No abstract

No abstract

Abstract. A Hidden Markov Model (HMM) is used to classify an out of sample observation vector into either of two regimes. This leads to a procedure for making probability forecasts for changes of regimes in a time series, i.e. for turning points.Instead of maximizing a likelihood, the model is estimated with respect to known past regimes. This makes it possible to perform feature extraction and estimation for different forecasting horizons. The inference aspect is emphasized by including a penalty for a wrong decision in the cost function. The method is tested by forecasting turning points in the Swedish and US economies, using leading data. Clear and early turning point signals are obtained, contrasting favourable with earlier HMM studies.Some theoretical arguments for this are given.

[1] This paper describes a new method for estimating snow albedo for satellite retrieval of surface UV irradiance and daily UV doses over snow-covered terrains. The method is based on combining satellite and meteorological analysis data. The satellite data exploited in this work are the measured reflectivities of the Total Ozone Mapping Spectrometer/ Nimbus 7 instrument that coincides with the European Centre for Medium-Range Weather Forecasts ERA-15 reanalyzed meteorological data. We compared satellite-retrieved UV daily doses to the ground-based measurements of two Finnish and five Canadian sites. The comparison clearly showed that the new snow albedo approach improves the accuracy of the satellite-retrieved UV doses.

No abstract

Spent nuclear fuel from Finnish power plants is planned to be deposited deep in the crystalline bedrock in Olkiluoto, Finland. The bedrock and more specifically the elemental composition of ground water, which is composed of the fracture water and the matrix pore water, needs to be well characterized to assess the risks inherent to the long term safety of the site. To this end, it is valuable to investigate elemental composition of the matrix pore water since it tends to conserve hydrogeological signals for longer time spans compared to open fracture waters.In this study, the chloride concentration of matrix pore water in veined gneiss (VGN) and pegmatitic granite (PGR) samples were investigated. Chloride was out-diffused from the naturally saturated rock cores into deionized water. Chloride pore diffusion coefficients were derived by modelling the chloride breakthrough curves obtained from the out-diffusion experiments. Two component modelling gave best fit to the experimental results. There two diffusion coefficients were (9±2)×10-11 m2/s and (0.5±0.1)×10-11 m2/s for PGR and (2.5±0.5)×10-11 m2/s and (0.4±0.1)×10-11 m2/s for VGN. Porosity distribution and total porosities of the rock samples were studied with the C-14-PMMA autoradiography. Porosity for PGR was found to be 0.6 % with large mineral transecting fissures, and porosity for VGN was found to be 0.7 % with highly porous mineral clusters connected to each other via grain boundaries and intragranular pores. The findings here show that heterogeneity has to be taken into account in modelling to find better agreement with the experimental results. C-14-PMMA autoradiography results indicate dual-component behavior for diffusion in PGR and VGN which were used in the modelling.

No abstract

The spatial porosity and mineral distribution of geological materials strongly affects transport processes in them.X-ray micro computed tomography (X-μCT) has proven to be a powerful tool for characterizing the spatial mineral distribution of geological samples in 3-D. However, limitations in resolution prevent an accurate characterization of pore space especially for tight crystalline rock samples and 2-D methods such as C-14polymethylmethacrylate (C-14-PMMA) autoradiography and scanning electron microscopy (SEM) are needed.The spatial porosity and mineral distributions of tight crystalline rock samples from Äspö, Sweden, and Olkiluoto, Finland, were studied here. The X-μCT were used to characterize the spatial distribution of the main minerals in 3-D. Total porosities, fracture porosities, fracture densities and porosity distributions of the samples were determined using the C-14-PMMA autoradiography and characterization of mineral-specific porosities were assisted using chemical staining of rock surfaces. SEM and energy dispersive X-ray spectroscopy (EDS) were used to determine pore apertures and identify the minerals. It was shown that combination of the different imaging techniques creates a powerful tool for the structural characterization of crystalline rock samples. The combination of the results from different methods allowed the construction of spatial porosity, mineral and mineral grain distributions of the samples in 3-D. These spatial distributions enable reactive transport modeling using a more realistic representation of the heterogeneous structure of samples. Furthermore, the realism of the 2 spatial distributions were increased by determinig the densities and porosities of fractures and by the virtual construction heterogeneous mineral distributions of minerals that cannot be separated by X-μCT.