Global climate change is a current problem that is discussed in many publications. The data provides only limited possibilities for the detection of future consequences. Surface water dynamics are described with some examples, but these give only a small insight into the complexity of this system. Using remote sensing methods, past and present dynamics can be analyzed. Nevertheless, the number of data sets with global data series is very small. Due to modifications of the "System Earth" caused by climate change, the global surface water dynamics is also changing. In high latitudes this is still poorly understood and an analysis of the data can provide an important insight into future consequences. Therefore, it is important to monitor Arctic surface water dynamics. In this master's thesis, the datasets of PICKENS et al. (2020) and PEKEL et al. (2016) are compared with respect to their results in surface water dynamics for Arctic permafrost areas from 2000 to 2020. In regards to this, both datasets were homogenized using the program "Google Earth Engine" and examined in an accuracy analysis. Any areas where both datasets show different results are detected and checked for possible causes. For this purpose, different reference datasets were used, which visually cover the topography, vegetation cover, landscape classification and permafrost content as indicators. Subsequently, in the program "Microsoft Excel", the data from the accuracy analysis were processed for each study area and analyzed in regards to the results. This shows that permanent water body areas in each study area were classified almost identically between the data sets. The largest differences between the results of PICKENS et al. (2020) and PEKEL et al. (2016) were found for seasonal water areas. The "Producer and Consumer Accuracy" calculations even yielded accuracy values below 65% in one study area. In these scenic areas, the problem of mixed pixels exists where, depending on pixel size and current water level, accurate classification becomes difficult. Also other factors, like seasonal events, cloud formations or technical errors cannot be excluded and have a share in the differences between the results of the data sets. Based on topography, the forest cover, the landscape classes and the permafrost content, differences between the data sets were recognized and distinctive shapes determined.