ABSTRACT

This study on coastal zone environment management with emphasis on mangrove ecosystems was carried out by the National Research Council of Thailand and the Royal Forest Department with the support of the Economic and Social Commission for Asia and the Pacific. The primary objective of the study was to identify an effective approach for sustainable mangrove forest management using remote sensing and GIS technology.

The study was conducted in three phases: (a) assessing forest land use in 1973, 1987, 1993 and 1998, based on visual interpretation of satellite imagery at the scale of 1:50,000; (b) evaluating forest land-use area and change using GIS; and (c) developing a plan for sustainable forest land use based on additional relevant thematic data.

The major results of this study were (a) the compilation of a relevant thematic database, (b) assessment of forest land use in 1973, 1987, 1993 and 1998 and (c) assessment of forest distribution in 1987, 1993 and 1998, as well as an assessment of the change of land use between 1987-1993 and 1993-1998 and the development of a forest land-use plan.

In conclusion, remote sensing appears to be a significant tool for assessment and monitoring of coastal zone resources, especially mangrove forest. In addition, planning and management of forest land use is easily and effectively conducted using GIS. However, the integration of remote sensing and GIS for the development of mangrove forest management plans by natural resource managers and planners is necessary.

Content

1. Introduction

Thailand's coastal lands are richly endowed with natural resources, such as mangrove forests, fertile soil, minerals and beautiful scenery. Coastal seas also support coral reefs, sea grass beds and diverse fish stocks. The coastal zone of Thailand covers an area of about 17,000 sq km with a total coastal length of about 2,614 km. The western coast faces the Andaman Sea, while the eastern coast faces the Gulf of Thailand. Almost one third of the country borders the sea, and about 76 per cent of the total population live within a few kilometers of the coast.

Mangrove forests are one of the most valuable coastal resources, important for their multiple economic, ecological, scientific and culture resources for present and future generations. They are an important component of Thailand's coastal zone ecosystem. In addition, mangrove forests are utilized as a source of fuelwood and pole production; they provide a nursery and breeding ground for many commercial fish and prawn species. They also protect against coastal erosion and provide a habitat for some protected species of birds and animals.

In recent decades, the coastal zone has been subjected to the effects of growing population and economic pressure manifested by a variety of activities such as aquaculture, tin mining, waste disposal, construction and industrialization. Major coastal resources in Thailand, especially mangrove forests, beaches, coral reefs and marine water have deteriorated or been degraded. In addition, the area of mangrove forest in Thailand has decreased rapidly because of the expansion of shrimp farming.

In order to use coastal resources on a sustainable basis, a proper management planning process is necessary. Remote sensing and GIS are the major tools for undertaking the plan. Since 1972, Thailand has used remotely sensed satellite data for assessment of natural resources in many fields, proving that remote sensing is a useful tool for assessing and monitoring the country's coastal resources, especially mangrove forests. Mangrove forest assessment and monitoring have been conducted continuously in Thailand since 1979.

In 1987, the Royal Forest Department used remotely sensed data as baseline information for mangrove land-use zoning and defined mangrove forests into three principal zones: Preservation, Economic A and Economic B, covering an area of 427, 1,997 and 1,300 sq km, respectively. However, from a recent study it was found that shrimp farms occupied 44.80 sq km, or 11 per cent, of the Mangrove Forest Preservation Zone; 229 sq km, or 12 per cent, of Mangrove Forest Economic Zone A; and 376 sq km, or 29 per cent, of Mangrove Forest Economic Zone B (National Research Council of Thailand 1991).

In 1990, high-resolution data of Landsat TM and SPOT PLA (the High-resolution Visible scanner in panchromatic mode), which can provide more effective information than Landsat MSS, was used for monitoring of mangrove forest land-use zones. It was found that the Landsat TM bands 4-5-3/R-G-B with linear stretching technique can be used to differentiate mangrove forests from other vegetation, and SPOT PLA shows a better pattern for shrimp farms and built-up areas. In addition, the combination of TM4-TM5-PLA/R-G-B via digital classification can easily depict the mangrove forest zonation that corresponds to the dominant species (Ratanasermpong and Silapathong 1990).

Furthermore, thanks to the development of computer technology and a decrease in the price of computer equipment, GIS has been rapidly developed. It also has been proven that an effective plan can be easily and effectively conducted under GIS, as shown in the study by Ongsomwang (1998), which utilized GIS as a tool for assessment of forest potential in Thailand.

Therefore the integration of remote sensing and GIS for mangrove forest management is considered an important tool for the development of effective plans by natural resource managers and planners. The main objective of this study is to demonstrate an effective approach for sustainable mangrove forest management by using remote sensing and GIS technology.

Content

2. Description of the study area

Ao Sawi - Thung Kha is located between the latitudes 10°15 and 10°30 N and the longitudes 99°7 and 99°20 E, at the mouth of Khlong Sawi and Khlong Thung Kha, in Chumphon Province on the south-eastern seaboard of Thailand. It has a total surface area of about 675 sq km. A location map of the study area is shown in figure 1.

The study site has a tropical monsoon climate with two main seasons (wet and dry). The average annual rainfall is 2,010 mm. The peak rainfall month is in November, while the driest month is in March. The wet season is from May to November. The mean annual temperature is 26.90°C, and the highest relative humidity is in October. Tides are diurnal (one high and one low tide a day) with an average amplitude of 1.04 m. The monthly mean sea level is highest in November and at the lowest in March. The monthly discharge from the rivers of Chumphon is at its highest in August and lowest in March. The monthly run-off is highest in November.

The present land use and land cover in the study area are composed of mangrove forest, tropical evergreen forest, freshwater swamp forest, paddy fields, rubber plantations, oil palm plantations, mixed orchards, coconut plantations, villages, wetland, wasteland, water bodies and shrimp farms.

Content

3. Geographic database contents

Geographic databases, which consist of base and thematic maps and materials relevant to the study area, were compiled and established using a GIS system. The geographic database contents are summarized in the following sections.

3.1 Maps

3.1.1 Base maps

Royal Thai Survey Department topographic maps from 1973, at a scale of 1:50,000, were used as base maps for the geographic database of the study area. Map sheet numbers used were 4729 III and 4829 III and IV.

3.1.2 Thematic maps

The type and source of the thematic maps that were compiled in this study are summarized in table 1.

3.2 Images

3.2.1 Satellite imagery

Satellite data used are listed in table 2.

3.2.2 Aerial photographs

Panchromatic aerial photographs at the scale of 1:15,000 and 1:50,000 taken on 3 March 1995 by the Royal Thai Survey Department were used as reference information.

¹"MLA" refers to the High-resolution Visible scanner in multispectral mode.

3.3 Ground photographs

Co lour photographs taken in November 1998 during a seven-day field survey were used as sample areas for image interpretation and classification.

3.4 Data dictionary

The compiled and generated geographic databases are summarized as shown in table 3.

Content

4. Description of methodology

The main tasks were to classify the forest land use in 1987, 1993 and 1998 based on visual interpretation of satellite images at the scale of 1:50,000. The assessment of land use/land cover and change was then conducted using GIS. Finally, the proposal for a sustainable forest land-use plan was developed based on extracted and relevant thematic data. A schematic flowchart methodology is shown in figure 2. The details of each main step are described below.

4.1 Digital image processing

In this study, satellite images and data were interpreted and analyzed for land-use/land-cover classification by visual interpretation and digital image classification, respectively. The practical sequence for data processing and analysis of satellite data used is as follows.

4.1.1 Geometric correction

Geometric correction of the Landsat TM data acquired on 10 March 1998 was first registered to a Universal Transverse Mercator (UTM) projection by using a linear transformation matrix with a nearest neighbor resampling for 25 x 25 m pixel size. This rectified image was used as a reference image for correction of geometric error of Landsat TM acquired on 25 December 1987 and SPOT MLA acquired on 22 February 1993 as image-to-image registration.

4.1.2 Image enhancement

In this step, image enhancement techniques were temporarily applied to improve the signature appearance on the images for visual interpretation. In addition, temporary enhanced data was used to identify a training area for digital image classification.

4.1.3 Image classification

There are several alternative approaches for decision and classification. In this step, visual interpretation and automated classification were applied to extract forest land-use categories.

4.1.3.1 Visual interpretation of satellite images

Satellite images were visually interpreted for land-use/land-cover classification based on the appearance signature in the image and background knowledge. The band combination applied for Landsat TM images acquired on 25 December 1987 and 10 March 1998 was bands 4, 5 and 3 (RGB), while the band combination for SPOT MLA acquired on 22 February 1993 was bands 3, 2 and 1 (RGB).

4.1.3.2 Automated classification

Satellite data that have been subjected to processing and feature extraction were classified by assigning individual pixels to specific classes. First, unsupervised classification with a clustering classifier was performed for a qualitative interpretation. Then, supervised classification with a maximum likelihood classifier was conducted based on ground information derived from aerial photographs at the scale of 1:15,000 acquired in 3 March 1995 and the ground survey in November 1998.

4.1.4 Ground verification

The preliminary land-use/land-cover maps using visual interpretation and supervised classification were produced for ground verification by using subjective random sampling. The location of checked spots was identified by the Global Positioning System (GPS) and ground photographs were also recorded. In addition, a sample plot for the study of the plant profile was also allocated in each stratum of mangrove forest.

4.1.5 Land-use/land-cover maps

The final maps of land use/land cover in 1987, 1993 and 1998 by visual interpretation and land use/land cover in 1998 by automated classification were produced.

4.2. Establishment of geographic databases

The geographic databases were established under the vector GIS of the PC ARC/INFO system. In practice, all thematic maps were firstly digitized as digital coverage of the ARC/INFO system with the UTM coordinate system. The procedure to create digital coverage under the PC ARC/INFO system was composed of the following seven steps:

• Step 1. Preparation of the map sheet for digitizing.
• Step 2. Digitization the coverage.
• Step 3. Identification and correction of digitizing errors.
• Step 4. Definition of features and building of topology.
• Step 5. Identification and correction of topology errors.
• Step 6. Assigning attributes to coverage features.
• Step 7. Identification and correction of attribute coding errors.

In practice, master ground control points (GCP) were firstly designed and generated in UTM coordinates and then digitized for each thematic layer. The digitizing tolerance under PC ARC/INFO as coverage was defined as in table 4.

4.3 Analysis of GIS

Analysis of GIS was performed under vector GIS of PC ARC/INFO. Four main outputs were derived based on the compiled and extracted geographic databases:

(a) Land use and land cover in 1973, 1987, 1993 and 1998;

(b) Forest distribution in 1973, 1987, 1993 and 1998;

(c) Change in land use and land cover between 1987-1993 and 1993-1998;

(d) Mangrove forest land-use plan.

The basic technique employed in the GIS analysis was overlay operation, in which two layers were input to extract a new layer. For example, for change in land use and land cover between 1987 and 1993, "the two inputs layers were land use and land cover in 1987 and 1993". The new layer was "change of land use and land cover during 1987 and 1993".

Furthermore, in the case of mangrove forest land-use planning based on mangrove forest land-use zonation and land use/land cover in 1998, the definition of mangrove forest land-use zones and characteristics of land use and land cover were applied to define the new mangrove forest land-use classes.

Content

5. Results

The major results are the compilation of relevant thematic databases and assessment of forest land use and forest distribution in 1973, 1987, 1993 and 1998, as well as changes in land use and land cover between 1987-1993 and 1993-1998 and development of a proposed forest land-use plan. The details of main results are summarized as follows.

5.1 Assessment of land use/land cover

The land-use and land-cover categories in 1987, 1993 and 1998 were extracted by visual interpretation of satellite images at a scale of 1:50,000, as shown in figures 3 to 5, respectively. Categories consist of (a) agriculture area, (b) built-up area, (c) forest area, (d) water bodies, (e) wasteland, (f) wetland and (g) shrimp farms. The land-use and land-cover categories in 1973, directly interpreted from a topographic map at a scale of 1:50,000, are composed of (a) agriculture area, (b) built-up area, (c) forest area, (d) water bodies, (e) wasteland and (f) wetland.

The visual interpretation key from Landsat TM imagery in 1987 is shown in table 5. The definition and characteristic of each land-use/land-cover category is described as follows:

(a) Agriculture area: The agriculture areas consist of paddy fields, cash and field crops and permanent crops such as fruit trees and rubber and oil palm plantations. Perennial crops, including paddy, cash and field crops, are mostly found on flat plains, as shown in figure 6, while permanent crops, including fruit trees, rubber and oil palm, are found on undulating to mountainous terrain (figure 7);

(b) Built-up area: Built-up areas are composed of settlements and industry, mostly found on the flat area and located near the rivers, as shown in figure 8;

(c) Forest area: Forest area is defined as the permanent natural forest and reforested area. The main natural forest includes mangrove, tropical evergreen and freshwater swamp forest, as shown in figures 9 to 11, respectively. Mangrove is found on the tidal flat along the shoreline. Freshwater swamp forest is situated near the mangrove and upward towards the mainland and is inundated by fresh water. Tropical evergreen forest is mostly found on the hills and in mountainous areas;

(d) Water bodies: Water bodies include ponds, lakes, canals, rivers and sea;

(e) Wasteland: Wasteland is defined as unused areas such as old laterite pits, abandoned clear-cut forest areas and abandoned shrimp farms, as shown in figure 12;

(f) Wetland: Wetlands are situated on the flat areas and are inundated by fresh or brackish water as a pond or lake. Wetlands are occupied by grass, as shown in figure 13;

(g) Shrimp farm: Shrimp farms include traditional and intensive shrimp farms, as shown in figure 14 and 15, respectively. They are mostly found on the tidal flats in former mangrove areas.

The components of land use and land cover in 1973, 1987, 1993 and 1998 are summarized and compared in tables 6 and 7 and figure 16. In addition, the distribution of land use and land cover in 1973, 1987, 1993 and 1998 are shown in figures 17 to 20. Furthermore, land use and land cover in 1998 use digital image classification with a maximum likelihood classifier.

The results show six main land-use and land-cover classes in the study area between 1973 and 1998, including agriculture area, built-up area, forest area, water bodies, wasteland and wetland. In fact, shrimp farms were not present in 1973. The change of area in each land use and land cover can be explained in each category as follows:

(a) Agriculture area: Agriculture areas are situated in river basins or flat to undulating land. The extent of agriculture area increased during 1973 to 1987 by about 10,051 ha. However, agriculture area decreased between 1987-1993 and 1993 and 1998 by 1,752 and 974 ha, respectively;

(b) Built-up area: The extent of built-up areas dramatically increased between 1973 and 1993. The area increased from 93 ha in 1973 to 530 and 1,293 ha in 1987 and 1993, respectively. However, it decreased slightly during 1993-1998 by 8 ha;

(c) Forest area: Forest area was the largest component of land use and land cover, excluding water bodies, in 1973. The extent of forest dramatically decreased between 1973 and 1987 by an area of 13,785 ha. In addition, it was found that forest area continued to decrease between 1987-1993 and 1993-1998 by an area of 1,976 and 238 ha, respectively;

(d) Water bodies: The extent of water bodies decreased between 1973-1993 and 1993-1998 by an area of 358 ha and 87 ha, respectively. However, water bodies increased between 1987 and 1993 by 25 ha;

(e) Wasteland: The area of wasteland increased between 1973 and 1998, from 1,065 ha in 1973 to 3,010 ha, 3,731 ha and 4,577 ha in 1987, 1993 and 1998, respectively:

(f) Wetland: The area of wetland decreased during 1973 and 1987, by approximately 64 ha. However, no change in wetland area was recorded between 1987 and 1998;

(g) Shrimp farm: In 1973, no shrimp farms existed in the study area. The increases in shrimp farm area between the years 1973-1987, 1987-1993 and 1993-1998 were 1,775 ha, 2,219 ha and 460 ha, respectively.

5.2 Forest distribution

Based on visual interpretation of satellite images at the scale 1:50,000, forest area was further classified into three classes according to the their signature appearance on the image and their characteristics. The major forest types are mangrove, evergreen and freshwater swamp forest. The distributions of forest in 1973, 1987, 1993, 1998 are shown in figures 22 to 25. In addition, mangrove forests are further classified based on crown cover into three classes, namely mangrove, with more than 75 per cent coverage, mangrove with 50-75 per cent coverage and mangrove with less than 50 per cent coverage.

The component of forest distribution in 1987, 1993 and 1998 are summarized and compared as shown in tables 8 and 9. The characteristics and changes of forest classes are described as follows:

(a) Mangrove with >75 per cent coverage: These mangrove forests consist mostly of Rhizophora apiculata and R. mucronata situated along the river channels and behind the pioneer species zone. The extent of mangrove forest with density of more than 75 per cent increased during 1987 to 1998 and covered an area of 290 ha, 488 ha and 519 ha, in 1987, 1993 and 1989, respectively. The main cause of the increase in area is natural succession and reforestation (figure 26);

(b) Mangrove with 50-75 per cent coverage: These mangrove forests consist mostly of pioneer species, Avicennia alba and A. officinalis, situated on the outer fringe facing the sea. The extent of mangrove with density between 50 to 75 per cent decreased slightly during 1987 to 1998. These areas covered an area of 1,267 ha, 1,140 ha and 1,091 ha, in 1987, 1993 and 1989, respectively. The main causes of decrease are natural succession by Rhizophora spp. and die-back caused by sand deposits (figure 27);

(c) Mangrove with <50 per cent coverage: These mangrove forests consist mostly of mixed species, including Bruguiera cylindrica, Ceriops tagal, Excoecaria agallocha, Ficus sp., Heritiera littoralis, Hibiscus tiliaceus, Intsia bijuga, Xylocarpus granatum, and X. moluccensis, situated behind the Rhizophora species zone. The extent of mangrove with density of less than 50 per cent decreased from 1987 to 1998, covering an area of 3,846 ha, 1,900 ha and 1,696 ha, in 1987, 1993 and 1989, respectively. The main cause of the decrease is the expansion of shrimp farms;

(d) Evergreen forest: Evergreen forests situated in the mountainous area decreased between 1987 and 1998. These forests covered an area of 2,691 ha, 2,583 ha and 2,565 ha, in 1987, 1993 and 1989, respectively. The main cause of the decrease is the expansion of agriculture areas;

(e) Freshwater swamp forest: This forest consists mostly of Melaleuca leucadendra situated on flat areas behind the mangrove forest. The extent of swamp forest increased during 1987-1993 and decreased between 1993 and 1998. This forest type covered an area of 517 ha, 551 ha and 527 ha, in 1987, 1993 and 1989, respectively.

5.2.1 Plant profile

During field work in November 1998, sample plots of mangrove forest were established based on mangrove forest zonation, which included mangrove forest zone I (Pioneer species), mangrove forest zone II (Rhizophora species) and mangrove forest zone III (Mixed species). The appearance of mangrove forest zonation on the satellite images is easily recognized by visual interpretation, as shown in figure 28. In each sample plot, data gathered on number, species, diameter-base-height (DBH) and height of each tree and in selected three plots, the plant profile was drawn. The characteristics of each zone are describe as follows:

(a) Mangrove forest zone I (pioneer species): Mangrove forest zone I is the outermost zone facing the sea, where the mud has a very soft and wet consistency. It is usually submerged about 4 to 8 hours by the flood tide. The main species of this zone are Avicennia alba and Avicennia officinalis, which are recognized as pioneer species for mangrove forests. The average height of trees is about 10.47 m and DBH is 18.37 cm. The crown cover is about 50-70 per cent. The plant profile of mangrove forest zone I is shown in figure 29;

(b) Mangrove forest zone II (Rhizophora species): Mangrove forest zone II is the zone situated along river channels or inner muddy land behind the zone of pioneer species. The boundary of mangrove forest zone I and II is obviously separated by the trees height different, as shown in figure 30. The main species in this zone are Rhizophora apiculata and Rhizophora mucronata. In general, mangrove forest zone II is the most productive in terms of forest utilization such as charcoal and fuelwood. The average height of the tree is about 11.02 m and DBH is 10.12 cm. The crown cover is more than 70 per cent. The plant profile of mangrove forest zone II is shown in figure 31;

(c) Mangrove forest zone III (mixed species): Mangrove forest zone III is located on hardened soil in the inner zone behind the Rhizophora species zone. This area is covered by the sea water only during the high tide. Species found vary frequently depending on the site. The main species include Bruguiera spp., Ceriops spp., Excoecaria agallocha, Ficus spp., Heritiera littoralis, Hibiscus tiliaceus, Intsia bijuga, Lumnitzera spp., and Xylocarpus spp. The average height of trees is about 7.74 m and DBH is 10.75 cm. The crown cover is less than 50 per cent. The plant profile of mangrove forest zone III is shown in figure 32.

Data from the sample plots for number, species, DBH and height of trees are summarized in tables 1 to 7 in the appendix.

5.3 Changes in land use and land cover

The study of changes in land use and land cover that was conducted under vector GIS of PC ARC/INFO was conducted for only two periods, "land use and land cover between the years 1987 and 1993 and land use and land cover between the years 1993 and 1998". In fact, this analysis utilized land-use and land-cover data in 1987, 1993 and 1998 extracted by the project, based on visual interpretation of satellite images at the scale of 1:50,000. The details of changes in land use and land cover are separately summarized below.

5.3.1 Changes in land use and land cover between 1987 and 1993

The changes in land use and land cover between 1987 and 1993 were determined by overlay operation, as shown in figure 33, and as the coincident matrix in table 10. The results show the change in land-use and land-cover categories from 1987 to 1993. This result can be simplified as patterns of major and minor changes based on the amount of area changed, as shown in figure 34. Changes in land use and land cover between 1987 and 1993 are briefly described as follows:

(a) Agriculture area: Large areas of agriculture land were converted to wasteland (1,456 ha) and built-up area (663 ha) from 1987 to 1993. Over the same period, agriculture areas were converted to shrimp farms (180 ha) and water bodies (30 ha), a minor change;

(b) Built-up area: The built-up area did not change between 1987 and 1993;

(c) Forest: Large areas of forest were converted to shrimp farms, wasteland and agriculture in the 1987-1993 period covering areas of about 1,291 ha, 343 ha and 296 ha, respectively. In contrast, forest areas converted to built-up area and water bodies in the 1987-1993 period covered an area of 84 and 8 ha, respectively, a minor change;

(d) Water bodies: Fifteen ha of water bodies covering mud flats were reforested in the 1987-1993 period;

(e) Wasteland: Areas of wasteland were converted to shrimp farms (787 ha) and agriculture area (282 ha) during the 1987-1993 period. At the same time, areas of wasteland were changed to built-up area (15 ha), forest area (14 ha) and water bodies (1 ha) in 1993;

(f) Wetland: The area of wetland did not change during the 1987-1993 period;

(g) Shrimp farm: During the 1987-1993 period, shrimp farms were converted to wasteland area (22 ha) and forest area (16 ha), a minor change.

5.3.2 Changes in land use and land cover between 1993 and 1998

The changes in land use and land cover between 1993 and 1998, conducted by overlay operation, are shown in figure 35 and shown in table 11 as the coincident matrix. The results show the change of land-use and land-cover categories from 1993 to 1998. These results can be also be simplified as patterns of major and minor change based on the amount of changed area, as shown in figure 36. Land use and land cover changes between 1993 and 1998 are briefly described as follows:

(a) Agriculture area: From 1993 to 1998, 1,261 ha of agriculture area were changed to wasteland, a major change. Over the same time, some agriculture areas were changed to shrimp farms, built-up areas and water bodies covering areas of 198 ha, 145 ha, and 8 ha, respectively;

(b) Built-up area: From 1993 to 1998, built-up areas were changed to agriculture area, wasteland and water bodies covering an area of 150, 8 and 1 ha, respectively, a minor change;

(c) Forest: From 1993 to 1998, 518 ha of forest were converted to shrimp farms, a major change. In contrast, forest areas were converted to wasteland (91 ha) and built-up areas (6 ha), a minor change;

(d) Water bodies: From 1993 to 1998, 96 ha of water bodies that covered mud flats were replanted;

(e) Wasteland: From 1993 to 1998, 489 ha of wasteland were changed to agriculture area. Over the same time, areas of wasteland were changed to shrimp farms (90 ha) and forest area (26 ha);

(f) Wetland: The area of wetland did not change between 1993 and 1998;

(g) Shrimp farm: From 1993 to 1998, 225 ha of shrimp farm were changed to forest area. Over the same time, 91 ha of shrimp farms were converted to wasteland, a minor change.

5.4 Mangrove forest land-use plan

The mangrove forest land plan is based on the combination of the Cabinet Resolution on Mangrove Forest Land Use Zonation in 1987, which was designated by the Royal Forest Department, and land-use and land-cover types in the study area in 1998 that were extracted by visual interpretation of satellite images. The Cabinet's Mangrove Forest Land Use Zonation in 1987 represents zonation of mangrove forest land utilization in the coastal areas, as shown in figure 37. In contrast, the actual land use and land cover in the study area in 1998 represent the social and economic factors that have affected the natural resources in the study area, especially mangrove forests. Hence, the combination of two factors will help decrease the persistent dispute between local people and government agencies regarding land use in the coastal zone. The combination of mangrove forest land-use zonation in 1987 and land use and land cover in 1998 as a coincident matrix is shown in table 12. The results show specific land-use activities that have taken place in mangrove areas after declaration of the Mangrove Forest Land Use Zonation on 15 December 1987 by the Cabinet Resolution. The details of land-use activities that have occurred in the three different mangrove forest land-use zones are summarized as follows:

(a) Preservation Zone: Preservation Zones are defined under the Cabinet's Mangrove Forest Land Use Zonation in 1987 as natural resource preservation areas, including forests and islands. In 1998, the Preservation Zone in the study area was occupied by agriculture, forest, wasteland and shrimp farms, which covered areas of 24.33, 1,321.75, 51.39 and 71.43 ha, respectively. The proportion of an existing mangrove forest in the Preservation Zone is approximately 89.98 per cent. The major land use not prescribed under the Mangrove Forest Land Use Zonation in 1987 is shrimp farming, representing 4.86 per cent of the zone;

(b) Economic A Zone: Economic A Zones are defined by the Cabinet's Mangrove Forest Land Use Zonation in 1987 as areas in which to utilize mangrove forests under the sustainable management approach, such as mangrove forest concessions. In 1998, the Economic A Zone of the study area was occupied by agriculture, built-up areas, forest, water bodies, wasteland and shrimp farms, which covered area of 99.84, 43.12, 1,316.93, 0.02, 144.44 and 936.10 ha, respectively. The proportion of existing mangrove forest in the Economic A Zone is about 51.84 per cent. The major land use not prescribed in the Mangrove Forest Land Use Zonation in 1987 is shrimp farming, representing approximately 36.85 per cent of the zone;

(c) Economic B Zone: Economic B Zones are defined by the Cabinet's Mangrove Forest Land Use Zonation in 1987 as areas in which to utilize mangrove forests for development purposes. The Economic B Zone of the study area was occupied by agriculture, built-up areas, forest, water bodies, wasteland, wetland and shrimp farms, which covered areas of 537.57, 48.10, 396.27, 12.51, 371.31, 7.25 and 1,730.46 ha, respectively. The major land-use and land-cover type is shrimp farming representing about 55.76 per cent. Mangrove forests make up approximately 12.77 per cent of the zone.

Based on the combination of mangrove forest land-use zonation in 1987 and land use and land cover in 1998, a mangrove forest land-use plan has been developed using the cabinet's definition of the mangrove forest land-use zones and land-use and land-cover information for the study area, as shown in table 13. The distribution of mangrove forest land-use type in the plan is presented in figure 38. In addition, a comparison of mangrove forest land-use area based on the plan is summarized in table 14. The characteristic of each forest land-use type in the plan is described as follows:

(a) Agriculture area: Agriculture area in the plan is derived from the data on agriculture area in 1998 within the Economic A and Economic B Zones and wasteland in 1998 outside the area considered by the Mangrove Forest Land Use Zonation. The agricultural area in the plan is increased to about 3,982 ha to mitigate against pressure from people who need the land and to increase local people's incomes. The area of agricultural land will increase by 3,981.82 ha or 39.56 per cent of the study area or 67.60 per cent of the total land area;

(b) Built-up area: The built-up area in the plan is directly derived from the data on built-up areas in 1998. The area of built-up land is 1,284.58 ha or about 1.90 per cent of the study area or 3.25 per cent of the total land area;

(c) Forest area: The forest area in the plan is directly derived from the data on forest area in 1998. The extent of forest area cover totals 6,397.09 ha or 9.48 per cent of the study area or 16.20 per cent of the total land area;

(d) Reforestation: Reforestation areas designated in the plan are derived from the data on agriculture areas in 1998 within the Preservation Zone, wasteland in 1998 within the Preservation and Economic A and Economic B Zones, and shrimp farm in 1998 within the Preservation and Economic A Zones. The areas of reforestation in the plan cover 1,602.75 ha or 2.37 per cent of the study area or 4.06 per cent of the total land area;

(e) Water bodies: The area of water bodies in the plan is directly derived from the data on water bodies area of land use and land cover in 1998. The area of water bodies is 27,999.58 ha or 41.48 per cent of the study area;

(f) Wetland: The area of wetlands in the plan is directly derived from the data on wetland area in 1998. Wetlands cover an area of 66.48 ha or 0.10 per cent of the study area or 0.17 per cent of the total land area;

(g) Shrimp farm: The area of shrimp farm in the plan is derived from data on shrimp farm area in 1998 within the Economic B Zone and areas outside the area considered by the Mangrove Forest Land Use Zonation. The area of shrimp area in the plan decreased by 1,008 ha. The extent of shrimp area covers an area of 3,446.63 ha or 5.11 per cent of the study area or 8.73 per cent of the total land area.

Content

6. Conclusion

In conclusion, remote sensing appears to be a significant tool for assessment and monitoring of coastal zone resources, especially mangrove forests. In addition, planning and management of forest land use is easily and effectively conducted using GIS. However, the integration of remote sensing and GIS for the development of mangrove forest management plans by natural resource managers and planners is necessary.

References

Ratanasermpong, S. and C. Silapathong, 1990. Mangrove forest zonation by using high-resolution satellite data. In Proceedings of the 11th Asian Conference on Remote Sensing, Guangzhou, China.

National Research Council of Thailand, 1991. Remote Sensing and Mangroves Project (Thailand): Final Report. IDRC/NRCT/RFD.

Suwit Ongsomwang, 1998. Forest Potential Assessment by Using GIS. Forest Resources Assessment Division, Forest Research Office, Royal Forest Department, Bangkok (in Thai).

Annex I: Data from sample plots

Annex II: Visual interpretation key of satellite imagery

Annex III: Information in WMF files

Content | Top