Vegetation Analysis Using Remote Sensing

Argentinian Spatial Agency (CONAE)
Cordoba National University (UNC)
Vegetation analysis using remote sensing
Master in Emergency Early Warning and Response Space Applications.

Mario Gulich Institute, CONAE. Argentina
Author: Biol. Lanfri Sofı́a
September, 2010
Cordoba, Argentina
2
Contents
Abstract i
1 Introduction 1
2 Vegetation analysis by remote sensing 3

2.0.1 Spectral signatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1 Vegetation analysis from optical data (visible and near-infrared reflectance) . 10
2.1.1 Vegetation indices summary . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.2 Normalized Difference Vegetation Index . . . . . . . . . . . . . . . . . 20
2.1.3 Vegetation analysis from hyperspectral sensors . . . . . . . . . . . . . 22
2.2 Vegetation analysis from microwave data . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Passive microwave data . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.2 Active microwave data . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.3 Subsection: Vegetation estimation techniques using SAR data . . . . . 27
3 Conclude remarks 35
3
4 CONTENTS
List of Figures
2.1 Generalized spectral signatures for some common cover types (dry bare soil,
vegetation and water). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Absorption characteristics of different pigments in the visible range. . . . . . 6
2.3 Characteristic absorption of different pigments in the visible range . . . . . . 6
2.4 Typical reflectance behavior of the vegetation components. . . . . . . . . . . 7
2.5 Idealized reflectance patterns for soil and herbaceous vegetation. . . . . . . . 8
2.6 Spectral signatures of different types of soil coverage . . . . . . . . . . . . . . 9
2.7 Spectral reflectance of vegetation at different phisiological conditions. . . . . . 10
2.8 An infrared/red ratio plot showing the distribution, in spectral space of com-
ponents of a rangeland scene . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.9 Vegetation analysis from optical data (visible and NIR reflectance). . . . . . . 11
2.10 Relationship between N DV I and total dry biomass according to the work of
Tucker et al. (1985). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.11 Vegetation analysis from microwave data. . . . . . . . . . . . . . . . . . . . . 23
2.12 Wavelength regions of the electromagnetic spectrum corresponding to mi-
crowave data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.13 April monthly mean values for NDVI in (A), A parameters using the low
frequency pair in (B), and the high frequency pair in (C), B parameters using
the low frequency pair in (D), and and the high frequency pair in (E). . . . . 26
2.14 Tipical scatterers contribution of forest structure. . . . . . . . . . . . . . . . . 28
2.15 Multitemporal surface responses at different polarizations of ENVISAT SAR
sensors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.16 Vegetation class set table and Tubuai SAR vegetation classification. . . . . . 31
2.17 The first order (no multiple scattering) dominant scattering contributions over
forest canopies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5
6 LIST OF FIGURES
Abstract
Monitoring vegetation is important to the entire environment because of the carbon stored in
their biomass and the potential for carbon exchange between this one and the atmosphere. In
the same way, understanding land-cover conversion and dynamics is one of the key research
topics because changes in land surface affect water, energy, nutrient, and carbon cycle in
a region and its surrounding ecosystems. Satellite imagery provides the large spatial and
temporal scales necessary to address vegetation analysis in an appropriate way. Numerous
remote sensing studies have been involved in the mapping and analysis of relative vegetation
cover at various scales, utilizing indices based on optical spectral behavior of vegetation.
Although optical remote sensing can be considered a relatively powerful mapping tool, there
are recognized limitations of these data due to cloud interference, atmospheric attenuation
and some constraints in its use for vegetation discrimination. An interesting alternative
to optical sensors is the microwave spectral region where Synthetic Aperture Radar (SAR)
systems operate. This work is a review of published information about diverse methodologies
for vegetation characterization by satellite observations. It presents a comparison of the
relationship between both optical and microwave spectral sensors (from visible to microwave
wavelengths) with vegetation study. There are exposed some examples about the utilization
of different remote sensing perspectives for this scope.
i
ii ABSTRACT
Chapter 1
Introduction
The study of land vegetation coverage, in any morphological or structural type, is widely
recognized. Vegetation analysis is important for diverse applications such as food produc-
tion study (Pieper, 1988[128]; Kennedy, 1989 [85]), land degradation (Pickup, 1996 [126];
Imeson and Lavee, 1998 [71]) and ecosystems recovery (Shoshany, 2001[156]). Moreover,
the study of vegetation coverage has great importance for identifying the effects of climate
variability, natural phenomena and anthropogenic factors on the environment (Tueller, 1987
[173]; Woodwell et al., 1984 [181]).
Monitoring vegetation is important to the entire environment because of the carbon stored in
their biomass and the potential for carbon exchange between this one and the atmosphere
(Running et al., 1995[145]; Luckman et al., 1998)[97]. Therefore, vegetation properties
are key elements in the study of the global carbon cycle and ecosystems. In this sense,
monitoring global vegetation properties from space can improve our understanding of land
surface processes and their interactions with the atmosphere, biogeochemical cycle, and
primary productivity.
In the same way, understanding land-cover conversion and dynamics is one of the key research
topics because changes in the land surface affect water, energy, nutrient, and carbon cycle
in a region and its surrounding ecosystems (Roberts et al., 2003[139]). In consequence,
the importance of vegetation in studies of global climate and biogeochemical cycles is well
recognized (Sellers and Schimel, 1993[152]; DeFries et al., 1995 [32]).
Traditional sampling techniques for estimating vegetation conditions, based on field collec-
tion data, are time-consuming, costly, and not generally applicable to inaccessible regions.
An alternative approach is remote sensing - the science and art of obtaining information
about an object, area, or phenomenon through the analysis of data acquired by a device
that is not in contact with the object, area, or phenomenon under investigation (Lillesand
and Kiefer, 1979[95]). Land surface parameterizations have traditionally derived from land
cover data sets (in situ surveys) (Matthews, 1983[101]) or from vegetation indices calculated
from remote sensing data (Sellers et al., 1994[153], 1996[151]). Satellite imagery provides
the large spatial and temporal scales necessary to manage this analysis in an appropriate
way. For instance, physical and physiological parameters of vegetation required for studies
of global climate can be obtained from satellite remote sensing (Myneni et al., 1995[110]).
Numerous remote sensing studies have been involved in the mapping and analysis of vegeta-
tion at various scales, utilizing indices based on optical spectral behavior of vegetation (e.g.
Justice and Hiernaux 1986[80], Tucker et al., 1986[171]; Smith et al., 1990[160]; Shoshany et
al., 1995 [157]). Satellite instruments measure solar radiation reflected by vegetation at any
given wavelength intervals. Of these, the red (ρRED : 0.6 − 0.7 µm) and near-infrared (N IR)
(ρN IR : 0.75 − 1.35 µm) channels have been found to be most relevant in the vegetation
remote sensing (Myneni et al., 1995[110]).
Plants absorb red radiation and scatter N IR radiation resulting in a large difference between
ρN IR and ρRED ; whereas in contrast, for bare soil, ρN IR is similar to ρRED . This difference
1
2 CHAPTER 1. INTRODUCTION
between plant and soil reflectance is often enhanced by computing a ratio of visible and
N IR reflectance. Then, the measured spectral reflectance data are usually compressed
into vegetation indices (V Is ) (Myneni et al., 1995[110]), which the most common V I is
the normalized difference vegetation index (N DV I) (e.g., Rouse et al., 1973[143]; Jackson,
1983[74]; Purevdorj et al., 1998[132]), calculated from the red and N IR channels.
More than a dozen such indices are reported in the literature and shown a good correlation,
for instance, with vegetation amount (Tucker, 1979 [169]), the fraction (f AP AR) of absorbed
Photosyntetically Active Radiation (P AR) (Asrar et al., 1984 [5]), unstressed vegetation,
photosynthetic capacity (Sellers et al., 1992 [150]), and seasonal atmospheric carbon dioxide
variations (Tucker et al., 1986[168]).
Although optical remote sensing can be considered a relatively robust mapping tool, there
are recognized limitations of these data due to cloud interference, atmospheric attenuation
and some constraints in its use for vegetation discrimination. For instance, optical data
are related only to the top millimeters of the canopy and therefore, they present some
restrictions to despite structural differences of the vegetation. An interesting alternative
to optical sensors is the microwave spectral region where Synthetic Aperture Radar (SAR)
systems operate. SAR sensors can penetrate into clouds and smoke, and are independent of
solar elevation and azimuth angles.
Experimental and theoretical investigations have shown that microwave backscattering (σ °),
the amount of energy returned to the radar antenna per unit area (Raney, 1998[134]), from
land surfaces is sensitive to vegetation features (e.g., Le Toan et al., 1989[91], 1997[92];
Ulaby and Dobson, 1989[177]; Ferrazzoli et al., 1992[39], 1997[40]). Because of the SAR’s
canopy penetration capability, significant relations between radar backscatter and vegeta-
tion parameters such as leaf area index (LAI) and biomass from empirical and theoretical
experiments have been published (Ulaby et al., 1984[176]; Paloscia et al., 1999[118]; Inoue
et al., 2002[73]).
The current level of experience in operational use of SAR data is limited compared to the use
of visible and infrared data. In addition, one of the major feebleness of SAR data for land-
cover and land-use mapping and monitoring is the relatively strong soil roughness and soil
moisture effects on the backscattered energy from leaves, stalks, and trunks. Because of such
complexities and uncertainties of radar sensors, there is a tendency to use the SAR-Optical
synergism to take advantage of favorable aspects of both sensors simultaneously (Smith et
al., 1995[159]; Moran et al., 2002[105]). Whereas the optical spectral responses of vegetation
canopies originate from molecular resonances and scattering at the micrometer scale, the
SAR backscatter results from plant morphology, i.e. the geometric and bulk dielectric
properties of the vegetation components. In the same way, depending of the wavelength,
and at difference of optical energy, SAR penetrate well into the canopy. These differences
also support the complementary nature of the two sensor types.
Chapter 2 presents a general approach to the vegetation analysis using remotely sensed data.
The first section 2.0.1 of this chapter is a description about the structure and function of
vegetation and its consequent reflectance properties (i.e. spectral signatures). Section 2.1
is a summary of the most used vegetation indices that is possible to obtain from optical
data, with special mention to Normalized Difference Vegetation Index (N DV I). Section
2.2 describes some characteristics of the analysis of vegetation using microwave data.
Chapter 2
Vegetation analysis by remote

sensing
Analyzing vegetation using remotely sensed data requires knowledge about the structure
and function of vegetation and its reflectance properties. Knowing which region of the
electromagnetic spectrum to consider and understanding the factors that influence the spec-
tral response of the interest features are critical to correctly interpreting the interaction of
electromagnetic radiation with the surface.
In general terms of vegetation analysis by remote sensing, vegetation can be divided into
the follow general categories (following to Asner, 1998[4]):
Plant Foliage
Canopies
Non-Photosynthetic Vegetation
Plant Foliage This category comprises plant foliage, including leaves, needles, and other
green materials that vary widely in both shape and chemical composition. The most
important leaf components that affect their spectral properties are: pigments, water,
carbon and nitrogen.
Canopies Although leaf reflectance properties, controlled by features of pigments, water,

and carbon, play a significant role in reflectance at the canopy level, the amount of
foliage and the architecture of the canopy are also meaningful in determining the
scattering and absorption properties of vegetation canopies. Different ecosystems,
such as forest, grassland, or agricultural field, have different reflectance properties,
even though the properties of individual leaves are usually quite similar. Vegetation
with mostly vertical foliage, such as grass, reflects light differently than foliage with
more horizontally-oriented foliage, seen frequently in trees and tropical forest plants.
The variation in reflectance caused by different canopy structures, much like individual
leaf reflectance, is highly variable with wavelength.
At the canopy level, there are a variety of vegetation properties of interest to scientists,
and many of these have direct effects on canopy reflectance properties. The two most
significant are leaf area index (LAI) and leaf angle distribution (LAD). LAI is the
green leaf area per unit ground area, which represents the total amount of green
vegetation present in the canopy. LAI has the strongest effect on overall canopy
reflectance. Whereas LAD describes the overall variety of directions in which the
leaves are oriented, but is often simplified by specifying the mean leaf angle (MLA)
and making assumptions about the actual distribution. The M LA is the average of
the differences between the angle of each leaf in a canopy and an horizontal line.
3
4 CHAPTER 2. VEGETATION ANALYSIS BY REMOTE SENSING
Non-Photosynthetic Vegetation: NPV Ecosystems can also contain senescent or dead

vegetation. Examples include grasslands, shrublands, savannas, and open woodlands,
which collectively cover over half of the global vegetated land surface. This material
is often called non-photosynthetic vegetation because it could be truly dead or simply
dormant. Included in the NPV category are woody structures in many plants, such as
tree trunks, stems, and branches.
2.0.1 Spectral signatures

For any given material, the amount of solar radiation that is reflected (absorbed, trans-
mitted) varies with wavelength. This important property of matter allows us to separate
distinct cover types based on their response values for a given wavelength. When the re-
sponse (reflectance) characteristics of a certain cover type is plotted against wavelength,
this plot is termed the spectral signature of that cover. The diagram 2.1 illustrates the
spectral signatures for some common cover types.
Figure 2.1: Generalized spectral signatures for some common cover types (dry bare soil,
vegetation and water).
Water
Longer wavelength visible and near infrared radiation is absorbed more by water than shorter
visible wavelengths (Fig. 2.1). Thus water typically looks blue or blue-green due to stronger
reflectance at these shorter wavelengths, and darker if viewed at red or N IR wavelengths.
If there is suspended sediment present in the upper layers of the water body, then this
will allow better reflectivity and a brighter appearance of the water. The apparent color of
the water will show a slight shift to longer wavelengths. Suspended sediment can be easily
confused with shallow (but clear) water, since these two phenomena appear very similar.
Chlorophyll in algae absorbs more of the blue wavelengths and reflects the green, making
the water appear more green in color when algae is present. The topography of the water
surface (rough, smooth, floating materials, etc.) can also lead to complications for water-
related interpretation due to potential problems of specular reflection and other influences
on color and brightness.
Soil
Soils usually exhibit monotonic increases in reflectance throughout visible and N IR regions
(Condit, 1970[24]; Stoner and Baumgardner, 1981[162]; Price, 1990[130]). High soil water
and high organic matter contents generally cause lower reflectance while dry, smooth surfaced
soils tend to be brighter (Daughtry, 2001[29]). Occurrence of specific minerals in soil have
been associated with unique spectral features (e.g., higher red reflectance in the presence of
5
iron oxides). In the SWIR, soil spectra display more features than those observed in shorter
wavelengths but are still dominated by water content, litter, and minerals (Gausman et
al., 1975[50]; Henderson et al., 1992[58]; Daughtry, 2001[29]). The presence of crop residue
causes significant changes in reflectance properties compared to bare soil.
Soils tend to have high reflectance in all bands, however this of course is dependent on factors
such as the color, constituents and specifically the moisture content. Water is a relatively
strong absorber of all wavelengths, particularly those longer than the red part of the visible
spectrum. Therefore, as a soils moisture content increases, the overall reflectance of that soil
tends to decrease. Soils rich in iron oxide reflect proportionally more of the red than other
visible wavelengths and therefore appear red (rust color) to the human eye. A sandy soil
on the other hand tends to appear bright white in imagery because visible wavelengths are
more or less equally reflected, when slightly less blue wavelengths are reflected this results
in a yellow color.
Vegetation
The most important leaf components that affect their spectral properties are:
Pigments
Water
Carbon
Nitrogen
Pigments There are three main categories of leaf pigments in plants: chlorophyll, carotenoids,
and anthocyanins. These pigments serve a variety of purposes, and are critical to the
function and health of vegetation, though the relative concentrations of these pigments
in vegetation can vary significantly. Vegetation with a high concentration of chloro-
phyll is generally very healthy, as chlorophyll is linked to greater light use efficiency or
photosynthetic rates. Conversely, carotenoid and anthocyanin pigments often appear
in higher concentrations in vegetation that is less healthy, typically due to stress or
senescence (dormant or dying vegetation that appears red, yellow, or brown).
Chlorophyll, the most well-known and most important pigment, causes the green color
of healthy plant leaves. It is primarily responsible for photosynthesis, the process
by which plants take up carbon dioxide (CO2 ) from the atmosphere and convert it
into organic forms such as sugar and starch. Chlorophyll concentrations in leaves are
broadly correlated with photosynthetic rates. Chlorophyll-a and -b pigments are the
most closely associated with photosynthesis. In autumn, there is less chlorophyll in
the leaves; therefore there is less absorption and proportionately more reflection of the
red wavelengths, making leaves appear red or yellow.
Carotenoids are a group of pigments containing alpha-carotene, beta-carotene, and
xanthophyll pigments (for example, zeaxanthin). Carotene is the yellow-orange pig-
ment found in tree leaves as they change from green to brown (as seen during autumn).
Carotenoid pigments have multiple functions, but they are generally found in higher
concentrations in plant leaves that are either stressed (by drought or nutrient deple-
tion), senescent, or dead. Carotenoids assist the process of light absorption in plants,
and help protect plants from the harmful effects of very high light conditions.
Anthocyanins also have multiple functions, but are typically related to changes in
foliage. These are pigments abundant in both newly forming leaves and leaves un-
dergoing senescence; and also serve to protect leaves from damage due to ultraviolet
radiation.
As a group, leaf pigments only affect the visible portion of the spectrum (400 nm to
700 nm), and the effects vary depending upon the pigment type. Figs. 2.2 and 2.3
shows the absorption of each pigment type as a function of wavelength throughout the
visible range.
Other components (such as phosphorus, calcium) are significant to plant function,
but they do not directly contribute to the spectral properties of leaves, and therefore
cannot be directly measured using remotely sensed data.
Figure 2.2: Absorption characteristics of different pigments in the visible range.
Figure 2.3: Characteristic absorption of different pigments in the visible range
Water
Water is critical for many plant processes, in particular, photosynthesis. Generally, for
example vegetation of the same type with greater water content is more productive
and less prone to burn.
Plants of different species inherently contain different amounts of water based on their
leaf geometry, canopy architecture, and water requirements. Among plants of one
species, there is still significant variation, depending upon leaf thickness, water avail-
ability, and plant health.
7
Leaf water affects plant reflectance in the N IR and shortwave infrared regions of the
spectrum (Fig. 2.4). Water has maximum absorptions centered near 1400 and 1900
nm, but these spectral regions usually cannot be observed from airborne or space-based
sensors due to atmospheric water absorption, preventing their practical use in the
creation of V Is . Water features centered around 970 nm and 1190 nm are pronounced
and can be readily measured from hyperspectral sensors, whereas are generally not
sampled by multispectral sensors.
Figure 2.4: Typical reflectance behavior of the vegetation components.
Carbon Plants contain carbon in many forms, including sugars, starch, cellulose, and lignin.
Sugars and starch are immediate products of photosynthesis; they are moved to other
locations in plants to construct cellulose and lignin. Cellulose is primarily used in the
construction of cell walls in plant tissues. Lignin is used for the most structurally robust
portions of plants, such as leaf vacuoles, veins, woody tissue, and roots. Cellulose and
lignin display spectral features in the shortwave infrared range of the spectrum (Fig.
2.4).
Nitrogen Leaves contain nitrogen in the chlorophyll pigment, proteins, and other molecules.
Nitrogen concentrations in foliage are linked to maximum photosynthetic rate and net
primary production. V Is sensitive to chlorophyll content (which is approximately 6%
nitrogen) are often broadly sensitive to nitrogen content as well. Some proteins that
contain nitrogen affect the spectral properties of leaves in the 1500 nm to 1720 nm
range.
A distinctive feature in the scattering spectrum of a green leaf is the chlorophyll absorption
maximum at about 0.69 µm. The lack of absorption in the adjacent N IR region (0.85 µm)
results in a strong absorption contrast across the 0.65-0.85 µm wavelength interval (Fig. 2.1
and 2.4).
Absorption centered at about 0.65 µm (visible red) is controlled by chlorophyll pigment in
green-leaf chloroplasts that reside in the outer or palisade leaf, but absorption occurs to a
similar extent in the blue. With these colors thus removed from white light, the predominant
but diminished reflectance of visible wavelengths is concentrated in the green. Thus, most
vegetation has a green-leafy color.
The internal structure of healthy leaves act as excellent diffuse reflectors of N IR wavelengths.
There is strong reflectance between 0.7 and 1.0 µm (near IR) in the spongy mesophyll cells
located in the interior or back of a leaf, within which light reflects mainly at cell wall/air
space interfaces, much of which emerges as strong reflection rays. The intensity of this
reflectance is commonly greater (higher percentage) than from most inorganic materials, so
vegetation appears bright in the N IR wavelengths.
Whereas vegetation strongly reflects light in the N IR and SW IR − 1 range of the spectrum,
it strongly absorbs photons in the visible and short wave infrared (SW IR − 2) ranges. This
results in a much shallower penetration of photons into the canopy in these wavelengths
and therefore, V Is using spectral data from the visible and SW IR − 2 are very sensitive to
upper-canopy conditions.
Respect to N P V , it is composed largely of the carbon-based molecules lignin, cellulose,
nitrogen and starch. As such, it has a similar reflectance signature to these materials and
also can be confused with mineral absorptions (Clark, 1997[22]). In many canopies, much
of the N P V is obscured below a potentially closed leaf canopy; the wavelengths used to
measure N P V (shortwave infrared) are often unable to penetrate through the upper canopy
to interact with this N P V . As such, only exposed N P V has a significant effect on the
spectral reflectance of vegetated ecosystems. When exposed, N P V scatters photons very
efficiently in the shortwave infrared range, in direct contrast to green vegetation which
absorbs strongly in the shortwave infrared range.
Idealized reflectance patterns for herbaceous vegetation and soil are compared in Fig 2.5;
where is possible to see that dead or dormant vegetation has higher reflectance than liv-
ing vegetation in the visible spectrum and lower reflectance in the N IR. Soil has higher
reflectance than green vegetation and lower reflectance than dead vegetation in the visible,
whereas, in the N IR, soil typically has lower reflectance than green and dead vegetation
(Tappan, 1980[164]).
Figure 2.5: Idealized reflectance patterns for soil and herbaceous vegetation.
Spectral signatures of different types of surface coverage as a result of the experimentation

of Lang et al., 2002 [89] are presented in Fig 2.6. Whereas reflectance patterns for healthy
green vegetation, stressed vegetation, and severely stressed vegetation are displayed in Fig
2.7. In the visible region on the electromagnetic spectrum, the three spectral signatures
look similar. However, in the N IR region of the spectrum, the spectral signatures look very
different from each other. In this region, healthy vegetation has the highest reflectance value
(Knipling, 1970[88]; Bauer, 1975 [11]).
Plant stress and/or normal end-of-season senescence typically result in lower chlorophyll
concentrations that allow expression of accessory leaf pigments such as carotenes and xan-
thophylls. This has the effect of moving the green reflectance peak (normally located near
550 nm) towards longer wavelengths, increasing visible reflectance (Adams et al., 1999 [1]),
and causing the tissues to appear chlorotic. At the same time, N IR reflectance decreases
proportionately less than the visible increases. With increasing stress, the abrupt transition
or red edge that is normally seen between visible and N IR in green vegetation begins to
shift towards shorter wavelengths and, in the case of senescent vegetation, may disappear
entirely.
9
(a) (b)
(c) (d)
(e) (f)
Figure 2.6: Spectral signatures of different types of soil coverage

Optical properties of leaves in a third region of the solar spectrum, the middle- or shortwave-
infrared (SW IR, 1300 to 2500 nm), are strongly mediated by water in tissues (Fig 2.4).
Reflectance in this region is relatively high for vigorously growing vegetation but decreases
as tissues dehydrate. However, research suggests such drought-induced decreases in SW IR
reflectance are not sufficiently large over biologically significant changes in plant water con-
tent for the practical use of this wavelength interval in the diagnosis of water stress in the
field (Bowman, 1989[14]; Carter, 1991[16]).
Figure 2.7: Spectral reflectance of vegetation at different phisiological conditions.
If the N IR reflectance of an object (or landscape component) is ploted against its corre-
sponding red reflectance, for example as shown in Fig. 2.8, it is possible to see clusters
representing the typical reflectance response of each kind of object (water, shrub, grass-
land, snow, ice, soil). It is important to mention that in this plot type (known as ‘Tasseled
Cap’) the soil determine a line termed ‘the soil line’. The objects located in areas of the
graph corresponding to high N IR reflectance and low red reflectance, would represent ‘dense
vegetation coverage’.
2.1 Vegetation analysis from optical data (visible and

near-infrared reflectance)
The location of optical data (visible and N IR reflectance) in the electromagnetic spectrum
is shown in Fig. 2.9.
2.1.1 Vegetation indices summary
Investigators have developed techniques for qualitatively and quantitatively assessing the
vegetation from spectral measurements. Vegetation indices (V Is ) take advantage of vege-
tation’s reflective contrast (as mentioned before) between the N IR and visible red (V IS)
wavelengths (Tucker, 1979[169]), sometimes with additional channels included; these are one
of the most widely used remote sensing measurements (Cracknell, 2001[25]). Generally, are
analyzed the reflectance at 660 nm in the red range of the spectrum and the reflectance
at 870 nm in the N IR range. At higher vegetation vigor the more high contrast between
reflectance values in these regions.
V Is to monitor terrestrial landscapes by satellite sensors were first developed in the 1970s
and have been highly successful in assessing vegetation condition, foliage cover, phenology,
and processes such as evapotranspiration (ET ) and primary productivity. They are highly
2.1. VEGETATION ANALYSIS FROM OPTICAL DATA (VISIBLE AND NEAR-INFRARED REFLECTANCE)11
Figure 2.8: An infrared/red ratio plot showing the distribution, in spectral space of compo-
nents of a rangeland scene
Figure 2.9: Vegetation analysis from optical data (visible and NIR reflectance).
correlated with the photosynthetically active biomass, chlorophyll abundance, and energy
absorption (reviewed in Myneni et al., 1995[110]). Then, these indices enable, for instance,
the estimation of biomass percentage cover, absorbed P AR and LAI.
Most formulas of vegetation indices are based on ratios or linear combination and exploit
differences in the reflectance patterns (Fig. 2.1 and 2.4) of green vegetation and other
objects (Perry and Lautenschlager, 1984[125]). Ratio V I may be the simple ratio of any two
spectral bands, or the ratio of sums, differences or products of any number of bands. Linear
combinations are orthogonal sets of n linear equations calculated using data from n spectral
bands. If V Is are ratios, they can be easily cross-calibrated across sensor systems, ensuring
continuity of data sets for long-term monitoring of the land surface and climate-related
processes.
According to Myneni et al., 1995[110] most indices can be expressed in the form kρ0 , i.e. a
coefficient k times the derivate of surface reflectance ρ0 .
Following a descriptive list of the principal Vegetation indices in literature is exposed. The
list is divided in accordance with the feature analyzed by the corresponding index. ρx is the
reflectance at an x wavelength ( in nm).
Structural indices
Chlorophyll indices
Water indices
Structural indices Structural indices are measures of the general quantity and vigor
of green vegetation. These indices provide a measure of the overall amount and quality
of photosynthetic material in vegetation, which is essential for understanding the state
of vegetation for any purpose.
RVI Ratio vegetation index.

Assumes that all the isovegetation lines 2.8 are crossing in one point.
ρN IR
RV I = (2.1)
ρRED
This approach do not have normal distribution.
Reference: Jordan, 1969 [79]; Rouse et al., 1974[144]; Jackson and Huete, 1991[78]
NRVI Normalized Ratio Vegetation Index.
RV I − 1
N RV I = (2.2)
RV I + 1
Reference: Baret and Guyot, 1991 [8]
PVI Perpendicular vegetation index.
Assumes that the perpendicular distance between one pixel and the soil line of a
ρRED /ρN IR space (Fig. 2.8) is directly related to vegetation coverage.
P V I = sin(a) ρN IR − cos(a) ρRED (2.3)
where a is the angle between the soil line of the ρRED /ρN IR space.
Reference: Richardson and Wiegand, 1977 [137]
DVI Difference Vegetation Index.
It is a particular case of PVI, used when the soil line (Fig. 2.8) has 1 as pending.
DV I = ρN IR − ρRED (2.4)
Reference: Lillesand and Kiefer, 1987 [94]; Richardson and Everitt, 1992[136];
Jordan, 1969[79]
AVI Ashburn Vegetation Index.

It was developed specifically for MSS sensor.
AV I = ρN IR − ρRED (2.5)
Reference: Ashburn, 1978 [3]

NDVI Normalized Difference Vegetation Index.
N DV I is one of the oldest, most well known, and most frequently used V Is
This index estimates green vegetation cover and will be more deeply explained in
section 2.1.2.
ρN IR − ρRed
N DV I = (2.6)
ρN IR + ρRed
where rhoN IR , rhoRed .
The value of this index ranges from -1 to 1. The common range for green vege-
tation is 0.2 to 0.8.
Reference: Rouse et al., 1974 [144]

RDVI Renormalized Difference Vegetation Index.
It was proposed to combine the advantages of the Difference Vegetation Index 2.4
and the N DV I for low and high LAI values, respectively.
ρ800 − ρ670
RDV I = p (2.7)
(ρ800 + ρ670 )
Reference: Rougean and Breon, 1995[142]
SR Simple Ratio Index.
This index estimates green vegetation cover.
ρN IR
SR = (2.8)
ρRed
The value of this index ranges from 0 to more than 30. The common range for
green vegetation is 2 to 8.
Reference: Jordan, 1969 [79]; Rouse et al., 1974 [144]

MSR Modified Simple Ratio.
It was suggested as an improvement over RDV I (2.7) in terms of sensitivity
to vegetation biophysical parameters through its combination with the Simple
Ratio (2.8). SR and M SR are considered more linearly related to vegetation
parameters.
ρN IR
ρRed − 1
M SR = (2.9)
( ρρNRed
IR 0.5
) + 1
Reference: Chen, 1996 [18]
TVI Transformed Vegetation Index.
The aim of this index is to eliminate negative values and to normalize the N DV I
histogram. √
T V I = N DV I + 0.5 (2.10)
Reference: Deering et al., 1975 [31]
CTVI Corrected Transformed Vegetation Index.
This transformation eliminate negative values, but change the distribution of
values.
(N DV I + 0.5) p
CT V I = |N DV I + 0.5| (2.11)
|N DV I + 0.5|
Reference: Perry and Lautenschlager, 1984 [125].
TTVI Thiam Transformed Vegetation Index.

p
T T V I = |N DV I + 0.5| (2.12)
Reference: Thiam, 1997 [165]
IPVI Infrared Percentage Vegetation Index.
ρN IR
IP V I = (2.13)
ρN IR + ρRed
Reference: Crippen, 1990 [26]
SAVI Soil Adjusted Vegetation Index.
It was developed in order to reduce the effect of background soil.
ρN IR − ρRed
SAV I = (1 + L) (2.14)
ρN IR + ρRed + L
soil-adjustment factor, L, is where the lines of any vegetation cross the soil line.
It is determinate by the relative percentage of vegetation and the soil darkness.
The value of factor L is critical in the minimization of soil optical properties
effects on vegetation reflectance. Huete (1988)[61] suggested an optimal value of
L = 0.5 to account for first-order soil background variations.
Reference: Huete, 1988 [61]; Qi et al., 1994[133].
TSAVI Transformed Soil Adjusted Vegetation Index.
a (ρN IR − a ρRed − b)
T SAV I = (2.15)
ρRed + a b + X (1 − a)2
where a y b are line soil parameters and X is 0.08.
a = slope of soil line.
b = intercept of soil line.
Reference: Baret et al., 1989[9]; Baret and Guyot, 1991[8]

MSAVI Modified Soil Adjusted Vegetation Index.
Improved SAVI with self-adjustment factor L. It was developed in order to de-
scribe more adequately this soil-vegetation system and to minimize the effect of
bare soil.
ρN IR − ρRed
M SAV I = (1 + L1) (2.16)
ρN IR + ρRed + L1
The L1 factor is fitted to the image data.
L = 1 − ( 2 a N DV I W DV I)W DV I = ρN IR − (aρRed ) (2.17)
W DV I is the Weighted Difference Vegetation.

Reference: Qi et al., 1994[133]
MSAVI2 Modified Soil Adjusted Vegetation Index.
p
(2 (ρN IR ) + 1) − (2 ρN IR + 1)2 − 8(ρN IR − ρRed )
M SAV I2 = (2.18)
2
Reference: Qi et al., 1994[133]
OSAVI Optimized Soil Adjusted Vegetation Index.
( 1 + 0.16)(ρ800 − ρ670 )
OSAV I = (2.19)
ρ800 + ρ670 + 0.16
Reference: Rondeaux and Baret, 1996 [141]
GESAVI Generalized SAVI.
Reference: Gilabert et al., 2002 [51]

VARIgreen Visible atmospherically resistant index

ρGreen − ρRed
V ARIgreen = (2.20)
ρGreen + ρRed − ρBlue
Reference: Gitelson et al. (2001)

EVI Enhanced Vegetation Index.
It was developed to optimize the vegetation signal with improved sensitivity in
high biomass regions and improved vegetation monitoring through a de coupling
of the canopy background signal and a reduction in atmosphere influences.
ρN IR − ρRed
EV I = (G) (2.21)
ρN IR + (C1) (ρRed ) − (C2) (ρBlue ) + L
where ρN IR ,ρRed ,and ρBlue are atmospherically corrected or partially atmosphere

corrected (Rayleigh and ozone absorption) surface reflectance in near infrared, red
and blue bands respectively, G is a gain factor, C1, C2 are the coefficients of the
aerosol resistance term, which uses the blue band to correct for aerosol influences
in the red band, and L functions as the soil adjustment factor.
The EV I is more functional on N IR reflectance than on Red absorption, and

therefore it does not ”saturate” as rapidly as N DV I in dense vegetation, and it
has been shown to be highly correlated with photosynthesis and plant transpi-
ration in several of studies (Huete et al., 2008[63]). The EV I is one of the two
V Is available from the MODIS sensors and it is increasingly used in phenological,
productivity and evapotranspiration (ET ) studies.
Reference: Huete et al., 2008[63]

PD54 It was developed for 4 and 5 MSS Bands.
Reference: Pickup et al., 1993[127]
ARVI Atmospherically Resistant Vegetation Index.
Atmospheric effects like absorption and scattering may considerably reduce the
interpretation of remote sensed data. The resistance of this index to atmospheric
effects, compared with N DV I, is accomplished by a self-correction process for
the atmospheric effect on the red channel. The difference between the blue and
red channels minimizes the effects of atmospheric scattering caused by aerosols
in the red channel. This index was developed to minimize atmospheric effects in
MODIS.
ρN IR − (ρRed ρBlue )
ARV I = ρRed ρBlue = ρRed − γ (ρBlue − ρRed ) (2.22)
ρN IR + ρRed ρBlue
γ is an atmospheric self correcting factor which depends on aerosol types.

Reference: Kaufman and Tanré, 1992 [82]
SARVI Soil and Atmospherically Resistant Vegetation Index.
It minimizes both canopy background and atmospheric effects.
(ρ800 − ρrb )
SARV I = (1 + L) (2.23)
ρ800 + ρrb + L
where
ρrb = ρRed − γ(ρBlue − ρRed ) (2.24)
Reference: Kaufman and Tanre, 1992

GEMI Global Environmental Monitoring Index.

This index is aimed to resist atmospheric effects in AVHRR data.
ρRed − 0.125
GEM I = η(1 − 0.25η) − (2.25)
1 − ρRed
where
[2 (ρN IR − ρRed ) + 1.5 ρN IR + 0.5 ρRed ]
η= (2.26)
ρN IR + ρRed + 0.5
Reference: Pinty and Verstraete, 1991 [129]

GVI In this index, n bands are used in order to find the perpendicular to the line
soil.
−0.2848 T M 1−0.2435 T M 2−0.5436 T M 3+0.7243 T M 4+0.0840 T M 5−0.1800 T M 7

(2.27)
Reference: Kauth and Tomas, 1976 [84]; Crist and Cicone, 1984[27]
WDVI
W DV I = ρN IR − s (ρRed ) (2.28)
MND It was built in order to include the slope of the best fit line for lands with low
vegetation cover. The MND maximizes the contrast in ‘greenness’ while assigning
a value of zero for no vegetation or dry vegetation.
Reference: Paltridge and Barber, 1988[120]
Chlorophyll indices
G Greenness Index
ρ554
G= (2.29)
ρ677
MCARI Modified Chlorophyll Absorption Reflectance Index.
Kim et al. (1994) developed the Chlorophyll Absorption Ratio Index (CARI)
which measures the depth of chlorophyll absorption at 670 nm relative to the
green reflectance peak at 550 nm and the reflectance 700 nm. The CARI was,
then, simplified by Daughtry et al. (2000)[30] to obtain the Modified Chlorophyll
Absorption Ratio Index (MCARI), which is quantified by the following equation:
ρ700
M CARI = {(ρ700 − ρ670 ) − 0.2 (ρ700 − ρ550 )( )} (2.30)
ρ670
Reference: Daughtry et al., 2000[30]
MCARI 1 Modified Chlorophyll Absorption Reflectance Index 1.
M CARI 1 = 1.2 {2.5 (ρ800 − ρ670 ) − 1.3 (ρ800 − ρ550 )} (2.31)

Reference: Haboudane et al., 2004[54]
MCARI 2 Modified Chlorophyll Absorption Reflectance Index 2.
1.5 {2.5 (ρ800 − ρ670 ) − 1.3 (ρ800 − ρ550 )}

M CARI 2 = p √ (2.32)
(2 ρ800 + 1)2 − (6 ρ800 − 5 ρ680 ) − 0.5
TCARI Transformed Chlorophyll Absorption Reflectance Index.
ρ700
T CARI = 3 {(ρ700 − ρ670 ) − 0.2 (ρ700 − ρ550 )( )} (2.33)
ρ670
TVI Triangular Vegetation Index.

Inspired by the general idea of CARI (2.30), Broge and Leblanc (2000) [15]
developed the Triangular Vegetation Index (T V I), which is meant to characterize
the radiant energy absorbed by leaf pigments in terms of the relative difference
between red and N IR reflectance in conjunction with the magnitude of reflectance
in the green region. T V I is determined as the area defined by the green peak, the
N IR shoulder, and the minimum reflectance in the red region. It is formulated
as:
T V I = 0.5 {120(ρ750 − ρ550 ) − 200 (ρ670 − ρ550 )} (2.34)
The general idea behind T V I is based on the fact that the total area of the triangle
(green, red, infrared) will increase as a result of chlorophyll absorption (decrease
of red reflectance) and leaf tissue abundance (increase of N IR reflectance) (Broge
and Leblanc, 2000). While this remains true, it is important to notice that the
increase of chlorophyll concentration also results in the decrease of the green
reflectance, leading, therefore, to a relative decrease of the triangle area. Further-
more, although there is no chlorophyll absorption beyond 700 nm, chlorophyll
indirect effects on the vegetation reflectance curve remain observable around the
‘red edge’ position up to 750 nm. In fact, as chlorophyll content increases, its
absorption feature broadens and causes the red-shift of the red-edge reflectance.
Consequently, the canopy reflectance at 750 nm is still influenced by leaf chloro-
phyll content.
Reference: Broge and Leblanc, 2000[15]
MTVI1 Modified Triangular Vegetation Index 1.
M T V I1 = 1.2 {1.2 (ρ800 − ρ550 ) − 2.5 (ρ670 − ρ550 )} (2.35)

Reference: Haboudane et al., 2004 [54]
MTVI2 Modified Triangular Vegetation Index 2.
1.5 (1.2 (ρ800 − ρ550 ) − 2.5 (ρ670 − ρ550 ))

p √ (2.36)
(2 ρ800 + 1)2 − (6 ρ800 − 5 ρ670 − 0.5
ZTM Zarco-Tejada and Miller.
ρ750
ZT M = (2.37)
ρ710
Reference: Zarco-Tejada et al., 2001[182]
RARS a Chlorophyll-A concentration.
ρ675
RARSa = (2.38)
ρ700
RARS b Chlorophyll-B concentration.
ρ675
RARSb = ρ700 (2.39)
ρ650
Reference: Datt, 1998[28]
Chl NDI Chlorophyll normalized difference index.
ρ750 − ρ705
RARSb = (2.40)
ρ750 + ρ705
Reference: Datt, 1998[28]; Gitelson and Merzlyak, 1994[52]; Gitelson, 1996[53]
NPCI Normalized Pigment Chlorophyll Ratio Index.
CCCI Canopy chlorophyll content index.
PSSR Pigment Specific Simple Ratio.
ρ800 − ρ470
P SSR = (2.41)
ρ800 + ρ470
Reference:Blackburn and Steele (1999)
PSSRa Pigment Specific Simple ratio Chl a.
ρ800
P SSRa = (2.42)
ρ680
Reference:Blacburn (1998)
PSSRb Pigment Specific Simple ratio Chl b.
ρ800
P SSRb = (2.43)
ρ635
PRSSc Pigment Specific Simple ratio Cars.
ρ800
P SSRc = (2.44)
ρ470
PSNDa Pigment Specific Normalized Difference a.
ρ800 − ρ680
P SN Da = (2.45)
ρ800 + ρ680
PSNDb Pigment Specific Normalized Difference b.
ρ800 − ρ635
P SN Db = (2.46)
ρ800 + ρ635
Water indices
NDWI Normalized Difference Water Index.
ρ860 − ρ1240
N DW I = (2.47)
ρ860 + ρ1240
tation is -0.1 to 0.4. Reference: Gao, 1996 [49]
SRWI Simple Ratio Water Index.
ρ858
SRW I = (2.48)
ρ1240
Reference: Zarco-Tejada et al., 2003[183]
PWI Plant Water Index.
ρ970
PWI = (2.49)
ρ900
Reference: Peñuelas et al., 1997[124]
WBI Water Band index.

As the water content of vegetation canopies increases, the strength of the absorp-
tion around 970 nm increases relative to that 900 nm.
ρ900
W BI = (2.50)
ρ970
The common range for green vegetation is 0.8 to 1.2.
Reference: Peñuelas, 1993[122]; Peñuelas et al., 1995[123]; Peñuelas, 1997[124];
Gamon and Qiu, 1999[48]
MSI Moisture Stress Index.
As the water content of leaves in vegetation canopies increases, the strength of the
absorption around 1599 nm increases. Absorption at 819 nm is nearly unaffected
by changing water content, so it is used as the reference.
ρ1599
M SI = (2.51)
ρ819
The value of this index ranges from 0 to more than 3. The common range for
green vegetation is 0.4 to 2.
Reference: Hunt and Rock, 1989[67]
WAI Water absorption index.
ρ895
W AI = (2.52)
ρ972
NDII Normalized Difference Infrared Index.
ρ819 − ρ1649
N DII = (2.53)
ρ819 + ρ1649
Reference: Hardinsky, 1983[56]
GVMI Global Vegetation Moisture Index.
(ρN IR + 0.1) − (ρSW IR + 0.02)

GV M I = (2.54)
(ρN IR + 0.1) + (ρSW IR + 0.02)
Another physiological plant features
PRI Photochemical Reflectance Index.

It is a reflectance measurement that is sensitive to changes in carotenoids pig-
ments (particularly xanthophyll pigments) in live foliage. Carotenoids pigments
are indicative of photosynthetic light use efficiency,or the rate of carbon diox-
ide uptake by foliage per unit energy absorbed. As such, it is used in studies
of vegetation productivity and stress. Applications include vegetation health in
evergreen shrublands, forests, and agricultural crops prior to senescence.
ρ531 − ρ570
P RI = (2.55)
ρ531 + ρ570
tation is -0.2 to 0.2.
Reference: Gamon, 1992[47]
NDNI Normalized Difference Nitrogen Index.
It estimates the relative amounts of nitrogen contained in vegetation canopies.
Reflectance at 1510 nm is largely determined by nitrogen concentration of leaves,
as well as the overall foliage biomass of the canopy.
1 1
log ( ρ1510 ) − log ( ρ1680 )
N DII = 1 1 (2.56)
log ( ρ1510 ) + log ( ρ1680 )
The value of this index ranges from 0 to 1. The common range for green vegetation
is 0.02 to 0.1.
Reference: Serrano, 2002[154]
NI Nitrogen Index.
ρ800 − ρ600
NI = (2.57)
ρ800 + ρ600
Reference: Ma (1996)
SIPI Structure Insensitive Pigment Index.
It is a reflectance measurement that is sensitive to the ratio of bulk carotenoids (for
example, alpha-carotene and beta-carotene) to chlorophyll while is less sensitive
to variation in canopy structure (for example, leaf area index).
ρ800 − ρ445
SIP I = (2.58)
ρ800 − ρ680
The value of this index ranges from 0 to 0.2. The common range for green
vegetation is 0.8 to 1.8.
Reference: Penuelas et al. (1995)[121]
2.1.2 Normalized Difference Vegetation Index
The Normalized Difference Vegetation Index (N DV I), calculated from the red and N IR
channels (Eq. 2.6), has been extensively used for vegetation studies. N DV I is mainly
dependent on the green leaf material of the vegetation cover (Tucker, 1979[169]; Huete,
1988[61]; Myneni et al., 1995[110]; Gitelson et al., 1996[53]) and has been found to be
correlated with the fraction of P AR absorbed by green vegetation (Asrar et al., 1984 [5];
Hatfield et al., 1984[57]; Choudhury, 1987 [20]; Baret and Guyot, 1991[8]; Asrar et al., 1992
[6]; Myneni and Williams, 1994 [113]; Friedl et al., 1995 [41]; Myneni et al., 1997[111]; Cohen
et al., 2003 [23]; Moulin et al., 1997[108]) and hence energy absorption of plant canopies
(Myneni and Ganapol, 1992[109]; Sellers et al., 1992[150]), to Leaf Area Index (LAI) (e.g.,
Asrar et al., 1984[5]; Begue and Myneni, 1996[13]; Chen and Cihlar, 1996[19]; Fassnacht
et al., 1997; Turner et al., 1999; Haboudane et al., 2004[54]); chlorophyll concentration in
leaves (Yoder and Waring, 1994; Gitelson and Merzlyak, 1997; Broge and Leblanc, 2000[15];
Daughtry et al., 2000[30]; Dawson et al., 2003; Sims and Gamon, 2003), above-ground
biomass (Todd et al., 1998; Labus et al., 2002; Foody et al., 2003), net primary productivity
(NPP) (Ruimy et al., 1994; Hunt, 1994), fractional vegetation cover (Purevdorj et al., 1998;
Gitelson et al., 2002), and vegetation water content (Tucker, 1980[170]; Ceccato et al., 2001)
and to carbon uptake and release by vegetation (Fung et al., 1987[44]), as well as providing
information about vegetation phenology (Moulin et al., 1997[108]).
For instance :
In Tucker et al., 1985[172] N DV I explained 80% of biomass accumulation (Fig 2.10).
Myneni et al. 1997[111] showed that N DV I was near-linearly related to the chlorophyll
content of single soybean leaves and curvilinearly related to the chlorophyll content
of soybean canopies (because surface leaves intercept more light than leaves deeper in
the canopy).
Sellers (1996[151]) used a canopy radiative transfer model to show that N DV I is

near-linearly related to area-averaged net carbon assimilation and plant transpiration,
even at different values of fractional cover and LAI. The Radiative Transfer Equation
(RTE) describes in mathematical form, energy exchanges when radiation with a given
wavelength λ is propagated according to the Ω direction through a volume element.
The availability of the N DV I data for two decades has motives a large number of studies
from regional to global scales (e.g., Tucker et al., 1985[172]; Myneni et al., 1998[112]; DeFries
et al., 1999[33]), relating N DV I to vegetation physiology and structure.
N DV I was one of the most successful of many attempts to simply and quickly identify
vegetated areas and their ”condition”, and it remains the most well-known and used index
to detect live green plant canopies in multispectral remote sensing data. One of the reason of
this extensive use of N DV I being that it employs basic spectral bands available in practically
all remote sensing systems and it is computationally very efficient (Walthall et al., 2004[180]).
Normalization reduces the effect of sensor calibration degradation by approximately 6% of
the overall index value (Holben et al., 1990[59]; Kaufman and Holben, 1993[81]). However,
N DV I presents the drawback of over-estimating the percentage of vegetative cover at the
beginning of the growth season and of under-estimating it at the end of the season (Cyr,
1993).
Among the disadvantages of the N DV I, the follow issues can be mentioned (Huete et al.,
2002 [62]), and all these factors introduce uncertainty in quantitative assessments.
The nonlinearity of ratio-based indexes
The influence of the atmosphere (the actual composition of the atmosphere with respect
to water vapor and aerosols)
Scaling problems
Asymptotic (saturated) signals over high biomass conditions
Sensitivity to background variations
Sensitivity to clouds (deep, thin, shadow)
Sensitivity to soil effects (moisture state, color)
Sensitivity to anisotropic effects (geometry of the target)
Sensitivity to spectral effects (different instruments).
A major limitation of the N DV I and similar indices is that the optical sensors can only
monitor a very thin layer of the canopy. They cannot provide information on woody biomass,
Figure 2.10: Relationship between N DV I and total dry biomass according to the work of
Tucker et al. (1985).
and total above-ground live carbon, which are of great interest to carbon cycle modeling,
and ecological applications.
Because N DV I can be affected by soil reflectance (Major et al., 1990[100]; Elvidge and Lyon,
1985[37]; Huete and Tucker, 1991[65]; Huete et al., 1985[66]; Todd and Hoffer, 1998[166]), soil
adjusted V Is have been developed to take into account the local soil reflectance. However,
these V Is were found to be highly correlated to N DV I such that the choice of one V I over
another essentially resulted in a scaling of N DV I (Lyon et al., 1998[98]).
Examples of modifications of the N DV I in order to account for the soil background are
indices such as the SAVI (Eq. 2.14) or the WDVI (Eq. 2.28) index (Hall et al., 2003
[138]). Moreover, new indexes have been developed to compensate for both the effect of
the atmosphere (Kaufman and Tanré, 1996 [83]) and soil (Huete, 1988[61]), particularly the
‘enhanced’ vegetation index (EVI: Eq. 2.21) improved sensitivity in high biomass regions
and then vegetation monitoring through a decoupling of the canopy background signal and
a reduction in atmospheric influences (Liu and Huete, 1995[96]; Huete et al., 1997[64]).
2.1.3 Vegetation analysis from hyperspectral sensors

Hyperspectral (i.e., reflectance for many contiguous narrow wavelength bands) approaches
have been proposed to detect water-, nutrient-, and pest-induced stress in plants while min-
imizing unwanted signals from varying soil conditions or biomass amounts. These methods
commonly use derivative analysis, peak fitting procedures, and ratio analysis to associate
spectral features with a particular stress (Horler et al., 1983 [60]; Demetriades-Shah et al.,
1990[34]; Chappelle et al., 1992 [17]; Osborne et al., 2002 [117]). When functional rela-
tionships between hyperspectral and plant properties cannot be envisioned using simple or
multiple regressions, more sophisticated statistical approaches such as principal component,
neural net, fuzzy, and partial least-squares regression analysis have been employed (Kimes
et al., 1998 [87]). Spectral mixing techniques (McGwire et al., 2000 [102]) draw on a li-
brary of ‘pure’ hyperspectral signatures of scene components (endmembers) to decompose
images into their separate constituents (e.g., sunlit and shaded soil, healthy and stressed
plant areas).
Indices based on derivate

1DGVI First order Derivative greenness vegetation indices.
λn 0 0
X ρ (λi ) − ρ (λj )
DGV I = (2.59)
∆λi
λ1
0
where λi and λj are the wavelengths at the midpoints of bands i and j, ρ is the
first derivate of the reflectance, λ1 is the start of a DGVHI (Derivative greenness
Hyperion vegetation indices) waveband, and λn is the end of a DGHVI waveband.
0
Where i and j are band numbers, λ = center of wavelength, and ρ = first derivate
reflectance.
Reference: Elvidge and Chen, 1995[36]
2DGVI It was found superior to the 1DGVI index (first order). Reference: Elvidge
and Chen, 1995[36]
For instance, Jacquemoud (1993 [146]) has taken a different approach to the study of high
spectral resolution remote sensing of vegetation by using a canopy reflectance model
(called P ROSP ECT ) and a reflectance spectrum to investigate the potential of estimating
canopy parameters. As another example, Gamon et al. (1993 [46]) used AVIRIS data to
study the distribution, physiology, and productivity of vegetation at the Stanford University
Jasper Ridge Biological Preserve in California. A spectral mixture analysis and the
normalized difference vegetation index (N DV I) calculated from narrow AV IRIS spectral
bands were employed to evaluate the spatial patterns of vegetation type, productivity, and
physiological activity in an annual grassland.
2.2. VEGETATION ANALYSIS FROM MICROWAVE DATA 23
2.2 Vegetation analysis from microwave data

The location of microwave data in the electromagnetic spectrum is shown in Fig. 2.11. The
different microwave bands (wavelength regions) are presented in Fig. 2.12
Figure 2.11: Vegetation analysis from microwave data.
Figure 2.12: Wavelength regions of the electromagnetic spectrum corresponding to mi-

crowave data.
Conventional visible-near infrared indices are often limited by the effects of atmosphere,
background soil conditions, and saturation at high levels of vegetation. For global vegetation
characterization, lower-resolution space-borne sensors operating in the microwave part of the
spectrum have also shown some ability to characterize the land surface at spatial resolutions
compatible with climatological applications.
Microwave responses to the land surface include contributions from the vegetation and from
the underlying surface. A review of research directed toward understanding of the mech-
anisms responsible for the microwave emission and backscattering of soil and vegetation is
presented by Ulaby et al., 1986[174].
Vegetation absorbs, emits, and scatters microwave radiation, being vegetation radiative prop-
erties mainly controlled by the dielectric properties of vegetation components, their density,
and the relative size of vegetation components with respect to the wavelength. Dielectric
properties of vegetation are closely related to their water content, whereas increasing vege-
tation density usually reduces the emissivity polarization difference (Choudhury, 1989[21])
and increases the backscattered signal (Frison and Mougin, 1996[42]). Typically, vegetation
has large emissivities and low emissivity polarization differences.
The radar backscatter from vegetated surfaces is controlled by two sets of parameters:
geometric parameters related to the structure of vegetation and soil. Geometrical

features of plant constituents affect scattering in a different manner, according to
frequency and polarization.
dielectric parameters related to plants moisture content and underlying soil surface.
Dielectric characteristics of vegetation material are strongly influenced by moisture
content over a wide range of the microwave spectrum.
It is well known that passive microwave sensors are sensitive to variations in vegetation
properties in a relatively thick layer of the canopy and therefore, N DV I mainly responds to
a thin layer of the canopy (leaves). On the other hand, both passive and active microwave
observations respond to the absorption/emission and scattering by vegetation elements, in-
cluding woody parts, and are not directly sensitive to the photosynthesis activity. Optical
systems, which proved their effectiveness in sensing leaf parameters, are not able to sense the
woody biomass. Microwave measurements therefore complement the VIS-NIR observations
by indicating structural density (Prigent and Aires, 2001[131]).
2.2.1 Passive microwave data

In comparison to optical sensors, passive microwave sensors can be used both day and night,
can penetrate clouds (all-weather), and are less affected by atmospheric conditions. These
sensors have a long period of record, and can be useful, for instance, as either a primary or
complementary tool in assessing the impacts of global climate change on carbon cycling and
ecosystem variability, or the effects of ecosystems change on climate.
For example, an L-band (21 cm, 1.4 GHz: Fig. 2.12) passive microwave radiometer measure
the intensity of the microwave emission from the land surface in terms of a brightness
temperature. This brightness temperature is related to the surface soil moisture (Jackson et
al., 1995 [77]). Any vegetation present is usually examined as noise and its effect is estimated
using a single parameter proportional to the vegetation water content (Jackson, 1993 [76]).
A vegetation canopy will scatter and absorb microwave emission from the soil and it will
also contribute with its own emission.
Pioneer research of the sensitivity of passive microwave to vegetation used linear combina-
tions of channels (e.g. Choudhury and Tucker, 1987[20]). However, these simple indices
have been shown to be contaminated by variations in atmospheric parameters and surface
temperature (Prigent and Aires, 2001[131]). For instance, microwave emissivities calcu-
lated from SSM/I between 19 and 85 GHz (Fig. 2.12) using ancillary data to remove
atmospheric contributions, show some ability to characterize vegetation types (Prigent and
Aires, 2001[131]).
Some of the earliest works in deriving vegetation information from passive microwave data
(i.e. Choudhury and Tucker, 1987[20]) showed that microwave polarization difference tem-
peratures (M P DT ) at 37 GHz (Fig. 2.12) were highly correlated to N DV I in arid and
semi-arid regions and related to variations in leaf water content (Njoku and Li, 1999 [114]).
From microwave radiative transfer theory and field measurements it is know that M P DT is
not only affected by the vegetation properties but also by the surface effective reflectivity (soil
moisture and roughness) and the physical temperature. In order to minimize the physical
temperature effects, Becker and Choudhury (1988 [12]) proposed the normalized microwave
polarization difference index (M P DI) (Eq. 2.60).
(T Bv − T Bh)
M P DI = C (2.60)
(T Bv + T Bh)
for a given frequency, where C is a scale factor. Here T Bv and T Bh are the brightness
temperature for horizontal (h) and vertical polarization (v). This is also referred to as the
normalized polarization index (P I) (Paloscia and Pampaloni, 1992 [119]). The use of
the M P DI helps to account for the effects of surface temperature variations on brightness
temperature (T b), hence leaving a quantity that is largely dependent on the emissivity
contributions from water surfaces, vegetation and soil moisture within the satellite footprint.
Then, it was also derived a microwave vegetation index based on the difference in normalized
brightness temperature (normalized by thermal infrared measurements) at two frequencies
δ Tn = Tn (f 2) − Tn (f 1) (2.61)
This approach was used for detecting biomass and water conditions of agricultural crops
using data at 10 GHz and 36 GHz (Fig. 2.12).
The microwave vegetation indices described above can be useful if all other perturbing factors
are uniform. However, depending upon the sensor frequencies and the level of vegetation
present they can be significantly affected by soil emission variations resulting from soil
moisture and surface roughness conditions. This problem can limit the value of such a
product in global vegetation monitoring.
Another index, defined as the microwave emissivity difference vegetation index (EDV I: (Eq.
2.62)).
EDV I = 2 (T Bp (f 1) − T Bp (f 2))/(T Bp (f 1) + T Bp (f 2)) (2.62)
was proposed by Min and Lin (2006 [103]). It was intended for application to dense forest
conditions using 19 GHz and 37 GHz observations, where both measurements do not ‘see’
the ground surface. It was demonstrated that the EDV I (Eq. 2.62) was more sensitive to
evapotranspiration than the N DV I.
Microwave vegetation indices (M V Is ) in general, show less seasonal variability than N DV I
since the woody part of vegetation exhibits less seasonal variation while the larger seasonal
variations of the N DV I are mainly due to the seasonal changes of the leafy part of the
vegetation.
Microwave emission model
The microwave emission model that can be used to derive the microwave vegetation indices is
the ω − τ model derived from a 0 th-order radiative transfer solution (Ulaby et al., 1982[43]).
The radiative transfer model at a given frequency can be described as:
TBp (f ) − Ve (f )
εsp (f ) = (2.63)
Vt (f )
ε is the emissivity
superscripts s indicate the soil component
subscript p is for polarization status
f = frequency
T is temperature
Ve (f ) = vegetation emission component
Vt (f ) = vegetation transmission component
TBp = brightness temperature at a given polarization p.
In Shi et al., 2008 [155] is used the relation: (Eq. 2.64)
TBp f2 = Ap (f1 , f2 ) + Bp (f1 , f2 ) TBp (f1 ) (2.64)
Eq. (2.64) indicates that brightness temperature observations at a given polarization p

observed with two adjacent AM SR − E frequencies can be described as a linear function.
The intercept Ap and slope Bp of this linear function are the microwave vegetation indices
that are defined in that study. They are independent of the underlying soil/ surface signals
and dependent only on vegetation properties such as vegetation fraction cover, temperature,
biomass, water content, and characteristics of the scatter size, shape and orientation of
vegetation canopy.
For vegetated surfaces when the sensor can partially “see” through the vegetation canopy
at both AM SR − E frequencies, the A parameter will increase and the B parameter will
decrease as the vegetation canopy becomes thicker. This is because the frequency dependence
of the vegetation emission component increases as frequency increases for the same type
of vegetation canopy. However, the A parameter is not only affected by the vegetation
properties but also by the surface physical temperature. The B parameter, on the other
hand, exhibits the opposite behavior.
In that work, for evaluating whether the microwave and optical sensors derived vegetation
information can be used synergistically in monitoring vegetation phenology, they compared
the observed global patterns for different seasons and the seasonal variations. Figs. 2.13
show the mean monthly values for April 2003 for the N DV I (A) derived by MODIS, the
AMSR-E derived M V IA parameters using the low frequency pair (6.925 GHz/10.65 GHz)
in (B) and the high frequency pair (10.65 GHz/18.7 GHz) in (C). The corresponding MVI
B parameters are shown in (D) and (E), respectively.
Figure 2.13: April monthly mean values for NDVI in (A), A parameters using the low
frequency pair in (B), and the high frequency pair in (C), B parameters using the low
frequency pair in (D), and and the high frequency pair in (E).
In general terms, it was shown that the M V Is can have a large dynamic range for a given
N DV I value, which results from differences in vegetation fractional cover, structure, sizes,
and water content or wet biomass, especially due to the differences in the woody part of
the vegetation that optical sensor have virtually no sensitivity to. From these results the
authors concluded that the these microwave vegetation indices can provide new and com-
plementary information (to N DV I) on vegetation that can improve our ability to monitor
global vegetation and ecosystem properties from space.
Stress-Degree-Day SDD Reference:Idso et al., 1977 [69]
Crop Water Stress Index CWSI Reference: Idso et al., 1981 [68]; Jackson et al., 1981
[75]
Non-water-stressed baselines Reference: Idso, 1982[70]

Thermal Kinetic Window TKW Reference: Mahan and Upchurch, 1988 [99]
Water Deficit Index WDI Reference: Moran et al., 1994 [104]
2.2.2 Active microwave data
Active microwave systems are useful because they are not constrained by either night (dark-
ness) or cloud cover. They create their own signals that are reflected back from the earth’s
surface giving imagery that has utility for vegetation mapping and strong capabilities for
looking at terrain features.
Active microwave backscattering observation are not affected by variations in atmospheric
conditions and do not require significant preprocessing (Prigent and Aires, 2001[131]). There-
fore, scatterometers appear to be very promising instruments for land surface characteriza-
tion, due to their high sensitivity to vegetation. In the same way, the significant progress
in radar technology, has made available a large amount of high-resolution, multifrequency,
multipolarization backscatter data.
The problem of backscatter coefficient (σ o ) sensitivity to forest biomass has received par-
ticular attention and many experimental and theoretical studies have concentrated on this
subject. Although experiments have been carried out in several countries, under different
environmental conditions, a significant agreement on some main features is observed. In
particular, most of the results indicate that at P (0.45 GHz) and L (1.2 GHz) bands (σ o ) of
dense forests is much higher than that of agricultural fields, particularly at HV polarization.
2.2.3 Subsection: Vegetation estimation techniques using SAR data
Synthetic Aperture Radar (SAR) takes advantage of the Doppler history of the radar echoes
generated by the forward motion of the spacecraft to synthesize a large antenna. This allows
high azimuth resolution in the resulting image despite a physically small antenna. As the
radar moves, a pulse is transmitted at each position. The return echoes pass through the
receiver and are recorded in an echo store.
SAR measure the spatial distribution of surface reflectivity in the microwave spectrum.
The radar transmits a pulse and then measures the time delay and strength of the reflected
echo (i.e., phase and amplitude measurements), where the ratio of scattered and incident
microwave energy is termed the radar backscatter (σ o ). The scattering behavior of the
SAR signal is governed by the dielectric properties of both soil and vegetation, and the
geometric configuration of the scattering elements (soil roughness, leaves, stalks, and fruit)
with respect to the wavelength, direction, and polarization of the incident wave. Whereas
the optical sensors are sensitive to plant structure at the micrometer scale, the SAR responds
to the centimeter and decimeter scales.
SAR systems have the advantages of cloud penetration, all-weather coverage, high spatial
resolution, day/night acquisitions, and signal independence of the solar illumination angle.
SAR data has high spatial resolution and provides fine-scale vegetation information within
the lower resolution radiometer footprint (Entekhabi et al., 2004).
SAR backscatter increases much for forest areas due to strong volume scattering by tree
branches or trunks (Fig. 2.14). SAR backscattering signal is sensible to the structural prop-
erties of the vegetation (volume, crown complexity). Of particular importance to SAR are
the vertical stems of plants and the trunks of trees because wave propagation and backscat-
tering from these media are polarization dependent. Thus different polarization combina-
tions like HH (Horizontal transmit- Horizontal receive), HV (Horizontal-Vertical) and VV
(Vertical-Vertical) will provide three different views of the vegetation canopy structure (Fig.
2.15). For some vegetation conditions, the phase differences between the different polariza-
tion backscattering may contain useful information about the vegetation cover.
Figure 2.14: Tipical scatterers contribution of forest structure.
Figure 2.15: Multitemporal surface responses at different polarizations of ENVISAT SAR

sensors.
A number of studies have shown that the above-ground biomass density of forest may be
monitored using SAR remote sensing at L-band (23 cm) and longer wavelengths (Le Toan
et al., 1992 [90]; Ranson and Sun, 1997 [135]). For instance, there appears to be a general
correspondence between the JERS-1 and AVHRR derived biomass density maps.
Physical models that describe radar backscatter from crops and soils, such as the wa-
ter cloud model (Attema and Ulaby, 1978[7]), the Integrated Equation Model (IEM )
(Fung et al., 1992[45]; Oh et al., 1992[116]), and the Michigan microwave canopy scatter-
ing (M IM ICS) model (Ulaby et al., 1990[178]) have been inverted to determine important
agricultural variables such as crop green leaf area index (GLAI) and surface soil moisture
(e.g., Altese et al., 1996[2]; Sano et al., 1998[149]).
As examples of SAR utilization in vegetation analysis I present the following
works:
In the microwave region, C-band SAR wavelength (1-6 cm), it is generally assumed
that (σ o ) is directly related to surface roughness, soil moisture condition, and vegeta-
tion density. This can be expressed by the water cloud model (Eq. 2.65): backscat-
tered by the whole canopy (σ o ) is the sum of the contribution of the vegetation (σv ),
and that of the underlying soil (σs ). The latter is attenuated by the vegetation layer.
Thus, for a given incidence angle (θ)
σ o = σvo + τ 2 σso (2.65)
where τ 2 is a function of green leaf area index (GLAI), σv in a function of τ 2 and

GLAI, and σs is a function of volumetric soil moisture content and surface roughness
(Ulaby et al., 1984[176]). Then, there are derived:
σvo = A V1E cos θ (1 − τ 2 ) (2.66)
σso = C + D hv (2.67)
cm 3 2
hv is volumetric soil moisture content ( cm 3 ), and V1 is a descriptor of the canopy. τ
is the two-way attenuation through th canopy, expressed as
−2BV2
τ 2 = exp ( ) (2.68)
cos θ
V2 is a second canopy descriptor. The canopy descriptors have been associated with
GLAI (Ulaby et al., 1984[176]). Then, we assumed V1 = V2 = GLAI.
In Moran et al., 1998[106] measurements of soil moistures were made in order to evalu-
ate the sensitivity of Ku- and C- band σ o (from ERS-1) to soil moisture. The approach
is based on that the Ku-Band σ o is sensitive to vegetation density and insensitive to
soil moisture content, and the C-band σ o is sensitive to soil moisture, through atten-
uated by increasing vegetation (Moran et al., in press). In Moran et al., 2002[105],
Landsat TM sensor measurements of reflectance and surface temperature (Ts ) were
used to interpret the ERS-2 C-Band SAR backscatter (σ o ). Theoretical response of
optical and SAR measurements to changes in plant/soil condition are presented in the
follow Table (2.2.3).
Change in plant/soil condition σo Ts ρRed ρN IR SAVI

Increase in surface roughness ↑ ↓ ↓
Decrease in green vegetation biomass ↑ ↑ ↑ ↓ ↓
Increase in surface soil moisture content ↑ ↓ ↓ ↓
Increase in plant litter ↑ ↑ ↑
where Ts is surface temperature, ρRed and ρN IR are surface reflectance in the red and
NIR spectrum, and SAV I is the soil adjusted vegetation index (Eq. 2.14).
The only surface condition, in non vegetated areas, that substantially affected surface
temperature was a change in surface soil moisture condition. Using SAV I (sensitive
only to changes in green vegetation biomass) and Ts (sensible to soil moisture in non
vegetated areas) as image filters, it was possible to identify all the pixels in the SAR
image associated with dry, bare soil. Then, the dominant influence on the magnitude
of SAR backscatter was surface roughness (related to tillage).
The results indicate that optical imagery obtained coincident with SAR imagery al-
lowed a better understanding of the interaction of the SAR signal.
The high radar sensitivity to volumetric moisture (Ulaby, 1974[175]) and the differ-
ences in water content between green vegetation and the uppermost soil layer, suggest
that SAR backscatter may allow better estimation of herbaceous aerial aboveground
biomass (AAB).
In the study of Svoray and Shoshany (2002 [163]) it is suggested that the synergy of
ERS-2 SAR data and Landsat TM data can be applied on the estimation of AAB of
herbaceous vegetation in a semi-arid environment.
Results indicate that in its modified form, the water-cloud model could provide
important information for a wide range of environmental applications in desert ringes.
Stoll, 2005 propose a new empirical vegetation index for polarimetric SAR data:
SARvi (Eq. 2.69). This index exploits the ability of AirSAR data to discrimi-
nate different kind of vegetated areas thanks to the polarimetric properties of the
vegetation: in the non-vegetated areas, the C-VV band has high values and the L-HV
band has lower values. On the contrary, on highly vegetated area, the C-VV band has
low values and L-HV band has higher values.
LHV − CV V
SARvi = 100 { + 1} (2.69)
LHV + CV V
SARvi minimum value occurs for non-vegetated areas (∼ 0) and its maximum value
represents the highly vegetated areas (∼ 200) The SARvi index is more correlated to
the vertical hierarchy of the vegetation and can be used for a first blind vegetation
study, without real ground truth mission. However, optical imagery obtained coinci-
dent with SAR imagery can be useful in order to have a better understanding of the
interactions of the SAR signal.
A simple class set (Fig. 2.16) has previously (Stoll, 2004[161]) been defined on a
vertical hierarchy of the vegetation. This class set is based on the relative height of
the vegetation. SAR backscattering signal is sensible to the structural properties of
the vegetation (volume, crown complexity, etc).
The statistics of the SARvi index according to the various ground truth classes show
its ability to better discriminate the class set.
Sano et al., (2005)[148] developed a comparative analysis of SAR backscatter coeffi-

cients (σ o ) with two vegetation indices: N DV I and EV I over the five most dominant
physiognomies. The results of the discriminant analysis suggest that the SAR and V Is
provide different types of information and complementary. A combination of dual tem-
poral radar backscatter with N DV I and EV I at a time indicated a higher capability
to identify vegetation types.
The highest backscattering values are related to the evergreen gallery forests. The
LHH transmitting and receiving waves of the JERS-1 SAR incident in an arboreal
canopy cover produce stronger and multiple scattering from trunks and stalks, known
as doublebounce scattering (Leckie and Ranson 1998[93]). The lowest values cor-
responded to the grass-dominated cerrado classes, that is, cerrado grassland, shrub
cerrado, and grassland with termite mounds.
As another example, Oguro et al., (2001 [115])described that combination of the use of
multi-temporal vegetation indices N DV I and EV I together with SAR data was found
to be effectively (error less than 2%) in rice conditions analysis. The combination of
N DV I and EV I is expected to extract the most information about changes in rice
conditions due to flowering and maturing, while SAR backscatter information is less
useful after the rice planting stage. SAR data are useful especially for the area with
high cloud coverage for monitoring of rice planted area.
In Shupe & Marsh (2004 [158]) maximum likelihood and ANN classifiers are used on a
combination of Landsat TM, ERS-1 SAR, and elevation data producing satisfactory
results in vegetation estimation. In the same way, in Dwyer et al., 2000[38] training
sets were selected in the sets of SAR data imagery for a number of classes based on
differences in backscatter and coherence response. Class names were assigned using
information in the topographic and land cover maps and knowledge of the regions. A
standard maximum likelihood classification was carried out. Some classes were then
merged and a majority filtering was carried out to remove small pixel agglomerations
in the final classified images. The utility of the coherence product, generated from
ERS tandem data, for forest mapping and monitoring has been shown.
In positions of biomass sampling data on radar images, the backscattering signatures

can be extracted. Regression relationships can be developed between the cube root
of total biomass and the averaged radar signature, for instance such as the method of
Ranson and Sun (1997)[135].
In Paloscia et al., 1999 [118] the sensitivity of backscattering coefficient, measured by

ERS-1 and JERS-1 radars, to vegetation biomass is discussed and compared with the
best results achieved using multifrequency polarimetric JPL-AIRSAR data. Experi-
mental results show that measurements with JERS-1/L-band and ERS-1/C -band
SAR provide the means for detecting vegetation growth. In particular, the C -band
signal of ERS radar was found to be very well correlated to forest woody volume.
Figure 2.16: Vegetation class set table and Tubuai SAR vegetation classification.
At C band, the backscattering σ o is generally higher than at L band. In addition,

two-dimensional diagram shows that herbaceous crops (lowest σ o ) can be separated
from forests and vineyards (highest σ o ) and from bare soils (intermediate σ o ).
Previous investigations had pointed out that biomass retrieval could be satisfactorily
achieved only after a correct classification of agricultural fields since different plant
geometries produce different biomass sensitivities (Baronti et al., 1995 [10]). On the
other hand, classification was better accomplished by using multifrequency polarimet-
ric SAR measurements. Indeed, it has been shown that P -band data are effective in
discriminating between broad land surface categories, L-band data make possible
the identification of well-developed “broad leaf” crops (sunflower and corn), while
C -band data are useful for discriminating different kinds of herbaceous crops, even
in the case of moderate growth (Ferrazzoli et al., 1997 [40]).
It is possible to see that the trend is better accomplished at L- band HV polarization
for herbaceous crops and at P -band HV polarization for forests. Unfortunately, these
configurations are not available from satellites. As for C -band data from ERS, the
relationship between backscattering and herbaceous biomass depends on crop type:
once broad leaf crops are separated from small leaves, an appreciable sensitivity is
noted. C -band data are also useful for detecting forest biomass, especially if clear-cut
areas have previously been separated using both ERS and JERS data.
In Moreau and Le Toan, (2003[107]) ERS-SAR C-band VV data were analyzed as a
function of plant biomass at homogeneous totora and bofedal areas. Backscattering
coefficient was found to be related to the dry or humid biomass of rice (expressed in
kg/m2 ) by a logarithmic relationship:
o
σdb = a log n(biomass) + b (2.70)
where a and b are varying with the vegetation type or species.

In Luckman et al (1998)[97] image regions or polygons corresponding to the homoge-
neous forest areas sampled on the ground were located in SAR images. For all of these
polygons, the mean and variance in backscatter were extracted from the geocoded
JERS-1 data. The backscatter model defined was:
σ o = a − e−bB+c (2.71)
where B is the biomass density. a, b, c are constants, and the incidence angle is assumed
not to vary significantly. The parameter a corresponds exactly to the saturation value
of σ o and b describes the gradient of the low biomass-density part of the curve, and
C relates to the residual backscatter at zero biomass.
Forest parameter estimation using polarimetric SAR data

Many researchers reported algorithms to estimate forest parameters using polarimetric SAR
data (Le Toan et al., 1992[90]; Dobson et al., 1995[35]). However, these algorithms cannot
be applied on all types of forests without additional information on the forest type and
environmental conditions, since radar measurements depend on the tree structure and envi-
ronmental conditions.
Estimation of forest biomass using SAR data can be complicated by topography that in-
fluences radar backscatter (van Zyl, 1993[179]), particularly through local incidence angle,
shadowing, and effects on radar backscattering can be complex.
It is necessary to develop a method to minimize the dependence of these three factors. On the
other hand, its responses from forest areas saturate as biomass increases (Imhoff, 1995[72]).
A successful algorithm must reduce the effects caused by these problems. If forest biomass
is estimated and the forest type is known, tree heights and other forest parameters can be
derived using allometric equations.
The polarimetric parameters can be divided into two classes: The absolute polarimetric
parameters are polarimetric backscattering cross sections and eigenvalues. The relative
parameters include the HH and VV correlation coefficient, entropy, anisotropy, and the
radar vegetation index. Relative polarimetric parameters are less sensitive to both incidence
angle and environmental condition effects. Polarimetric backscattering cross sections (σpqrs )
can be written as:
σpqrs = F (ε) cos θi Ipqrs (2.72)
where F (ε) is a factor that represent environmental changes and Ipqrs is a polarimetric
measure that does not depend on the incidence angle and the environmental changes. This
means that the relative polarimetric parameters are not very sensitive to the incidence angle
and environmental changes.
The radar vegetation index (RVI) is defined as Eq. (2.73):
8(ρhv )
RV I = (2.73)
(ρhh + ρvv + 2ρhv )
where ρhvhv is the cross-polarization backscattering cross section and ρhhhh and ρvvvv are
co-polarization backscattering cross sections. The backscattering cross sections saturate as
forest biomass increases. Relative polarimetric parameters saturate faster than the absolute
polarimetric parameters.
Polarimetric responses from a forest could vary significantly depending upon the tree struc-
ture. For instance, Kim and van Zyl, (1999 [86]) examined a classification technique based
on the strength of three dominant scattering mechanisms: surface scattering, double bounce
scattering, and randomly oriented thin cylinders scattering. Their polarimetric responses
are shown in the following six equations:
Surface scattering conditions:
Re(σhhvv ) > σhvhv (2.74)
σvvvv > σhhhh (2.75)
Double bounce scattering conditions:
−Re(σhhvv ) > σhvhv (2.76)
σvvvv < σhhhh (2.77)
Randomly oriented thin cylinders scattering conditions:
σvvvv = σhhhh (2.78)
σhhhh
σhhvv < σhvhv = (2.79)
3
In Saatchi and Moghaddam (1999[147]) it was studied the use of a semi-empirical

algorithm for estimating boreal forest type biomass from polarimetric SAR measure-
ments. The algorithm is based on a two layer radar backscatter model that stratifies
the forest canopy into crown and stem layers and separates the structural and biomet-
ric attributes of forest stands. In this study, the authors use an alternative approach,
by first estimating the forest canopy moisture content and then using conversion fac-
tors between dry and wet weight to estimate the above ground woody biomass. The
estimation of canopy water content is performed by using a semi-empirical model for
boreal type forests developed by Saatchi and Moghaddam (1999)[147]. The model is
based on analytical simplifications of a two layer forest backscatter model in order to
separate structural and dielectric parameters in forest crown and stem layers.
It is assumed that over forest canopies, the total backscattering coefficients consist of
three dominant scattering mechanisms: crown volume scattering, crown-ground scat-
tering, and trunk-ground scattering. Fig. 2.17 illustrates the first order (no multiple
scattering) dominant scattering contributions.
In general, the total backscattering coefficient measured by SAR is given by:
o o o o
σpq = σpqc + σpqcg + σpqtg (2.80)
where p and q represent the electromagnetic wave polarizations of the received and
transmitted radar signals, and c, cg, and tg represent the crown, crown-ground, and
trunk- ground scattering mechanisms respectively. In the above expression, they have
assumed that the forest canopy consists of two layers (crown and trunk) and the direct
contribution from the soil surface is small compared to other scattering mechanisms.
Vegetation features estimation using interferometric SAR data

Although the interest in interferometry as a tool to study forests is more recent, forest pa-
rameters also have been estimated using SAR interferometry. The interferometric corre-
lation coefficient has been used for estimating tree heights (Rodriguez et al.,[140]; Treuhaft
et al., 1996[167]). The fundamental concept of this approach is that the interferometric cor-
relation will decrease if more scattering objects are present in a pixel. A simple parameter
to measure this decorrelation is the angular scattering range of a pixel to an interferometric
antenna. However, in order to relate the tree height to the interferometric correlation coef-
ficient accurately, one must assume a scattering profile within a pixel. Since several simple
functions have been used to describe the scattering profile, it is necessary to understand the
variation of estimated parameters for these different scattering profiles. Most of the existing
models for radar backscatter of forests are based on the Radiative Transfer Theory which
cannot be used to interpret the phase information. However, some insights on the depen-
dency of interferometric data to forest parameters can be derived from the these theoretical
models. Coherence has been shown to be efficient to discriminate forest areas from clear-cuts
in a temperate forest environment. The coherence is generally low over forested areas and
high over open fields which makes discrimination between forest and non-forest possible.
Figure 2.17: The first order (no multiple scattering) dominant scattering contributions over
forest canopies.
Chapter 3
Conclude remarks
Vegetation can be divided, in a general perspective, into plant foliage (with reflectance
patterns determined by pigments, water, carbon and nitrogen); canopy (characterization
depending on amount of vegetation and architecture, LAI, LAD) and non photosynthetic
vegetation (dead or dormant vegetation with spectral signature similar to mineral objects).
Vegetation has a typical spectral signature originated by its characteristics structural com-
pounds. Stress or senescent vegetation with greater amounts of carotenoids and anthocyanin,
and decreasing amount of chlorophyll concentration, can determine a shift of the red edge
into shorter wavelengths. Simultaneously, water content in vegetation affects the reflectance
response in the N IR and shortwave infrared region of the spectrum, whereas cellulose and
lignin influences the reflectance in shortwave infrared. All these observations drive to the
development of indices that analyze and summarize characteristics of the vegetation. Most
formulas of vegetation indices are based on ratios or linear combination and exploit dif-
ferences in the reflectance patterns of green vegetation and other objects. As ratios, they
can be easily cross-calibrated across sensor systems, ensuring continuity of data sets for
long-term monitoring of the land surface and climate-related processes. In general terms
optical vegetation indices can be divided into the follows categories: indices based on the
normalized difference (improving the linearity); soil–line vegetation indices (improving the
resistance to soil and atmospheric effects) and vegetation indices based on more than two
discrete bands. Another classification is about the particular characteristic that the indice
resumes: structural, chlorophyll or pigments, water content, etc.
Conventional visible-near infrared indices have the benefit of present both frequent coverage
and high spatial resolution. However, it is possible to mention two challenges: the effects of
the atmosphere and the background (surface underlying the vegetation) signals. Indices that
try to reduce the background effect can be increase their sensitivity to atmosphere variations.
N DV I, obtained from optical sensors, was one of the most successful of many attempts to
simply and quickly identify vegetated areas and their ”condition”, and it remains the most
well-known and used index to detect live green plant canopies in multispectral remote sens-
ing data. Modifications of the N DV I have been around, for instance, in order to account
for the soil background such as the SAV I or the W DV I. Moreover, new indexes have been
developed to compensate for both the effect of the atmosphere and soil, particularly the ‘en-
hanced’ vegetation index (EV I) improved sensitivity in high biomass regions and therefore
vegetation monitoring through a decoupling of the canopy background signal and a reduc-
tion in atmospheric influences. On the other hand, SAV I and P V I would proportionate
better LAI and percent cover estimations.
Multiple combinations of vegetation structural and functional characteristics produce similar
reflectance signatures, as have been observed in multi-spectral remote sensing instruments.
Also, at the beginning of the vegetation season, thin vegetation covers have a tendency
of being over-estimated by the vegetation indices, probably due to intense chlorophyllian
activity during early growth stages. In this sense, hyperspectral observations provide addi-
tional degrees of freedom that allow improved resolution of vegetation land cover maps and
35
36 CHAPTER 3. CONCLUDE REMARKS
estimates of physiological properties e.g., LAI, f P AR, Chla + b and H2 O. Imaging spec-
troscopy opens the door to measurements that are directly indicative of ecosystem structure
and functioning, such as chlorophyll concentration and dry C loading.
For global vegetation characterization, lower-resolution space-borne sensors operating in the
microwave part of the spectrum have triggered less interest than their visible and N IR coun-
terparts. However, these instruments have also shown some ability to characterize the land
surface at spatial resolutions compatible with climatological applications. Passive and ac-
tive microwave observations respond to the absorption/emission and scattering by vegetation
elements (dielectric properties, density, size), including woody parts, and are not directly
sensitive to the photosynthesis activity. Optical systems, which proved their effectiveness
in sensing leaf parameters, are sensitive to molecular resonance and scatter at micrometer
scale; and therefore are not able to sense the woody biomass. Microwave measurements
therefore complement the VIS-NIR observations by indicating structural density. In this
way, diverse authors suggest that optical imagery obtained coincident with SAR imagery
allowed a better understanding of the interaction of the SAR signal.
Active microwave systems are useful because they are not constrained by either night (dark-
ness) or cloud cover. Particularly, SAR is sensitive to plant morphology, geometric (struc-
ture) and dielectric properties (due principally to moisture content). To accurately predict
vegetation backscattering, physical models of a canopy should take into consideration several
details of soil, agricultural crops, and forests such as the density, dimension, orientation, and
shape of stems, leaves, branches, and trunks. Estimation of forest biomass using SAR data
can be complicated by topography that influences radar backscatter, particularly through
local incidence angle, shadowing, and effects on radar backscattering can be complex. It
is necessary to develop a method to minimize the dependence of these three factors. On
the other hand, its responses from forest areas saturate as biomass increases. A successful
algorithm must reduce the effects caused by these problems. If forest biomass is estimated
and the forest type is known, tree heights and other forest parameters can be derived using
allometric equations.
It is important to consider that many reported algorithms to estimate forest parameters
using polarimetric SAR data cannot be applied on all types of forests without additional
information on the forest type and environmental conditions, since radar measurements
depend on the tree structure and environmental conditions.
As final recommendations in order to select a methodology for vegetation analysis it would
be appropriate to consider the following questions:
At which scale we want to work? (local, regional, global, etc). This will determine the
spatial resolution of the necessary satellite data.
What is the vegetation characteristic that must be analyzed? (i.e. Structural or ‘gree-
ness’, Chlorophyll or pigments, Water content; woody parts, geometry, vegetation
types). There exist indices calculated from the spectral optical range, which empha-
size particularly the estimation of density, or specific pigments as chlorophyll, or water
content or photosynthetic activity. Hyperspectral sensor permit to evaluate for exam-
ple: the spatial patterns of vegetation type, productivity, and physiological activity.
Optical sensors can only monitor a very thin layer of the canopy, then they cannot pro-
vide information on woody biomass. On the other hand, vegetation microwave analysis
allow the estimation of vegetation elements, including woody parts, tree height, tree
structure, biomass, volume, crown complexity, leaf area index.
There exists indices that consider in a more complete way the compensation for both
the effect of the atmosphere and soil background? (alternative indices). For instance,
NDVI can be affected by soil reflectance; and soil adjusted V Is (such as SAVI or
WDVI) have been developed to take into account the local soil reflectance.
It is available data of an optical (i.e. multiespectral, hyperspectral) or microwave (i.e.

passive or active) sensor?
37
If optical data is available, what bands or wavelengths are present? (i.e. in the range
of blue, green, red or hyperspectral)
If radar data is available, what range of microwave spectrum is present? (i.e. band L,
C, X, etc).
It would be adequate to select two diverse not functionally equivalent indices in order
to complement the information of both approaches. For instance, NDVI has been the
most recognized and used index by the scientific community. Nevertheless, the same
one presents diverse issues as: the nonlinearity of ratio-based indexes, the influence of
the atmosphere, scaling problems, asymptotic (saturated) signals over high biomass
conditions, sensitivity to background variations, sensitivity to clouds, sensitivity to soil
effects, sensitivity to anisotropic effects and sensitivity to spectral effects. Therefore,
the complementacion of this accepted index with others that are constructed in an
alternative way might making possible well-fundamented conclusions.
38 CHAPTER 3. CONCLUDE REMARKS
Bibliography
[1] Adams, M., Philpot, W., and W.A., N. Yellowness index: an application of
spectral second derivatives to estimate chlorosis of leaves in stressed vegetation. Int.
J. Remote Sens. 20 (1999), 3663–3675. 8
[2] Altese, E., Bolognani, O., Mancini, M., and Troch, P. A. Retrieving soil
moisture over bare soil from ers 1 synthetic aperture radar data: sensitivity analy-
sis based on a theoretical surface scattering model and field data. Water Resources
Research 32 (1996), 653–661. 29
[3] Ashburn, P. The vegetative index number and crop identification. In The LACIE
Symposium, Proceedings of the Technical Session (1978), pp. 843–856. 13
[4] Asner, G. Biophysical and biochemical sources of variability in canopy reflectance.
Remote Sensing of Environment 64 (1998), 234–253. 3
[5] Asrar, G., Fuchs, M., Kanemasu, E., and Hatfield, J. Estimating absorbed
photosynthetic radiation and leaf area index from spectral reflectance in wheat. Agron,
J. 76 (1984), 300–306. 2, 20
[6] Asrar, G., Myneni, R. B., and Choudhury, B. J. Spatial heterogeneity in veg-
etation canopies and remote sensing of absorbed photo synthetically active radiation:
A modeling study. Remote Sensing of Environment 41 (1992), 85−103. 20
[7] Attema, E. P. W., and Ulaby, F. T. Vegetation modeled as a water cloud. Radio
Science 13, 2 (1978), 357–364. 29
[8] Baret, F., and Guyot, G. Potentials and limits of vegetation indices for lai and
apar assessment. Remote Sensing of Environment 35 (1991), 161−173. 12, 14, 20
[9] Baret, F., Guyot, G., and Major, D. J. Tsavi: A vegetation index which
minimizes soil brightness effects on lai and apar estimation. Proc. IGARSS’89 and
12th Canadian Symposium on Remote Sensing, Vancouver, Canada (1989), 1355 –
1358. 14
[10] Baronti, S., Del Frate, F., Ferrazzoli, P., Paloscia, S., Pampaloni, P.,
and Solimini, D. Sar polarimetric features of agricultural areas. Int. J. Remote
Sensing (1995), 2639–2656. 32
[11] Bauer, M. The role of remote sensing in determining the distribution and yield of
crops. Advances in Agron. 27 (1975), 271–304. 8
[12] Becker, F., and Choudhury, B. J. Relative sensitivity and microwave polarization
difference index for vegetation and desertification monitoring. Remote Sens. Environ.
24 (1988), 297– 331. 24
[13] Begue, and Myneni. 20
[14] Bowman, W. D. The relationship between leaf water status, gas exchange, and
spectral reflectance in cotton leaves. Remote Sensing of Environment 30 (1989), 249–
255. 10
39
40 BIBLIOGRAPHY
[15] Broge, N. H., and Leblanc, E. Comparing prediction power and stability of
broadband and hyperspectral vegetation indices for estimation of green leaf area index
and canopy chlorophyll density. Remote Sens. Environ 76 (2000), 156 – 172. 17, 20
[16] Carter, G. A. Primary and secondary effects of water content on the spectral
reflectance of leaves. American Journal of Botany 78 (1991), 916–924. 10
[17] Chappelle, E., Kim, M., and McMurtrey, I. Ratio analysis of reflectance spectra
(rars): An algorithm for the remote estimation of the concentrations of chl a, b and
carotenoids in soybean leaves. Remote Sensing Environ. 39 (1992), 239–247. 22
[18] Chen, J. Evaluation of vegetation indices and modified simple ratio for boreal appli-
cations. J. Remote Sens 22 (1996), 229 – 242. 13
[19] Chen, J., and Cihlar, J. Retrieving leaf area index of boreal conifer forests using
landsat tm images. Remote Sensing of Environment 55 (1996), 153−162. 20
[20] Choudhury, B. J. Relationship between vegetation indices, radiation absorption, and
net photosynthesis evaluated by a sensitivity analysis. Remote Sensing of Environment
22 (1987), 209−233. 20, 24
[21] Choudhury, B. J. Monitoring global land surface using nimbus-7 37 ghz data; theory
and examples. Int.j. Remote Sens. 10 (1989), 1579–1605. 23
[22] Clark, R. N., King, T., Ager, C., and Swayze, G. Vegetation species and stress
indicator mapping in the san luis valley, colorado using imaging spectrometer data.
Remote Sensing of Environment (1997). 8
[23] Cohen, W. B., Maierpserger, T. K., Gower, S. T., and Turner, D. P.
An improved strategy for regression of biophysical variables and landsat etm+data.
Remote Sensing of Environment 84 (2003), 561−571. 20
[24] Condit, H. R. The spectral reflectance of american soils. Photogramm. Eng. 36
(1970), 955–966. 4
[25] Cracknell, A. P. The exciting and totally unanticipated success of the avhrr in
applications for which it was never intended. Advances in Space Research 28 (2001),
233–240. 10
[26] Crippen. 14
[27] Crist, E. P., and Cicone, R. C. Application of the tasseled cap concept to
simulated thematic mapper data. Photogramm. Eng. Remote Sens. (1984). 16
[28] Datt, B. Remote sensing of chlorophyll a, chlorophyll b, chlorophyll a 1 b, and total
carotenoid content in eucalyptus leaves. Remote Sensing of Environment 66 (1998),
111–121. 17
[29] Daughtry, C. Discriminating crop residues from soil by shortwave infrared re-
flectance. Agron. J. 93 (2001), 125–131. 4, 5
[30] Daughtry, C. S. T., Walthall, C. L., Kim, M. S., Brown de Colstoun, E.,
and McMurtrey III, J. E. Estimating corn leaf chlorophyll concentration from leaf
and canopy reflectance. Remote Sens. Environ. 74 (2000), 229 – 239. 16, 20
[31] Deering, D. W., Rouse, J. W., Haas, R. H., and Sehell, J. A. Measuring
’forage production’ of grazing units from landsat mss data. In Proceedings of the 10th
International Symposium Remote Sensing of Environment (1975), vol. 2, pp. 1169–
1178. 13
[32] DeFries, R., Field, C., Fung, A., Justice, C., Los, S., Matson, P.,
Matthews, M., Mooney, H., Potter, C., Prentice, K., Sellers, P., Town-
shend, J., Tucker, C., Ustin, S., and Vitousek, P. Mapping the land surface for
global atmosphere biopshere models: toward continuous distributions of vegetation’s
functional properties. J. Geophysical Research 100, 20 (1995), 867–882. 1
BIBLIOGRAPHY 41
[33] Defries, R. S., and Townshend, J. R. G. Global land cover characterization

from satellite data: From research to operational implementation. Global Ecology and
Biogeography 8, 5 (1999), 367–379. 21
[34] Demetriades-Shah, T., Steven, M., and Clark, J. High resolution derivative
spectra in remote sensing. Remote Sens. Environ 33 (1990), 55–64. 22
[35] Dobson, M. C., Ulaby, F. T., Pierce, L. E., Sharik, T. L., Bergen, K. M.,
Kellndorfer, J., Kendra, J. R., Li, E., Lin, Y. C., Nashashibi, A., Sara-
bandi, K., and Siqueira, P. Estimation of forest biophysical characteristics in
northern michigan with sir-c/x- sar. IEEE Trans. Geosci. Remote Sens 33, 4 (1995),
877–895. 32
[36] Elvidge, C., and Chen., Z. Comparison of broad-band and narrow-band red and
near-infrared vegetation indices. Remote Sens. Environ 54 (1995), 38–48. 22
[37] Elvidge, C., and R. Lyon, . Influence of rock-soil spectral variation on the assess-
ment of green biomass. Remote Sensing of Environment 17 (1985). 22
[38] ERS-ENVISAT Symposium, Göteborg. An operational forest mapping tool using

spaceborne SAR data (2000). 31
[39] Ferrazzoli, P., Paloscia, S., Pampaloni, P., Schiavon, G., Solimini, D.,
and Coppo, P. Sensitivity of microwave measurements to vegetation biomass and
soil moisture content: A case study. IEEE Transactions on Geoscience and Remote
Sensing 30 (1992), 750–756. 2
[40] Ferrazzoli, P., Paloscia, S., Pampaloni, P., Schiavon, P., Sigismondi, S.,
and Solimini, D. The potential of multifrequency polarimetric sar in assessing agri-
cultural and arboreous biomass. IEEE Transactions on Geoscience and Remote Sens-
ing 35 (1997), 5–17. 2, 32
[41] Friedl, M. A., Davis, F. W., Michaelsen, J., and Moritz, M. A. Scaling and
uncertainty in the relationship between ndvi and land surface biophysical variables:
An analysis using a scene simulation model and data from fife. Remote Sensing of
Environment 54 (1995), 233−246. 20
[42] Frison, X., and Mougin, x. 23
[43] F.T. Ulaby, R.K.Moore, A. F. Microwave Remote Sensing, vol. 2. Addison Wesley,
Reading, MA,, 1982. 630–634. 25
[44] Fung. 20
[45] Fung, A. K., Li, Z., and Chen, K. S. Backscattering from a randomly rough
dielectric surface. IEEE Transactions on Geoscience and Remote Sensing 30 (1992),
356–369. 29
[46] Gamon, J. A., Field, C. B., Roberts, D. A., Ustin, S. L., and Valentini,
R. Functional patterns in an annual grassland during and aviris overflight. Remote
Sensing of the Environment 44, 2-3 (1993), 239–253. 22
[47] Gamon, J. A., Penuelas, J., and Field, C. B. A narrow-waveband spectral

index that tracks diurnal changes in photosynthetic efficiency. Remote Sens. Environ.
41 (1992), 35–44. 19
[48] Gamon, J. A., and Qiu, H. L. Handbook of Functional Plant Ecology. Marcel
Dekker, Inc., New York, 1999, ch. Ecological applications of remote sensing at multiple
scales, p. 805–846. 19
[49] Gao, B. Ndwi—a normalized difference water index for remote sensing of vegetation
liquid water from space. Remote Sensing of Environment 58 (1996), 257 – 266. 18
42 BIBLIOGRAPHY
[50] Gausman, H. W., Gerbarman, A. H., and Wiegand, C. L. Use of erts — 1

to detect chlorotic grain sorghum. Photogram. Eng. and Remote Sensing 41 (1975),
177–181. 5
[51] Gilabert, M. A., González-piqueras, J., Garcı́a-haro, F. J., and Meliá,
J. A generalized soil-adjusted vegetation index. Remote Sensing of. Environment 82
(2002), 303−310. 14
[52] Gitelson, A., and Merzlyak, M. N. Quantitative estimation of chlorophyll using
re ectance spectra: Experiments with autumn chestnut and maple leaves. Journal of
Photochemistry and Photobiology (B). 247–252. 17
[53] Gitelson, A. A., Kaufman, Y., and Merzlyak, M. Use of a green channel in
remote sensing of global vegetation from eos-modis. Remote Sensing of Environment
58 (1996), 289−298. 17, 20
[54] Haboudane, D., Miller, J. R., Pattey, E., Zarco-Tejada, P. J., and Stra-
chan, I. Hyperspectral vegetation indices and novel algorithms for predicting green
lai of crop canopies: Modeling and validation in the context of precision agriculture.
Remote Sensing of Environment 90 (2004), 337−352. 16, 17, 20
[55] Haboudane, D., Miller, J. R., Tremblay, N., Zarco-tejada, P. J., and Dex-
traze, L. Integrated narrow-band vegetation indices for prediction of crop chlorophyll
content for application to precision agriculture. Remote Sensing of. Environment 81
(2002), 416−426. 16
[56] Hardinsky, M. A., and Lemas, V. e. a. The influence of soil salinity, growth
form, and leaf moisture on the spectral reflectance of spartina alternifolia canopies.
Photogrammetric Engineering and Remote Sensing 49 (1983), 77–83. 19
[57] Hatfield, J. L., Vauclin, M., Vieira, S. R., and Bernard, R. c. Surface tem-
perature variability patterns within irrigated fields. Agricultural Water Management
8 (1984), 429–437. 20
[58] Henderson, T., Baumgardner, M., Franzmeier, D., Stott, D., and Coster,
D. High dimensional reflectance analysis of soil organic matter. Soil Science Society
of America Journal 56, 3 (1992), 865–872. 5
[59] Holben, B. N., Kaufman, Y. J., and Kendall, J. D. Noaa-11 avhrr visible and
near-ir inflight calibration. Int. J. Remote Sens., forthcoming (1990). 21
[60] Horler, D., Dockray, M., and J., B. The red edge of plant leaf reflectance.
International Journal of Remote Sensing 4 (1983), 273–288. 22
[61] Huete, A. A soil adjusted vegetation index (savi). Remote Sensing of Environment
25 (1988), 295−309. 14, 20, 22
[62] Huete, A., Didan, K., and Miura, T. Overview of the radiometric and biophysical
performance of the modis vegetation indices. Remote Sens. Environ 83 (2002), 195–
213. 21
[63] Huete, A., Didan, K., van Leeuwen, W., Miura, T., and Glenn, E. Land
Remote Sensing and Global Environmental Change: NASA’s Earth Observing System
and the Science of ASTER and MODIS 2008. 2008, ch. MODIS vegetation indices.
15
[64] Huete, A., Liu, H. Q., Batchily, K., and Van Leeuwen, W. A comparison of
vegetation indices over a global set of tm images for eos-modis. Remote Sens. Environ.
59, 3 (1997), 440–451. 22
[65] Huete, A., and Tucker, C. Investigation of soil influences in avhrr red and near-
infrared vegetation index imagery. International Journal of Remote Sensing 12 (1991),
1223–1242. 22
BIBLIOGRAPHY 43
[66] Huete, A. R., Jackson, R. D., and Post, D. F. Spectral response of a plant
canopy with different soil backgrounds. Remote Sensing of Environment 17 (1985),
37−53. 22
[67] Hunt, E. R., and Rock, B. N. Detection of changes in leaf water content using
near- and middle-infrared reflectances. Remote Sensing of Environment 30 (1989), 43
– 54. 19
[68] Idso, S., Jackson, R., P.J., P., Reginato, Jr., R., and Hatfield, J. Nor-
malizing the stress degree day parameter for environmental variability. Agricultural
Meteorology 24 (1981), 45–55. 26
[69] Idso, S., Jackson, R. D., and Reginato, R, J. Remote sensing of crop yields.
Science 196 (1977), 19–25. 26
[70] Idso, S. B., Reginato, R. J., and Farah, S. M. Soil-and atmosphere-induced

plant water stress in cotton as inferred from foliage temperatures. Water Resour Res
18 (1982), 1143– 1148. 27
[71] Imeson, A. C., and Lavee, H. Soil erosion and climate change: the transect ap-
proach and the influence of scale. Geomorphology 23 (1998), 219–227. 1
[72] Imhoff, M. C. Radar backscatter and biomass saturation: Ramifications for global
biomass inventory. IEEE Trans. Geosci. Remote Sens 33, 2 (1995). 32
[73] Inoue, Y., Kurosu, T., Maeno, H., Uratsuka, S., Kozu, T., Dabrowska-
Zielinska, K., and Qi, J. Season-long daily measurements of multifrequency (ka,
ku, x, c and l) and full-polarization backscatter signatures over paddy rice field and
their relationship with biological variables. Remote Sens. Environ 81 (2002), 194–204.
2
[74] Jackson, R. D. Spectral indices in n-space. Remote Sens. Environ 14 (1983), 409–
421. 2
[75] Jackson, R. D., Idso, S., Reginato, R., and Pinter, P. Canopy temperature
as a crop water stress indicator. Water Resources Research 17 (1981), 1133–1138. 26
[76] Jackson, T. J. Measuring surface soil moisture using passive microwave remote
sensing. Hydrol. Processes 7, 139-152 (1993), 239–253. 24
[77] Jackson, T. J., Le Vine, D. M., Swift, C. T., Schmugge, T. J., and Schiebe,
F. R. Large area mapping of soil moisture using the estar passive microwave ra-
diometer in washita’92. Remote Sensing of Environment 53, 27-37 (1995), 239–253.
24
[78] Jackson, T. J., and Schmugge, T. J. Vegetation effects on the microwave emission
of soils. Rem. Sensing of Env 36 (1991), 203–212. 12
[79] Jordan, C. Derivation of leaf-area index from quality of light on the forest floor.
Ecology 50 (1969), 663–666. 12, 13
[80] Justice, C. O., and Hiernaux, P. H. Y. Monitoring the grasslands of the sahel
using noaa avhrr data: Niger 1983. International Journal of Remote Sensing 11 (1986),
1475–1497. 1
[81] Kaufman, Y. J., and Holben, B. N. Calibration of the avhrr visible and near-
ir bands by atmospheric scattering, ocean glint and desert reflection. International
Journal of Remote Sensing 14 (1993), 21−52. 21
[82] Kaufman, Y. J., and Tanre, D. Atmospherically Resistant Vegetation Index

(ARVI) for EOS-MODIS. IEEE Transactions on Geoscience and Remote Sensing
30 (1992), 261–270. 15
44 BIBLIOGRAPHY
[83] Kaufman, Y. J., and Tanre, D. Strategy for direct and indirect methods for
correcting the aerosol effect on remote sensing: from avhrr to eos-modis. Remote
Sensing of Environment 55 (1996), 65−79. 22
[84] Kauth, R. J., and Thomas, G. S. The tassel cap-a graphic description of the
spectral-temporal development of agricultural crops as seen by landsat. In Proceedings
of the Symposium on Machine Processing of Remotely Sensed Data (1976), U. W.
L. I. LARS, Purdue, Ed., no. 76 in IEEE Catalogue No. 76, Ch. 1103-1 MPRSD,
pp. 1169–1178. 16
[85] Kennedy, P. J. Monitoring the phenology of tunisian grazing lands. International
Journal of Remote Sensing 10 (1989), 835–845. 1
[86] Kim, Y.-J., van Zyl, J., and Chu, A. Preliminary results of single pass polari-
metric sar interferometry. IEEE 1999 International Geoscience and Remote Sensing
Symposium, Hamburg, Germany, (1999). 33
[87] Kimes, D. S., Nelson, R. F., Manry, M. T., and Fung, A. K. Attributes of
neural networks for extracting continuous vegetation variables from optical and radar
measurements. International journal of remote sensing 19, 14 (1998), 2639 –2663. 22
[88] Knipling, E. B. Physical and physiological bases for the reflection of visible and
near-infrared radiation from vegetation. Remote Sensing Environ. 1 (1970), 155–159.
8
[89] Lang, M., Kuusk, A., Nilson, T., Lukk, T., Pehk, M., and Alm, G. Re-
flectance spectra of ground vegetation in sub-boreal forests. Tartu Observatory, Esto-
nia. (2002). 8
[90] Le Toan, T., Beaudoin, A., Riom, J., and Guyon, D. Relating forest biomass
to sar data. IEEE Trans. Geosci. Remote Sens 30, 2 (1992), 403–411. 29, 32
[91] Le Toan, T., Laur, H., and Mougin, E. Multitemporal and dual-polarization
observations of agricultural vegetation covers by x-band sar images. IEEE Transactions
on Geoscience and Remote Sensing 27 (1989), 709–718. 2
[92] Le Toan, T., Ribbes, F., Wang, L. F., Floury, N., Ding, K. H., Au Kong,
J., Masaharu, F., and Kurosu, T. Rice crop mapping and monitoring using ers-1
data based on experiment and modeling results. IEEE Transactions on Geoscience
and Remote Sensing 35, 1 (1997), 41–56. 2
[93] Leckie, D. G., and Ranson, K. J. Forestry applications using imaging radar.
Principles and Applications of Imaging Radar, Vol. 23 (1998), 435–509. 30
[94] Lillesand, and Kiefer. 12
[95] Lillesand, T. M., and Kiefer, R. W. Remote Sensing and Image Interpretation.
John Wiley and Sons, NY, 1979. 1
[96] Liu, H., and Huete., A. A feedback based modification of the ndvi to minimize
canopy background and atmospheric noise. IEEE Trans. Geosci. Remote Sens. 33
(1995), 457–465. 22
[97] Luckman, A., Baker, J., Honzak, M., and Lucas, R. Tropical forest biomass
density estimation using jer sar: Seasonal variation, confidence limits, and application
to image mosaics. Remote Sensing of Environment 63 (1998), 126. 1, 32
[98] Lyon, J. D., Yuan, G., Lunetta, R. S., , and Elvidge, C. D. A change detec-
tion experiment using vegetation indices. Photogrammetric Engineering and Remote
Sensing 64 (1998), 143–150. 22
[99] Mahan, J., and Upchurch, D. Maintenance of constant leaf temperature by
plants. 1. hypothesis limited homeothermy. Environmental and Experimental Botany
28 (1988), 351–357. 27
BIBLIOGRAPHY 45
[100] Major, D. J., Baret, F., and Guyot, G. A ratio vegetation index adjusted for
soil brightness. Int. J. Remote Sens 11 (1990), 727–740. 22
[101] Matthews, E. Global vegetation and land use: New high-resolution data bases for
climate studies. J. Clim. Appl. Meteorol 22 (1983), 474–486. 1
[102] McGwire, K., Minor, T., and Fenstermaker, L. Hyperspectral mixture mod-
eling for quantifying sparse vegetation cover in arid environments. Remote Sensing
Environ. 72 (2000), 360–374. 22
[103] Min, Q., . L. B. Remote sensing of evapotranspiration and carbon uptake at harvard
forest. Remote sensing of environment 100 (2006), 379−387. 25
[104] Moran, M. Irrigation management in arizona using satellites and airplanes. Irrigation
Science 15 (1994), 35–44. 27
[105] Moran, M. S., Hymer, D. C., Qi, J., and Kerr, Y. Comparison of ers-2 sar and
landsat tm imagery for monitoring agricultural crop and soil conditions. Remote Sens.
Environ. 79 (2002), 243–252. 2, 29
[106] Moran, M. S., Vidal, A., Troufleau, D., Inoue, Y., and Mitchell, T. A.
Ku- and c-band sar for discriminating agricultural crop and soil conditions. IEEE
Transactions on Geoscience and Remote Sensing 36 (1998), 265– 272. 29
[107] Moreau, S., and Le Toan, T. Biomass quantification of andean wetland forages
using ers satellite sar data for optimizing livestock management. Remote Sensing of
Environment 84 (2003), 477– 492. 32
[108] Moulin, S., Kergoat, L., Viovy, N., and Dedieu, G. Global-scale assessment
of vegetation phenology using noaa/avhrr satellite measurements. Journal of Climate
10, 6 (1997), 1154–1170. 20
[109] Myneni, R. B., and Ganapol, B. D. Remote sensing of vegetation canopy photo-
synthetic and stomatal conductance efficiencies. Remote Sensing of. Environment 42
(1992), 217−238. 20
[110] Myneni, R. B., Hall, F. G., Sellers, P., and Marshak, A. L. The interpre-
tation of spectral vegetation indexes. IEEE Transactions on Geoscience and Remote
Sensing 33, 2 (1995), 481–486. 1, 2, 12, 20
[111] Myneni, R. B., Nemani, R. R., and Running, S. W. Estimation of global leaf
area index and absorbed par using radiative transfer models. IEEE Transactions on
Geoscience and Remote Sensing 35 (1997), 1380−1393. 20
[112] Myneni, R. B., Tucker, G. A., and Keeling, C. D. Interannual variations in

satellite-sensed vegetation index data from 1981 to 1991. J. Geophys. Res. 103 (1998),
6145–6160. 21
[113] Myneni, R. B., and Williams, D. L. On the relationship between fapar and ndvi.
Remote Sensing of Environment 49 (1994), 200−211. 20
[114] Njoku, E. G., and Li, L. Retrieval of land surface parameters using passive mi-
crowave measurements at 6-18 ghz. IEEE Transactions on Geoscience and Remote
Sensing 37, 1 (1999), 79–93. 24
[115] Oguro, Y., Suga, Y., Takeuchi, S., Ogawa, M., Konishi, T., and Tsuchiya,
K. Comparison of sar and optical sensor data for monitoring of rice plant around
hiroshima. Advances in Space Research 28 (2001), 195–200. 31
[116] Oh, Y., Sarabandi, K., and Ulaby, F. T. An empirical model and an inver-
sion technique for radar scattering from bare soil surfaces. IEEE Transactions on
Geoscience and Remote Sensing 30 (1992), 370–381. 29
46 BIBLIOGRAPHY
[117] Osborne, S., Schepers, J., Francis, D., and M.R., S. Use of spectral radiance
to estimate in-season biomass and grain yield in nitrogen- and water-stressed corn.
Crop Sci. 42 (2002), 165–171. 22
[118] Paloscia, S., Macelloni, G., Pampaloni, P., and Sigismondi, S. The potential
of c- and l-band sar in estimating vegetation biomass: The ers-1 and jers-1 experiments.
IEEE Trans.Geosci. Remote Sens 37 (1999), 2107–2110. 2, 31
[119] Paloscia, S., and Pampaloni, P. Microwave vegetation indexes for detecting
biomass and water conditions of agricultural crops. Remote Sens. Environ 40 (1992),
15–26. 24
[120] Paltridge, G., and Barber., J. Mornitoring grassland dryness and fire potential
in australia with noaa/avhrr data. Rem. Sens. of Environ (1988). 16
[121] Penuelas, J., Baret, F., and Filella, I. Semi-empirical indices to assess
carotenoids/chlorophyll a ratio from leaf spectral reflectance. Photosynthetica 31
(1995), 221–230. 20
[122] Penuelas, J., Filella, I., Biel, C., Serrano, L., and Save, R. The reflectance
at the 950 – 970 nm region as an indicator of plant water status. International Journal
of Remote Sensing 14, 10 (1993), 1887 – 1905. 19
[123] Penuelas, J., Fillela, I., Biel, C., Serrano, L., and Save, R. The reflectance
at the 950-970 region as an indicator of plant water status. International journal of
remote sensing 14 (1995), 1887–1905. 19
[124] Penuelas, J., Pinol, J., Ogaya, R., and Filella, I. Estimation of plant water
concentration by the reflectance water index wi (r900/r970). International Journal of
Remote Sensing 18 (1997), 2869–2875. 18, 19
[125] Perry, C. R., and Lautenschlager, L. F. Functional equivalence of spectral

vegetation indices. Remote Sens. Environ. 14 (1984), 169–182. 12, 13
[126] Pickup, G. Estimating the effects of land degradation and rainfall variation on
productivity in rangelands: an approach using remote sensing and models of grazing
and herbage dynamics. Journal of Applied Ecology 33 (1996), 819–832. 1
[127] Pickup, G., Chewings, V. H., and Nelson, D. J. Estimating changes in vegeta-
tion cover over time in arid rangelands using Landsat MSS data. Remote Sensing of.
Environment 43 (1993), 243–263. 15
[128] Pieper, R. D. Vegetation science applications for rangeland analysis and manage-
ment, Handbook of vegetation science. P. T. Tueller, Dordrecht, Holland, Kluwer, 1988,
ch. Rangeland vegetation productivity and biomass, pp. 449–469. 1
[129] Pinty, B., and Verstraete, M. M. Extracting information on surface proper-

ties from bidirectional reflectance measurements. Journal of Geophysical Research 96
(1991), 2865–2874. 16
[130] Price, J. C. On the information content of soil reflectance spectra. Remote Sens.
Environ. 29 (1990). 4
[131] Prigent, X., and Aires, X. 24, 27
[132] Purevdorj, T. S., Tateishi, R., Ishiyama, T., and Honda, Y. Relationships
between percent vegetation cover and vegetation indices. Int. J. Remote Sens 19
(1998), 3519–3535. 2
[133] Qi, J., Chehbouni, A., Huete, A. R., Keer, Y. H., and Sorooshian, S. A
modified soil vegetation adjusted index. Remote Sens. Environ. (1994). 14
BIBLIOGRAPHY 47
[134] Raney, R. K. Manual of Remote Sensing, 3 ed., vol. 2. F. M. Henderson and

A. J. Lewis, Eds., John Wiley and Sons, 1998, ch. Radar fundamentals: Technical
perspective. Principles and Applications of Imaging Radar, pp. 9–130. 2
[135] Ranson, K. J., and Sun, G. Mapping of boreal forest biomass from spaceborne
synthetic aperture radar. Journal of Geophysical Research 102, 24 (1997), 29599 –
29610. 29, 31
[136] Richardson, A., and Everitt, J. Using spectral vegetation indices to estimate
rangeland productivity. Geocarto In ternational (1992), 63–77. 12
[137] Richardson, A. J., and Wiegand, C. L. Distinguishing vegetation from soil

background information. Photogram. Eng. Remote Sens. 43 (1977), 1541–1552. 12
[138] R.J., H., Davidson, D. P., and Peddle, D. R. Ground and remote estimation of
leaf area index in rocky mountain forest stands, kananaskis, alberta, canadian. Journal
Of Remote Sensing 29, 3 (2003), 411–427. 22
[139] Roberts, D. A., Keller, M., and Soares, J. V. Studies of land-cover, land-
use and biophysical properties of vegetation in the large scale biosphere atmosphere
experiment in amazonia. Remote Sensing of Environment 87 (2003), 377–388. 1
[140] Rodriguez, E., Michel, T. R., and Harding, D. J. Interferometric measurement

of canopy height characteristics of coniferous forests. 34
[141] Rondeaux, G., and Baret, F. Optimization of soil-induced vegetation indices.

Remote Sensing of. Environment 55 (1996), 95−107. 14
[142] Rougean, J. L., and Breon, F. M. Estimating par absorbed by vegetation from
bidirectional reflectance measurements. Remote Sens. Environ. 51 (1995), 375 – 384.
13
[143] Rouse, J. W., Haas, R. H., Schell, J. A., and Deering, D. W. Monitoring
vegetation systems in the Great Plains with ERTS, vol. 1. Third ERTS Symposium,
NASA, 1973. 2
[144] Rouse, J. w., Haas, R. H., Schell, J. A., Deering, D. W., and Harlan, J. C.
Monitoring the vernal advancement and retrogradation (greenwave effect) of natural
vegetation. NASA/GSFC Type III Final Report (1974). Greenbelt, Maryland. 12, 13
[145] Running, S., Loveland, T. R., Pierce, L. L., Nemani, R. R., and Hunt, E. R.
A remote sensing based vegetation classification logic for global land cover analysis.
Remote Sensing of Environment 51 (1995), 39–48. 1
[146] S., J. Inversion of the prospect+sail canopy reflectance model from aviris equivalent
spectra: theoretical study. Remote Sens. Environ. 44 (1993), 281–292. 22
[147] Saatchi, S., and Moghaddam, M. Retrieval of forest canopy parameters from
polarimetric sar data: I. theory. IEEE Trans. Geosci. Remote Sensing (1999). 33
[148] Sano, E. E., Ferreira, L. G., and Huete, R. A. Synthetic aperture radar (l band)
and optical vegetation indices for discriminating the brazilian savanna physiognomies:
A comparative analysis. Earth Interactions 9, 15 (2005), 570–581. 30
[149] Sano, E. E., Moran, M. S., Huete, A. R., and Miura, T. C- and multiangle
ku-band synthetic aperture radar data for bare soil moisture estimation in agricultural
areas. Remote Sensing of Environment 64 (1998), 77– 90. 29
[150] Sellers, P., Berry, J. A., Collatz, G., Field, C., and Hall, F. Canopy re-
flectance, photosynthesis and transpiration, iii. a reanalysis using improved leaf models
and a new canopy integration scheme. Remote Sensing Environ. 42 (1992), 187–216.
2, 20
48 BIBLIOGRAPHY
[151] Sellers, P., Randall, D., Collatz, G., Berry, J., Field, C., Dazlich, D.,
Zhang, C., Collelo, G., and Bounoua, L. A revised land surface parameteri-
zation (sib2) for atmospheric gcms. part i: Model formulation. Journal of Climate 9
(1996), 676–705. 1, 20
[152] Sellers, P. J., and Schimel, D. Remote sensing of the land biosphere and biogeo-
chemistry in the eos era: science priorities, methods and implementation. Global and
Planetary Change 7, 4 (1993), 279–297. 1
[153] Sellers, P. J., Tucker, C. J., Collatz, G. J., Los, S. O., Justice, C. O.,
Dazlich, D. A., and Randall, D. A. A global 1-deg by 1-deg ndvi data set for
climate studies, 2, the generation of global fields of terrestrial biophysical parameters
from ndvi. Int. J. Remote Sens 15 (1994), 3519–3545. 1
[154] Serrano, L., penuelas, J., and Ustin, S. L. Remote sensing of nitrogen and
lignin in mediterranean vegetation from aviris data: decomposing biochemical from
structural signals. Remote sensing of Environment 81 (2002), 355–364. 20
[155] Shi, C. X. A study on soil moisture remote sensing data assimilation based on
ensemble kalman filter (enkf) (in chinese), ph.d. diss. Inst. of Atmos. Phys., Chin.
Acad. of Sci.Beijing (2008), 177. 25
[156] Shoshany, M. Satellite remote sensing of natural mediterranean vegetation: A review

within an ecological context. Progress in Physical Geography 24 (2001), 153–177. 1
[157] Shoshany, M., Kutiel, P., and Lavee, H. Seasonal vegetation cover changes as
indicators of soil types along a climatological gradient: a mutual study of environ-
mental patterns and controls using remote sensing. International Journal of Remote
Sensing 16 (1995), 2137–2151. 1
[158] Shupe, S. M., and Marsh, S. E. Cover- and density-based vegetation classifications
of the sonoran desert using landsat tm and ers-1 sar imagery. Remote Sensing of
Environment 93 (2004), 131–149. 31
[159] Smith, A. M., Major, D. J., McNeil, R. L., Willms, W. D., Brisco, B., and
Brown, R. J. Complementarity of radar and visible-infrared sensors in assessing
rangeland conditions. Remote Sens. Environ. 52 (1995), 173–180. 2
[160] Smith, M. O., Ustin, S. L., Adams, J. B., and Gillespie, A. R. Vegetation
in deserts: I. a regional measure of abundance from multispectral images. Remote
Sensing of Environment 31 (1990), 1–51. 1
[161] Stoll, B., and P. Capolsini, A. IGARSS04. Anchorage, Alaska, 2004, ch. Simple
Class-Set Based Vegetation Classification of a South Pacific Island (Moorea Island,
French Polynesia) Using Both AirSAR and MASTER Data. 30
[162] Stoner, E. R., and Baumgardner, M. Characteristic variation in reflectance of

surface soils. Soil. Sci. Soc. Am. J. 45 (1981), 1161–1165. 4
[163] Svoray, T., and Shoshany, M. Sar-based estimation of areal aboveground biomass
( aab ) of herbaceous vegetation in the semi-arid zone : a modi cation of the water-
cloud model. International Journal of Remote Sensing 23, 19 (2002), 4089–4100. 30
[164] Tappan, G. The Monitoring of Green Vegetation Cover in the Kansas Flint Hills
from Landsat Data. KARS, Univ. of Kansas, Lawrence, 1980. 8
[165] Thiam. 14
[166] Todd, S. W., and Hoffer, R. M. Responses to spectral indices to variations in

vegetation cover and soil background. Photogramm. Eng. Remote Sens 64 (1998),
915–921. 22
BIBLIOGRAPHY 49
[167] Treuhaft, R. N., Madsen, S. N., Moghaddam, M., and van Zyl, J. J. Veg-
etation characteristics and underlying topography from interferometric radar. Radio
Science 31, 6 (1996), 1449–1485. 34
[168] Tucker, C., Fung, Y., Keeling, C. D., and Gammon, R. H. Relationship
between atmospheric co2 variations and a satellite-derived vegetation index. Nature
319 (1986), 195–199. 2
[169] Tucker, C. J. Red and photographic infrared linear combination for monitoring
vegetation. Remote sensing Environ 8 (1979), 127–150. 2, 10, 20
[170] Tucker, C. J., Holben, B. N., Elgin, J. H., and McMurtrey, J. E. Relation-
ship of spectral data to grain yield variations. Photogramm. Eng. Remote. Sens. 46
(1980), 657 – 666. 20
[171] Tucker, C. J., Justice, C. O., and Prince, S. D. Monitoring the grasslands of
the sahel: 1984–1985. Remote Sensing of Environments 17 (1986), 235–249. 1
[172] Tucker, C. J., Townshend, J. R. G., and Goff, T. E. African land-cover

classification using satellite data. Science 227 (1985), 369–375. 20, 21
[173] Tueller, P. T. Remote sensing science applications in arid environments. Remote

Sensing of Environment 23 (1987), 143–154. 1
[174] Ulaby, F., Moore, R., and Fung, A. Microwave remote sensing–active and pas-
sive. volume iii, from theory to applications. Artech House (1986). 23
[175] Ulaby, F. T. Radar measurementsof soil moisture content. IEEE Transactions

Antennas Propagation 22 (1974), 257–265. 30
[176] Ulaby, F. T., Allen, C. T., Eger, G., and Kanemasu, E. Relating the mi-
crowave backscattering coefficient to leaf area index. Remote Sens. Environ 14 (1984),
113–133. 2, 29
[177] Ulaby, F. T., and Dobson, M. C. Handbook of Radar Statistics for Terrain,
vol. 27. Artech House, Inc., Dedham, Massachusetts, 1989, ch. Multitemporal and
dual-polarization observations of agricultural vegetation covers by X-band SAR im-
ages, pp. 709–718. 2
[178] Ulaby, F. T., Sarabandi, K., McDonald, K., White, M., and Dobson, M. C.
Michigan microwave canopy scattering model. International Journal of Remote Sens-
ing 11 (1990), 1223– 1253. 29
[179] van Zyl, J. The effect of topography on radar scattering from vegetated areas. IEEE
transactions on geoscience and remote sensing 31 (1993), 153–160. 32
[180] Walthall, C., Dulaney, W., Anderson, M., Norman, J., Fang, H., and
Liang, S. A comparison of empirical and neural network approaches for estimating
corn and soybean leaf area index from Landsat ETM+ imagery. Remote Sens. Environ
92 (2004), 465–474. 21
[181] Woodwell, G. M., Hobbie, J. E., Houghton, R. A., Melillo, J. M., Moore,
B., Park, A. B., Peterson, B. J., and Saver, G. R. The Role of Terrestrial
Vegetation in the Global Carbon Cycle: Measurement by Remote Sensing. SCOPE 23
(Woodwell, G. M., Ed.), Wiley, New York, 1984, ch. Measurement of changes in the
vegetation of the earth by satellite imagery, pp. 221–240. 1
[182] Zarco-Tejada, P. J.and Miller, J. R., Noland, T. L., Mohammed, G. H.,

and Sampson, P. H. Scaling-up and model inversion methods with narrow-band
optical indices for chlorophyll content estimation in closed forest canopies with hyper-
spectral data. IEEE Trans. Geosci. Remote Sens. 39, 7 (2001), 1491 – 1507. 17
50 BIBLIOGRAPHY
[183] Zarco-tejada, P. L. J., Rueda, C. A., and Ustin, S. L. Water content estimation
in vegetation with MODIS reflectance data and model inversion methods. Remote
Sensing of. Environment 85 (2003), 109−124. 18

Vegetation Analysis Using Remote Sensing

Uploaded by

Copyright:

Available Formats

Vegetation Analysis Using Remote Sensing

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vegetation Analysis Using Remote Sensing

Uploaded by

Copyright:

Available Formats

Argentinian Spatial Agency (CONAE)

Cordoba National University (UNC)

Vegetation analysis using remote sensing

Master in Emergency Early Warning and Response Space Applications.

Author: Biol. Lanfri Sofı́a

2 Vegetation analysis by remote sensing 3

Vegetation analysis by remote

Canopies Although leaf reflectance properties, controlled by features of pigments, water,

Non-Photosynthetic Vegetation: NPV Ecosystems can also contain senescent or dead

2.0.1 Spectral signatures

Figure 2.2: Absorption characteristics of different pigments in the visible range.

Figure 2.3: Characteristic absorption of different pigments in the visible range

Figure 2.4: Typical reflectance behavior of the vegetation components.

Spectral signatures of different types of surface coverage as a result of the experimentation

Figure 2.6: Spectral signatures of different types of soil coverage

Figure 2.7: Spectral reflectance of vegetation at different phisiological conditions.

2.1 Vegetation analysis from optical data (visible and

2.1.1 Vegetation indices summary

RVI Ratio vegetation index.

P V I = sin(a) ρN IR − cos(a) ρRED (2.3)

AVI Ashburn Vegetation Index.

Reference: Ashburn, 1978 [3]

Reference: Rouse et al., 1974 [144]

Reference: Jordan, 1969 [79]; Rouse et al., 1974 [144]

TTVI Thiam Transformed Vegetation Index.

Reference: Baret et al., 1989[9]; Baret and Guyot, 1991[8]

L = 1 − ( 2 a N DV I W DV I)W DV I = ρN IR − (aρRed ) (2.17)

W DV I is the Weighted Difference Vegetation.

Reference: Gilabert et al., 2002 [51]

VARIgreen Visible atmospherically resistant index

Reference: Gitelson et al. (2001)

where ρN IR ,ρRed ,and ρBlue are atmospherically corrected or partially atmosphere

The EV I is more functional on N IR reflectance than on Red absorption, and

Reference: Huete et al., 2008[63]

γ is an atmospheric self correcting factor which depends on aerosol types.

Reference: Kaufman and Tanre, 1992

GEMI Global Environmental Monitoring Index.

Reference: Pinty and Verstraete, 1991 [129]

−0.2848 T M 1−0.2435 T M 2−0.5436 T M 3+0.7243 T M 4+0.0840 T M 5−0.1800 T M 7

M CARI 1 = 1.2 {2.5 (ρ800 − ρ670 ) − 1.3 (ρ800 − ρ550 )} (2.31)

1.5 {2.5 (ρ800 − ρ670 ) − 1.3 (ρ800 − ρ550 )}

TVI Triangular Vegetation Index.

T V I = 0.5 {120(ρ750 − ρ550 ) − 200 (ρ670 − ρ550 )} (2.34)

M T V I1 = 1.2 {1.2 (ρ800 − ρ550 ) − 2.5 (ρ670 − ρ550 )} (2.35)

1.5 (1.2 (ρ800 − ρ550 ) − 2.5 (ρ670 − ρ550 ))

CCCI Canopy chlorophyll content index.

PSSR Pigment Specific Simple Ratio.

NDWI Normalized Difference Water Index.

WBI Water Band index.

(ρN IR + 0.1) − (ρSW IR + 0.02)

PRI Photochemical Reflectance Index.

2.1.2 Normalized Difference Vegetation Index

 In Tucker et al., 1985[172] N DV I explained 80% of biomass accumulation (Fig 2.10).

 Sellers (1996[151]) used a canopy radiative transfer model to show that N DV I is

 The nonlinearity of ratio-based indexes

 Asymptotic (saturated) signals over high biomass conditions

 Sensitivity to background variations

 Sensitivity to clouds (deep, thin, shadow)

 Sensitivity to soil effects (moisture state, color)

In Tucker et al., 1985[172] N DV I explained 80% of biomass accumulation (Fig 2.10).

Sellers (1996[151]) used a canopy radiative transfer model to show that N DV I is

The nonlinearity of ratio-based indexes

Asymptotic (saturated) signals over high biomass conditions

Sensitivity to background variations

Sensitivity to clouds (deep, thin, shadow)

Sensitivity to soil effects (moisture state, color)

Sensitivity to anisotropic effects (geometry of the target)

Sensitivity to spectral effects (different instruments).

Indices based on derivate

geometric parameters related to the structure of vegetation and soil. Geometrical

Sano et al., (2005)[148] developed a comparative analysis of SAR backscatter coeffi-

In positions of biomass sampling data on radar images, the backscattering signatures

In Paloscia et al., 1999 [118] the sensitivity of backscattering coefficient, measured by

Surface scattering conditions:

Double bounce scattering conditions:

Randomly oriented thin cylinders scattering conditions:

In Saatchi and Moghaddam (1999[147]) it was studied the use of a semi-empirical

It is available data of an optical (i.e. multiespectral, hyperspectral) or microwave (i.e.