Assessment of spatio-temporal variations in lake surface using landsat imageries and Google Earth Engine

Abstract

The purpose of this research is to determine the lake's water surface area and the Lake Surface Water Temperature (LSWT), as well as to give a long-term analysis of the spatial and temporal relationships between these variables. Lake Burdur, Lake Egirdir, and Lake Beysehir were the three lakes in Türkiye's Lakes Region that were taken into consideration to be used as testing sites. Lakes are crucial to the ecosystem and ecological stability. For the preservation and development of such sensitive regions, continual monitoring using accurate and dependable sources is required. The development of remote sensing technology provides a significant opportunity for the determination, monitoring, and assessment of the current state of lakes. Specifically, the improvement of satellite images spatial and spectral resolutions contributes to this potential. Mapping water surface area is essential for agriculture, economic, and the structure and function of the ecosystem, as well as for the extent of lakes' water surfaces. Remote sensing methods were used to comprehend the long-term variations in the water surface area of Lake Burdur, Egirdir, and Beysehir. Google Earth Engine (GEE) is a cloud-based platform that applies operations to Landsat satellite images, was used to perform the analysis. In the study, two satellite sensors; Landsat 5 TM and Landsat 8 OLI/ TIRS, were employed to cover as much of the research period from 2000 to 2021 as possible. Landsat 8 OLI/TIRS gives coverage from 2013 through 2021, while Landsat 5 TM provides coverage from 2000 to 2011. There was no available satellite data for 2012. Also, ERA5-Land and TerraClimate monthly satellite dataset were used in this study to show whether there is correlation between water surface area and LSWT with meteorological parameters (temperature, evaporation and precipitation). The first chapter discusses the relevance of lakes to the ecosystem and global climate, as well as the information on water body extraction using satellite images acquired by remote sensing was presented. This chapter also discusses the study's scope and objectives. The second chapter presents the role of remote sensing and satellite data in extracting and monitoring the spatio-temporal changes of water surfaces. The third chapter introduces the study area and clarified all of its features, in addition to providing the satellite data utilized in the analysis. Thus, a total of 606 Landsat images were collected and analyzed through GEE cloud platform. In the fourth chapter, the Normalized Difference Water Index (NDWI) was utilized to extract the water pixels in order to determine the surface water extent from the image inventory. Thus, the procedure of calculating surface water area and LSWT has been explained in detail via the used equations and GEE code. The accuracy assessment for the largest and smallest surface water area for each lake was evaluated using the random point selection method and confusion matrix. Moreover, the water surface area changes of the three lakes and the LSWT variable, as well as their relationship with one to another, were analyzed to clarify the effects of an increase or decrease in LSWT over the extent of the lake's surface water. In the fifth chapter, the daily, seasonal, and annual variations in each lake surface area across the research period were presented. The findings from the study indicate that by using NDWI, water pixels can be extracted rather accurately, with an overall accuracy of more than 98% for the surface area of all lakes. Between 2000 and 2021, the water surface area value of Lake Burdur decreased significantly by more than 22%, while for Lake Egirdir it has dropped slightly by less than 4%, and for Lake Beysehir has not changed noticeably, it has decreased by roughly less than 1%. After analyzing the LSWT values, it can be established that they increased in all lakes except for Beysehir over the study time period. Thus, Burdur Lake's LSWT increased by more than 2.1°C, Egirdir Lake heightened by more than 0.3°C, and Beysehir Lake decreased by more than 1.4°C. The obtained results were evaluated with meteorological parameters and our findings showed that human-induced activities were more dominant than climate effects over the lakes. In the sixth chapter, the study's findings were reviewed and discussed.