Estimation of Snow Depth in Northern Sweden: Using Gaussian Markov Random Fields

Abstract: The goal of this dissertation is to estimate the amount of water in snowpacks at the areas Kultsjön and Rensaren in northen Sweden. A model is built using spatial statistics, and the estimations will be done by Integrated Nested Laplace Approximations (INLA). Then a validation process was done by removing data from the data set to create a validation set, and compare the estimations done by INLA to a simple linear regression. Finally the model is applied to the areas of interest, and the results are analysed.