Assessing spatiotemporal eco-environmental vulnerability by Landsat data

An indicator of quantifying eco-environmental vulnerability was established by synthesizing 12 variables, mainly retrieved from satellite data with incorporation of analytical hierarchy process (AHP). Six vulnerability levels of potential, slight, light; medium, heavy, and very heavy were graded to depict changes of vulnerability over temporal and spatial scales. The proposed approach was employed to study spatiotemporal eco-environmental vulnerability with Landsat data acquired in 1989, 2003, and 2014 for the Thua Thien - Hue Province, Vietnam. Over the time periods of 1989-2003 and 2003-2014, both heavy and very heavy vulnerability levels exhibit an increasing trend in both magnitude and spatial size: The former raised from 5.9% in 1989, to 7.9% in 2003, and 15% in 2014; and the later increased from 1.2% in 1989, to 3.2% in 2003, and 7.3% in 2014. Both levels mainly appeared on urbanized area, bare land, semi-bare land, agricultural land, and sparse forests. In contrast, there was a significant decline in potential vulnerability level with 36.4% in 1989, 30.9% in 2003, and 19.2% in 2014, while the remaining vulnerability levels slight, light, and medium fluctuated slightly, increased in 2003 and decreased in 2014. Supporting reasons for such changes include: (1) deforestation, agriculture intensification, construction of three hydro-electric projects during the period 2003-2014; and (2) significant expansion of urbanized area leading to differences in thermal signatures in urban areas as compared with rural areas. The findings demonstrate that eco-environmental vulnerability is primarily exaggerated by anthropogenic activities through land use/land cover (LULC) changes and further enhanced by natural processes including disasters in the Thua Thien - Hue Province of Vietnam. The correlation between land surface temperature (LST) and Normalized Difference Built-up Index (NDBI) is found to be positively correlated with 0.87, 0.89, and 0.84 for 1989, 2003, and 2014, respectively. In contrast, LST-Normalized Difference Vegetation Index (NDVI) is found negatively correlated with respect to the spatiotemporal trend of environmental vulnerability with -0.81, -0.81, and -0.76 in 1989, 2003, and 2014, respectively. In addition, areas having potential, slight, and medium thermal environmental levels are decreased from 1989 to 2003 to 2003-2014. At the regional scale, increased anthropogenic activities through land's modification have intensified the eco-environmental vulnerability in the study area. The currently proposed methodology is feasible for evaluating long-term eco-environmental changes processes by using remote sensing data, and valid for the other regions.