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 the Natural Sal Forest Cover in Central Bangladesh by Temporal Landsat Imagery 1972-2010

Hasan Muhammad Abdullah
Assistant Professor
Department of Agroforestry and Environment, Bangabandhu Sheikh Mujibur Rahman Agricultural University, Bangladesh

Abstract/ Executive Summary:

Modhupur Sal (Shorea robusta) forest, the largest natural Sal forest area of Bangladesh, is a crucial ecological resource situated in the central part of the country. This unique natural forest belongs to tropical moist deciduous forest type. It harbors a good number of native flora and fauna. Unfortunately this forest is heavily encroached. Current study investigated and quantified the destruction in areal extent of natural Sal forest of Modhupur using Landsat satellite data of 1972, 1989 and 2010. Post classification change analysis technique was used to quantify the natural Sal forest coverage to know the loss of forest cover. The result showed that the extent of original natural Sal forest was depleted by 64% and 31% in the year of 2010 and 1989 respectively compared to that of 1972. The study presented a successful usage of freely available Landsat data to monitor forest cover change which could be followed for low cost but effective monitoring of forests and other natural resources in Bangladesh and other parts of the globe. Alarming evidence of mass destruction of natural sal forest revealed by this study is indicative to emphasize in taking adequate measures for protection and regeneration of natural Sal forest in Modhupur for environmental sustainability.

Keywords: Tropical Forest, Shorea robusta, Remote Sensing, Deforestation.

Location: Modhupur, Tangail, Bangladesh

Start Date: 12-01-1989

End Date: 30-01-2010

Program Area: Conservation and Biodiversity

Non-commodity: Tree

Objectives:

To quantify the natural depletion in the Modhupur of central Bangladesh since 1972 using satellite imagery.

Methodology:

Geologically, the study area falls in a terrace from 1 to 10 m above the adjacent flood plains. Landsat MSS1 and TM5 data were acquired for this study. Three Landsat data comprised of MSS1 (23 November 1972) and TM5 (12 January 1989 and 30 January 2010) were obtained and employed in this study. Landsat thermal bands were not included to detect land cover change. The images were geometrically corrected using a pre-rectified Landsat TM image of 1989 as a reference to eliminate any possible inherent unsystematic errors present in the data. At least 50 well distributed ground control points were used in the image to image registration process. The root mean square error (RMSE) varied from 0.24–0.59 pixels. Finally, a first-order polynomial fit was applied and all the data were resampled to 30 m pixel size using the nearest neighbor method. The Bangladesh Transverse Mercator System (BTM) was used as the coordinate system which is an area-specific standard UTM projection system for Bangladesh. A detailed and comprehensive field survey was conducted to explore the current forest condition to collect reference land cover/use data and their corresponding geographic coordinates by Global Positioning System (GPS) in the study area. The fieldwork was carried out during 16 – 26 January 2009. Eighty five (85) reference points were visited in the field and geo-referenced using a handheld GPS. This information was then brought into GIS and overlaid to the images for the selection of training areas for classification and subsequent accuracy assessment. Forty five (45) reference data points were used to train the classification algorithm, the remaining forty (40) reference data points were used for classification accuracy purposes.

Classification of land use cover is an important task before monitoring and mapping forest depletion to reduce errors by other classes on the final forest class map. Hydrid approach of image classification was followed with the help of ERDAS Imagine 2010 software to discriminate natural sal forest area. ArcGIS 9.3 software was also used to carry out related GIS operations. Forest and other non forest areas were classified on the three images using Iterative Self Organizing Data Analysis (ISODATA) clustering algorithm. Initially one hundred clusters were generated. Then natural sal forest class or other non natural sal forest class was assigned at each cluster after carefully checking with ground reference data. Finally natural sal forest class was extracted from each of the three classified imagery. An accuracy assessment of natural sal forest class was done in the form of an error matrix by comparing classified output with the ground truth information of field survey sites collected using a GPS and other ancillary data including high resolution GoogleEarth images of 03 January 2001, and 14 October 2003, forest maps and plantation journals available from local Forest Department Offices. Accuracy assessment of 1972 imagery was not performed due to lack of Sal forest map of that time. A kappa coefficient was also computed. Post classification change detection algorithm was used. Maps of remaining Modhupur Sal forest were prepared from classified Landsat imagery showing remnant natural Sal forest coverage for 1972, 1989 and 2010.

Source of Information: Proceedings of Research Review-2013 and Programme Planning Workshop-2014 on Forestry Research in Bangladesh held on 12-13 March 2014, at Bangladesh Agricultural Research Council, Farmgate, Dhaka-1215, Bangladesh.

Article Link (Online Journal):

Funding Source:

Total Budget (Tk.) :

Achievement/Findings:

Monitoring land cover changes using multi-temporal remotely-sensed data provides an effective and accurate evaluation of human impact on the environment. The study presented a proof of the alarming trend of drastic depletion of natural Sal forest coverage in Modhupur in spite of several attempts to protect Sal forest in Bangladesh. This simple and low cost monitoring technique advantaged from the freely accessible Landsat data resources would be helpful for decision maker, forest planner, natural resource manager for sustainable management of this invaluable forest and other natural resources of Bangladesh and different developing nations. Scope of the current research was limited to quantify the depletion of natural sal forest coverage in Modhupur. Further research using advanced remote sensing sensors and data analysis techniques supported by well-designed extensive ground investigation is recommended to quantify the loss in ecosystem and its consequences from this depletion of natural sal forest.

Knowledge Management: Report/Proceedings