|Short-term training programme on GIS (Geographic Information System) for fisheries (1993)|
Md. Obaidul Quader
Principal Scientific Officer SPARRSO, Agargaon, Dhaka-1207
1.1 Bangladesh is situated in the tropical region in the northern hemisphere and is bounded by north latitudes 20°34 and 26°38 and east longitudes 88°01 and 92°41. The tropic of Cancer passes through approximately the middle of Bangladesh. Some of the areas in these range of latitudes are very sensitive to aridity. But the country falls within the active monsoon belt and receives normal rainfall from 1266 mm to 6496 mm from west to east. Eighty five percent of the total annual rainfall is in the monsoon period (June - October), 13 percent in the summer (March to May) and 2 percent in the winter (November to February).
The catchment area of the Ganges, Meghna and Brahmaputra river system is about 1.61 million sq. km. The annual load carried by these three river systems and others into the Bay of Bengal is estimated to be around 2.4 billion tons. The most of the Bengal delta (Bangladesh part only) was formed by the network of rivers and their tributaries and distributaries, numbering about 230 with a total length of about 24140 km (BBS 1981). Rainfall occurs relatively almost round the year. The huge amount of water which flows through Bangladesh to the Bay of Bengal every year, keeps the countrys topographical and morphological features dynamic. So, the 30 years revision period of Survey of Bangladesh maps is considered to be too long a gap for Bangladesh. Therefore, updating of water bodies in the map by using satellite imagery would provide uptodate information on water resources in general and fisheries in particular. The water bodies are rivers, haors (tectonic depressions), baors (oxbow lakes), beels (low lying areas) and dighis (big ponds), etc.
1.1 Objective of the Study: The aim of this study was to collect data on water bodies using satellite imagery and aerial photographs and to make an estimate of total water bodies usable for fish production and irrigation facilities in the study area (Sylhet, Jessore and Natore). This study also showed the capability of Landsat TM and SPOT (HRV) data. Another main purpose of this study was to compare 1984 SPARRSO inventory of water bodies of the whole country based on infra red aerial photo and Landsat MSS data with latest available Landsat TM and SPOT data.
Due to shortage of time and manpower only three districts were selected for updating the water bodies using aerial photo and satellite data. From 1984 inventory of water bodies it was found that single band enlargement of Landsat MSS and TM data was not feasible, because of scanning and distortion of the enlarged image which hinders interpretation. So, the methodology has been changed to using infrared aerial photography as a base map with topo sheets of 1:50,000 scale and updated by SPOT digital mosaic maps of same scale. This study will also help better planning and management of fisheries resources and irrigation facilities in these three districts of Bangladesh.
1.3 Data Sources: (a) 30 topo sheets of 1:50,000 scale from the Survey of Bangladesh covering the study area, (b) Landsat TM image (9 feb 1989) from the National Research Council of Thailand (NRC), Landsat TM (FCC) of bands 457 which were prepared at SPARRSO, (d) SPOT mosaic print of the same scale and area of the topo sheets from the Bangladesh Water Development Board, (e) Questionnaires supplied through the Water Development Board and the Directorate of Fisheries in the study area through their contact persons.
2.1 Mapping of Water bodies using infra red aerial photo and SPOT mosaic prints: The district boundary of the three districts were taken from topo sheets of the Survey of Bangladesh. All the infrastructures - roads, markets, villages, towns, and prominent structures etc. which are relevant to water bodies management were also taken from topo sheets. Then water bodies were traced from infrared aerial photos of February 1984. In this way base maps of all three districts were prepared. These base maps were updated. All water bodies and other uptodate information were incorporated from digital SPOT mosaic prints of March 1989. These prints were prepared toposheetwise into 1:50,000 scale. In this way eleven maps of three districts were prepared.
In the maps the water bodies were classified into following categories: Perennial river/canal, seasonal river/canal, perennial lake/bil/haor/baor, seasonal lake/bil/haor/baor, pond. These were identified and classified considering the drainage pattern, landuse and soil condition from the tonal variation of aerial photos and satellite images.
Other infrastructures represented in the maps are as follows: International boundary, district boundary, district/thana headquarters, mixed urban areas, villages, main road, cart track/paved roads, railways, airports, ferrries, embankments etc.
3. DIGITAL IMAGE PROCESSING AND GIS
3.1 Data load: Loading and selecting the study area were done with the module AUTOLOAD and SUBSET OF ERDAS system for a portion of Sylhet district. (Sylhet city and adjacent areas)
3.2 Geometric correction: The TM CCT containing full scene was mounted on tape drive and a portion of the study area of Sylhet district was loaded to the hard disk of the PC based ERDAS system from CCT. This raw data was needed to be geometrically corrected before digital analysis. The geo-reference point was taken using Lambert Conformal Conic grid coordinates from topographic sheet at the scale of 1:50,000. Twelve control points from the topographic maps were used which appear clearly in image such as cross roads, pond, river etc. to correct the raw data and the root mean square of the cross was less than 0.2.
3.3 Enhancement: Image enhancement is made to improve the visual interpretability of an image by increasing the apparent distinction between the features in the scene. Enhancement techniques can range from relatively simple to rather complex procedure for computer processing data, but in every case the final product is an Image which is then given to be analysed and interpreted to obtain the desired information.
In the study following enhancement techniques were used. Histogram equalization: For the enhancement of the pixel value (Intensity) of water bodies and other than features.
Filtering: For the removal of noise due to the small features on the image.
Maximum likelihood classification and False Colour Composite (FCC) of bands 457 for relevant features by supervised classification technique (Fig. 2).
3.4 Determining Chlorophyll Contents in a part of Sylhet District Module used for chlorophyll estimation:
1) In Chla = 12.05 + 6.40 Ln TM2 (Lathrop and Lillesand, 1986) TM2 means Landsat TM band 2.
2) Chla = A + B * ln((TM2 - TM1)(TM2 + TM1)) + TM5 (Li. et al, 1991)
Where, A & B are constants and, TM2 & TM1 mean, Landsat TM bands 2 & 1 respectively.
With module 2 qualitative analysis of chlorophyll concentration level of the study area has been made. But their statistics could not be taken due to non availability of data with the ERDAS software. In module 1 turbidity could not be removed because of small size and shallowness of the study area.
3.4. Application of GIS.
Planning for suitable locations for increasing inland fish production is mostly carried out at various administrative levels; eg. districts, divisions or national head quarter or concerned ministry, though it might be local fisheries officers who would advice on specific sites for fish production, and who would need informations to help them with this.
An objective of this study was to show that GIS can provide rational and easy way to fulfill these requirements, especially in view of the rapid developments of low cost software and hardware.
A portion of finalized map of Sylhet district was digitized in PC based ARC/INFO system with proper available layers and attribute data. Nine layers were chosen for the map.
* Surface water layers allow users easily to locate the interesting water body and identify this type (Seasonal water body, single line river, double side river).
* Road layer shows the suitability of aquaculture fish transport and marketing and irrigation management.
* Town/Villages layer which also indicate the potential market for selling the product and population density also is an important factor for storage.
* Water bodies classified from TM data on ERDAS system and connected to ARC/INFO coverage were taken as another data source for GIS and compared with the digitized map.
Roads, permanent water bodies, villages, rivers were buffered with distance 1 km and cities buffered with 5 km. The potential ponds were determined with the relationship among the permanent water bodies (as water source), roads (as availability of transportation), cities/town (as source of manpower and markets), (Fig. 4) share areas falls on the buffer of roads and rivers considered as the potential plans where new ponds could be constructed for fish culture and iceplant for preserving the fishes.
4. RESULTS AND DISCUSSIONS:
Counting of total number of water bodies in each class was done. However, measuring of total areas of each class could also be made. Due to devastating flood of 1988 some of the areas of these three districts seem to be still inundated or partly dried in 1989 SPOT images. This creates confusion among seasonal water bodies and flood waters or soil moistures in the study area. After ground truthing this confusion had been eliminated. Some of the permanent water bodies were also confusing in low lying areas. This also has been eliminated by ground truthing. Some differences were also found in 1984 inventory and 1989 SPOT data. 30% of the questionnaires which were sent to the officials of the Directorate of Fisheries and the Bangladesh Water Development Board were returned duly filled in. These information have been given in Table (1-2). 20 sheets of SPOT image (1:50,000 scale) prints of same date covering the study area were collected from Water Development Board for rechecking before preparing the final maps. Final checking has been made with ground truthing and 11 maps were prepared. In the Sylhet areas some one line rivers/ canals are not continuous. This is due to siltation or drying up or covering with vegetation. Minimum size of 22mm ponds were considered for tracing in the map from aerial photo and SPOT prints.
Total number of water bodies in three districts areas are as follows:
Jessore: River - 6, Bil - 300, Pond - 39, Baor - 20
Natore: Rivers - 9, Lakes - 7, ponds - 179, Bil - 252
Sylhet: River - 31, Khal - 21, Bil - 633, Haor -17, Pond - 40
Digital image processing using Landsat TM CCTs of 1990 and 1991 using the PC based computer system of SPARRSO and application of GIS ARC/INFO has been made digitizing the map for a part of Sylhet. Qualitative analysis of chlorophyll concentration level has been made using the module of Li et. al (1991).
GIS study has been made in a part of Sylhet district including Sylhet town. The results show very useful information on locations of potential areas for constructing fish ponds and suitable site for iceplant for preserving the fishes.
Digital mosaic map of SPOT data is better than single or multiband TM data for mapping of water bodies into 1:50,000 scale because of its resolution. Single band MSS or TM data are not good for mapping into 1:50,000 scale because of scanning and distortion which hinders interpretations. Multiband TM data is good for chlorophyll and other water quality studies. For landuse classification TM band 457 is found to be good for digital analysis. Questionnaires on fisheries also give useful information on fish stock, fish marketing and possible water pollution in the study area. For further study following recommendations are made:
1) Areas should also be measured for each water class district wise and different classes of water bodies should also be counted separately for IR aerial photo and SPOT data from 11 maps prepared for three districts and to be prepared for future studies.
2) In future studies water quality and chlorophyll concentration study should be made in the field level extensively as this data are scarcely available now.
3) Water quality studies with chlorophyll concentration using satellite data should be developed.
4) GIS should be done for the whole country showing the potential areas for fish culture and suitable site for ice plant (Cold storage) for preserving the fishes.
1. BBS, 1981, Small Area Atlas of Bangladesh, Sylhet District, p6.
2. BBS, 1981, Small Area Atlas of Bangladesh, Rajshahi District, p5.
3. BBS, 1981, Small Area Atlas of Bangladesh, Jessore District, p7.
4. Lathrop, R.G. and T. M. Lillesend, 1986. Use of Landsat TM Data to Assess Water Quality in Green Bay and Central Lake, Michigan. Photogrammetric Engineering and Remote Sensing, Vol. 52, No 5, pp. 671-680.
5. Li, J., L. Zhang, M. Jin and Q. Huang, 1991. The use of Landsat TM Data to Quantify Chlorophyll and Lake Weed. Asian Pacific Remote Sensing Journal, Vol. 4, no. 1, pp. 1-14. ESCAP, Bangkok.