Detection of vegetation changes in cold regions using a combination of radar and optical satellite data

The extremely sensitive ecosystem of the cold regions is going through a rapid climate change, and the projections show an increase in temperature. Considering that the temperature influences the length of the growing season as well as composition, biomass and plant distribution, studying the vegetation is crucial to analyse the consequences of seasonal variability in mountain and Arctic areas. Notably, changes in the start and end of the growing season may have a considerable impact on these ecosystems. Unfortunately, the method in which data are generally collected, such as field data, does not allow a large spatial or temporal scale analysis of these ecological responses. Conversely, the remote sensing measurement of environmental parameters by satellite sensors facilitates this type of analysis, and has proved to be crucial in ecological studies. There are, however, challenges of processing optical data, namely clouds and their shadows, which interferes with remote sensing studies. Cloud detection in the Arctic and alpine areas is especially demanding since cloud-contaminated conditions are frequent. Hence, optical data require corrections according to different environmental conditions in order to be enabled in vegetation studies. As opposed to optical images, radar data can provide a systematic data source of land surface and cover changes that are insensitive to cloud cover and hour of acquisition, even though signal processing is challenging. The overall aim of the present project is to monitor vegetation seasonal dynamics in mountainous and Arctic region by synergic use of optical and radar satellite data. We expect that a combination of radar and optical data allow more accurate analysis of these changes. To analyse these seasonal variations three different study areas have been chosen: Adventdalen valley (Svalbard archipelago), Dovrefjell National Park (Southern part of Norway) and Val di Mazia-Matscher Tal (Südtirol, Italy). Specifically, the research in these areas will require the accomplishment of the following tasks: determine the onset/end of the growing seasons; map vegetation communities; biomass estimation. To reach the aim of this study, we will use time-series from Sentinel 1 and Sentinel 2 satellites. To assess the accuracy of satellite processed products, we use proximal sensing information and field measurements