Robust parametric indirect estimates of the expected cost of a hospital stay with covariates and censored data.

We consider the problem of estimating the mean hospital cost of stays of a class of patients (e.g., a diagnosis-related group) as a function of patient characteristics. The statistical analysis is complicated by the asymmetry of the cost distribution, the possibility of censoring on the cost variable, and the occurrence of outliers. These problems have often been treated separately in the literature, and a method offering a joint solution to all of them is still missing. Indirect procedures have been proposed, combining an estimate of the duration distribution with an estimate of the conditional cost for a given duration. We propose a parametric version of this approach, allowing for asymmetry and censoring in the cost distribution and providing a mean cost estimator that is robust in the presence of extreme values. In addition, the new method takes covariate information into account.

The effect of age at the first mating and herd size on the lifetime productivity of sows

The effect of age at the first mating and herd size were evaluated in the reference Spanish Databank (BDporc) of 37 698 sows born between 1991 and 1995 and with individual lifetime records. The data included dates of births at entrance and culling, first mating, repetitive mating and conception, first farrowing and weaning records. Individual records were validated before the analysis by screening them through a tolerance “filter” in order to eliminate the extreme values from the analysis. The total database of the sows was classified in 7 classes according to age at the first mating (< 210, 210–220, 221–230, 231–240, 241–250, 251–270, and > 270 days) and in 6 classes of herd size (< 200, 200–300, 301–400, 401–600, 601–800, and > 800 sows). The total number of litters and number of weaned piglets obtained from each sow during the lifetime production were significantly (P < 0.05) greater for gilts between 221 and 240 d of age at the first mating. There was a significant (P < 0.001) effect of the herd size on the reproductive performance of the sow, and the best performance was obtained with herds with 401 to 600 sows compared to < 200 or > 800 sow-herds. Furthermore, a significant (P < 0.001) interaction between age at the first mating and herd size was detected and can be associated with a particular pattern for the herd size class 401–600 sows with the best performances obtained for the sows first mated at less than 200 days. For the other herd sizes, the results indicated that sows mated for the first time at the right age, 221–240 days, are more productive, both in the number and size of the parities throughout lifetime production.

Use of spectrophotometry as a tool to quantify the sporulation of Penicillium allii in garlic lesions

A technique to measure the concentration of Penicillium allii conidia in damp chamber experiments by spectrophotometry was developed. A negative linear correlation (R²=0.56) was observed between transmittance at 340 nm and the concentration of P. allii conidia in water agar 0.05%. The equation that relates transmittance (T) with concentration (conidia mL-1) (y) is: y = 9.3 10(6) - 86497 T. The method was assayed by inoculating 43 P. allii strains in two garlic cultivars. The method proved to be more rapid than the traditional use of a hemocytometer with an improved accuracy. The CV of the number of conidia per hemocytometer reticule was of 35.04%, while the transmittance CV was of 2.73%. The extreme values chosen for T were 40 and 80 because the sensitivity of the method decreased when concentrations of conidia were out of this range.

Characteristics of urban heat island (UHI) in a high latitude coastal city - a case study of Turku, SW Finland

The term urban heat island (UHI) refers to the common situation in which the city is warmer than its rural surroundings. In this dissertation, the local climate, and especially the UHI, of the coastal city of Turku (182,000 inh.), SW Finland, was studied in different spatial and temporal scales. The crucial aim was to sort out the urban, topographical and water body impact on temperatures at different seasons and times of the day. In addition, the impact of weather on spatiotemporal temperature differences was studied. The relative importance of environmental factors was estimated with different modelling approaches and a large number of explanatory variables with various spatial scales. The city centre is the warmest place in the Turku area. Temperature excess relative to the coldest sites, i.e. rural areas about 10 kilometers to the NE from the centre, is on average 2 °C. Occasionally, the UHI intensity can be even 10 °C. The UHI does not prevail continuously in the Turku area, but occasionally the city centre can be colder than its surroundings. Then the term urban cool island or urban cold island (UCI) is used. The UCI is most common in daytime in spring and in summer, whereas during winter the UHI prevails throughout the day. On average, the spatial temperature differences are largest in summer, whereas the single extreme values are often observed in winter. The seasonally varying sea temperature causes the shift of relatively warm areas towards the coast in autumn and inland in spring. In the long term, urban land use was concluded to be the most important factor causing spatial temperature differences in the Turku area. The impact was mainly a warming one. The impact of water bodies was emphasised in spring and autumn, when the water temperature was relatively cold and warm, respectively. The impact of topography was on average the weakest, and was seen mainly in proneness of relatively low-lying places for cold air drainage during night-time. During inversions, however, the impact of topography was emphasised, occasionally outperforming those of urban land use and water bodies.

Étude du maximum et des hauts points de la marche aléatoire branchante inhomogène et du champ libre gaussien inhomogène

Voir la bibliographie du mémoire pour les références du résumé. See the thesis`s bibliography for the references in the summary.

Mixing compositions and other scales

Theory of compositional data analysis is often focused on the composition only. However in practical applications we often treat a composition together with covariables with some other scale. This contribution systematically gathers and develop statistical tools for this situation. For instance, for the graphical display of the dependence of a composition with a categorical variable, a colored set of ternary diagrams might be a good idea for a first look at the data, but it will fast hide important aspects if the composition has many parts, or it takes extreme values. On the other hand colored scatterplots of ilr components could not be very instructive for the analyst, if the conventional, black-box ilr is used. Thinking on terms of the Euclidean structure of the simplex, we suggest to set up appropriate projections, which on one side show the compositional geometry and on the other side are still comprehensible by a non-expert analyst, readable for all locations and scales of the data. This is e.g. done by defining special balance displays with carefully- selected axes. Following this idea, we need to systematically ask how to display, explore, describe, and test the relation to complementary or explanatory data of categorical, real, ratio or again compositional scales. This contribution shows that it is sufficient to use some basic concepts and very few advanced tools from multivariate statistics (principal covariances, multivariate linear models, trellis or parallel plots, etc.) to build appropriate procedures for all these combinations of scales. This has some fundamental implications in their software implementation, and how might they be taught to analysts not already experts in multivariate analysis

North Atlantic Oscillation response to transient greenhouse gas forcing and the impact on European winter climate: a CMIP2 multi-model assessment

This study investigates the response of wintertime North Atlantic Oscillation (NAO) to increasing concentrations of atmospheric carbon dioxide (CO2) as simulated by 18 global coupled general circulation models that participated in phase 2 of the Coupled Model Intercomparison Project (CMIP2). NAO has been assessed in control and transient 80-year simulations produced by each model under constant forcing, and 1% per year increasing concentrations of CO2, respectively. Although generally able to simulate the main features of NAO, the majority of models overestimate the observed mean wintertime NAO index of 8 hPa by 5-10 hPa. Furthermore, none of the models, in either the control or perturbed simulations, are able to reproduce decadal trends as strong as that seen in the observed NAO index from 1970-1995. Of the 15 models able to simulate the NAO pressure dipole, 13 predict a positive increase in NAO with increasing CO2 concentrations. The magnitude of the response is generally small and highly model-dependent, which leads to large uncertainty in multi-model estimates such as the median estimate of 0.0061 +/- 0.0036 hPa per %CO2. Although an increase of 0.61 hPa in NAO for a doubling in CO2 represents only a relatively small shift of 0.18 standard deviations in the probability distribution of winter mean NAO, this can cause large relative increases in the probabilities of extreme values of NAO associated with damaging impacts. Despite the large differences in NAO responses, the models robustly predict similar statistically significant changes in winter mean temperature (warmer over most of Europe) and precipitation (an increase over Northern Europe). Although these changes present a pattern similar to that expected due to an increase in the NAO index, linear regression is used to show that the response is much greater than can be attributed to small increases in NAO. NAO trends are not the key contributor to model-predicted climate change in wintertime mean temperature and precipitation over Europe and the Mediterranean region. However, the models' inability to capture the observed decadal variability in NAO might also signify a major deficiency in their ability to simulate the NAO-related responses to climate change.

When can we expect extremely high surface temperatures?

In the Essence project a 17-member ensemble simulation of climate change in response to the SRES A1b scenario has been carried out using the ECHAM5/MPI-OM climate model. The relatively large size of the ensemble makes it possible to accurately investigate changes in extreme values of climate variables. Here we focus on the annual-maximum 2m-temperature and fit a Generalized Extreme Value (GEV) distribution to the simulated values and investigate the development of the parameters of this distribution. Over most land areas both the location and the scale parameter increase. Consequently the 100-year return values increase faster than the average temperatures. A comparison of simulated 100-year return values for the present climate with observations (station data and reanalysis) shows that the ECHAM5/MPI-OM model, as well as other models, overestimates extreme temperature values. After correcting for this bias, it still shows values in excess of 50°C in Australia, India, the Middle East, North Africa, the Sahel and equatorial and subtropical South America at the end of the century.

Accessing extremes of mid-latitudinal wave activity: methodology and application

A statistical methodology is proposed and tested for the analysis of extreme values of atmospheric wave activity at mid-latitudes. The adopted methods are the classical block-maximum and peak over threshold, respectively based on the generalized extreme value (GEV) distribution and the generalized Pareto distribution (GPD). Time-series of the ‘Wave Activity Index’ (WAI) and the ‘Baroclinic Activity Index’ (BAI) are computed from simulations of the General Circulation Model ECHAM4.6, which is run under perpetual January conditions. Both the GEV and the GPD analyses indicate that the extremes ofWAI and BAI areWeibull distributed, this corresponds to distributions with an upper bound. However, a remarkably large variability is found in the tails of such distributions; distinct simulations carried out under the same experimental setup provide sensibly different estimates of the 200-yr WAI return level. The consequences of this phenomenon in applications of the methodology to climate change studies are discussed. The atmospheric configurations characteristic of the maxima and minima of WAI and BAI are also examined.

Reconstruction of gusty wind speed time series from autonomous data logger records

The collection of wind speed time series by means of digital data loggers occurs in many domains, including civil engineering, environmental sciences and wind turbine technology. Since averaging intervals are often significantly larger than typical system time scales, the information lost has to be recovered in order to reconstruct the true dynamics of the system. In the present work we present a simple algorithm capable of generating a real-time wind speed time series from data logger records containing the average, maximum, and minimum values of the wind speed in a fixed interval, as well as the standard deviation. The signal is generated from a generalized random Fourier series. The spectrum can be matched to any desired theoretical or measured frequency distribution. Extreme values are specified through a postprocessing step based on the concept of constrained simulation. Applications of the algorithm to 10-min wind speed records logged at a test site at 60 m height above the ground show that the recorded 10-min values can be reproduced by the simulated time series to a high degree of accuracy.

An experimental test of response consistency in contingent valuation

We compare hypothetical and observed (experimental) willingness to pay (WTP) for a gradual improvement in the environmental performance of a marketed good (an office table). First, following usual practices in marketing research, subjects’ stated WTP for the improvement is obtained. Second, the same subjects participate in a real reward experiment designed to replicate the scenario valued in the hypothetical question. Our results show that, independently of the degree of the improvement, there are no significant median differences between stated and experimental data. However, subjects reporting extreme values of WTP (low or high) exhibit a more moderate behavior in the experiment.

Towards a general theory of extremes for observables of chaotic dynamical systems

In this paper we provide a connection between the geometrical properties of the attractor of a chaotic dynamical system and the distribution of extreme values. We show that the extremes of so-called physical observables are distributed according to the classical generalised Pareto distribution and derive explicit expressions for the scaling and the shape parameter. In particular, we derive that the shape parameter does not depend on the cho- sen observables, but only on the partial dimensions of the invariant measure on the stable, unstable, and neutral manifolds. The shape parameter is negative and is close to zero when high-dimensional systems are considered. This result agrees with what was derived recently using the generalized extreme value approach. Combining the results obtained using such physical observables and the properties of the extremes of distance observables, it is possible to derive estimates of the partial dimensions of the attractor along the stable and the unstable directions of the flow. Moreover, by writing the shape parameter in terms of moments of the extremes of the considered observable and by using linear response theory, we relate the sensitivity to perturbations of the shape parameter to the sensitivity of the moments, of the partial dimensions, and of the Kaplan–Yorke dimension of the attractor. Preliminary numer- ical investigations provide encouraging results on the applicability of the theory presented here. The results presented here do not apply for all combinations of Axiom A systems and observables, but the breakdown seems to be related to very special geometrical configurations.

Simulation of long-term thermal characteristics of three Estonian lakes

A one-dimensional surface energy-balance lake model, coupled to a thermodynamic model of lake ice, is used to simulate variations in the temperature of and evaporation from three Estonian lakes: Karujärv, Viljandi and Kirjaku. The model is driven by daily climate data, derived by cubic-spline interpolation from monthly mean data, and was run for periods of 8 years (Kirjaku) up to 30 years (Viljandi). Simulated surface water temperature is in good agreement with observations: mean differences between simulated and observed temperatures are from −0.8°C to +0.1°C. The simulated duration of snow and ice cover is comparable with observed. However, the model generally underpredicts ice thickness and overpredicts snow depth. Sensitivity analyses suggest that the model results are robust across a wide range (0.1–2.0 m−1) of lake extinction coefficient: surface temperature differs by less than 0.5°C between extreme values of the extinction coefficient. The model results are more sensitive to snow and ice albedos. However, changing the snow (0.2–0.9) and ice (0.15–0.55) albedos within realistic ranges does not improve the simulations of snow depth and ice thickness. The underestimation of ice thickness is correlated with the overestimation of snow cover, since a thick snow layer insulates the ice and limits ice formation. The overestimation of snow cover results from the assumption that all the simulated winter precipitation occurs as snow, a direct consequence of using daily climate data derived by interpolation from mean monthly data.

Estimating the number of ozone peaks in Mexico City using a non-homogeneous Poisson model

In this paper, we consider the problem of estimating the number of times an air quality standard is exceeded in a given period of time. A non-homogeneous Poisson model is proposed to analyse this issue. The rate at which the Poisson events occur is given by a rate function lambda(t), t >= 0. This rate function also depends on some parameters that need to be estimated. Two forms of lambda(t), t >= 0 are considered. One of them is of the Weibull form and the other is of the exponentiated-Weibull form. The parameters estimation is made using a Bayesian formulation based on the Gibbs sampling algorithm. The assignation of the prior distributions for the parameters is made in two stages. In the first stage, non-informative prior distributions are considered. Using the information provided by the first stage, more informative prior distributions are used in the second one. The theoretical development is applied to data provided by the monitoring network of Mexico City. The rate function that best fit the data varies according to the region of the city and/or threshold that is considered. In some cases the best fit is the Weibull form and in other cases the best option is the exponentiated-Weibull. Copyright (C) 2007 John Wiley & Sons, Ltd.

Using stochastic volatility models to analyse weekly ozone averages in Mexico City

In this paper we make use of some stochastic volatility models to analyse the behaviour of a weekly ozone average measurements series. The models considered here have been used previously in problems related to financial time series. Two models are considered and their parameters are estimated using a Bayesian approach based on Markov chain Monte Carlo (MCMC) methods. Both models are applied to the data provided by the monitoring network of the Metropolitan Area of Mexico City. The selection of the best model for that specific data set is performed using the Deviance Information Criterion and the Conditional Predictive Ordinate method.