Study on the seasonal changes of land cover and their impact on surface albedo in the northwestern part of Bangladesh using remote sensing
Abstract:In the present work, the dynamic behaviour of surface albedo of a coupled soil-vegetation-atmosphere (SVA) system has been studied in relation to landform characteristics, seasonally varying surface cover, and hydro-meteorological condition using multi-temporal Landsat TM data. Two adjacent, sharply contrasted and physiographically different land areas, (1) the semi-arid Barind Tract area and (2) the flood plain areas of the rivers Ganges and Mahananda situated in the northwestern part of Bangladesh, have been selected for this study. Spatio-temporal analysis of Landsat TM-derived spectral signatures manifests quite contrasted response characteristics for different physiographic land units in the two study sites and shows a seasonal trend that is consistent with field observations. Coupling of satellite-derived land cover information with landform and hydro-climatic data reveals that the observed variability in surface cover and its condition, as well as seasonal transformation of surface categories (e.g. soil, vegetation, and water) over time and space, are a function of landform characteristics and the prevailing hydro-climatic condition of the area. Particularly, depth of flooding in relation to land elevation during the wet season in the flood plain areas and soil moisture deficit during the dry season in the Barind Tract area seemed to be the two major determinative variables governing land use of the area. Climatic variables, particularly precipitation, appear to be the main driving force that brings out the dynamic changes in the geo-environmental condition in both study sites. The changes in vegetation cover in response to the rainfall amount show a much higher sensitivity of the Barind tract areas in comparison to the flood plain areas. Land cover information as derived from multidate Landsat TM data demonstrates a relatively high change (about 68 per cent) in the aerial extent of vegetation cover in the Barind Tract and a small change (only about 8 per cent) in the flood plain areas between summer and the rainy season. The flood plain areas are found to be moderately vegetated throughout different seasons, which is in contrast to the semi-arid Barind Tract area that remains either mostly bare (about 90 per cent) or highly vegetated (vegetation cover greater than 75 per cent), depending on the season, though both sites receive an almost similar amount of seasonal and annual precipitation. Variation in soil moisture and vegetation condition, as well as transformation of bare land into a vegetated or seasonally flooded area in response to seasonally varying precipitation, introduce sequential variability in the surface absorption level, thereby provoking seasonal rhythmic variation in the surface albedo over the area. It is found that the transformation of bare land into a crop area in the Barind Tract reduces the surface albedo values by about 60-75 per cent. Relatively higher spectral responses in summer over most of the semi-arid area depict a deficit in moisture content in the soil layer, as compared to that in the relatively moist flood plain area, except the sandy area, whereas transformation of land into water due to seasonal flooding, especially in the flood plain areas in the wet season, associates a decrease in albedo from 13 per cent in summer to about 6 per cent in winter. Increased albedo, particularly over the river area in summer, indicates the presence of suspended sediments originating from erosion due to surface runoff and nearby riverbank shifting. As a whole, surface albedo seems to be a land type-dependent seasonal indicator of the area.
Document Type: Research Article
Affiliations: 1: Digital Modelling and Hydro Environmental Analysis & Research Team (DIMO-HEART), Bangladesh Space Research and Remote Sensing Organization (SPARRSO), Bangladesh 2: Digital Modelling and Hydro Environmental Analysis & Research Team (DIMO-HEART), School of Computer Science & Engineering, University of Information Technology & Sciences (UITS), Dhaka 1212, Bangladesh 3: Department of Applied Physics and Electronics, University of Rajshahi, Rajshahi, Bangladesh
Publication date: 2007-01-01