Study of Spatial Distribution and Temporal Evolution of Snow Cover and Snow Depth, using reanalysis data over Karakoram, Himalaya and Kunlun Mountains

Snow cover and snow depth are essential for regional climate dynamics, impacting billions of people, and influencing water availability, hydropower and climatic balance. Studying spatial distribution and investigating the interannual, annual and seasonal variation in snow across this region is therefore fundamental. If assessed on a long-term basis, it is also useful for monitoring Climate Change and the connected impacts. The most recent Intergovernmental Panel on Climate Change (IPCC) report indicates the continuing decline in seasonal snow duration and glacial mass, with reductions in snow covered areas and snow volumes decreasing in most regions of the Himalayas, further supported by recent studies that already identified snowdiminishing in recent decades. Therefore, this thesis aims to investigate the temporal evolution and the spatial distribution of snow cover and snow depth in the Kunlun, Himalaya and Karakoram Mountains, latitudes 20°N to 40°N andlongitudes 70°E to 100°E, with the reanalysis produced by ECMWF-ERA 5 Land from 1951 to today. ERA5-Land reanalysis dataset provided by the European Centre for Medium- Range Weather Forecasts (ECMWF) offers high resolution climate data with a spatial resolution of 9 km and hourly data resolution. Snow cover and snow depthdata were processed using RStudio, ideal to manage and manipulate the extensive dataset. The spatial distribution of snow cover and snow depth was analyzed on annual, seasonal, and monthly basis, and maps were generated for specific subperiods(e.g., 1991–2023 and 2000–2023). For temporal analysis, trends in snow cover were achieved using linear regression and the Mann-Kendall test to determine their significancy and calculated for the mean regional series and for individual grid points to identify regional patterns in snow cover evolution. The analysis of the annual mean snow cover distribution from 1951 to 2023 shows that 48.6% of (or 51% if we account the sea) of the gridded area can be considered snow free, while 49% is considered snow-covered, meaning it each grid point has from 1% to 100% of snow coverage. Although the snow peaking close or exactly of 100% coverage is more present in the western region, there are more snow-covered grid points in the eastern part of the area, specifically where the Karakoram Mountain range is. Conversely, areas with no snow cover correspond to the lower areas in altitude, and mainly bellow latitude 30º N, including most of India, and especially the central part of the country. Snow depth distributions follow a similar spatial pattern, with the deepest snow found in the highest snow-covered areas, those with coverage ranging 90-100%, particularly concentrated in the western regions. In the seasonal analysis, Winter consistently showed the highest snow accumulation, followed by Spring, and Autumn while Summer exhibited minimal snow presence. Autumn, as the snow accumulation season, showed more grid points with at least 1% snow cover compared to the other seasons. Monthly analysis further supported the seasonal trends, with February usually showing the highest snow cover, while April marks the decrease, and August records the least snow. Snow depths show a more consistent pattern, with less variations seasonally, but still, some in lower snow depths, while deeper snow (>1m) remained stable through the year. Lastly, examining individual years, the year with the highest snow cover was 1957, with an average annual snow cover of 19.9 %, in contrast, the year with the lowest snow cover was 2016, with 13.7% average annual snow cover. The temporal analysis revealed a statistically significant negative trend in snow cover from 1951 to 2023, with a decline of -0.28% per decade. Moreover, the rate of decline has accelerated in the most recent decades (1991–2023), moving to - 0.60% per decade, statistically significant as well. However, in the 2000-2023 period, the snow cover trend of -0.49% per decade is no longer statistically significant, with a p-value > 0.05 confirmed with the Mann-Kendall test, likely due to the shorter dataset duration with respect to the high interannual variability observed in the data. For the regional approach towards temporal evolution,statistically significant slopes were filtered in relation to not significant ones, resulting in 23% of the whole area presenting significant change in snow cover between 1951 and 2023. The strongest significant negative trends, for individual grid points, were observed in the eastern region of the study area, while the Karakoram and western Himalayas showed less pronounced negative changes in snow, with some areas even suggesting positive trends, though most of these positive slopes were proven not statistically significant. Given the size of the study area, this work could benefit from a more local separation of regions, anddetailed examination of the regional trends in order to understand snow dynamics in the Himalayas.