A study of land transformation in Savar Upazila, Bangladesh, 1915-2001 : an integrated approach using remote sensing, census, map and field data
The aim of this research is to analyse the process of land transformation for an upazila (subdistrict)
of Dhaka, the Capital of Bangladesh, using image interpretation techniques and GIS approaches.
The analysis spansa time period at decenniali ntervals from 1951 (predevelopmenpt hase)t o 2001
(urbanisation stage). The upazila is considered in Bangladesh to be the main focal point of
government planning, land management, environment and development policies. Savar Upazila
has been selected as the study area because of its rapidly changing population density, socioeconomicp
henomenaa nd land use/coverc hanges ince the independencefr om the Great Britain.
The major sources of high resolution (up to a 2.4 metre resolution) remotely sensed data are
panchromatic and infrared aerial photography (1953,1984 and 1990); CORONA K114 and K114b
spy satellite panoramic film (1962 and 1972) and IRS-ID panchromatic imagery (2000); and
DGPS based GCP coordinates of 2001. Digitally converted conventional and Historical
administrative, settlement, planning, revenue and topographic maps have been digitised and used
at a large-scale of up to 1: 3,960. Moreover, the enhanced 1951,1961,1974,1981,1991 and 2001
population censuses are used in parallel to help interpret images and related factors. In-depth
survey and participatory approaches were used during the fieldwork at plot level to help interpret
and develop a weighted land cover model and to understand factors responsible for change. In grid
format, compatible to temporal image data, detailed land value data were converted into gridded
format compatible with a decadal time-series of imagery and from the field for the early 1950s to
2001 integrated with mauza maps. The attribute data have been used as if it was an image data
layer in order to visualise land value data.
The results illustrate that significant and visible land transformations and population change have
occurred over the last half-century from a completely river-dependent countryside to a modem
road network orientation, with a change from the byde (low-lying flood-prone area) to chala
(highland flood-free area) land economies. A spatio-temporal land transformation index is
proposed to explain the complex micro-scale change that have taken place using land use and land
value data. Remotely sensed data provides the context with which to interpret past and present
land cover and land use. This study shows the importance of modem GIS techniques to integrate
enhanced field and secondary data/maps with raster data as well as for mapping historical images
and information on the future development, land reform/management, environmental study and
planning in Bangladesh.