By:
S.P.S. Kushwaha
The problems involved in maintaining a sustained supply of forest resources
for the present day needs and future demands of the mankind have made the
forest managers conscious about the compelling need for rational utilization
of these resources. The conventional methods of forest resources assessment
and monitoring are time and cost-intensive. Many a time they do not match
with the resource dynamism and hence, become obsolete by the time the
results are available. The advent of remote sensing and geographic
information system (GIS), and global positioning system (GPS) technologies
has revolutionized the forest resources assessment, monitoring, management
and has reduced the time and cost considerably. Remote sensing is the
science and art of obtaining information about an object, area or phenomenon
through analysis of data acquired by the device that is not in contact with
the object, area or phenomenon under investigation (Lillesand and Kiefer,
2006). It is also said to be the practice of deriving information about the
earth’s land and water surfaces using images acquired from an overhead
perspective, using electromagnetic radiation in one or more regions of the
electromagnetic spectrum, reflected or emitted from the earth’s surface.
Some of the highlights of the remote sensing technology are stable sensing
platforms, synoptic coverage, high frequency of observations, and real-time
images available in spatial form and on multiple scales. It is perhaps the
only technology, which allows retrospective evaluation of forest resources.
Remote sensing, as such a multidisciplinary discipline, is utilized in the
inventory and monitoring of the forest resources. All remote sensing systems
including aerial cameras capture radiation (signature) in different
wavelengths reflected or emitted by the land and water surface features and
record it either directly on the film as in case of air photos or on a
digital medium like tape, and other digital media. As no two objects in
nature are theoretically ditto, their signatures are unique. This property
of the objects is exploited in remote sensing to differentiate the objects
from one another. The reflectance from vegetation is controlled by leaf
pigments, cell structure, and the leaf water content. The radiation absorbed
in red region is primarily used for photosynthesis. In healthy vegetation,
both absorption and reflectance are more pronounced. Diseased and senescent
vegetation shows lesser absorption as well as reflection in red and
near-infrared regions and a higher overall reflectance in blue and green
regions. These spectral properties of vegetation are exploited to detect
their type and condition through image interpretation. Forests are one of
the most conspicuous terrestrial features on the planet earth.
Identification and mapping of forests using remote sensing techniques is
thus, relatively easy. Satellite remote sensing is extremely popular in
forest surveys.
Satellite Remote Sensing
During the World War-II, the use of electromagnetic spectrum was extended
from visible and infrared to microwave regions and this is considered as a
major milestone in the history of remote sensing. The beginning of
space-based remote sensing dates back to 1891, when Germans developed
rocket-propelled camera systems. But it was in 1957, when Sputnik-1 took the
first photograph of the earth from satellite. Systematic earth observation
from space started with the launch of Explorer-1 in 1959 and meteorological
satellite, TIROS-1 in 1960. The launch of Earth Resources Technology
Satellite (ERTS) in 1972 dawned a new era in remote sensing. This was the
first satellite available for systematic and repetitive observations of
earth’s land resources. Landsat-1, 2 and 3 carried multispectral sensors
operating in 0.5-1.1µ wavelength range and had 79m x 57m spatial resolution.
The satellite data provided by these satellites facilitated the
identification and mapping of broad forest types and canopy densities.
Landsat-4 and 5 carried both MSS and Thematic Mapper (TM) sensors and
provided low as well as medium resolution (30m x 30m pixel) imagery. The
Landsat TM imagery changed the foresters’ outlook about the forests. The
availability of 20m and 10m resolution imagery from French satellite, SPOT
later significantly advanced remote sensing applications and brought image
interpretation close to virtual reality.
The Indian Space Research Organization (ISRO) launched a number of Indian
Remote Sensing Satellites (IRS) with capabilities similar to contemporary
earth observation satellites. The launch of Indian Remote Sensing Satellite
(IRS-1A) in March 1988 marked a new era in the history of satellite remote
sensing programme in India. Subsequently, the IRS-1B, IRS-P2, IRS-P3,
IRS-P4, IRS-1C, IRS-1D, TES and IRS P6/Resourcesat-1, and Resourcesat-2 were
launched. Both IRS-1A and 1B carried 72.5m and 36.25m spatial resolution
sensors on-board and provided not only the continuity of satellite data from
Landsat Programme (of U.S.) to indigenous one but also an opportunity for
Indian scientific community to test their data for forest resources
inventory and monitoring. While the earlier two sensors are meant to
facilitate in locale-specific intensive resource inventories, the WiFS
(Wide-Field Sensor) and AWiFS (Advanced Wide-Field Sensor) data facilitated
the assessment of the land and water features and phenomena encompassing
large areas. The Indian Remote Sensing Programme has come a long way and is
all set to grow further through more advanced sensors with improved
capabilities of information generation.
The availability of satellite data in digital form provides a whole lot of
flexibility with respect to its use. Single band black and white as well as
false colour composite (FCC) images can be used for interpretation. Since
single imagery covers large area, the broad forest features encompassing
larger areas could be studied. The use of multi-date imagery provides
information on phenological conditions of the forests i.e., evergreen or
deciduous as well as on the extent and the distribution of forests with the
passage of time. Over time, the visual interpretation has given way to
digital interpretation. It has been found that a combination of supervised
and unsupervised techniques i.e. hybrid methods yield better results. In
India the first attempt to categorize forest cover types by digital
classification of satellite data was made in 1978 for Nagaland delineating
temperate evergreen, tropical evergreen, tropical semi-evergreen, tropical
deciduous, bamboo, degraded forests, shifting cultivation and permanent
cultivation. Many studies on forest cover mapping have been done in India so
far using satellite imagery. NRSA carried out the first-ever forest cover
monitoring using 1972-75 and 1980-82 timeframe satellite imagery, showing
very high rate of deforestation (NRSA, 1983). In general, forest cover
mapping has been attempted more frequently than type mapping. Selecting a
time of the year when maximum differences occur due to phenological changes
such as leaf fall, flowering etc. improves the capability of satellite data
in forest type delineation.
Microwave Remote Sensing
The use of microwave imagery in forestry/vegetation sciences is mainly
driven by the fact that microwave remote sensing is capable of providing
data at any time of the day/night and more than that is the capability of
microwaves to penetrate the atmosphere under virtually all conditions.
Depending on the wavelengths involved, microwave energy can see through
haze, light rain, fog, snow, clouds and smoke. Hence, microwave remote
sensing has some edge over optical remote sensing. Moist vegetation returns
more signal than dry vegetation. Also like-polarized (HH or VV) sensing
penetrates vegetation more than cross-polarized (HV or VH) microwave remote
sensing. Likewise more energy is returned from crops having their rows
aligned in the azimuth direction than from those aligned in the range
direction of radar. Radar imagery has been widely used in qualitative and
quantitative forest stratification. The backscattering from forested areas
has been found to be dominated by tree crowns consisting of foliage and
branches. The K- and C-band have been found to be sensitive to low biomass
level while P-band is sensitive to high biomass forests. The backscattering
from broadleaved forest stands is normally more than that from coniferous
forests. Many studies have related various forest stand parameters like tree
height, density, age, timber volume, biomass etc., with radar backscatter
with varied degrees of success.
LiDAR Remote Sensing
LiDAR (Light Detection And Ranging) remote sensing is a breakthrough
technology for forest resources inventory. It offers a great potential for
forest conservation and management. The advantage of using LiDAR is that it
provides three-dimensional data. If cautiously planned, LiDAR can form one
of the most scientific and accurate means of forest management. The various
advantages of LiDAR technology are higher accuracy, weather independence,
capability of canopy penetration, lesser time needed for data acquisition
and processing, minimum user interference. Besides, laser-derived images
help in terrain visualization. As vertical component (z-axis) measurement is
the backbone of LiDAR technology, this characteristic is exploited in a very
straight forward way for tree height estimation. Tree canopy height is
obtained by subtracting the elevations of the first and last returns.
Vegetation height when coupled with species composition and site quality
information, serves as an estimate of stand age or successional stage. Like
simple height estimate, the vertical distribution of laser returns provides
basis to classify vegetation, and to estimate other important canopy
characteristics such as canopy cover, crown volume (foliage, trunk, twigs,
branches etc.). LiDAR data provides input for estimation of aboveground
biomass with high accuracy. The combination of LiDAR and satellite remote
sensing data could be very useful for describing biodiversity and monitoring
changes in biodiversity.
Forest Type and Canopy Density Mapping
The utility of the remote sensing data for forest canopy density mapping and
monitoring on various scales is well established by now. The nationwide
forest studies carried out by NRSA and FSI have amply demonstrated the scope
of satellite remote sensing in forest mapping and monitoring. Multispectral
data such as that from IRS LISS-4 with 5.8m, IKONOS with 1m, and QuickBird
with 0.61m spatial resolutions provide an unprecedented opportunity to
monitor the forests. Barring single species dominated forests and forest
plantations, majority of the Indian forests are highly heterogeneous. This
makes their differentiation, delineation and mapping a difficult task.
Problem gets further compounded in case of forests that are located in hilly
and mountainous regions on account of topographic effects. Forest type
mapping from satellite imagery has been attempted in past with varied degree
of success. Local level studies backed by intensive ground truth have
generally resulted in more number of forest type categories. Except for the
small scale map prepared by Champion and Seth in 1968, showing sixteen major
forest groups, India did not have a forest type map until recently. This
year, the Forest Survey of India, for the first time, has come out with a
satellite image-based forest map with 180 types.
Wildlife Habitat Evaluation
Remote sensing can be applied to wildlife habitat inventory, evaluation and
wildlife census. Remote sensing not only provides spatial data but also
allows us to compare temporal variations in the spatial data, essential for
wildlife management. While ground survey methods such as counting animals,
trapping, collection of droppings, investigations of feeding sites as well
as ground mapping of habitats will always be useful, remote sensing can
supplement or considerably replace tedious ground surveys. Ground methods
have limitations as whole area can not be accessed in one go in many of the
cases and the information collected may not be as accurate as is possible
through remote sensing. The GIS-based modeling of species-habitat
relationships is one of the popular forms of habitat suitability analysis
(Singh and Kushwaha, 2011). The GIS output is a map depicting habitat
suitability for any wild animal. The map can guide decisions regarding
habitat preservation priorities, forest/land management practices, or sites
for reintroduction of endangered species. As more and more data and
information is being collected day by day by wildlife managers, world over,
a need is being felt to develop a Wildlife Information System (WILIS),
integrating wildlife species databases in spatial/non-spatial formats,
habitat suitability models and rules/guidelines for habitat evaluations. A
WILIS could be operational at national level with linkages to regional and
local databases. One particular study done in Chilla Sanctuary showed US$
100 as the cost of per square kilometer habitat evaluation.
Timber Volume/Growing Stock Inventory
Timber volume and the total growing stock are the key information required
for the forest planning and management. Remote sensing data facilitates in
the stratification of forests, which in-turn reduces the sampling error and
allows the growing stock assessment with fewer samples. Most of the
territorial forest divisions have working/management plans and these are
revised every ten years. Conventional methods have limitations and hence,
space technology should be adopted to achieve objectives envisaged in the
National Forest Policy. The satellite image-based forest stratification can
be correlated to the actual on-ground timber volume/growing stock or biomass
using two-stage inventory design (Köhl and Kushwaha, 1994). For large areas,
it is advisable to go for multi-phase sampling techniques. In a multi-phase
design, visually or digitally classified imagery makes the first stage
followed by large-scale image interpretation and ground measurements. The
multi-phase sampling design reduces the cost considerably. Overtime, high
resolution satellite imagery has significantly replaced aerial photos in
forest in growing stock and biomass inventory.
Plant Richness Assessment
India, with a geographical area of 2.4 per cent of the world, has about 8
per cent the world’s total biodiversity. The country is very rich in
biodiversity with 45,000 plant and 75,000 animal species. Hence, it is
called as a mega-diversity region. The threat to biodiversity in India is
due to the over-exploitation of plant and animal resources as well as their
habitats resulting in the fragmentation of habitat and creation of
considerable landscape heterogeneity (Singh and Kushwaha, 2008). Remote
sensing and GIS can contribute immensely in biodiversity assessment at
regional to global scales. Remote sensing can play a very useful role in
assessment of bio-rich areas, which could then be related to actual values
of biodiversity, got through ground methods. This can be done through
landscape characterization wherein a digitally or visually classified image
is taken as direct input (Behera et al., 2005). Many landscape parameters
viz., porosity, patch size and shape, interspersion and juxtaposition etc.
have been said to have a direct relationship with a variety of vegetation
features like biodiversity, physiognomy, composition and other stand
parameters. Geospatial analysis of a forest landscape can take care of many
factors that persist within each ecosystem. Vegetation type maps are prime
inputs for biodiversity assessment at landscape level. The impact of scale
on biodiversity assessment is significant; higher scale results in better
assessments (Kushwaha et al, 2005).
Forest Fire Risk Assessment
More than ninety five percent of the forest fires in India are man-made.
Forest understorey is burnt year after year by local people for deriving
benefits like cattle grazing, fodder, accessibility for timber/firewood
extraction etc. Recurrent burning hampers forest regeneration and arrests
plant succession. About 65 percent of the India’s 64 million ha forests are
broadleaved deciduous and ca. 5 percent are coniferous; thus, about 70
percent deciduous forests happen to be fire-prone. The estimated average
annual tangible loss due to forest fires in the country is of the order of
approx. US$ 100 million. The Himalayan coniferous forest comprising of fir,
spruce, cedar, chirpine and blue pine etc. are highly prone to fire. The
factors governing fire occurrence and spread are: fuel loading (type and
moisture), temperature, humidity, wind (both speed and direction), slope,
aspect and the accessibility. A rapid assessment of the forests for their
proneness to fire could be worked out using a combination of information
generated from remote sensing and other means. The information on wind is
generally seldom available because of the poor network of meteorological
stations in India. A dynamic fire spread model using a combination of
biophysical parameters including wind direction, speed and live fire events
could be worked out to predict the fire spread.
Forest Litigation
Remote sensing and GIS applications have assisted the judiciary for quite
some time now to decide on the disputed cases of land ownership in the
country. In one such case between Sanjay Gandhi National Park authority,
Mumbai and the encroachers, wherein considerable area along the park
periphery had been encroached by local people over past several decades, the
High Court at Mumbai, based on the evidence produced by remote sensing gave
the verdict in favour of the Sanjay Gandhi National Park and directed
Government of Maharashtra to evict the encroachers. In another similar case
in Maharashtra involving farmers of Dhule district and the forest
department, wherein farmers had encroached during past three decades
sizeable reserved forest area and were practicing agriculture, the remote
sensing and GIS-based evidence showed that the disputed area actually was
well within the reserved forest boundary and hence, the High Court in this
case too decided in favour of the Forest department. In yet another similar
case, where fisher folks had encroached a part of the Jambu island in
Sunderban Biosphere Reserve for fish drying, the Supreme Court of India gave
the verdict in favour of the Biosphere authorities after untiring efforts of
the Director, Sunderbans Biosphere Reserve by producing concrete evidence
through time-series analysis of satellite images.
National Forest Information System (NAFIS)
Responsibility for 64 million ha of dense, open forests and forest
plantations across the country falls under a great many jurisdictions and
management agencies including government organizations and the individuals.
Forestry information is gathered in different ways, for different uses and
is stored in different formats and locations. As a result, accessing and
integrating this information is extremely complex. There is need for a
national forest information system (NAFIS) for effective information and
communication management. NAFIS is envisaged as an internet information
gateway to forest information resources from across the country. The NAFIS
should address these significant and wide-ranging differences by adopting
international standards and by building a distributed network of servers and
applications which allow access to forestry information held by independent
agencies. The NAFIS should provide web tools ranging from simple portrayal
to sophisticated analyses, to users from anywhere in India or abroad, which
means that users can discover, interact, integrate and display the authentic
and accurate information on India’s forests and forestry. NAFIS should
promote standards for enhancing the interoperability of multiple information
resources.
Future Prospects
The need for higher spatial, spectral and radiometric resolutions for forest
types and species (associations) has been emphasized over time. Gradually,
the mapping and monitoring scenario is expected to be a lot better than ever
before. The spatial, spectral, and the radiometric resolutions have already
improved considerably. Currently many of the sensors provide spectral
resolutions of the order of about 10 nanometers per band. Sensing using
continuous spectra is expected to help not only in better species
identification, association/formation and forest/vegetation type level
mapping but also result in higher accuracy in timber volume and biomass
estimations by highlighting the subtle differences in the physiognomy of the
vegetation. The hyperspectral imagery is providing insight into the state of
biodiversity and the vegetation continuum across ecosystems and the
landscapes in an unprecedented manner. Ground penetration radars are already
helping the scientific community in belowground biomass Assessment.
Forestry and Ecology Department, Indian Institute of Remote Sensing,
Indian Space Research Organisation, Dehradun 248001, Uttarakhand, India.
<[email protected]>