No description available.

No description available.

Η παρούσα μεταπτυχιακή διατριβή εντάσσεται ερευνητικά στην περιοχή της Στατιστικής Θεωρίας Αποφάσεων και ειδικότερα στη βελτίωση των εκτιμητών του μέτρου της διασποράς για την αντίστροφη κανονική κατανομή IG(μ,λ), όπου μ είναι η παράμετρος θέσης, ενώ το λ είναι η παράμετρος κλίμακος και εκφράζει το αντίστροφο μέτρο της διασποράς. Στο Κεφάλαιο 1 παρουσιάζονται κάποιοι χρήσιμοι, για την πορεία της μελέτης μας, ορισμοί και θεωρήματα, στο Κεφάλαιο 2 μελετάται το μοντέλο της αντίστροφης κανονικής κατανομής, δίνονται τα χαρακτηριστικά μεγέθη αυτής και υπολογίζονται γνωστοί εκτιμητές για το μέτρο της διασποράς. Στο Κεφάλαιο 3, εξετάζεται το πρόβλημα εκτίμησης του μέτρου της διασποράς , τόσο ως προς την τετραγωνική συνάρτηση ζημίας , όσο και προς την συνάρτηση ζημίας εντροπίας . Υπό ορισμένες συνθήκες, αποδεικνύεται η μη αποδεκτικότητα, του βέλτιστου αναλλοίωτου εκτιμητή, κατασκευάζοντας καλύτερους εκτιμητές τύπου Stein [1964, Ann. Inst. Statist. Math.]. Η μέθοδος κατασκευής αυτών των εκτιμητών παρουσιάστηκε στην εργασία των N. Pal and B. Sinha [1989, Statistical data analysis and inference]. Στο Κεφάλαιο 4, και για το ίδιο πρόβλημα εκτίμησης που πραγματεύεται το προηγούμενο κεφάλαιο,κατασκευάζονται καλύτεροι εκτιμητές από το βέλτιστο αναλλοίωτο εκτιμητή χρησιμοποιώντας την μεθοδολογία των Brewster and Zidek [1974, Ann. Statist.]. Η μέθοδος κατασκευής αυτών των εκτιμητών παρουσιάζεται στην εργασία των B.MacGibbon and G.Shorrock [1997, Statist. Probab. Lett.]. / The present postgraduate thesis is placed among the area of Statistical Decision Theory and especially we give improved estimators of dispersion of an inverse Gaussian distribution IG(μ,λ), where μ is the mean and λ is a parameter, known as inverse measure of dispersion. In Chapter 1 are some useful definitions and theorems are presented, in Chapter 2 the model of inverse Gaussian distribution is studied, we give some properties of the model and known estimators for the measure of dispersion , are presented. In Chapter 3, we examine the problem of estimating the measure of dispersion under quadratic and entropy losses. Under certain conditions, we derive improved estimators (Stein [1964, Ann. Inst. Statist. Math.]) of the best affine equivariant estimator for . The method of construction of these estimators is presented in the paper of N. Pal and B. Sinha [ 1989, Statistical data analysis and inference ]. In Capital 4, and for the same problem of estimating , improved estimators of the best affine equivariant estimator are derived, using the methodology of Brewster and Zidek [1974, Ann. Statist. ]. This method of construction of these estimators is presented in the paper of B. MacGibbon and G. Shorrock [1997, Statist. Probab. Lett. ].

No description available.

No description available.

Chlorine is the primary method for disinfection of drinking water in the United States; however, growing concerns about the potential hazards associated with carcinogenic chlorine disinfection by-products have resulted in increased efforts to develop alternative methods of water disinfection. In addition, it is sometimes difficult to maintain an adequate concentration of free chlorine throughout a drinking water distribution system due to intrusion events and the presence of biofilms.Silver and copper are widely used as environmental biocides and as clinical antimicrobial agents. Copper has been extensively used as an algaecide for many years, and is reported to be one of the most toxic metals to heterotrophic bacteria in aquatic environments. In this study, silver and copper, both individually and in combination, were able to significantly reduce the numbers of the bacterial pathogens Salmonella typhimurium, Escherichia coli, Listeria monocytogenes, and Mycobacterium fortuitum in aqueous solutions. Silver and copper together could potentially be used as a secondary disinfectant in water treatment because their residual effect is long lasting and no harmful byproducts are generated.Besides water, produce can also be a source of microbial pathogens. Fruits and vegetables are the third leading source in microbial foodborne outbreaks, accounting for more than 10,000 cases of foodborne illness in the U.S. between 1990 and 2001. Alternative produce sanitizers in addition to chlorine were evaluated against foodborne pathogens inoculated onto lettuce. Overall, Fit® (citric acid, grapefruit oil extract) and chlorine yielded the greatest reductions of the study organisms.A large number of compounds are available for use as disinfectants; however, some are highly toxic, corrosive and produce harmful by-products. Natural antimicrobial products are another possible alternative. The efficacy of a natural peptide-based antimicrobial, Absolute Fx, was assessed against bacteria and viruses. Absolute Fx effectively inactivated the study organisms.

I den har rapporten kartläggs och bedöms omfattningen av värmeförluster från kulvertar i Landstinget Kronobergs sekundära fjärrvärmenät på Sigfridsområdet. Teoretiska stationära värmeförlustberäkningar har utförts med Petter Wallenténs samband från 1991 samt med det via internet tillgängliga beräkningsprogrammet Logstor Calculator 2.1. Beräkningar över energibesparingen vid ett byte av de äldsta kulvertarna i det sekundära nätet till nya kulvertar visar att tvillingkulvertar i isolerklass 2 är att föredra framför enkelrör både i isolerklass 2 såväl som 3. Rapporten behandlar miljöaspekter kring fjärrvärme samt möjligheter och konsekvenser kring en energieffektivisering av distributionen genom sänkta systemtemperaturer.

The spatial distribution of four species of rorqual whales was studied along the north shore of the Gulf of St. Lawrence during the months of June to October from 1989 to 2000. A research effort of 6511 hours at sea yielded 849 sightings of blue whales (Balaenoptera musculus), 5291 of finback whales (Balaenoptera physalus), 3822 of humpback whales (Megaptera novaeangliae) and 6489 of minke whales (Balaenoptera acutorostrata). Spatial and behavioural data were collected at sea using inflatable boats, and Global Positioning Systems (GPS) were used to obtain accurate positions. These data were plotted and analysed using a Geographic Information System (GIS) to test the hypotheses that patterns of distribution were not random, were associated with bathymetry and reflected specific differences in habitat use. The resulting maps illustrated the clustered distribution of rorqual whales linked to sea-bottom topography, probably associated with areas of local upwelling and increased productivity. Blue and fin whales shared almost the same distribution, humpback whales were found in slightly deeper, offshore waters whereas minke whales were more abundant in shallower waters. Little attention has been given until now to local patterns of distribution in the area and such information can be useful for practical management considerations. The results emphasise the importance of scale in ecological studies of marine mammals and the need for further research using additional oceanographic parameters, in order to better understand habitat selection.

No description available.

Cast iron pipes were installed broadly in North American water utilities. Many of these cast iron pipes are corroded and are continuous sources of Fe(II) ions in drinking water distribution systems. Recent studies have reported that soluble or particulate iron decreases water quality in distribution systems. In this study, an array of bench scale experiments were conducted to evaluate the impact of most common water parameters (e.g., pH, PO4, Cl2, and DOM) on the oxidation rate of Fe(II) ions, and on the formation of iron suspense in synthetic water samples. This was accomplished using a 24 full factorial design approach at a 95% confidence level. This study demonstrated that a lower content of iron suspension color, turbidity, and smaller particle size would appear to be obtained in presence of a phosphate based corrosion inhibitor at a pH value of 6.5 compared to a pH value of 8.5. To investigate the impact of Fe(II) ions, phosphate, pH and reaction time, and their interaction on DBPs formation in water samples, this study was conducted following an experimental design approach. Considering all the significant (α = 0.05, p < 0.05) factors, mathematical models for HAAs and THMs prediction were developed using 80 experiments. The models’ adequacy was checked thorough the statistical and graphical diagnostics. Different sources of natural water samples collected from three main water treatment plants in Halifax, Canada, were used to validate the models. This study suggested that the models’ performance were found to be excellent under a wide range of studied variables. Consequently, the most predominant iron oxides (goethite and magnetite) were used to investigate their impact on chlorine decay and DBPs formation study. Goethite and magnetite were also used for the adsorption of DBPs precursor (DOM). The DOM adsorption data illustrated to fit well with the Langmuir adsorption isotherm, indicating monolayer coverage. Molecular weight (MW) distributions of DBPs precursor (DOM) revealed that the higher molecular weight fractions adsorbed preferentially onto goethite followed by magnetite surface. The change of MW distribution of DOM was found to be in reasonable agreement with the change of DBPs formation in iron-water systems.