radio Science, 31, 1449–1485.

Treuhaft, R. N., & Siqueira, P. R. (2000). Vertical structure of vegetated

land surfaces from interferometric and polarimetric radar. Radio

Science, 35, 141–177.

Ulaby, F. T., Dobson, M. C., Pierce, L. E., Kellndorfer, J. M., & Sarabandi,

K. (1995). Scene classification and biomass estimation using SIR-C/X-

SAR. Proceedings of the international symposium on the retrieval of

biophysical and geophysical parameters from SAR data for land

applications (pp. 53–57). Toulouse7 CNES/IEEE.

United States Geological Survey (2003). Shuttle Radar Topography

Mission documentation: SRTM Topo. http://edcftp.cr.usgs.gov/pub/

data/srtm/Documentation/SRTM_Topo.txt

Vogelmann, J. E., Howard, S. M., Yang, L., Larson, C. R., Wylie, B. K., &

van Driel, J. N. (2001). Completion of the 1990’s National Land Cover

Data Set for the conterminous United States. Photogrammetric

Engineering and Remote Sensing, 67, 650–662.

Vona, J. (2001). Modeling slash pine (Pinus elloitti) parameters using

synthetic aperture radar. MS thesis, The University of Georgia, Athens.

http:\\edcftp.cr.usgs.gov\pub\data\srtm\Documentation\SRTM_Topo.txt

Vegetation height estimation from Shuttle Radar Topography Mission and National Elevation Datasets

Documents

Vegetation height estimation from Shuttle Radar Topography Mission and National Elevation Datasets