Remote Sensing and GIS Techniques for monitoring The Coastal Environmenal Changes: A Case Study in Tuticorin Coast, India
M. Thanikachalama, S. Ramachandranb
a Department of Civil, Adhiyaman College of Engineering, Anna University, Hosur-635 109
b Institute for Ocean Management, Anna University, Chennai-600 025
The Agency for Assessment and Application of Technology
Graduate student of Department of Geodetic Engineering, Gadjah Mada University
The coastal ecosystems are now highly disturbed and threatened due to rapid increase of population and developmental activities along the coast. It is necessary to protect coastal ecosystem to ensure sustainable development. This requires information on coastal wetlands and geomorphology. Remote sensing data have been found to be extremely useful for provide information on this aspects. IRS LISS-II & III satellite data (1988 and 1998), Survey of India Tographic map (SOI 1969) and Naval Hydrographic Chart (NHO, 1975) has been used to generate the coastal ecosystem, coastal land use and land cover, coastal landforms, shoreline and bathymetry maps on 1:50,000 scale. This information has been used for the assessment of present status and identification the changes in coastal ecosystem. The analysis of multi-date coral reef maps showed that 2 kmē of reef area in Tuticorin coast have been lost over a period of ten years. The changes in coastal land use/land cover indicate that the major changes occurred in mangrove, crop land, fallow land, agriculture plantation, forest plantation, scrub land, sandy area and tanks. Various coastal landform units have been identified and grouped in to marine, fluvio-marine, fluvial, aeolian and biogenic landforms. The analysis of multi-date shoreline maps showed that 0.84 and 2.73 kmē of the mainland coast and 0.35 and 0.13 kmē areas of island coast have been eroded and accreted, respectively, in Tuticorin coast. The analysis of multi-date bathymetry data indicates that, the depth of seafloor has decreased along the coast and around the islands. The average reduction of depth in seafloor has been estimated as 0.31m over a period of twenty-four years. Most of the coastal geomorphic features such as spits, back swamp, mud flat, seaward migration of shore line, reduction of seafloor depths and some of the increased land cover features such as mangrove and mud flat etc are indicate that the coast of Tuticorin going on emerging by tectonic movement. The coastal ecosystem particularly coral reef ecosystem is very severely affected by anthropogenic and natural factors. The validation by ground truth has also confirmed these results
Coastal environment play a vital role in nation's economy by virtue of their resources, productive habitats and rich biodiversity. India has a coastline of 7,516 km and nearly 250 million people live within a distance of 50 km from the coast. The coastal zone is endowed with a variety of coastal ecosystems like mangroves, coral reefs, lagoons, sea grass, salt marsh, estuary etc. Coastal ecosystems are important for millions of people around the world as they provide subsistence. The coastal ecosystems are now highly disturbed and threatened due to rapid increase of population and developmental activities along the coast. In the state of Tamil Nadu, between the year 1988 and 1998, 25.56 kmē of coral reefs and 2.16 kmē of seaweeds were lost in Gulf of Mannar (Thanikachalam and Ramachandran 2002a, 2002b, 2002c and 2003). Between the year 1986 and 1993, 0.36 kmē area of mangrove in Pichavaram was lost and nearly 2500kmē of the mangrove were lost in entire India between 1986 and 1994 (Krishnamoorthy 1995). Apart from the anthropogenic activities, natural causes are also play an important roll in coastal environment changes. In Gulf of Mannar coast, between the year 1969 and 1998, 4.34 and 23.49 kmē of mainland coast and 4.16 and 3.31kmē of island coast were eroded and accreted due to the combined action of anthropogenic and natural agents (Thanikachalam and Ramachandran 2002b and 2003). Current approaches to the monitoring and management of coastal ecosystem are not capable for sustainable development. The modern scientific tools of remote sensing, GIS and GPs are extremely valuable in coastal environmental studies. Availability of repetitive, synoptic and multi-spectral data from various satellite platforms, have helped to generate information on varied aspects of the coastal and marine environment (Nayak 2002). The present study has been attempted to monitor the coastal environment changes in Tuticorin coast using remote sensing and GIS techniques.
2. STUDY AREA
The cost of Tuticorin (Figure 1) is the part of Gulf of Mannar Biosphere Reserve, is situated in between the latitude of 8š 45' N and 9š02'31''N and the longitude of 78š07'17''E and 78š19'18''E. This geographical area runs from the Mouth of Vaippar River to Tuticorin Harbor. The coast of Tuticorin encompasses 4 small islands located an average distance of 4km away from the mainland. These islands are built up of calcareous framework of dead corals and coral reefs. The area is endowed with a combination of ecosystem including mangroves, coral reefs, sea grass and seaweeds.
3. MATERIALS AND METHOD
To fulfill the objectives of this study, five types of approaches have been attempted, such as (i) interpretation of multi-date optical remote sensing data for detection and mapping of changes in coral reefs ecosystem, (ii) interpretation of multi-date remote sensing data for mapping and change detection in Coastal Land-use/Land cover, (iii) interpretation of remote sensing data for mapping the coastal landform, (iv) interpretation of multi-date remote sensing data for mapping and changes in shore line (iv) and analysis of the multi-date bathymetry data for see floor changes.
Figure 1. IRS LISS-III Image of Tuticorin Coast
3.1 Coral Reef Mapping
Geocoded FCC of IRS LISS-II (April 1988) and IRS LISS-III (May 1998) images on 1:50,000 scale were visually interpreted based on image characteristics, various coral reef categories in the Gulf of Mannar have been identified and mapped. In the present study, the classification system developed by Space Application Center for the national coral reef mapping project (Anjali Bahuguna and Nayak, 1994) has been adopted. After identification and delineation, an accuracy test based on probability of success/failure of occurrence (Nayak et al. 1991) was made for sample points. The study area map on 1:50,000 scale was divided into grids and intersecting points of each grid within the study area were taken as sample points for validation of classified satellite data in ground truth checking. Over the ground, out of the 112 sample points, 95 points have been found to be correctly interpreted and thus giving an accuracy of about 84.82 percent. The geo-referenced coral reef maps were digitized, edited, assigned corresponding labels and projected using ARC-INFO GIS. Finally a coral reef map was generated using intercept statistics of coral reef classes in the maps.
3.2 Coastal Land use/Land cover Mapping
Geocoded multi-date (IRS 1A LISS - II of April 1988 and 1D LISS - III of May 1998) FCC on 1:50,000 scale were visually interpreted based on image characteristics, and various land use / land cover categories were identified and mapped around Tuticorin coast. The basic information like transport network, tanks, rivers etc are transferred from Survey of India (SOI) topographic sheet. After identification and delineation, an accuracy test was made for 150 sample points on SOI topographic sheet. The study area map at 1:50, 000 scale was divided into grids and intersecting points of each grid within the study area were taken as sample points for validation of classified satellite data in ground truth checking. Over the ground, out of the 150 sample points, 139 points were found to be correctly interpreted gave an accuracy of about 92 per cent. The georeferenced multi-date land use maps of 1988 and 1998 were digitized in ARC/INFO and were overlaid using TIC coordinates of the study area. Digitized maps were edited and labeled assigned to the polygons. Finally a temporal land use/ land cover change map was generated using intercept operation of ARC/VIEW.
3.3 Coastal Landform Mapping
Geocoded FCC of IRS1D LISS - III imagery on 1:50,000 scale were visually interpreted based on image characteristics, and various coastal geomorphic categories were identified and mapped along the coast of Gulf of Mannar. In the present study, an image interpretation key indicating tone/colour, size and pattern developed by Space Applications Center, Ahmedabad (SAC 1991) was mad. After identification and delineation, an accuracy test was made for 118 sample points on SOI topographic sheet. The study area map at 1:50, 000 scale was divided into grids and intersecting points of each grid within the study area were taken as sample points for validation of classified satellite data in ground truth checking. Over the ground, out of the 118 sample points, 105 points were found to be correctly interpreted giving an accuracy of about 90 per cent. The georeferenced geomorphology map was digitized, edited, labeled and projected using ARC-INFO GIS. Finally a coastal geomorphology map was generated using intercept operation of ARC/VIEW. The area statistics of coastal geomorphology in the map were generated.
3.4 Shoreline Change Mapping
Geocoded FCC of IRS LISS - III (May 1998) satellite data on 1:50,000 scale belonging to low tide period and Survey of India Topographic map (SOI 1969) on 1: 50000 scale were used to prepare shore line change map. The low tide line (as shore line) from SOI topographical sheet and satellite image were extracted and mapped on 1: 50,000 scale by visual techniques. After identification and delineation, coastal villages, some monuments, lighthouse etc were selected as sample points on SOI topographic sheet for ground truth checking. During the time of ground truth study the entire coast was checked with local people and some necessary corrections were carried out on primary shoreline map. The georeferenced multi date shoreline maps were carefully digitized in ARC/INFO and were overlaid using TIC coordinates of the study area. Digitized maps were edited and labels assigned to the polygons. Finally a temporal shoreline change map was generated using intercept operation of ARC/VIEW, and summary statistic was generated for erosion and accretion areas using STATISTIC program of ARC/INFO. The changes were estimated for a period of 30 years between 1969 and 1998.
3.5 Bathymetry Mapping
Bathymetry map of study area on 1: 50,000 scale was prepared manually using 1975 Naval Hydrographic Chart. The prepared bathymetry map was digitized into ARC-INFO and a digital bathymetry map was prepared. The prepared bathymetry map was updated during fieldwork, which was carried out during April 1999. The depth of the sea bottom was measured (within 10m depth) using eco-sounder (ODEM) and Global Position System (GPS) along the Tuticorin coastal region. The depth values are recorded at a particular location with reference to chart datum (1975). The measured depths values were tide corrected with respect to time and then converted with respect chart datum (0.2m accuracy). Measured tide table from the Tuticorin port was used for final data conversion to chart datum.
4. RESULTS AND DISCUSSION
4.1 Changes in Coral Reefs
Coral reefs are considered as one of the most important critical resources for various ecological, environmental and socio-economic reasons. Coral ecosystem face many threats, of which some are natural like storms and waves particularly tropical storms and cyclones that cause major damage to reefs. The majority of damage to coral reefs around the world has been through direct anthropogenic stress (Grigg and Dollar 1990). In Tuticorin Coast, there are 4 low sandy islands, fringing reefs along the windward side of the islands protects the island from direct wave action. Morphology of these islands is very dynamic. Thanikachalam and Ramachandran (2002b and 2003) carried out extensive work on islands of Gulf of Mannar, and concluded that the Tuticorin group of islands are migrating towards landward side (13.21m/year), whereas Vembar, Keelakkarai and Mandapam group of islands are migrating towards seaward side (3.18, 32.9 and 74.88m/year). This is caused due to natural as well as anthropogenic factors. As a result of above, coral reef degradation is very severe in Gulf of Mannar. Thanikachalam and Ramachandran (2002b and 2003) estimated that 25.52kmē area of coral reef was lost over a period of ten years (1988-1998). The coral reefs of Tuticorin group of Islands in Gulf of Mannar have been damaged due to the basting and dredging activities result in high sedimentation on the coral reef, discharge of effluents from petrochemical industries along the coast and fly ash discharges from thermal plants (Ramanujam and Mukesh 1998 ; Thanikachalam and Ramachandran (2002b and 2003)).
The total coral reef area in Tuticorin coast based on the present study (1998) is about 3.70 kmē, of which reef area covers 3.50 kmē and reef vegetation covers 0.28 kmē. The analysis of multi-date satellite data indicates that nearly 2 kmē area of coral reef was lost over a period of ten years between 1988 and1998 (Figure 2a and 2b). Through remote sensing the demarcation and estimation of the extent of dead and live coral is not possible. However, during the time of ground truth it was identified and estimated that nearly 67.2% of the corals were dead corals, 19% of coral reefs are directly removed by coral mining and the remaining 13% are the live corals.
4.2 Land use/land cover changes
Multi temporal satellite data used in the present study enabled to observing the land use and land cover changes in the study area from 1988 to 1998 (Figure 3a and 3b). Over the past 10 years, areas of some land use classes have increased, areas of some classes have decreased and some categories have changed in to another category in the study area. These changes have taken place due to the increase in population in towns and villages along the coast and various other developmental activities. The major land use/land cover changes has occurred in the following classes: (1) cropland area has increased from 116.8 to 124.1 kmē, (2) fallow land decreased has from 8.2 to 3.2 kmē, (3) agricultural plantation has reduced from 4.03 to 1.5 kmē, (4) forest plantation has increased 7.2 kmē, (5) scrublands has decrease from 82.45 to 55.2 kmē, (6) sandy area has reduced to 0.55 kmē, (7) tanks has increased from 12.2 to 14.8 kmē and (8) the mangrove area has increased from0.65 to 1.4 kmē.
Figure 2. Coral reef maps of Tuticorin (a) 1988 and (b) 1998
Figure 3. Coastal Land use/ Land cover maps of Tutcorin
(a) 1988 and (b) 1998
4.3 Coastal Erosion and Accretion
In the coastline between Vaippar River and Tuticorin Harbor, erosion and accretion areas have been identified. The areas of erosion and accretion have been estimated as 0.84 and 2.73 kmē, respectively, over a period of 30 years (1969 to1998). Along this shore, twelve accretion and seven erosional sites have been identified (Figure 4). The estimated average rates of accretion and erosion in this area are in the order of 4.3 and 2.73 m/year, respectively.
Figure 4. Shoreline change map between 1969 and 1998
There are 4 islands in Tuticorin coast. Fringing reef along the windward side of the islands protects the islands from direct wave action. Morphology of sandy islands is very dynamic. Ramanujam et al., (1995), Ramanujam and Mukesh (1998) and Thanikachalam and Ramachandran (2002b ans 2003) have studied the morphological variation of Tuticorin group of islands. The comparison of 1969 and 1998 maps of islands of Tuticorin showed changes in their shape, size and location and these have been caused by erosion and accretion of shore (Figure 4). The total area of erosion and accretion were calculated as 0.35 kmē and 0.13 kmē, respectively, during the 30-years period. In the study region the waves are in the northeast and southwest direction and wind direction is similar to the wave direction. Mining of stony corals from the reef area, especially from Tuticorin group of islands, for building, industrial and chemical purposes have destabilised the formation of Tuticorin group of islands (Ramanujam et al., 1995). Hence the waves hit directly on south, southeast and southwest shores of these islands, causing erosion. These eroded sediments are then transported by wave-induced currents and deposited at the leeward sides of these islands. By such repeated processes, the windward sides of the island get reduced and leeward sides of the islands are accreted. Hence size, shape and location of these islands have changed. All islands in Tuticorin group have been migrating towards mainland. It is estimated that (1) 528.74 m of Van Island, (2) 118 m of Koswarai Island and (3) 137 m of Kariya Shulli migrated towards mainland between 1969 to 1998, According to Ramanujam et al., (1995) the landward migration of islands in Tuticorin region are caused by sea level variation and mining of reef material. Vilangu Shuli Island is one of the islands in Tuticorin group, situated at 6.25 km from Sippikkulam. The entire part of this island was eroded due to the direct action of waves; it may be caused by coral reef mining. Ramanujam et al., (1995) showed that the direct wave attack on this island eroded the whole area below the sea level. The area of erosion was estimated as 0.06 kmē.
4.4 Seafloor Changes
Any changes in seafloor may be the result of sea-level variation or due to a change in the elevation of land surface. Changes in absolute water-surface levels are worldwide due to the interconnectivity of the oceans and are termed as eustatic changes. Changes in the absolute level of the land are localized. They may be due to tectonic adjustments or due to adjustments caused by the distribution of weight on the land surface. As and when sedimentation or ice build-up occurs, such changes happen and are known as isostatic. A rise in the sea level or down warping of land would involve the opposite movements of sea and land. Synonymous with positive and negative changes are the forms of sea-level transgression and regression, although in many cases these terms also refer to the horizontal movement of the shoreline associated with vertical changes of sea level.
Recent bathymetry map of 1999 have been compared with bathymetry map of 1975 (Figure 5a and 5b). It reflects that the depth of seafloor decreased along the mainland coast and around the islands in the study area. It may be due to emerging of land due to tectonic movement. In very few places particularly at river mouths and in island areas, the sea floor level has increased, which may be due to erosion caused by anthropogenic activities. The average reduction in the depth has been calculated as 0.31m over a period of 24 years (1975-1999).
Figure 5. (a) Bathymetry map of Tuticorin Neare shore area durine1975 and (b) 1999
The depth of sea floor has reduced along 8 transects and increased along 7 transects around the Vilangu Shuli and Van Islands. This reduction of depth may be caused by deposition of sediment due to ocean currents, whereas the increasing of sea floor depth may be due to erosion caused by anthropogenic activities (Coral mining). Evidences show that the entire Shuli Island was eroded and lies below the sea level. The average amount of reduction and increase of sea floor depth around the Vilangu Shuli Island were calculated as 0.08 m and 0.35 m over a period of 24 years. Around Van Island the decreasing and increasing trend of sea floor depth have been calculated as 0.88 and 0.62 m over a period of 24 years, this may be due to emerging of land or lowering of sea level by tectonic activities. Some of the field evidence such as seaward migration of shoreline, increasing the areal extent of mangroves swamp and mudflat, new spit formation, occurrence of bunch of beach ridges, swales and strandlines (Thanikachalam and Ramachandran, 2002b and 2003) in the Tuticorin coast proved that the coast is going on emerging by tectonic movement
4.5 Coastal Landform
The coastal plain between Vaippar and Tuticorin has various geomorphic units with different types of configurations (Figure 6). The geomorphic units, interpreted from remotely sensed data and checked subsequently through fieldwork, have been categorized into four genetic classes - marine, fluvio-marine, fluvial, aeolian and biogenic landforms.
Figure 6. Coastal geomorphology map of Tuticorin coast
4.5.1 Marine Landform
In the coastal zone various marine landform features such as beaches, spit, mudflat, offshore islands and coral been identified. Beaches are extensively developed along coast between Tuticorin and south of Sippikulam covered an area of 4.75 kmē. All along the shore the beach is observed to be gently sloping and marked with altered crusts and troughs that are formed due to wave action. Among the various depositional landform features the formation of spit is a significant feature of recent age. South of Tuticorin coastal area two spit formations have been observed with 0.75 to 2 km long and tongue shaped. It appears to have been built by the sediments brought by long shore current during southwest monsoon. As the Gulf of Mannar is on the lee of the northeast monsoon, there is no long-shore drift from the northeast that might be the cause for the inward curving of this spit (Ahmad 1972). It can be explained that the Tuticorin spit might have been the result of the long shore currents during monsoon and the sediments discharged by Tamiraparani River. Mudflat is a flat area containing a fluid to plastic mixture of finely derived particles of solid material mainly silt and clay water. They are always associated with silted environments like lagoons, estuaries and other embankments. The mudflats are very common in Vaippar River mouth and Kallar River mouth. The area covered by mudflat has been estimated to be 2.4 kmē. A chain of 4 islands has been observed along the offshore region of Tuticorin. They extend from East of Vaippar River Mouth to Tuticorin. All islands are made up of a calcareous framework of dead reef and sand. They have a low and narrow sandy coast. Around all offshore islands, well-developed coral reefs have been noticed. Geomorphologically, coral reefs in this area are of fringing type, though some patchy corals are also observed in Kariya Shuli Island.
4.5.1 Fluvio-Marine Landforms
Deltaic plains are very common landform along the coast of Tuticorin, which is predominantly controlled by Fluvio-Marine processes. The area around the river courses of Vaippar had vast deltas, but at present they are found to be inactive. The deltaic plains are marked by flat and vast areas, having vegetation cover. Number of tanks has been noticed on the deltaic plains. The total area has been estimated to be about 18.2 kmē.
4.5.2 Fluvial Landforms
Well-established rivers usually have their floors covered with alluvium, in which the normal flow channel is covered. The surface of low relief on the alluvium from the banks of the low-water channel to the base of the valley walls is called the flood plain of a river. Flood plain and their major morphologic subdivisions are primarily deposited landforms. Floodplains have been very clearly observed along the riverbanks of Kallar, Vaippar in Tuticorin coast. The total estimated area of flood plain in study region is 14.4 kmē.
4.5.3 Aeolian Landforms
Sand dunes and Teri dunes are the most common aeolian land forms in the coastal zone of Tuticorin area. Almost entire coastal plains in the study area are covered by sand dune. The area covered by dune complex has been estimated to be about 98.3 kmē.
In the coastal plains between south of Vaippar River and Tuticorin near Maravanmadam seven patches of teri dunes have been observed with a thick cover of vegetation. North of Panaiyur, oval shaped teri dunes with sparse vegetation have been observed. The areal extent of this dune has been calculated to be 6.27 kmē. Another two teri dunes have been observed near Kumarapuram. They cover an area of 2.67 kmē. Near Pandiyapuram, rounded dune complex has been observed with thick vegetation. The area of this dune has been estimated as 4.27 kmē. Near Milavittam small rounded patches of teri dune complex covering an area of 2 kmē has been observed. Two other teri dune complexes have been observed near Maravanmatam area with thick vegetation. They cover an area of 3.08 kmē. All teri dune complexes in this area are trending in the northeast to southeast direction.
4.5.4 Biogenic Landform
Back swamp is very common biogenic landform along the coast of Tuticorin. It occurs in marshy areas along the coast; they particularly occur at the edge of the tropical or sub-tropical seas, in bays lagoons and estuarine regions (Gerlech 1973). Small back swamp areas have been observed in the areas near the mouth of Korampallam odai around Tuticorin coast and cover an area of 1.3 kmē. These swamps are covered by mangrove vegetation.
Remote sensing data and GIS techniques are very useful for mapping the coastal environment changes such as coral reef, mangrove, coastal land use and land cover, shoreline and bathymetry. Thus remote sensing data helped in studying the coastal ecosystem on a regional scale and the time series data helped in identifying the problems and their causes. Remote sensing and GIS play an important role as decision-making tools, to assess the present ecological status and for future predictions to conserve and manage coastal ecosystem. As the coastal region is spread over a larger area, it is very difficult to assess the various ecological parameters status without satellite.
Ahmad, E., 1972. Coastal geomorphology of India. Orient Longman, New Delhi, 222 p.
- Anjali Bahuguna, and Nayak, S.R., (1994). Coral mapping of India. In: Scientific Note SAC/RAS/RASG/DOD-COS/SN/07/94), Space Application Center, Ahmedabad, India.
- Grigg, R.W. and Dollar, S.J., (1990). Natural and anthropogenic disturbance on coral reefs. Ecosystem of the world, Elsevier Science Publishing, New York, pp.439-452.
- Krishnamoorthy, R., 1995. Remote sensing of mangrove forest in Tamil Nadu coast, India. Ph.D. thesis, Anna University, 202p.
- Nayak, S.R., (2002). Application of remote sensing to coastal zone management in India. Paper presented in ISPRS Technical Commission VII Symposium on Remote Sensing and Environment Monitoring and ISRS Annual Convension held at Hyderabad, India.
- Nayak, S.R. et al., (1991). Manul for mapping of coastal wetlands/landforms and shoreline changes using satellite data. Technical Note: IRS-UP/SAC/MCE/TN/32/91, Space Application Center, Ahmedabad, India.
- Ramanujam, N. and Mukesh, M.V., (1998). Geomorphology of Tuticorin Group of Island. In: Biodiversity of Gulf of Mannar Marine Biosphere Reserve. (Eds.: M. Rajeswari, K. Anand, Dorairaj and A. Parida). M.S. Swaminathan Research Foundation, Chennai. pp.32-37.
- Ramanujam, N., Mukesh, M.V., Sabeen, H.M. and Preeja, N.B., (1995). Morphological variation in some islands in the Gulf of Mannar. Journal of Geogical Survey of India, Vol.45, pp.703-708.
- SAC, (1991). Manual for mapping coastal wetlands/landform and shoreline changes using satellite data. Space Application Center, Ahmedabad, India, No: IRS-UP/SAC/MCE/TN/32/91, 63p.
- Thanikachalam, M and Ramachandran, S., (2002a). Remote Sensing and GIS Techniques for Monitoring and Conservation of Coral Reef in Gulf of Manar, Southeast Coast of India. http://www.envis-eptri-org/images/abstracts.pdf
- Thanikachalam, M. and Ramachandran, S., 2002. Remote sensinf and GIS techniques for monitoring and conservation of coral reefs in Gulf of Mannar south east coast of India. http://www.envis-eptri-org/images/abstracts.pdf
- Thanikachalam, M and Ramachandran, S., (2002c). Conservation of coral reefs in Gulf of Mannar: A remote sensing and GIS approach. Indian society of Geomatics (ISG) Newsletter, Special Issue on Coastal & Marine Environmant, Vol. 8(2 & 3), pp.65-71.
- Thanikachalam, M.and Ramachandran, S., (2003). Shoreline changes along the coast of Gulf of Mannar and its Influence on Coral Reefs: A Remote Sensing and GIS Approach. Journal of the Indian Society of Remote Sensing, Vol.31, No.3, 2003.
Page 1 of 1