Effect of competing landuse practices on Chakaria Sundarbans mangrove in Bangladesh using Landsat imagery

9th IGRSM International Conference and Exhibition on Geospatial & Remote Sensing

地球科学;计算机科学

Prince, Husni Mobarak^1 ; Idrees, Mohammed Oludare^1 ; Shafri, Helmi Zulhaidi Mohd^1 ; Iqbal, Mehedi^2 ; Iqbal, Md Zaheer^3 ; Aziz, Md Tariq^3

Department of Civil Engineering, Faculty of Engineering, Universiti Putra Malaysia, UPM, Serdang

43400, Malaysia^1

Departmrnt of Geography and Environment, Jahangirnagar University, Savar Dhaka

1342, Bangladesh^2

Bangladesh Forest Department, Dhaka

1207, Bangladesh^3

关键词: Atmospheric effects;    Classification accuracy;    Classification process;    Human interference;    LANDSAT;    mangrove;    Supervised classification;    Support vector machine classifiers;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/169/1/012038/pdf
DOI  :  10.1088/1755-1315/169/1/012038

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

This paper quantifies the extent to which Chakaria Sundarbans mangrove has been depleted through human interference using Landsat imagery of 1972 and 2017. The images were corrected for radiometric and atmospheric effects. To improve the classification process, the Chakaria Sundarbans's Landsat 2017 image was pan-sharpened. The earlier image which comprises of the virgin forest was classified into three classes (water, mangrove, wetland) while the later was classified into four classes - waterbody, mangrove, pond scum and saltpan using supervised classification method and support vector machine classifier. Using the statistical bias adjustment, precise area estimates for each land cover class was obtained. The result shows that between 1972 and 2017, Chakaria Sundarbans mangrove forest has reduced by about 87.5% (from 6000.27 to 877.76 hectares). Currently, about 21% of the land is being used for salt mining, 45% for shrimp farming while the water body takes 26%. It is observed that the river has reduced in width; however, water surface area increased by 2%. The bias-adjusted overall classification accuracy yields 95.44% and 94.70% for classified maps of 1972 and 2017, respectively. Evidently, the mangrove has been completely lost to over-exploitation of resources.