Abstract (paper -196)
This paper describes the preparation of vegetation map of Ganges river basin that could be used for various hydrological analysis that could be useful for water resources planning, flood forecasting and disaster mitigation. This study focuses on vegetation mapping of the Ganges river basin covering from 70° E to 95° E and 35° N to 20° N using NOAA AVHRR.
Initial vegetation cover of the area was prepared by monthly maximum NDVI data. The reason for aggregate daily NOAA-AVHRR data for monthly average was the presence of considerable amount of cloud cover in this region. Having generated the basic vegetation map based on NDVI, it was further classified according to climatic and elevation zones. The final vegetation map represents vegetation classes that are interpreted considering their temporal climatic and altitudinal variation that are needed to be considered for hydrological analysis.
Introduction
At the United Nations Conference on Environment and Development held in Brazil in 1992, Agenda 21, an action program for addressing global environmental challenges while continuing to support sustainable economic development, was resolved. Agenda 21 clearly makes the case that baseline data on key environmental parameters is important. In this course, in 1992, the Ministry of Construction of Japan began to advocate the Global Mapping concept. The fundamental basis of this concept is to develop global scale geographic information through international cooperation. Since its inception in 1992 the Global Mapping concept has obtained support from many countries in the world. Efforts are now being made to implement the development of global scale map products with uniform accuracy and specification under the auspices of an International Steering Committee for Global Mapping (ISCGM). This will facilitate resolving global problems by forming and implementing policies for the issues such as river control, disaster countermeasure and food security. As continuation of development of global map, Ganges River Basin was selected following the successful completion of Mekong River Area.
The Ganges basin contains the largest river system on the subcontinent. The Ganges rises in the southern Himalayas on the Indian side of the Tibet border. Area covered in the study of global mapping of Ganges river basin(approximately 174000 square kilometer) is 20 N to 35 N latitudes and 75 E to 95 E longitudes. The water supply is dependent partly on the rains brought by the southwesterly monsoon winds from July to October, as well as on the flow from melting Himalayan snows, in the hot season from April to June.
In order to monitor global scale vegetation cover, satellite data with frequent repetitive coverage should be available. This could be achieved only with NOAA-AVHRR series of polar orbiting, meteorological satellites. A lot of research is going on to develop techniques for continental and global scale studies of land cover using AVHRR data. One of such technique is maximum value composting of normalized difference vegetation index (NDVI) for reducing cloud contamination and atmospheric effects. The seasonal variation of NDVI can be used to classify land cover /land use.
Significance of the study
Hydrological analysis, which could provide vital information for water resources planning, flood estimation, desertification analysis etc. requires land cover information as one of the significant parameters or a parameter derived form land cover status. In regional scale studies as in this present study, forest and vegetation cover should be considered with respect to climatic region, elevation condition as vegetation cover in different climatic regions or with the change of elevation could have different characteristics needed to be taken care in hydrological analysis. In order to develop a land cover map satisfying these issues, this study incorporated NOAA-AVHRR data with climatic and elevation information.
Data and Method
NOAA AVHRR maximum NDVI composite
The following flow chart (Fig 1) gives the complete overview of the methodology of analysis of the study.
At first, 10 days composite images of NOAA AVHRR based on maximum normalized difference vegetation index (NDVI) for year 1998 were collected form Asian Institute of Technology. In order to interpret temporal variation pattern, ten days NDVI composites were generated from these data, which is a method accepted globally(Moody & Stranfler, 1994). Due to frequent cloud cover in this region, 10 days temporal resolution of NOAA-AVHRR could not produce cloud-free composites. To overcome this situation, further data aggregation, monthly maximum NDVI was attempted. This step provided cloud free composite for the year 1998 except for the months June, July and August. As these three sense contained significant amount of cloud cover, they were excluded from subsequent analysis.
Multi-temporal NDVI data classification was carried out based on iso-class unsupervised classification method. With few initial evaluation of class distribution it was decided to terminate iso-class groups into 25. The resulting 25 classes were interpreted into vegetation classes using knowledge derived from temporal profiles, elevation, meteorological information as well as existing landuse maps. Basically, the classification scheme developed by International Geographic Information System Examination Committee for Global Mapping was adopted in this stage.
Elevation data
Elevation data freely available from National Imagery Mapping Agency (NIMA) was used for the study. The data is in 30 arc sec, similar to the working pixel NOAA data, but had to transform to the present co-ordinate system.
