Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Monitoring Landuse-landcover Changes in Dhaka City by Integrating Remote Sensing and Ground Based Observations

Rahul Bhattacharya
Department of Geology, University of Dhaka, Dhaka-1000, Bangladesh

Md. Bodruddoza Mia
Department of Geology, University of Dhaka, Dhaka-1000, Bangladesh

Md. Mostafizur Rahman
Department of Geology and Petroleum Geology, School of Geosciences, King’s College, University of Aberdeen, United Kingdom

A S M Woobaid Ullah
Department of Geology, University of Dhaka, Dhaka-1000, Bangladesh

Abstract/ Executive Summary:

We have employed remote sensing approach in this study to understand the LULC changes in Dhaka urban area for a period of 26 years. We have used Landsat TM, ETM+ and OLI/TIRS image and employed unsupervised classification to prepare LULC maps from 1989 to 2015. To validate the results fieldwork has also been followed. We have found a decrease of water bodies of 967.14 ha and a decrease of vegetated lands of 2641.59 ha. The rate of decrease is 37.20 ha per year and 101.60 ha per year for water bodies and vegetation respectively. At the same time the urban areas have been increased by an area of 4370.67 ha. The rate of increase of urban structures is 168.10 ha per year. Integrating the spatial changes, we have found a correlation between loss of vegetation and water body with increase of urban structures. Our field observation also provides accuracy of our satellite based observation with an overall accuracy of 82.61%.

Keywords: Dhaka city, Urbanization, Landuse Landcover, Remote Sensing

Location: Dhaka, Bangladesh

Start Date: 00-00-1989

End Date: 00-00-2015

Program Area: Knowledge Management

Non-commodity: Remote sensing, Data collection and collation

Objectives:

To determine the Monitoring Landuse-Landcover Changes in Dhaka City by Integrating Remote Sensing and Ground Based Observations.

Methodology:

We collected four Landsat satellite images from USGS archive. These images were radiometrically and geometrically corrected. All of these images individual bands have been layerstacked together except the thermal band. Layerstacking is useful for image classification and identification of individual LULC by color composite process. Finally the region of our study area has been separated from these layerstacked images. An extensive fieldwork has been performed following the LULC classification from satellite data. Handeheld GPS have been used for this purpose to locate and navigate observation points. All of the processing has been done using ERDAS IMAGINE 2014 and Arc GIS 10.2.1 geospatial software. Those software helped to achieve precision with the fast data computation process. We followed a detail study of the LULC both from satellite data and ground observations. The individual Landsat images have been classified and different LULC have been identified. These classes have been observed for a period of 26 years and spatio-temporal changes have been detected. Finally we correlated the results with some ground observations.The aim of our research was to perform change detection of LULC and hence image classification was useful. We employed unsupervised classification for classifying the individual images. There were several advantages to perform unsupervised classification in this study. But most of all, as there was lack of available ground data for past images used in this study; unsupervised classification would only provide reliable information about the spectral classes. The Iterative Self Organizing Data Analysis Technique or ISODATA has been followed using ERDAD IMAGINE software. We created 15 spectral classes at first but after merging them with each other by spectral characteristics observation, we got 4 representative information classes. Each of these classes has been observed for any changes over the study period. To evaluate the results of LULC changes and correlate with the image of 2015 to validate our results, an extensive field based observations have been performed. We used handheld GPS to locate different ground observation points and observed the LULC at those positions. We kept similarities of fieldwork with Landsat 8 image of 2015 in terms of image acquisition date and time. We took 23 observations and tried to correlate with the 2015 image to figure out the changes of LULC in our study area. The mixed pixels have been given special consideration for ground based LULC studies.

Source of Information: Dhaka University Journal of Earth and Environmental Sciences, Vol. 4, December 2015

Article Link (Online Journal):

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

The study reveals several changes of LULC in a period of 26 years. Though there was some limitation because of the coarser resolution of Landsat images and mixed pixels, we have found significant results from this research. We have figured out a decrease of water bodies in terms of percentages of 40.06%. We have also observed decrease of vegetation with a percentage of 39.81%. At the same time the urban areas have been increased by percent of 87.48%. The losses of vegetation and water bodies are mostly the impact of rapid and unplanned urbanization. The peripheral part of the city has affected due to losses of vegetation and water bodies whereas urban growth has been expanded in such areas. Field observation of LULC was also consistent with the satellite based observation of LULC and an overall accuracy of 82.61% has been found. The study can be further improved by using satellite images of finer resolution and integrating more ground data.

Knowledge Management: Journal