Project Report

1

1. INTRODUCTION

The term land cover signifies various attributes of the Earth’s surface including

forest, agriculture, settlements, urban landscapes and so on. However, the characteristics

of land cover have impacts on climate, hydrology, biogeochemical cycles, Diversity of

biological species and distribution of resources. More importantly, the impact of land

covers and its changes are getting importance day by days. The predominant land cover

class, which has direct implications over the biodiversity richness, is the forest cover. The

forest cover assessment, thus directly gives an idea of the carbon load and biological

species diversity of an area.

Biological diversity (referred to as biodiversity in short) is a part of our daily life

and livelihood, and constitutes resources, upon which humanity depends. It is not

uniformly distributed on the earth. Increasing human interventions on the ecosystems

have accelerated the process of biodiversity loss. As, it is fundamental to the fulfillment

of human needs. An environment, which is rich in biological diversity, offers the broadest

array of options for sustaining human welfare and for adapting to changes. Loss of

biodiversity has serious economic and social costs for any country. From the experiences

of the past, few decades have shown that, as industrialization and economic development

take place, the patterns of consumption, production and needs change, strain, alter and

even destroy ecosystems. If, one traces the linkages of biodiversity, it is observed that

human population and its interventions have been the major factors for enhancement in

the extinction process. If one overlay the global disturbance map on the biological

richness map of the world, one can see that Asia is the most threatened continent (Roy

P.S., 2002).

Biodiversity is the totality of genes, species, ecosystems and habitats in a region.

Biodiversity has direct consumptive value in agriculture, medicine and industry. Russian

scientist Nikolai I. Vavilov estimated that about 80,000 edible plants have been used at

one time or the other in human history, of which about 150 have even been cultivated on

a large scale. Today, merely 10 to 20 plant species provide 80-90% food requirements of

the world. In India, rural communities, particularly the tribal, obtain a considerable part

of their daily food from wild plants (Roy P.S., 2002). In this regard, the tropical forests

area including the mangroves is the highest valued ecosystem. In India, the mangrove

area used mostly for Non timber forest products (NTFP) like fuel, food, fodder,

construction materials for houses, net and cage preparation materials, honey, different

types of grasses for mat or baskets manufacture and also boat construction materials.

Another important use of mangrove woods in countries outside India like Sarawak

estuary of Indonesia is charcoal preparation (Bennet & Reynolds, 1993).

However, other than resource exploitation, the encroachment and reclamation of forest

land for agricultural fields and settlements to keep pace with the rapidly increasing

population pressure on the coastal zones, threats the mangroves of the world to a great

extent. The deforestation of tropical forests in South-East Asia was estimated as 1.4% per

year during 1980-1990 (Rudel et. al, 2000). On the same tract with this South East Asian

scenario, India lost 40 percent of its mangrove areas in the last century. The National

Remote Sensing Agency (NRSA) recorded a decline of 7 000 ha of mangroves in India

2

within the six-year period from 1975 to 1981. The government of India also reported a

loss of nearly 1,91,300 ha mangrove area in India during 1987 and 1997 (Kumar Rajiv,

2000). Though, the government reports showed an increase in mangrove vegetation in

Orissa coast (FSI, 1997 & 1999), several other research communications have reported a

reduction in the last decade due to huge population and developmental pressures along

with development of prawn farming aquaculture.

Coastal vegetation develops certain unique characteristics as the biota are constantly

under physiological stress caused by extreme environmental conditions, such as saline

ambience, periods of inundation and exposure, etc. The most dominating coastal

vegetation is the mangroves in tropical regions and marsh vegetations in temperate

countries. Today, mangrove forest comprise 15.8 million hectares, roughly less than half

of the original mangrove forest cover and are first declining further at an assumed rate of

2-8% per year, or 0.6% of all island forest in the world (Nayak & Bahuguna, 2001).

Bhitarkanika mangroves of Orissa, is the second largest mangrove patch of India at the

estuary of Bramhini and Baitarani river and extended between 20º33’50’’N & 20º48’N

latitude, and 86º46’50’’E & 87º05’E longitudinal extent under administrative jurisdiction

of Rajnagar Community Development block of Kendrapara district, Orissa. The

resourceful forest on Orissa coast and its adjacent areas had been designated as Wildlife

sanctuary in 1975 and National Park in 1998. The nearby seacoast, Gahirmatha, which

host Arribadas of endangered Olive Ridley turtle, has also been declared as Gahirmatha

(Marine) sanctuary for its majestic faunal biodiversity (Nayak A.K., 2002). The

mangrove has also been declared as a wetland of international importance or Ramsar Site

in 2002 for its wide range of avifauna population (Anon, GOI, 2003).

The climate of Bhitarkanika is tropical humid with average temperature range 33º to 18º

C and a mean humidity of 61 to 77%. The region receives a rainfall of 2000-2500 mm

annually which is distributed mostly between May and November. The district is situated

directly on the track of cyclonic storms, which frequently cross Orissa coast. These

storms occur between June and October. The estuarine vegetation of Bhitarkanika is

nourished mainly by the rivers Dhamra, Maipura, Patsala and Hansua, which are the

important distributaries of Brahmini and Baitarani Rivers. The fast flowing rivers become

slow while approaching the delta regions and deposit their silt content along the gradient.

As a result the riverbeds are raised and they break into many distributaries that radiate

across the plain. This topography is very conducive for luxuriant growth of mangroves

and associated fauna. The sanctuary has mangroves, tropical forest, scrub land, grass

lands, inter tidal mud flats, sandy beach, sand dunes and off shore islands (Kanvinde H.

S., 1997).

The macro level land use/ land cover assessment in the present day condition is done by

Remote Sensing techniques. The term ‘Remote Sensing’ can be defined as the technique

of sensing anything without having any physical contact with the sensed object. Such a

technique may refer to the use of different types of sensing procedures such as optical

sensing, microwave sensing, sonic sensing etc. However, of late, the most valued Remote

3

Sensing technique, especially in macro level studies is the Satellite remote sensing, where

the optical and long wave sensing technique is predominantly involved.

Since the launch of the first polar orbiting satellite Landsat in 1970s the satellite, Remote

Sensing has proved its efficiency and necessity mainly in the fields of geographical

mapping and reconnaissance surveys. The first Indian contribution in this field was the

launch of IRS 1A in 1988 with a multi-spectral Linear Image Scanning System known as

LISS I sensor having nearly 73 mts. resolution on ground. Since then, the country has

made huge advancement in this field with subsequent satellites like IRS 1B, IRS 1C, IRS

1D, IRS P4 (Ocean sat), IRS P6 (Resource sat) etc. and has reach a resolution as good as

nearly 1meter for panchromatic data and 5.5 mts. for multi-spectral data. The recent

advancements in international scenario has made a multi-spectral satellite imaging in 1

mt resolution from IKONOS satellite.

Earlier studies at the School of Oceanographic Studies, Jadavpur University by the

process of visual interpretation revealed significant rate of forest destruction at the

Bhitarkanika mangroves in last 30 years, possibly to cope up with the rapid increase in

local population (Mitra R., 2004). Further study with vegetation index calculations, to

assess particular vegetation distribution pattern of the forest, showed transitions in

mangrove forest cover of the area, which was witnessed by the presence of non-

mangrove vegetations within a dense mangrove patch. This study reported true mangrove

cover much lesser than that was estimated earlier (Mitra et. al. 2004).

The present study is thus, aimed in extension of the earlier one, to assess the pattern of

changes in mangrove cover of Bhitarkanika in a time scale of 12 years from 1989 to

2001. Here, to trace out the stages of transformation, Normalized Difference Vegetation

Index has been used for forest zonation.

4

2. Aims & Objectives

Remote sensing has been used for the reconnaissance survey for a long time. However,

multi-spectral data is proved to be the most useful incase of vegetation cover study,

because of the feasibility of band ratio study with Green, Red and Infrared, as vegetation

gives distinct reflectance patterns to these three bands.

Bhitarkanika being the second largest mangrove patch of India, on Orissa coastal area,

supports growths of mangrove and associate faunal diversity. However, in recent years,

with the increasing pressure on coastal zones through out the subcontinent, Bhitarkanika

is also facing severe threat of forest destruction. Apparently, though the forest cover was

not razed up vigorously, but a trend of forest transformation from dense mangroves to the

mixed-jungle and non-mangrove patches has been reported in earlier study using remote

sensing tools at the School of Oceanographic Studies (Mitra et. al. 2004).

It seemed necessary to study, whether, the mangrove distribution was as it is now in

recent past or there is a transformation going on? The only way to assess that fact was

thus, to study the distribution in time scale framework, and that too with the similar types

of records from past to the present. Therefore, the project was designed with the use of

IRS multi-spectral (LISS) data from 1989 to 2001, with an intermediate data at 1998.

The main aim of this study was thus set to assess the pattern of forest cover change with

special reference to the mangrove distribution at Bhitarkanika mangroves at Orissa, using

Remote Sensing techniques, such as vegetation index determination.

5

3. Review of Literatures

3.1. Biodiversity assessment:

The term Biodiversity can be defined as the existence of a wide number of species

or other taxa of plants, animals and microorganisms in a natural community or habitat or

of communities within a particular environment (Anon, OUP, 2000). The initiative for

biodiversity assessment was taken long back in 1991 with the UNEP Biodiversity

Country Studies Project (consisting of bilateral and Global Environmental Facility funded

studies in developing countries) implemented in cooperation from donor countries and

UNDP.

The agreed text of the Convention on Biological Diversity was adopted by 101

governments in Nairobi in May 1992, signed by 159 governments and the European

Union at the United Nations Conference on Environment and Development (UNCED)

held at Rio de Janeiro in June 1992. At present 120 governments are parties to this

convention. More than scores of nations and regions are engaged in developing their own

National Biodiversity Strategies or Action Plans.

Global Biodiversity Assessment (UNEP, 1995) estimates the total number of animal and

plant species to be between 13 and 14 million. It further records that so far only 1.75

million species have been described and studied. Ecosystem diversity has not been even

reasonably explored as yet. Hence, there seems to be wide gap of knowledge at global,

regional and local levels.

The Indian sub-continent is known for its diverse bioclimatic regions supporting

one of the richest floras and fauna. In a most recent attempt to map biogeographical

regions, Rodgers and Pawar (1988) attempted to define the biogeographical regions of

India. The sub-continent has ten biogeographical zones, viz., Trans-Himalayan,

Himalayan, Indian desert, Semi-Arid, Western Ghats, Deccan Peninsula, Gangetic Plains,

North East India, Islands and Coasts and not yet defined zones for aquatic (freshwater

and marine) ecosystems have been mapped. India is rich in endemic flora and fauna.

According to an estimate (BSI, 1983) about 30 percent plant species are endemic to India.

Recently, Ministry of Environment and Forests, Government of India has

launched a project viz., National Biodiversity Strategy and Action Plan (NBSAP), which

envisages the assessment and stock taking of biodiversity related information at various

levels, including distribution of endemic and endangered species and site-specific threats.

India, a mega-biodiversity country, while following the path of development, has been

sensitive to the needs of conservation. India’s strategies for conservation and sustainable

utilization of biodiversity in the past have comprised of by providing special status and

protection to biodiversity rich areas by declaring them as national parks, wildlife

sanctuaries, biosphere reserves and ecologically fragile and sensitive areas. It has helped

in reducing pressure from reserve forests by alternative measures of fuel wood and fodder

need satisfaction, by afforestation of degraded areas and wastelands, creation of ex situ

conservation facilities such as gene banks and eco-development. The challenges before

India are not only to sustain the efforts of the past but also further add to these efforts by

involving people in the mission.

6

3.2. The mangroves and forest cover of India

Mangrove resources are available in approximately 117 countries, covering an

area of 190,000 to 240,000 km2. Countries like Indonesia, Nigeria and Australia have the

largest mangrove areas. These ecosystems harbour 193 plant species, 397 fishes, 259

crabs, 256 molluscs, 450 insects and more than 250 other associated species. Mangrove

ecosystem has the highest level of productivity among natural ecosystems, and performs

several ecosystem services. According to one estimate, the mangrove forest cover in

India has reduced from 6000 km2 in 1953 to 2000–3000 km

2 in 1989 (Upadhyay et. al.,

2002).

The total length of Indian coastline is 7516.6 km, including the island territories,

but the Indian mainland coastline length is extended to 5700 km. Among the length, only

3979 km2 in covered with mangroves and nearly 62954.1 km

2 are landforms like mudflat,

beach/spit, shore/bar, coral reef, marsh vegetation, paleo-mudflat, lagoons / backwaters,

sand dunes, salt pans, flood prone areas etc. Most interestingly, these data on the basis of

Landsat TM / IRS LISS 11 FCC / SPOT FCC on 1:250000 scale do not tally with the

data generated from simply ground land survey on India coastal areas and particularly on

the mangrove zones. However, it is reported that in comparison to the total Indian

mangals, west coast mangals are sparse, less extended, patches and with very stunted

growth, which may be due to less fresh water supply from the upstream and soil of these

zones are not much silty-clay in comparison to the deltaic lands of the Ganga and the

Mahanadi (Naskar & Mandal, 1999). Moreover, the maximum mangrove reduction has

been reported to be as high as 44% in the west coast in comparison with that of the all

India estimate of 30% (Jagtap et al, 1994).

The mangrove ecosystem of east coast of India is comprises of five major deltas

and estuaries in four maritime states like Tamil Nadu, Andhra Pradesh, Orissa, West

Bengal. Among the five estuaries, Sundarbans at the estuary of Hooghly river in West

Bengal and Bhitarkanika at the Mahanadi delta of Orissa stand first and second largest

mangrove patch of India respectively. According to the biological diversity, Bhitarkanika

is known to the richest mangrove of India with nearly 71 mangrove and mangrove

associate species (Banerjee & Rao, 2001) with the dominating species being Avicennia

marina, A. officinalis, Rhizophora apiculata, Excoecaria agallocha, Phoenix and

Sonneratia sp. Paddy cultivation, salinity increase and slow regeneration of mangroves

have caused considerable damage. Mangroves are being cut down for fuel and building

material. Creeks have been blocked, there by resulting in the sweetening of the water

upstream killing the mangroves. Satellite data has revealed damage and cutting of

mangroves for agriculture, aquaculture etc. (Nayak & Bahuguna, 2001).

With the rapid decrease of mangrove vegetation (from 2210 km2 in 1975 to 1635

km2 in 1990-91) on the east coast of India, Bhitarkanika is also facing sever problem of

forest clearance (Jagtap et. al., 1993). While Mohanty et. al. (2001) reported tremendous

pressure on coastal zone of Orissa for the development of fisheries, aquacultures, port &

harbours and urban settlements, Das (1998) reported more than 30 sq.kms of

Bhitarkanika mangrove forest has been razed and the rest is threatened due to extensive

7

prawn farming in recent years. To be precise, an estimated 25,000 illegal settlers and

prawn farms from the sanctuary area are threatening the forest cover leading to a drastic

reduction in an average rate of 3 km2 per annum along with pollution of coastal water of

the favored habitat of the Olive Ridley.

It has been an accepted truth that the pressure on coastal zones and mangroves are

increasing with the increase in population over years. The main threats over the

mangrove forests are mostly related to human activities like over exploitation of food,

fuel and fodder forest resource, over fishing, grazing and land reclamation for agriculture

and aquaculture activities (Kathireson & Rajendran, 2003; Mitra & Pal, 2003). Mohanti

(1998) reported a 15% loss of mangrove wetland during a span of 18 years between 1975

and 1993 due to human interference.

Some efforts have been made to raise mangrove plantations in degraded forests in

Orissa, West Bengal, etc. by the respective Forest Departments. With adequate

protection, mangroves regenerate and cover the exploited area in a short period. This

process needs to be augmented with human intervention. For example, harvesting of

mangroves on sustained yield basis with defined period of rotation should be mandated to

ensure sufficient regeneration. Clear felling of mangroves promotes fast growing of

unwanted weed species, which inhibit mangrove regeneration. Natural regeneration and

recovery can also be promoted by reducing the anthropogenic pressure due to fuel-wood

demand on natural populations. There is an urgent need to undertake massive

afforestation programmes with improved fuel-wood yielding trees in the buffer areas, to

sustain the requirements of the local people. These plantations will act as a major energy

source to local communities in future. Plantation of fast-growing species should also be

taken up in the villages, which will reduce the dependence of coastal communities for

conventional firewood from mangrove forests (Upadhyay et. al., 2002).

3.3. Application spectral characteristics in Remote Sensing:

Remote sensing is 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 surface (Campbel J.B., 1996).

However, in most of the satellite remote sensing, electromagnetic spectrums are

used. The major divisions of the electromagnetic spectrum are in essence, arbitrarily

defined. Generally, the whole useful spectral range is divided as ultraviolet, visible

spectrum and Infrared spectrum and also the microwave energy.

In multi-spectral Remote Sensing, the sensors at satellites, receive spectral bands

at different ranges for building up an image. For IRS satellites in LISS Sensors mostly

Green, Red, near Infrared and Mid Infrared bands are used. This gives a wide scope of

data analysis, as the band properties vary enormously, for features like mangroves, non-

mangroves, dry sands or lands, water bodies etc. (Figure 3.1 & 3.2). the Blue bands,

however, are excluded using appropriate filters, as the high scattering potentials of short

wavelength blue bands may create noise in data.

8

020406080100120140160180

Band 1Band 1Band 1Band 1 Band 2Band 2Band 2Band 2 Band 3Band 3Band 3Band 3 Band 4Band 4Band 4Band 4MangroveMangroveMangroveMangrove Non-mangroveNon-mangroveNon-mangroveNon-mangrove Dry sandDry sandDry sandDry sand WaterWaterWaterWater

Figure 3.1 : Spectral characteristics of different earth signatures

0

20

40

60

80

100

120

140

160

180

MangroveMangroveMangroveMangrove Non-mangroveNon-mangroveNon-mangroveNon-mangrove Dry sandDry sandDry sandDry sand WaterWaterWaterWaterRefl

ecta

nce v

alu

e

Band 1Band 1Band 1Band 1 Band 2Band 2Band 2Band 2 Band 3Band 3Band 3Band 3 Band 4Band 4Band 4Band 4

Figure 3.2: The Characteristics of different spectral bands for different features.

9

3.4. Remote Sensing in Forestry Applications

Forestry is concerned with the management of forests for wood, forage, water,

wildlife and recreation. Forest of one type or another cover nearly one third of the Earth’s

land area. They are distributed unevenly and their resource values vary widely. Visual

image interpretation provides a feasible means of monitoring many of the world’s forest

conditions. We will be concerned principally with the application of visual image

interpretation, to tree species identification, studying harvested areas, timber cruising and

the assessment of disease and insect manifestations. The visual interpretation process of

tree species are more complex than for agricultural crop identification, as a given area of

forest land is often composed of more heterogeneous mixture of different species. In

contrast, the crop distributions in agricultural lands are more homogeneous for a wide

area.

The forestry application for the tree species identification and forest zonation has

been possible by the use of Aerial Photographs and even by high-resolution panchromatic

images. But, it is always better to use multi-spectral satellite imagery in forest cover

assessment or zonation. However, the lower resolution of the data imparts an additional

burden in distinguishing vegetation patches with different species composition. The

image characteristics of shape, pattern, shadow, tone and texture are used for

interpretation of tree species as individual tree species have their own characteristic

crown shape and size.

Importance of remotely sensed data for inventorying, mapping, monitoring and

recently for the management and development planning of the optimum sustainable

utilization of natural resources has been well established. Remote Sensing data has

proved to be extremely useful in providing information on various components of the

coastal environment viz., coastal wetland conditions, density wise mapping of

mangroves, coastal landforms and shoreline changes, tidal boundary, brackish water

areas, suspended sediment dynamics etc. The use of remote sensing in forestry

applications and vegetation cover assessment is getting so much of pace now a days, a

special Vegetation Sensor was launched on-board SPOT-4 satellite in March, 1998.

Though basic principle of the sensor is more or less similar to that of AVHRR, but it

differs a little on having a push-broom system of sensing and Short Wave Infra Red

(SWIR) band, which permits the study of vegetation water content. Study with one and

half years weekly data from the SPOT-4 Vegetation data enabled the researchers in broad

scale classification of dense humid forest, dry deciduous forest, mangroves, savanna etc.

with an accuracy of 87.8% (Philippe et. al., 2000).

The polar orbiting Indian Remote Sensing Satellite (IRS) has also been proved

useful in forestry application since its launch in 1988. The presence and absence of

mangroves and its aerial extent can be mapped at 1: 2,50,000 scale using a course

resolution satellite data such as LISS I data from IRS 1A/1B. High-resolution data

however, helps in getting information on the dominant community zones of mangroves.

Use of those data enabled investigators to classify mangrove forest of Indian coast on the

basis of crown density. Mangroves with >40% crown density are termed as dense (closed

10

forest), those having 10-40% crown density are termed as sparse (open forest) and

density <10% regarded as degraded forest, in density wise classification, which gives an

idea about the conditions of mangrove habitat also (Nayak & Bahuguna, 2001). However,

the property of stocking levels generally required at the working level is of 20% crown

cover intervals. Adoptions of spectral based stock maps appear get saturated beyond 40%

crown cover. However, in certain areas of Madhya Pradesh it is found, based on tone and

texture associated with spectral separability as a function of time has been found suitable

to provide 40-60% stocking level in addition. This study indicated greater scope for

future exploitation and utilization of various algorithms for optimizing time window and

spectral band may be feasible to explore satellite based forest stock map preparation. The

studies using NDVI as a function of stocking level needs review as it provides

information pixel level on the vigor of the canopy with less degree structural relationship

canopy closure. However, development of algorithm using textural function might proved

to be an effective method for quick assessment of stocking levels (Dutta et. al., 1994).

More recently, the IRS-P4 Ocean Colour Monitor (OCM) data with a spatial resolution

360mt x 236mt are being used to study the impact of different natural catastrophes like

cyclone on coastal environment, mangroves as well as on concentration and distribution

of Chlorophyll-a and suspended particles. The OCM of the IRS-P4 is optimally designed

for estimation of chlorophyll in coastal and oceanic waters, detection and monitoring of

phytoplankton blooms, studying the suspended sediment dynamics and other coastal

processes in respect to the time and space. Study with NDVI images of OCM data,

successfully traced the impact of Super Cyclone on coastal mangroves and its change

dynamics near Paradip port of Orissa in 1999 (Nayak et. al, 2001).

However, the accuracy of remote sensing change detection study is influenced by Sensor

systems, Environmental Characteristics and Geodetic controls. Accurate spatial

registration of the images is most essential for most of the change detection methods,

though precise geometric registration of image is often difficult to achieve when there is a

lack of accurate ground points (Singh A., 1989). Moreover, extremely wet or dry

conditions on one or more of the dates can result in serious change detection problems,

which can be mitigated, to some extent by referring the subsequent precipitation records

(Forse et al., 1997).

For the forestry application, as it has been already stated, the preferred data are

the multi-spectral images, which, in case of Indian satellite is available from the LISS

sensors. As, in a LISS III data with a spatial resolution of nearly 23.5 mts, a single pixel

may cover an array of several features and represents a cumulative band value, any

particular shape, size or pattern for a defined tree is hardly possible to identify. Thus, in

spite of the species identification, the species association study is preferred. The

differential band combination values for different vegetations due to higher absorbance

range of Red and high reflectance range of Infrared band has enabled the interpreter to

design various Vegetation Indices for proper and most authentic zonation of vegetation

associations. Normalized Difference Vegetation Index (NDVI), which was developed by

Rouse et. al. (1973), is the most widely used vegetation index (Forse et al., 1997).

In different studies, significant difference were found between the forest areas and other

crops with the help of NDVI and RVI, which helped in better enhancement of forest areas

(Md. Seeni & Salleh, 1992).

11

Though the index has been proved useful for forest cover assessment, Forse et al.

reported that the general land cover types appear to be mostly unrelated to the changes in

the NDVI. The NDVI values are describing changes in moisture content of the vegetation

rather than changes in land cover. The only exception appears to be the drop in NDVI

over time, which correlates with intrusion of shrimp farms into the marsh, and the

increasing NDVI values, which corresponds with the increasing area of mangroves in and

around the marsh. It is interesting to note the stability of the ‘dense humid forest’ and the

high NIR reflectance of the secondary complex denoting a very high photosynthetic

activity. The mangroves show a similar behavior to the ‘dense humid forest’ (Philippe et

al., 2000).

Any classification scheme based on NDVI values require revalidation of the

classification, which can be done using a confusion matrix following the processes

adapted by Congalton (1991). Using such confusion matrix, during assessment of

community-based restoration of Pichavaram mangrove was found useful in measurement

of accuracy of visual classification (Selvam et al. 1993).

3.5. Vegetation Indices

Vegetation indices are quantitative measures, based on digital values, that attempt

to measure biomass or vegetative vigor. Usually, a vegetation index is formed from

combinations of several spectral values that are added, divided, multiplied in a manner

designed to yield a single value that indicates the amount or vigor of vegetation within a

pixel. For large area vegetation monitoring, primarily the vegetation index was calculated

over the Advanced Very High Resolution Radiometer (AVHRR) data. Typically the

spectral bands used for this purpose have been the channel 1 visible band (0.58 to 0.68

µm) and channel 2, near-Infra Red band (0.73 to 1.10 µm). Various mathematical

combinations of AVHRR channel 1 and 2 data have been found to be sensitive indicators

of the presence and condition of green vegetation. These mathematical quantities are thus

referred to as vegetation indices. Two such indices have been calculated from AVHRR

data – a simple vegetation index (VI) and a normalized difference vegetation index

(NDVI). These indices are computed from the equations:

VI = Ch2 – Ch1

NDVI = Ch2 – Ch1 / Ch2 + Ch1

Where Ch1 and Ch2 represent data from AVHRR channels 1 and 2, preferably expressed

in terms of radiance or reflectance.

12

High values of the vegetation index identify pixels covered by substantial

proportions of healthy vegetation. Band ratios are quotients between measurements of

reflectance in separate portions of the spectrum. Ratios are effective in enhancing or

revealing latent information when there is an inverse relationship between two spectral

responses to the same biophysical phenomenon.

For living vegetation, the ratio strategy can be especially effective because of the inverse

relationship between vegetation brightness in the Red and Infra Red region. That is,

absorption of Red light (R) by chlorophyll, and strong reflection of Infra Red (IR)

radiation by Mesophyll tissues ensures that the red and near infrared values will be quite

different and that the Ratio (IR/R) will be high. Plant pigments such as chlorophyll

strongly absorbs much of the light at the wavelengths longer than 1.4 mm. This contrasts

with strong reflectance in the near infrared in the range from 0.75 mm through about 1.4

mm, at which plant materials are relatively transparent. Much attention in the Remote

Sensing of green vegetation is focused on the strong reflectance contrast between the

visible red and near infrared, which forms a strong step in the spectrum of green

vegetation which is often referred to as the “Red Edge” (Anon, CIT, 1999). Studies on

the photosynthetic efficiency of four rhizophoracean mangroves, Rhizophora apiculata,

R. mucronata, Bruguiera cylindrica and Ceriops decandra, randomly collected from

Pichavaram mangrove forest at southeast coast of India revealed that Bruguiera

cylindrica had higher stomatal conductance where the net CO2 fixation was also high. It

was concluded that the chlorophylls present in reaction centre and light harvesting

complex could be referred as "membrane-bound chlorophyll"; and also could be used as

an index to measure the photosynthetic productivity of mangrove species (Moorthy &

Kathiresan, 1999).

The high ‘penetrating’ capability of the near-infrared band through forested

canopies was the dominant factor in vegetation index sensitivity and performance. We

found that indices with higher weighing coefficients in the “near-infrared” to be the best

approach in extending vegetation index performance over forested and dense vegetated

canopies. (Huete et. al., 1997)

There have been two general approaches taken to develop indices for measuring

green vegetation cover based on the characteristics of the tasseled cap. The first approach

is to measure the distance between where the pixel plots in the tassel cap point from the

soil line. The soil line is used, because it is generally easier to find than the 100%

vegetation point. The Perpendicular Vegetation Index (PVI) of Richardson and Wiegand

(1977) assumes that the perpendicular distance of the pixel from the soil line in linearly

related to the vegetation cover. This index is calculated as follows:

PVI NIR red = - sin a (NIR) cos a (red)

Where, NIR is the near-infrared reflectance, red is the red reflectance and a is the angle

between the soil line and the near-infrared axis.

13

The above equation means that the isovegetation lines (lines of equal vegetation) would

all be parallel to the soil line. A special case of this is the simple Vegetation Index or

Vegetation Index (VI) described by Lillesand and Kiefer (1987) which has more recently

been christened deference vegetation index (DVI) by Richardson and Everitt (1992):

VI = DVI = NIR –Red

This case occurs when the soil line has a slop of 1.0.

The next possibility is to assume that the isovegetation lines all intersect at a

single point. As the first approximation, Jordan (1969) developed the ratio vegetation

index:

RVI = NIR / R

RVI itself is no longer generally used in remote sensing. Instead, an index known

as the normalized difference vegetation index (NDVI) is used. This index is functionally

identical to the RVI, and it can be written as:

NDVI = NIR - red / NIR +red = RVI-1 / RVI +1

Both RVI and NDVI basically measure the slope of the line between the origin of

red – NIR space and the red-NIR value of the image pixel. The only difference between

RVI and NDVI is the range of values of the two indices. The range from –1.0 to + 1.0 for

NDVI is easier to deal with than the infinite range for RVI (Anon, CIT, 1999).

Non-vegetated surfaces, including open water, manmade features, bare soil, and

dead or stressed vegetation, will not display this specific spectral response, and the ratios

will decrease in magnitude. Thus, the IR/R ration can provide a measure of the

importance of vegetative reflectance within a given pixel, which provides a tool of better

visualization of the image (Herndez-Cruz et. al., 2003)

The IR/R ratio is only one of many related measures of vegetation vigor and

abundance. The Green/Red (G/R) ratio is based on the same concepts as used for the

IR/R ratio, although it is considered less effective. Although ratios can be applied with

digital values from any Remote Sensing system, much of the research on this topic has

been conducted using Landsat MSS data. In this context, the IR/R ratio is implemented

for Landsat 4 & 5 as (MSS 4 / MSS 2), although some have preferred to use MSS 3 in

place of MSS 4.

Thus, the NDVI calculations with the MSS data can be formulated as:

NDVI = IR –R / IR + R = MSS4 – MSS2 / MSS4 + MSS2

This index in principle conveys the same kind of information as the IR/R and G/R ratios

but is defined to produce desirable statistical properties in the resulting values. Studies of

14

Tucker et. al.(1979) and Perry and Lautenschlager (1984) suggested that in practice there

are few differences between the many vegetation indices that have been developed.

Meza Diaz and Blackburn (2000) found major differences in the behavior of the different

spectral indices when correlated with mangrove leaf area index (LAI). NDVI, SAVI,

TSAVI, RVI and SAVI2 were the indices that were most strongly correlated with

mangrove LAI, for both the three individual species (Avicennia germinans, Rhizophora

mangle, Laguncularia racemosa) and for all species combined. The rest of the indices

showed very poor correlation with mangrove LAI, especially the three derivative-based

indices. It has been found that the background in this environment has a strong influence

on the spectral properties of the mangrove canopy, which affected the performance of the

spectral indices investigated in different ways.

Mangroves are intertidal, often grow in dense stands and have complex aerial root

systems, which make extensive sampling impractical with the difficulty of moving

through dense mangrove stands and the general inaccessibility of many mangrove areas

posing a major logistic problem. Luckily field measurements indicated that there is a

linear relationship between mangrove Leaf Area Index and NDVI, which can be obtained

from remotely sensed data. This means that a relatively modest field survey campaign

can be conducted to obtain LAI measurements in more accessible mangrove areas and

these used to establish a relationship to NDVI using regression analysis. Once this

relationship is known then NDVI values for the remainder of the mangrove areas can be

converted to LAI (Anon, 1999). Although we lack quantitative biomass data, by

observation it is difficult to qualitatively associate any component with increasing

biomass or a characteristic NDVI over the whole year. For instance, while mangroves

and dry prairie areas were strongly positive in the first component image, the

pine/hardwood complex just to the north, home of the endangered Florida panther, was

strongly positive in the third component image (Roberts, 1994).

Although NDVI has been accepted by most of the scientific schools as one of the best

vegetation index world wide, a few limitations of the index have also been revealed.

NDVI has been reported to be unable to highlight subtle differences in canopy density. It

has been found to improve by using power degree of the infrared response. The index

thus calculated has been termed as advanced vegetation index (AVI). It has been more

sensitive to forest density and physiognomic vegetation classes. AVI has been calculated

using following equation

AVI = {(B4 +1) x (256-B3) x B4/3}1/3

However, the biophysical spectral indices may require several steps of indices calculation

steps like Normalization of Landsat TM band, Temperature calibration, calculation of

Bar Soil Index, Canopy Shadow Index etc (Roy et. al., 1997).

15

3.6. Image Classification

Digital classification can be defined as the process of defining pixels to class or

more precisely land cover classes. Normally, multi-spectral data are used to perform the

classification indeed the spectral pattern present within the data for each pixel is used as

numerical basis of categorization. That is different feature types manifest different

combinations of DN values based on their inherent spectral reflectance and emitance

properties. These classes form regions on a map or an image; after classification, thus,

digital image is presented as a mosaic of uniform parcels, each identified by a colour or

symbol.

Image classification has an important role in remote sensing data analysis, as it depends

on the spectral properties of each pixel or a cluster of pixels. For a work, like forest cover

change assessment, two types of classification schemes can be followed, such as

Unsupervised and Supervised classification.

3.6.1. Unsupervised classification:

Unsupervised classification can be defined as the identification on natural groups,

or structures, within multi-spectral data. The notion of existence of natural, inherent

grouping of spectral values within a scene may not be intuitively obvious, but it can be

demonstrated that remotely sensed images are usually composed of spectral classes that

internally are reasonably uniform in respect to brightness in several spectral channels.

Unsupervised classification identifies spectrally homogenous classes within the data,

thus, these classes does not necessarily corresponds to the informal categories that the

analyst is intend to have. Spectral properties of specific information changes over time in

respect of seasonal or even diurnal variations, but being the classification scheme without

a knowledge support, unsupervised classification separates same feature information as

different class. Moreover, such classification provides little scope to the analyst to

analyze the data using spectral characteristics (Campbell J. B., 1996 and Lillesand &

Kiefer, 2000).

3.6.2. Supervised classification:

Supervised classification can be defined informally as the process of using

samples of known identity to classify pixels of unknown identity. The analyst in these

cases defines training areas (sample pixels located within a defined region) by identifying

regions on the image that can be clearly matched to areas of known identity on the image.

In supervised classification the analyst have several controls over the classes and thus,

can manipulate the data according to the interest of his or her study. Moreover, a

supervised classification considers the seasonal variations, as the training sets are

assigned manually, which includes the knowledge base. Thus, the classification possesses

more accuracy and truthfulness in practice (Campbell J. B., 1996 and Lillesand & Kiefer,

2000).

16

3.6.3. Accuracy assessment and confusion matrix:

the standard form for reporting site-specific error is the error matrix, sometimes referred

to as the confusion matrix because it identifies not only overall errors for each category

but also misclassification (due to confusion between categories) by category.

Compilation of an error matrix is required for any serious study of accuracy. The matrix

consists of n x n array, where ‘n’ represents the number of categories. It reveals the

results of comparison of evaluated and reference images. Thus, inspection of the matrix

shows how the classification represents actual areas on the landscape (Campbel J.B.,

1996).

However, the accurate classes, i.e. the frequency of the match evaluated and reference

classes are situated as the diagonal position in the matrix. In several studies this

confusion matrix has cropped huge success in measurement of classification accuracy

and also in identifying the ground features, which represent same characteristics of each

other.

17

4. METHODOLOGY

Mangrove forest is regarded as one of the most productive ecosystem of the world. Due

to presence of high water and chlorophyll content it gives a higher value of Normalized

Difference Vegetation Index (NDVI) than other coastal vegetations like Casuarina

plantations or woodlots. However, using this property, several forest cover change

detection studies have been done successfully with the help of Multi-spectral satellite

imageries.

4.1. Data used

The present forest cover change detection study has been done on the basis of forest

community assessment by NDVI values. All the satellite imageries used, were taken by

Indian Remote Sensing (IRS) satellite with Linear Image Scanning System (LISS)

sensors. Pre-monsoon data of 1989 (21st February), 1998 (13

th February) and 2001 (19

th

March) from IRS-1A (LISS I), IRS-1C (LISS III) and IRS-1D (LISS III) respectively,

have been used for the forest cover assessment.

4.2. Data processing, Classification procedures and analysis

The digital data supplied by the National Remote Sensing Agency (NRSA), Hyderabad

were either in Binary or Super structure format. Those data were first imported as ‘.img’

format and then analyzed and processed by the Remote Sensing software ERDAS

IMAGING 8.5. As the digital data did not have any real earth co-ordinates, all the three

data were geometrically corrected using ground control points such as road–road

intersection, jetty, canal–road intersection, etc. taken from the toposheet using the same

image- processing package.

The area of interest was chosen and scooped out of the whole image by a ‘sub-set’

operation on the same. Only the mangrove areas and villages adjacent to the forested

areas were considered as study area, as man-mangrove relationships are most

predominant at this part and it is the perfect area for studying any trend of forest

degradation.

Reconnaissance survey of Mangrove wetland of 1989, 1998 and 2001 were done by on-

screen visual interpretation method using different enhancement techniques like

histogram equalization, available with the software. Different classes of the mangroves

and adjacent areas, such as dense mangroves, degraded mangroves, young mangrove

stands, sand dune associated with mangrove wetlands, vegetation associated with sand

dunes, water body and dry land were then identified using visual interpretation keys such

as colour, tone, texture, pattern, size and shape. However, prior knowledge of the ground

reality was necessary for such type of identification process (Table 4.1).

18

As the author was frequenting to the area of study and it was very much possible to

gather ground reality from several on field points, Supervised classification was chosen

for the study. The software was made familiar to the tones and their combinations in

clusters of pixels by a few informal sets of defined pixel with known ground signature,

known as ‘Training sets’. The Training sets were made on the basis of field experience

and at a few ground truth points. For the data of past, as a ground truth verification was

not possible, the researcher had to rely on the visual interpretations on the basis of the

field experience. Primarily, total 42 different classes were defined and classified

accordingly. But, after ward, those were clustered together as per the need of this present

study following the vegetation index calculation.

During the classification, the software was instructed to calculate the minimum and

maximum values of each band of reflectance among all the pixels of a given ‘Training

set’, and then to ascertain an average with standard deviations, which will be the assigned

value for a certain class. The NDVI was calculated with the average values of each of

those 42 classes using the formula discussed in chapter 4.

Area of each class was calculated digitally by software itself. The total area for each of

the categories, however were calculated afterwards by simply adding the areas of each

class.

4.3. Radiometric correction

The data supplied to us were mostly corrected for haze and other atmospheric

disturbances on the reflectance. Further, radiometric correction was done during the data

processing and classification. The radiometric correction is done on the assumption that,

in a full scene there must be at least, one absolutely dark pixel, which does not have any

reflectance value for any band of the useful spectrum. Because of the atmospheric

scattering the imaging system records a non-zero DN value at this supposedly dark

shadowed pixel location. This represents the DN value that must be subtracted from a

particular spectral band to remove the first order scattering component (Singh et al,

1999). Thus, the darkest point on the scene is identified and processed accordingly.

19

Class Colour Shape Features No. of

training

sets

Mangroves Bright red Irregular Mangroves & new mangroves

11

Degraded

mangroves

Light brown

to pink

Irregular Dry and scrubby and sparse

mangrove

4

Succession Brownish red Irregular Acanthus sp., Poteratia sp. &

Excoecaria sp.

1

Mixed

jungle

Mixed red

and brown

Irregular Mangroves & non-mangroves

like Acacia sp., Zizypus, Ficus,

Prosopis sp. etc.

2

Non-

mangroves

Pink,

purplish red

Irregular

and

linear

Bank vegetations with weed

and grass, scrubs,

embankment Acacia sp.

Prosopis sp. etc.

3

Marshy

vegetation

Dark blue-

brownish red

Irregular Wet drench vegetation

predominantly with Sueda sp.

and Nalia grass.

1

Swamp

Greenish blue Irregular Wetlands with marshy

vegtations

1

Agricultur

al land

White to

greenish blue

& pink

Flat

irregular

Dry and wet agricultural lands

and lands with sparse weeds

and grasses

3

Sand

White(dry) &

light greenish

blue (wet)

Rough/s

mooth

Dry sand and wet sand 2

Water

body

Light to dark

blue and

black

Irregular,

regular

& linear

River, Channels, Sea,

potholes, aquaculture, shallow

water and sediment load.

8

Table 4.1: Details of classification and the description different classed and their ground

features.

20

5. Results and Discussion

Tree canopy is essential to environmental and economic health, providing

additional cooling, reducing energy needs, increasing property values, improving

air/water quality, reducing the cost of storm water control, and contributing to a more

beautiful, friendlier, and livable community. In lieu of assessing the canopy cover of

mangroves in Bhitarkanika, the study was designed in such a way, so that, a proper

assessment of mangrove and non-mangrove floral distribution can be done, even in the

inaccessible areas.

Analysis of satellite data has been carried out using different digital analytical

procedures, as it has been described previously. In the view of minimizing the seasonal

variations and the atmospheric disturbances cloudless pre-monsoon multi-spectral images

were used for all the three years data. However, the radiometric corrections were also

done to nullify the reflectance values from aerosols and hazes. Although, the band

combinations and spatial resolution were same for the IRS 1C and IRS 1D LISS III data

of 1998 and 2001 respectively, but it differed in case of 1989 data from IRS 1A satellite

by LISS I sensor (Ref. Annexure 1). With the use of various enhancement techniques

and classification schemes, that too on basis of pixel properties and the most widely used

band ratio indexing method like Normalized Difference Vegetation Index (NDVI), the

mapping procedure is expected to be quite efficient.

Supervised classification technique based on ground truth observation points and

revalidation of the classified areas represented significant variations in NDVI values of

mangroves, degraded mangroves, mixed jungle with both the mangrove and non-

mangrove plant species and non-mangrove patches. While, dense mangrove forest cover

gave NDVI at a higher side, the degraded forest area represented a NDVI value at the

lowest range (Table 5.1) among all the vegetation signatures. The other land use, land

cover classes, such as, marshy vegetations, swamp, agriculture, aquaculture, Sands,

Mudflats and water bodies were classified differently (Table 5.2). During ground truth

verification and revalidation at Bhitarkanika, it was found that patchy distribution of non-

mangrove plants like Prosopis juliflora, Acacia sp., Zizypus jujuba, Ficus sp. etc. are

present in close association with mangroves to form a mixed vegetation type (Table 5.3),

which may be indicative of an undesired transformation of mangrove ecosystem.

However, the NDVI study could distinctly separate such patches from the dense

mangrove covers. It was most interesting to note that in close observation of the

classified map, such mixed patches were found to be present at the areas nearer to the

forest edges, where human interactions are more (See Map 1, 2, & 3).

The NDVI values show significant differences for Mangroves and non-mangroves. While

the mangroves represented average NDVI between 0.398272±0.044 and 0.475683±0.029

for the three respective years, the maximum NDVI for mixed vegetations with

comprising of both the mangrove and non-mangrove species composition were reported

to be 0.301818 among all the data. The NDVI of non-mangroves including woodlot

plants, coconut and palm trees and typical river bank scrubs ranged between 0.023793

21

and 0.23811 with average values between 0.089342± 0.035 and 0.146132± 0.047 for the

three years. But most interestingly, the marshy land vegetations predominated by Saueda

sp. and mangrove grasses showed NDVI in the higher side and the highest value was

obtained from the young mangrove patches at the second or third sere of succession, for

all the data. Such clear trends in NDVI values for various signatures indicate the

efficiency of the process. Estimated accuracy level of the classified data, using Confusion

matrix process was more than 86% (Table 5.6). The highest amount of errors was

observed in cases of the class succession and degraded forest, probably due to their

similarity in ground characteristics i.e. muddy soil base with scattered vegetations.

However, merging of those two classes as a single signature resulted in an increased

accuracy of nearly 92%.

Area calculation for different land cover classes was done digitally in a composite way.

As, during classification, 12 different signatures were designated as mangroves to

enhance the accuracy of classification through NDVI calculation, all those classes

together have been classified as Mangrove and the composite area of those were

designated as the dense mangrove area. The same schemes were followed in case of other

features also. However, 10.53 sq.km. (>10%) decrease in dense mangrove cover has been

reported during 1989 and 2001. Perhaps there was a conversion towards the agricultural

fields, which has increased 9.9 sq.km during the same time scale. On the other hand, the

cultural lands as a whole has shown an increase perhaps in expense of the vegetation

covers and forestland. Most surprisingly, the pattern of changes, particularly in case of

forest cover, varied in nature for the two different time scale of 1989-1998 and 1998-

2001(Figure 5.1 & 5.2). While during the first nine years of 1990s, changes in land

coverage indicated towards a trend of forest destruction, in contrast, the forest health

were found to be revived during the last three years (Table 5.4 & 5.5). In addition, in

some areas with mixed or degraded mangroves in 1998, restoration of dense mangroves

has been reported. The observation, however, supports the claim of Forest department

regarding their honest efforts to protect and restore the mangrove forest cover following

establishment of Bhitarkanika National Park in 145 sq.kms, within the premises of

Wildlife sanctuary.

Trend analysis of the aforesaid observations set different relations of diverse land cover

types, especially in forested areas. Other than the general trend of forest to cultural land

(agriculture and aquaculture) conversion, several trends of forest conversion were also

found. Decreasing trends of dense mangrove were accompanied by an increase in mixed

or non-mangrove vegetations and vice-versa. On the other hand, the trends of changes in

degraded mangroves, agriculture and aquaculture were nearly parallel (Figure 5.2).

Setting these trends on an analytical purview it seems that the forest goes through a series

of changes before it gets wiped away. The forest degradation, invasion of non-mangroves

within dense mangrove, growth of non-mangroves and marshy vegetations are a few

probable intermediate stages of a forest transformation and subsequent clearance process.

22

23

2001 1998 1989

Class NDVI Area NDVI Area NDVI Area

Mangrove 1 0.45002 260.4717 0.43895 1276.916 0.404239 645.76

Mangrove 2 0.41912 151.9236 0.4903 541.8657 0.345439 933.12

Mangrove 3 0.37351 867.7287 0.47457 899.2944 0.419951 309.76

Mangrove 4 0.37046 1422.7164 0.48466 773.3961 0.404784 1242.88

Mangrove 5 0.33783 1511.0712 0.421 2374.718 0.356539 2011.52

Mangrove 6 0.41674 486.3888 0.47099 428.7249 0.452721 126.72

Mangrove 7 0.40035 306.6174 0.45798 404.0847 0.452164 146.56

Mangrove 8 0.34141 1450.9287 0.50479 89.8857 0.412148 343.68

Mangrove 9 0.42646 297.3591 0.48147 276.4368 0.342803 866.56

Mangrove 10 0.42796 541.1367 0.4653 86.6781 0.383513 1527.04

Mangrove 11 0.34365 1362.7197 0.48093 767.4912 0.412366 817.92

New mangrove 0.47177 86.9697 0.53725 293.6412 0.4204 828.16

Deg. mangrove 1 0.26583 496.3761 0.23529 329.7267 0.223732 798.08

Deg. mangrove 2 0.26589 1217.3571 0.18635 355.2417 0.17724 107.52

Deg. mangrove 3 0.1826 448.8453 0.21184 471.5172 0.157641 544

New Succession 0.14677 603.4662 0.17522 85.4388 0.215504 975.36

Mangrove scrub 0.1904 476.037 0.23113 1105.456 0.205841 673.28

Mixed jungle 1 0.22394 1244.9862 0.30182 1619.911 0.280815 862.08

Mixed jungle 2 0.26501 501.7707 0.25421 755.8272 0.21284 808.32

Bank vegetation 0.1091 395.9928 0.02379 953.8965 0.060694 359.04

Embankment vegetation 0.15743 783.675 0.16931 2650.644 0.096727 1141.76

Woodlots 0.20849 732.9366 0.23811 966.3624 0.136476 714.24

Floodplain vegetation 0.10951 1481.8383 0.09357 1255.192 0.063472 468.48

Dune vegetation 0.0618 515.1114 0.2722 250.776 0.225589 256

Marshy vegetation 0.28055 469.2573 0.36853 156.4434 0.304193 82.56

Swamp 0.01261 1618.0155 -0.0028 628.1064 0.057518 584.32

Dry Agricultural land 0.00105 2182.1886 -0.0114 1892.703 0.012335 1112.32

Agricultural land 0.00086 4771.4508 -0.0207 3640.991 0.047834 5767.68

Weedy Agriculture 0.06047 1915.3017 0.0485 2623.015 0.098678 1000.96

Dry sand -0.09346 308.8044 -0.0745 249.6096 -0.03664 853.12

Wet sand -0.204 932.5368 -0.1424 391.3272 -0.19837 380.16

Wetland -0.09124 289.8504 -0.0635 232.7697 -0.03906 424.32

Aquaculture 1 -0.23942 239.6952 -0.1817 54.6021 0 0

Deposition -0.17468 511.1019 -0.469 1692.446 -0.35039 2928

Sediment flow -0.39144 3268.6173 -0.259 1174.273 -0.3496 1615.36

Creeks/ mud 0.16873 251.7237 0.27172 957.3228 0.148289 645.76

Wet channel -0.02124 979.4844 -0.084 1056.758 0.041783 2903.68

River 1 -0.18939 2011.2381 -0.4462 5483.174 -0.19355 3414.4

River 2 -0.26862 2970.7479 -0.2769 1073.161 -0.03397 684.8

Shallow water -0.1631 520.3602 -0.1786 602.7372 -0.10634 850.56

Coastal sea -0.30075 5876.6877 -0.4078 6826.283 -0.32387 4679.04

Sea -0.32278 2010.7278 -0.3621 2105.498 -0.34107 4030.08

Table 5.2: Details of classification signatures with their NDVI values and area in Hector.

24

No. Scientific name Common Name

1. Ficus bengalensis Bat

2.

Ficus religiosa Aswath

3.

Syzigium cumini Jam

4.

Prosopis juliflora Kanta babla

5.

Acacia nilotica Babul

6. Thespetia pupalnea Habul

7. Strichnos potatorum Makal

8. Cocos nusifera

Coconut

9.

Mimusops elingii Bakul

10.

Borasus flabelicum Palm

11.

Pongamea glabra ---

12. Casuarina equisitifolia Jhau (on Dune)

13. Pandanus sp.

Kea (Back mangrove)

14.

Spinifix littoralis (Dune vegetation)

Table 5.3: The list of non-mangrove plants found in close association with

the mangroves in mixed jungle patches at Bhitarkanika.

25

Year Area Area Change Area Change Change

Land cover class 1989 1998 1989-98 2001 1998-2001 1989-2001

Mangrove 9799.68 8213.133 -1586.55 8746.03 532.899 -1053.65

Deg-man 2122.88 2261.941 139.061 2638.62 376.675 515.736

Succession 975.36 85.4388 -889.921 603.466 518.0274 -371.894

Mixed jungle 1670.4 2375.738 705.338 1746.76 -628.981 76.357

Non-mangrove 2683.52 5826.095 3142.575 3394.44 -2431.65 710.923

Dune vegetation 256 250.776 -5.224 515.111 264.3354 259.1114

Marsh vegetation 82.56 156.4434 73.8834 469.257 312.8136 386.697

Agriculture 7880.96 8156.708 275.748 8868.94 712.233 987.981

Aquaculture 0 54.6021 54.6021 239.695 185.0931 239.6952

Composite cover

Forest 11470.08 10588.87 -881.209 10492.8 -96.082 -977.291

Degraded/succession 3098.24 2347.38 -750.86 3242.08 894.7024 143.8422

Non-mangrove 3658.88 5911.534 2252.654 3997.91 -1913.62 339.0292

Cultural land 7880.96 8211.31 330.3501 9108.64 897.3261 1227.676

Table 5.4: The land cover area calculation on the basis of NDVI studies.

(Areas written in Hector.)

Land %

1989

Land %

1998

Change/yr

1989-98

Land %

2001

Change/yr

1998-2001

Change/yr

1989-2001

Mangrove 38.47 29.99 -0.94 32.13 0.71 -0.53

Deg-man 8.33 8.26 -0.01 9.69 0.48 0.11

Succession 3.83 0.31 -0.39 2.22 0.64 -0.13

Mixed jungle 6.56 8.68 0.24 6.42 -0.75 -0.01

Non-mangrove 10.54 21.28 1.19 12.47 -2.94 0.16

Dune veg. 1.01 0.92 -0.01 1.89 0.32 0.07

Marsh veg. 0.32 0.57 0.03 1.72 0.38 0.12

Agriculture 30.94 29.79 -0.13 32.58 0.93 0.14

Aquaculture 0 0.2 0.02 0.88 0.23 0.07

Table 5.5: The percentage of change in land cover area on the basis of NDVI studies.

26

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1989 1998 2001

% of

Area

Mangrove Deg_man Succession

Mixed jungle Non_mangrove Dune vegetation

Marsh vegetation Agriculture Aquaculture

Figure 5.1: The change in land cover area during 1989-2001 at Bhitarkanika

27

-25

-20

-15

-10

-5

0

5

10

15

20

25

1989-98 1998-2001

Ch

an

ge

in

are

a (

Sq

. k

ms

)

Forest Degraded forest Non forest Cultural land

Figure 5.2 : The profile of changes in land cover at Bhitarkanika

28

0

2000

4000

6000

8000

10000

12000

1989 1998 2001

Are

a (

ha

)

mangrove deg_man SuccessionMixed jungle Non_mangrove Marsh vegetationAgriculture

Figure 5.3: The trend of land cover changes at Bhitarkanika during 1989-2001

29

30

6. Conclusion

Bhitarkanika, the second largest mangrove patch of India, has been in the news headlines

for supporting one of the largest rookeries of endangered Olive Ridley turtles. But, it has

another matter of concern regarding its majestic floral and faunal diversity, which is

sometimes considered as one of the richest in India. Increasing human interference with

the increasing population pressure at the forest adjacent villages has been held

responsible for causing severe damages to the mangroves for its resourcefulness. Even,

the most predominant Estuarine crocodiles (Crocodylus porosus) are facing a threat of

habitat shrinkage over the rivers of and around Bhitarkanika.

The present study was designed with a goal of forest cover assessment with special

reference to mangrove and non-mangrove distribution. Time series analysis with the help

of vegetation index like Normalized Difference Vegetation Index (NDVI), revealed

yearly 0.5% loss in mangrove forest during 1989-2001. But, most interestingly, the rate

of loss was 0.9% during 1989 to 1998, which was reversed after the forest was declared

National Park. During 1998 to 2001 the mangrove regeneration rate was 0.7% a year,

which is perhaps indicating an honest effort of forest department of Orissa in preservation

of the valuable forest.

Although, the rate of forest destruction is much lower than the global average of 2%, the

reports of human interference within the forest premises and the trend of change in land

use and human lifestyle of adjacent areas are alarming for the mangroves. Mushrooming

of aquaculture ponds in expense of agricultural fields, over-fishing leading to a reduced

fish catch, unaccounted forest resource utilization by the locals etc. are driving the forest

ecosystem towards the upper limit of carrying capacity, exceeding which may create a

disastrous result for the ecosystem. An indication of such habitat threat was perceived

when 85% of crocodile population of the forest was found to be restricted in only 40%

river areas during 2000-01 crocodile census (Kar S., 2001).

However, besides the forest cover change assessment, the efficiency of Remote Sensing

techniques and supervised classification in such type of study was also assessed, which,

cropped a success. But, it has also been confirmed that, a ground truth verification and

knowledge based supervision is most important tool for such studies, which enhances the

accuracy levels a lot.

Thus, it can be concluded that use of band ratio analysis preceded by ground truth

verification and visual interpretation of imageries and followed by thorough revalidation

of the classified map with quite a large number of ground points is an useful integrated

process of such vegetation cover assessment study.

31

7. List of References:

Anon, (1999) Assessing Mangrove Leaf-Area Index (Lai) Using Casi Airborne Imagery,

Applications Of Satellite And Airborne Image Data To Coastal Management

(Session 2: Lession8), Bilko for Windows Module 7.

Anon, (1999) California Institute of Technology website (http://www.gps.caltech.Edu/~

arid/analysis/vegindex.html), Vegetation Indices, 1999.

Anon, (1999) Dictionary of Science, Oxford University Press.

Anon, (2003) Govt. of India Order on announcement of 13 new Wetlands of India to be

declared as RAMSAR sites (D.O. No. J-22012/68/89 – W (vol II); Ministry of

Environment and Forest, 16th

January.

Banerjee L.K. Rao, T.A. (2001) Flora of the Mahanadi Delta, Orissa (Flora of India

Series 4); Botanical Survey of India; pp. 307.

Bennet E. L., Reynolds C.J., (1993) The value of mangrove area at Sarawak,

Biodiversity and Conservation, 2, 359-375.

Campbell J. B., (1996) Introduction to Remote Sensing, 2nd

Ed. Taylor & Francis,

London, pp. 621.

Chadha S., Kar. C.S. (1998) Bhitarkanika: Myth and Reality; Natraj Publishers,

Dehradun.

Congalton, R. G., (1991) A review of assessing the accuracy of classifications of

remotely sensed data. Remote Sensing Environ., , 37, 34–46.

Das B.B. (1998) Present Status of Gahirmatha beach in Bhitarkanika Sanctuary, Orissa.

Marine turtle Newsletter, 79, 1-2.

Dutta C.B.S., Udayalakshmi V., Sadhasivaiah A. S., (1994) Role of Remote Sensing in

Forest Management, Proceedings of Asean Coeference on Remote Sensing -

Session Forestry.

Forest Survey of India, (1997) MoEF, Dehradun, State Forest Report,.

Forest Survey of India, (1999) MoEF, Dehradun, State Forest Report, 1999.

Forse J.L., Leach J.H., Choowaew S., Bishop I.D., (1997) Change detection analysis of

coastal land cover in Thailand using Landsat Thematic Mapper, Proceedings of

Northern Australian Geographic Information Systems (NAGIS) Conference,

Caims.

Harnandez-Cruz L.R., (2003) Remote Characterization of Black Mangrove areas in the

Estuarine bay of Bahia de Jobos, Puetro Rico, ADVANCE Institutional

Transformation Program, University of Puerto Rico at Humacao.

Web page: http://cuhwww.upr.clu.edu/~advance/participants%20Coloquio.htm.

32

Huete A. R., Liu H.-Q., & van Leeuwen, W. J. D. (1997) The use of vegetation indices

in forested regions: Issues of linearity and saturation. In Proceedings of IGARSS

'97 - International Geoscience and Remote Sensing Seminar (Vol. 4, 1966-1968).

Noordwijk, The Netherlands: ESA Publications.

Huete, A. R., Liu, H.-Q., & van Leeuwen, W. J. D. (1997) The use of vegetation indices

in forested regions: Issues of linearity and saturation in Proceedings of IGARSS

'97 - International Geoscience and Remote Sensing Seminar (Vol. 4).

Noordwijk, The Netherlands: ESA Publications. pp. 1966-1968.

Jagtap T. G., Untawale A. G., Inamdar S.N., (1994) Study of mangrove environment of

Maharashtra coast using Remote Sensing data, Ind. J. Mar. Sci. vol 23, June

1994. 90-93.

Jagtap T.G., A.G.Untawale, V.S.Chavan. (1993) Synopsis: Mangrove Ecosystem in

India; A need for protection. Ambio, vol. 22(4) June 1993.

Jordan, C. F. (1969) Derivation of leaf area index from quality of light on the forest

floor, Ecology_, vol. 50,. pp. 663-666.

Kanvinde H. S., (1997) Gender Dimensions in Biodiversity Management : India, Report

submitted to FAO Regional Office for Asia and Pacefic, Bangkok, Thiland. June

1997.

Kar S.K., Kar C.S., (2002) Bhitarkanika Mangrove Ecosystem and its Biodiversity – An

overview, Proceedings of National Workshop on Mangrove Conservation and

Restoration, 70-76.

Kar Sudhakar (2001) Annual Census of Salt water Crocodiles in Bhitarkanika Wildlife

Sanctuary. Crocodile Specialist Group newsletter vol. 20. No.3, July-Sept. 2001,

pp. 57-58.

Kathiresan K., Rajendran N., (2003) Conservation and management of mangrove

ecosystem in India, Seshaiyana, vol. 11, No. 1, p. 1-4.

Kumar R., (2000) Conservation and management of mangroves in India, with special

reference to the State of Goa and the Middle Andaman Islands, Unasylva (203),

Vol. 51- 2000/4.

Lillesand T.M., Kiefer R.W., (2000) Remote Sensing and Image Interpretation, 4th

Ed.,

John Wiley & Sons, Inc.

Lillesand, T. M. and Kiefer, R. W. (1987) Remote Sensing and Image Interpretation, 2nd

edition, John Wiley and Sons, New York, Chichester, Brisbane, Toronto,

Singapore, 721 p.

Md Seeni Ibrahim0, Salleh Azhar Jj., (1992) Detecting forest areas and crops using

vegetation indices, Proceedings of Asean Conference on Remote Sensing -

Session Agriculture/forestry.

33

Meza Diaz B., and Blackburn, G. A., (2000) The Relationships between Mangrove LAI

and Broadband and Derivative-Based Spectral Vegetation Indices: Evidence

from a field study, 26th

Annual Conference of Remote Sensing Agency,

University of Leicester, UK.

Mitra A., Pal S., (2002) The Oscillating Mangrove ecosystem and the Indian

Sundarbans, WWF, pp. 112.

Mitra R. (2004) Forest cover changes in the Bhitarkanika Wildlife sanctuary, Orissa,

Proceedings (Abstract) of Environmental Science section of 91st Session of

Indian Science Congress, Chandigarh. p. 13.

Mitra R., Hazra S., Santra S.C., (2004) ‘Mangrove distribution study using Remote

Sensing techniques, at Bhitarkanika National Park, Orissa’, International

Conference on Biogeochemistry of Estuaries-Mangroves and the Coastal Zone

Management, JNU, New Delhi, ENVIS Centre in Biogeochemistry Newsletter,

Vol. 9 No. 4. p. 25.

Mohanti M, (1998) Tropical coastal wetlands, Orissa coast, Eastern India and

management perspective: An overview, Proceedings of 5th

International

conference on Remote Sensing for marine and coastal environment.

Mohanty P. K., Pal, Supriyo Ranjan. Mishra Pravakar. Takashige Sugimoto, (2001),

Environmental Changes along Orissa Coast, Bay Bengal: Its monitoring and

management; Proceedings of The 5th

International Conference on the

Environmental Management of Enclosed Coastal Seas.(EMECS-Abstracts).

Moorthy P., and Kathiresan K. (1999), Photosynthetic efficiency in Rhizophoracean

mangroves with reference to compartmentalization of photosynthetic pigments.

Revista de Biologia Tropical. March/June 47(1-2): 21-25.

Naskar K R., Mandal R N., (1999), Ecology and Biodiversity of Indian Mangroves, Part

–I, Daya Pub. House, Delhi, pp. 361.

Nayak A. K., (2000) Conservation and Management of Mangrove forests of

Bhitarkanika, Orissa, Proceedings of National Workshop on Mangrove

Conservation and Restoration, pp. 30-42.

Nayak S., Bahuguna A. (2001), Application of Remote Sensing data to monitor

mangroves and other coastal vegetations of India, Indian Journal of Marine

Sciences, 30 (4), 195-213.

Nayak S.R., Sarangi R.K., Rajawat A.S., (2001) Application of IRS-P4 OCM data to

study the impact cyclone on coastal environment of Orissa, Current Science, vol.

80, No 9, pp. 1208-1212.

Philippe M., Valory G., Etienne B. (2000), Mapping the forest cover of Madagascar

with SPOT 4- VEGETATION data, Proceedings of Vegetation - 2000, Lake

Maggiore, Italy.

Richardson, A. J. , Everitt, J. H. (1992) Using spectra vegetation indices to estimate

rangeland productivity, Geocarto International_, vol. 1, pp. 63-69.

34

Richardson, A. J., Wiegand, C. L. (1977) Distinguishing vegetation from soil

background information, Photogrammetric Engineering and Remote Sensing, vol.

43, pp. 1541-1552.

Roberts M., Wells C., Doyle Thomas W. (1994), Component Analysis for Interpretation

of time series NDVI imagery, Proceedings of American Society for

Photogrammetry and Remote Sensing, Annual Convention & Exposition.

Roy P. S., (IIRS) (2002), Biodiversity characterization at landscape level in North-East

India using Satellite Remote Sensing and Geographic Information System; Jt.

Project of Dept. of Biotechnology and Dept. of Space.

Roy P.S., Miyatake S., Rikimaru A. (1997), Biophysical Spectral Response Modeling

Approach for Forest Density Stratification, Proceedings of Asean Conference on

Remote Sensing (Session Forestry).

Rudel T.K., Flesher K., Bates D., Baptista S., Holmgren P. (2000), Tropical

deforestation literature: geographical and historical patterns, Unasylva (203),

Vol. 51.

Selvam V., Ravichandran K. K., Gnanappazham L. and Navamuniyammal M. (2003),

Assessment of community-based restoration of Pichavaram mangrove wetland

using remote sensing data, Current Science, Vol. 85, No. 6, 25 Sept.2003.

Singh A. (1989) Digital change detection techniques using remotely-sensed data,

International Journal of Remote Sensing, vol. 10, 989-1003.

Singh S., Agarwal S., Joshi P.K., Roy P.S. (1999), Biome level classification of

vegetation in Western India – An application of Wide Field View Sensor (WiFS),

Proceedings of Joint Workshop of International Society for Photogrammetry and

Remote Sensing (ISPRS) Working Group I/1, I/3 & I/4 on Sensors and Mapping

from Space.

Singh, H.S. (2002) Marine Protected Areas of India: Status of Coastal wetland

conservation. World Comission on Protected Areas publication, January 2002.

Upadhyay V. P. Ranjan Rajiv, Singh J. S. (2002), Human–mangrove conflicts: The way

out, Current Science, Vol. 83, No. 11, pp.1328 – 36.

van Leeuwen W. J. D., Laing T. W., & Huete A. R. (1997) Quality assurance of global

vegetation index compositing algorithms using AVHRR data. In Proceedings of

IGARSS '97 - International Geoscience and Remote Sensing Seminar (Vol. 1,

341-343). Noordwijk, The Netherlands: ESA Publications.