International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
involves, fairly obviously, direct observation of vegetation
categories and even animal populations (Sidle ef al. 2002) from
remotely sensed images. Indirect ecological remote sensing
involves the derivation of environmental parameters from
remotely sensed images as proxies for ecological phenomena
(de Leeuw et al. 2002). Commonly, for instance, habitats are
derived from vegetation categories to infer the distribution of
animal populations.
Kerr and Ostrovsky (2003) describe three main areas of
ecological remote sensing. First, simple land cover
classification is useful for straightforward identification of
vegetation types and derivation of habitats. Second, integrated
ecosystem measurements are invaluable in providing estimates
of ecosystem function over large areas (entire ecosystems). In
particular, there has been considerable recent interest in using
remote sensing to derive biophysical parameters such as leaf
area index (LAI) and net primary productivity (NPP),
sometimes using normalised difference vegetation indices
(NDVI) (Goetz 2002). Third, change detection is essential for
ecological monitoring and, given the continuous and stable
nature of spaceborne image acquisition, remote sensing
provides an excellent source of data for this purpose (Coppin ef
al. 2004). Further, such temporal analysis can be extrapolated to
predict future ecological change, for instance to estimate the
result of anthropogenic land use practices on protected species.
While most ecological remote sensing involves the use of
optical (multispectral, and to a lesser extent, panchromatic)
imagery, various other image types are used. The recent
emergence of hyperspectral imagery is particularly relevant
here, given the ability of this fine spectral resolution imagery to
detect subtle differences between highly specific land cover
classes, typically vegetation categories or soil types (Turner ef
al. 2003). Another emerging technology, lidar (light detection
and ranging), is particularly useful for measuring height (Van
der Meer et al. 2002), which may then be incorporated in
further ecological analysis. Finally, radar imagery has been
used relatively widely for ecological investigation, and
Kasischke ef al. (1997) describe four main applications: (i) land
cover classification, (ii) woody plant biomass estimation, (iii)
flood analysis and (iv) temporal monitoring.
3. FINE SPATIAL RESOLUTION SPACEBORNE
IMAGERY
For much of its history, spaceborne remote sensing has been
constrained by technological limitations, and also by
governmental legislation. Specifically, during the Cold War, the
USA and Russian militaries restricted public access to fine
spatial resolution imagery. Given rapid technological
development and relaxing legislation when the Cold War ended,
various private and governmental organisations became
engaged in developing fine spatial resolution satellite sensors
by the mid 1990. Further, the imagery from these instruments
was intended for general public use, rather than being restricted
to the military (Aplin er al. 1997, Birk et al. 2003). (For
convenience, here, ‘fine spatial resolution imagery’ is defined
as imagery finer than the 10 m spatial resolution SPOT HRV
imagery (which was already available at the time the new
sensors were developed).)
Various fine spatial resolution satellite sensors have now been
developed (Table 1), and others are planned for development.
The first instrument to be operational was the 5.8 m spatial
326
resolution panchromatic sensor on board the Indian Remote
Sensing Satellite (IRS)-1C, launched in 1995. A successor, IRS-
ID, was launched two years later, with a slightly finer (5.2 m)
spatial resolution. The most well-known of the fine spatial
resolution satellite sensors, and regarded widely as the first of
this new ‘era’ of remote sensing, is IKONOS, launched in 1999.
(The IRS instruments had received little attention in large parts
of the world, notably the USA. Indeed, IKONOS, comprising 1
m and 4 m sensors (panchromatic and multispectral,
respectively) (Dial er al. 2003), does have markedly more
advanced technology that the IRS instruments.) In 2000, EROS-
Al was launched with a 1 m spatial resolution panchromatic
sensor, although this Saudi Arabia-based enterprise has also
received relatively little attention in the USA and elsewhere. A
significant event occurred in 2001, with the launch of
QuickBird, comprising the finest spatial resolution imagery
available currently (0.61 m and 2.44 m spatial resolution
panchromatic and multispectral imagery, respectively). Since
then, the latest SPOT satellite has been launched, including
more advanced instruments than previous SPOT missions (two
panchromatic sensors with spatial resolutions of 2.5 m and 5 m
respectively, and a 10 m spatial resolution multispectral sensor),
and, finally, OrbView-3 was launched in 2003, with instruments
similar to IKONOS.
Year of | Satellite Spatial resolution (m)
launch | sensor Panchromatic | Multispectral
1995 IRS-IC 5.8
1997 IRS-1D 5.2
1999 IKONOS 1 4
2000 Eros-Al 1
2001 QuickBird 0.61 2.44
2002 SPOT.HRG | 2.5.5 10
2003 OrbView-3 1 4
Table 1. Fine spatial resolution satellite sensors (IRS = Indian
Remote Sensing Satellite, SPOT = Systeme Pour l’Observation
de la Terre, HRG = High Resolution Geometry.)
Fine spatial resolution satellite sensor imagery has been used
for a range of ecological applications. There has been
considerable interest in forest analysis and, in fact, some of the
earliest published examples of IKONOS data exploitation
related to woodland. For instance, Franklin ef al. (2001)
demonstrated the value of texture measures derived from
panchromatic IKONOS imagery for distinguishing forest age
classes. Other studies have focused on tropical forests, and
IKONOS and QuickBird data have been demonstrated as an
accurate means of studying forest demographics (Clark ef af.
2004a), structure and dynamics (Read er al. 2003). However,
not all ecological investigation is focused on rural areas, and
IKONOS imagery has been used to aid certain urban analyses.
Greenhill et al. (2003), for instance, describe the benefit of
multispectral IKONOS imagery in determining the ecological
characteristics of suburban land. There has also been extensive
use of IKONOS imagery in hydrological applications such as
watershed management (Hall er al. 2004), and marine
applications like coral reef classification (Andréfouet et al
2003).
The benefit of fine spatial resolution imagery over coarse
spatial resolution imagery for ecological investigation is fairly
obvious. Generally, as spatial resolution increases (becomes
finer), the accuracy with which small objects are identified and
Int
ch
stt
co
ins
th:
(E
et
ap
(2
im
al.
ac
et
be
ar