Mixed Logit Model Performance and Distributional Assumptions: Preferences and GM foods

Using mixed logit models to analyse choice data is common but requires ex ante specification of the functional forms of preference distributions. We make the case for greater use of bounded functional forms and propose the use of the Marginal Likelihood, calculated using Bayesian techniques, as a single measure of model performance across non nested mixed logit specifications. Using this measure leads to very different rankings of model specifications compared to alternative rule of thumb measures. The approach is illustrated using data from a choice experiment regarding GM food types which provides insights regarding the recent WTO dispute between the EU and the US, Canada and Argentina and whether labelling and trade regimes should be based on the production process or product composition.

A practical approach to a multi-level analysis with a sparse binary outcome within a large surgical trial

Objectives: To assess the potential source of variation that surgeon may add to patient outcome in a clinical trial of surgical procedures. Methods: Two large (n = 1380) parallel multicentre randomized surgical trials were undertaken to compare laparoscopically assisted hysterectomy with conventional methods of abdominal and vaginal hysterectomy; involving 43 surgeons. The primary end point of the trial was the occurrence of at least one major complication. Patients were nested within surgeons giving the data set a hierarchical structure. A total of 10% of patients had at least one major complication, that is, a sparse binary outcome variable. A linear mixed logistic regression model (with logit link function) was used to model the probability of a major complication, with surgeon fitted as a random effect. Models were fitted using the method of maximum likelihood in SAS((R)). Results: There were many convergence problems. These were resolved using a variety of approaches including; treating all effects as fixed for the initial model building; modelling the variance of a parameter on a logarithmic scale and centring of continuous covariates. The initial model building process indicated no significant 'type of operation' across surgeon interaction effect in either trial, the 'type of operation' term was highly significant in the abdominal trial, and the 'surgeon' term was not significant in either trial. Conclusions: The analysis did not find a surgeon effect but it is difficult to conclude that there was not a difference between surgeons. The statistical test may have lacked sufficient power, the variance estimates were small with large standard errors, indicating that the precision of the variance estimates may be questionable.

Effect of citric acid and glycine addition on acrylamide and flavor in a potato model system

Acrylamide levels in cooked/processed food can be reduced by treatment with citric acid or glycine. In a potato model system cooked at 180 degrees C for 10-60 min, these treatments affected the volatile profiles. Strecker aldehydes and alkylpyrazines, key flavor compounds of cooked potato, were monitored. Citric acid limited the generation of volatiles, particularly the alkylpyrazines. Glycine increased the total volatile yield by promoting the formation of certain alkylpyrazines, namely, 2,3-dimethylpyrazine, trimethylpyrazine, 2-ethyl-3,5-dimethylpyrazine, tetramethylpyrazine, and 2,5-diethyl-3- methylpyrazine. However, the formation of other pyrazines and Strecker aldehydes was suppressed. It was proposed that the opposing effects of these treatments on total volatile yield may be used to best advantage by employing a combined treatment at lower concentrations, especially as both treatments were found to have an additive effect in reducing acrylamide. This would minimize the impact on flavor but still achieve the desired reduction in acrylamide levels.

Investigations on the effect of amino acids on acrylamide, pyrazines, and Michael addition products in model systems

Acrylamide and pyrazine formation, as influenced by the incorporation of different amino acids, was investigated in sealed low-moisture asparagine-glucose model systems. Added amino acids, with the exception of glycine and cysteine and at an equimolar concentration to asparagine, increased the rate of acrylamide formation. The strong correlation between the unsubstituted pyrazine and acrylamide suggests the promotion of the formation of Maillard reaction intermediates, and in particular glyoxal, as the determining mode of-action. At increased amino acid concentrations, diverse effects were observed. The initial rates of acrylamide formation remained high for valine, alanine, phenylalanine, tryptophan, glutamine, and Ieucine, while a significant mitigating effect, as evident from the acrylamide yields after 60 min of heating at 160 degrees C, was observed for proline, tryptophan, glycine, and cysteine. The secondary amine containing amino acids, proline and tryptophan, had the most profound mitigating effect on acrylamide after 60 min of heating. The relative importance of the competing effect of added amino acids for alpha-dicarbonyls and acrylamide-amino, acid alkylation reactions is discussed and accompanied by data on the relative formation rates of selected amino acid-AA adducts.

Effect of fecal water on an in vitro model of colonic mucosal barrier function

Fecal water (FW) has been shown to exert, in cultured cells, cytotoxic and genotoxic effects that have implications for colorectal cancer (CRC) risk. We have investigated a further biological activity of FW, namely, the ability to affect gap junctions in CACO2 cell monolayers as an index of mucosal barrier function, which is known to be disrupted in cancer. FW samples fi-om healthy, free-living, European subjects that were divided into two broad age groups, adult (40 +/- 9.7 yr; n = 53) and elderly (76 +/- 7.5 yr; n = 55) were tested for effects on gap junction using the transepithelial resistance (TER) assay. Overall, treatment of CACO2 cells with FW samples fi-om adults increased TER (+ 4 %), whereas FW from elderly subjects decreased TER (-5%); the difference between the two groups was significant (P < 0.05). We also measured several components of FW potentially associated with modulation of TER, namely, short-chain fatty acid (SCFA) and ammonia. SCFAs (propionic, acetic, and n-butyric) were significantly lower in the elderly population (-30%, -35%, and -21%, respectively, all P pound 0.01). We consider that FW modulation of in vitro epithelial barrier function is a potentially useful noninvasive biomarker, but it requires further validation to establish its relationship to CRC risk.

Synergistic effect of BSA on antioxidant activities in model food emulsions

Sunflower oil-in-water emulsions containing TBHQ, caffeic acid, epigallocatechin gallate (EGCG), or 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), both with and without BSA, were stored at 50 and 30degreesC. Oxidation of the oil was monitored by determination of the PV, conjugated diene content, and hexanal formation. Emulsions containing EGCG, caffeic acid, and, to a lesser extent, Trolox were much more stable during storage in the presence of BSA than in its absence even though BSA itself did not provide an antioxidant effect. BSA did not have a synergistic effect on the antioxidant activity of TBHQ. The BSA structure changed, with a considerable loss of fluorescent tryptophan groups during storage of solutions containing BSA and antioxidants, and a BSA-antioxidant adduct with radical-scavenging activity was formed. The highest radical-scavenging activity observed was for the isolated protein from a sample containing EGCG and BSA incubated at 30degreesC for 10 d. This fraction contained unchanged BSA as well as BSA-antioxidant adduct, but 95.7% of the initial fluorescence had been lost, showing that most of the BSA had been altered. It can be concluded that BSA exerts its synergistic effect with antioxidants because of formation of a protein-antioxidant adduct during storage, which is concentrated at the oil-water interface owing to the surface-active nature of the protein.

The effect of nonlinearity on the variational assimilation of satellite observations using a simple column model

Cloud imagery is not currently used in numerical weather prediction (NWP) to extract the type of dynamical information that experienced forecasters have extracted subjectively for many years. For example, rapidly developing mid-latitude cyclones have characteristic signatures in the cloud imagery that are most fully appreciated from a sequence of images rather than from a single image. The Met Office is currently developing a technique to extract dynamical development information from satellite imagery using their full incremental 4D-Var (four-dimensional variational data assimilation) system. We investigate a simplified form of this technique in a fully nonlinear framework. We convert information on the vertical wind field, w(z), and profiles of temperature, T(z, t), and total water content, qt (z, t), as functions of height, z, and time, t, to a single brightness temperature by defining a 2D (vertical and time) variational assimilation testbed. The profiles of w, T and qt are updated using a simple vertical advection scheme. We define a basic cloud scheme to obtain the fractional cloud amount and, when combined with the temperature field, we convert this information into a brightness temperature, having developed a simple radiative transfer scheme. With the exception of some matrix inversion routines, all our code is developed from scratch. Throughout the development process we test all aspects of our 2D assimilation system, and then run identical twin experiments to try and recover information on the vertical velocity, from a sequence of observations of brightness temperature. This thesis contains a comprehensive description of our nonlinear models and assimilation system, and the first experimental results.

Bayesian estimation of the double hurdle model in the presence of fixed costs

We present a model of market participation in which the presence of non-negligible fixed costs leads to random censoring of the traditional double-hurdle model. Fixed costs arise when household resources must be devoted a priori to the decision to participate in the market. These costs, usually of time, are manifested in non-negligible minimum-efficient supplies and supply correspondence that requires modification of the traditional Tobit regression. The costs also complicate econometric estimation of household behavior. These complications are overcome by application of the Gibbs sampler. The algorithm thus derived provides robust estimates of the fixed-costs, double-hurdle model. The model and procedures are demonstrated in an application to milk market participation in the Ethiopian highlands.

The effect of nonlinearity in CO2 heating rates on the attribution of stratospheric ozone and temperature changes

An analysis of the attribution of past and future changes in stratospheric ozone and temperature to anthropogenic forcings is presented. The analysis is an extension of the study of Shepherd and Jonsson (2008) who analyzed chemistry-climate simulations from the Canadian Middle Atmosphere Model (CMAM) and attributed both past and future changes to changes in the external forcings, i.e. the abundances of ozone-depleting substances (ODS) and well-mixed greenhouse gases. The current study is based on a new CMAM dataset and includes two important changes. First, we account for the nonlinear radiative response to changes in CO2. It is shown that over centennial time scales the radiative response in the upper stratosphere to CO2 changes is significantly nonlinear and that failure to account for this effect leads to a significant error in the attribution. To our knowledge this nonlinearity has not been considered before in attribution analysis, including multiple linear regression studies. For the regression analysis presented here the nonlinearity was taken into account by using CO2 heating rate, rather than CO2 abundance, as the explanatory variable. This approach yields considerable corrections to the results of the previous study and can be recommended to other researchers. Second, an error in the way the CO2 forcing changes are implemented in the CMAM was corrected, which significantly affects the results for the recent past. As the radiation scheme, based on Fomichev et al. (1998), is used in several other models we provide some description of the problem and how it was fixed.

Comparison of the radiative properties and direct radiative effect of aerosols from a global aerosol model and remote sensing data over ocean

A global aerosol transport model (Oslo CTM2) with main aerosol components included is compared to five satellite retrievals of aerosol optical depth (AOD) and one data set of the satellite-derived radiative effect of aerosols. The model is driven with meteorological data for the period November 1996 to June 1997 which is the time period investigated in this study. The modelled AOD is within the range of the AOD from the various satellite retrievals over oceanic regions. The direct radiative effect of the aerosols as well as the atmospheric absorption by aerosols are in both cases found to be of the order of 20 Wm−2 in certain regions in both the satellite-derived and the modelled estimates as a mean over the period studied. Satellite and model data exhibit similar patterns of aerosol optical depth, radiative effect of aerosols, and atmospheric absorption of the aerosols. Recently published results show that global aerosol models have a tendency to underestimate the magnitude of the clear-sky direct radiative effect of aerosols over ocean compared to satellite-derived estimates. However, this is only to a small extent the case with the Oslo CTM2. The global mean direct radiative effect of aerosols over ocean is modelled with the Oslo CTM2 to be –5.5 Wm−2 and the atmospheric aerosol absorption 1.5 Wm−2.

Representation of model error in a convective-scale ensemble prediction system

In this paper ensembles of forecasts (of up to six hours) are studied from a convection-permitting model with a representation of model error due to unresolved processes. The ensemble prediction system (EPS) used is an experimental convection-permitting version of the UK Met Office’s 24- member Global and Regional Ensemble Prediction System (MOGREPS). The method of representing model error variability, which perturbs parameters within the model’s parameterisation schemes, has been modified and we investigate the impact of applying this scheme in different ways. These are: a control ensemble where all ensemble members have the same parameter values; an ensemble where the parameters are different between members, but fixed in time; and ensembles where the parameters are updated randomly every 30 or 60 min. The choice of parameters and their ranges of variability have been determined from expert opinion and parameter sensitivity tests. A case of frontal rain over the southern UK has been chosen, which has a multi-banded rainfall structure. The consequences of including model error variability in the case studied are mixed and are summarised as follows. The multiple banding, evident in the radar, is not captured for any single member. However, the single band is positioned in some members where a secondary band is present in the radar. This is found for all ensembles studied. Adding model error variability with fixed parameters in time does increase the ensemble spread for near-surface variables like wind and temperature, but can actually decrease the spread of the rainfall. Perturbing the parameters periodically throughout the forecast does not further increase the spread and exhibits “jumpiness” in the spread at times when the parameters are perturbed. Adding model error variability gives an improvement in forecast skill after the first 2–3 h of the forecast for near-surface temperature and relative humidity. For precipitation skill scores, adding model error variability has the effect of improving the skill in the first 1–2 h of the forecast, but then of reducing the skill after that. Complementary experiments were performed where the only difference between members was the set of parameter values (i.e. no initial condition variability). The resulting spread was found to be significantly less than the spread from initial condition variability alone.

The effect of increased convective entrainment on Asian monsoon biases in the MetUM General Circulation Model

We demonstrate that summer precipitation biases in the South Asian monsoon domain are sensitive to increasing the convective parametrisation’s entrainment and detrainment rates in the Met Office Unified Model. We explore this sensitivity to improve our understanding of the biases and inform efforts to improve convective parametrisation. We perform novel targeted experiments in which we increase the entrainment and detrainment rates in regions of especially large precipitation bias. We use these experiments to determine whether the sensitivity at a given location is a consequence of the local change to convection or is a remote response to the change elsewhere. We find that a local change leads to different mean-state responses in comparable regions. When the entrainment and detrainment rates are increased globally, feedbacks between regions usually strengthen the local responses. We choose two regions of tropical ascent that show different mean-state responses, the western equatorial Indian Ocean and western north Pacific, and analyse them as case studies to determine the mechanisms leading to the different responses. Our results indicate that several aspects of a region’s mean-state, including moisture content, sea surface temperature and circulation, play a role in local feedbacks that determine the response to increased entrainment and detrainment.

The effect of model configuration on modelled hillslope–riparian interactions

The transfer of hillslope water to and through the riparian zone forms a research area of importance in hydrological investigations. Numerical modelling schemes offer a way to visualise and quantify first-order controls on catchment runoff response and mixing. We use a two-dimensional Finite Element model to assess the link between model setup decisions (e.g. zero-flux boundary definitions, soil algorithm choice) and the consequential hydrological process behaviour. A detailed understanding of the consequences of model configuration is required in order to produce reliable estimates of state variables. We demonstrate that model configuration decisions can determine effectively the presence or absence of particular hillslope flow processes and, the magnitude and direction of flux at the hillslope–riparian interface. If these consequences are not fully explored for any given scheme and application, the resulting process inference may well be misleading.