| Summary: | Abstract. Monitoring natural phenomena and environment on a large scale by in-situ measurements is slow, costly, and might be challenging on rough terrains. Satellite data offers an alternative to obtain continuous measurements on large areas. In order to achieve this, satellite data has to be trustworthy and exact. Arctic regions have a special and unforeseeable role for climate change effects. Glaciers act as a freshwater reservoir, and their albedo effect maintains a global balance that climate change disrupts. Snow cover extent, snow water equivalent, first and last days of snow, and amount of snowy days are important metrics when monitoring changes in snow behaviour and studies related to climate change. In this thesis two products offered by Copernicus Global Land Service will be analyzed: snow cover extent and snow water equivalent. This will be done by fitting in-situ measurement data to satellite data and using several different evaluation metrics. Data is obtained from three cities in Finland: Sodankylä, Oulu, and Vantaa. The results reveal that 91.3% of snow water equivalent satellite data variance can be explained by in-situ data according to Nash-Sutcliffe model efficiency coefficient. For snow cover extent the results were not as straightforward due to binarization — F1-score for the whole binarized dataset was 0.728, while for Oulu, Vantaa, and Sodankylä it was 0.356, 0.849, and 0.957, respectively. The results support that satellite data is trustworthy for snow water equivalent for every city, and also for snow cover extent, when evaluating the dataset as a whole.Lumen peittoalan ja vesiarvon analyysi Copernicus Global Land Service -palvelun aineistolle. Tiivistelmä. Luonnonilmiöiden ja ympäristön monitorointi laajassa mittakaavassa in-situ mittauksin on hidasta, kallista sekä vaikeakulkuisilla alueilla haastavaa, eikä näin saatu data ole jatkuvaa. Satelliittidata tuo mahdollisuuden kiertää nämä haasteet. Tämä edellyttää, että satelliittidata on luotettavaa ja paikkansapitävää. Arktisilla ...
|
|---|