THE USE OF MULTI-TEMPORAL SENTINEL SATELLITES IN THE ANALYSIS OF LAND COVER/LAND USE CHANGES CAUSED BY THE NUCLEAR POWER PLANT CONSTRUCTION
- 1ITU, Civil Engineering Faculty, Geomatics Engineering Graduate Program, 80626 Maslak Istanbul, Turkey
- 2Chemnitz University of Technology, Germany
- 3ITU, Civil Engineering Faculty, Geomatics Engineering Department, 80626 Maslak Istanbul, Turkey
Keywords: Multi-temporal Data, Change Detection, Nuclear Power Plant, Sentinel 1& 2
Abstract. Turkey, due to increased demand for energy has made plans for nuclear power generation since 1970. Sinop Nuclear Power Plant, which will be built on Sinop İnceburun peninsula at the Black Sea coast of Turkey, is one of the three different nuclear power plants planned to be built in Turkey. The Sinop Nuclear Power Plant consist of four different reactors. The construction of the first unit is expected to be completed by 2023, and the fourth unit is planned to be activated by 2028. On the other hand, the construction of the nuclear power plant will alter the land use at the actual plant site and its surroundings and hence may cause significant environmental changes. As an indicator, more than 650000 trees have been cut so far for the construction of nuclear power plant, and this can adversely affect the ecological balances of the region by endangering habitats and creating ecological damages. The aim of this study is to analyse the land use/land cover changes (LULC) in the forestry-dominated areas due to the construction of nuclear power plants using the multi-temporal Synthetic Aperture Radar (SAR) and optical satellite images. For this purpose, different change detection methods such as SAR intensity image differencing, supervised image classification method (Support Vector Machine algorithm) were applied to Sentinel 1 satellite image datasets (2016-2019) to evaluate the annual changes due to construction. In addition, a correlation analysis was performed between the canopy structure and vegetation biomass using Sentinel 2 NDVI dataset (2016-2019) and calibrated Sentinel 1 backscatter values. Furthermore, using the Google Earth Engine (GEE), the Landsat 8 NDVI time series of the affected forest area, generated at 8-day intervals, was used to validate changes in vegetation.