A Bayesian model averaging approach for estimating the relative risk of mortality associated with heat waves in 105 U.S. cities

Estimating the risks heat waves pose to human health is a critical part of assessing the future impact of climate change. In this article, we propose a flexible class of time series models to estimate the relative risk of mortality associated with heat waves and conduct Bayesian model averaging (BMA) to account for the multiplicity of potential models. Applying these methods to data from 105 U.S. cities for the period 1987-2005, we identify those cities having a high posterior probability of increased mortality risk during heat waves, examine the heterogeneity of the posterior distributions of mortality risk across cities, assess sensitivity of the results to the selection of prior distributions, and compare our BMA results to a model selection approach. Our results show that no single model best predicts risk across the majority of cities, and that for some cities heat-wave risk estimation is sensitive to model choice. Although model averaging leads to posterior distributions with increased variance as compared to statistical inference conditional on a model obtained through model selection, we find that the posterior mean of heat wave mortality risk is robust to accounting for model uncertainty over a broad class of models.     

The hydrogeomorphic approach to functional assessment of riparian wetlands: evaluating impacts and mitigation on river floodplains in the U.S.A.

1. The ’hydrogeomorphic‘ approach to functional assessment of wetlands (HGM) was developed as a synthetic mechanism for compensatory mitigation of wetlands lost or damaged by human activities. The HGM approach is based on: (a) classification of wetlands by geomorphic origin and hydrographic regime (b) assessment models that associate variables as indicators of function, and (c) comparison to reference wetlands that represent the range of conditions that may be expected in a particular region. In this paper, we apply HGM to riparian wetlands of alluvial rivers. 2. In the HGM classification, riverine wetlands are characterized by formative fluvial processes that occur mainly on flood plains. The dominant water sources are overbank flooding from the channel or subsurface hyporheic flows. Examples of riverine wetlands in the U.S.A. are: bottomland hardwood forests that typify the low gradient, fine texture substratum of the south-eastern coastal plain and the alluvial flood plains that typify the high gradient, coarse texture substratum of western montane rivers. 3. Assessment (logic) models for each of fourteen alluvial wetland functions are described. Each model is a composite of two to seven wetland variables that are independently scored in relation to a reference data set developed for alluvial rivers in the western U.S.A. Scores are summarized by a ’functional capacity index‘ (FCI), which is multiplied by the area of the project site to produce a dimensionless ’functional capacity unit‘ (FCU). When HGM is properly used, compensatory mitigation is based on the FCUs lost that must be returned to the riverine landscape under statutory authority. 4. The HGM approach also provides a framework for long-term monitoring of mitigation success or failure and, if failing, a focus on topical remediation. 5. We conclude that HGM is a robust and easy method for protecting riparian wetlands, which are critically important components of alluvial river landscapes.     

Using diatoms to assess the biological condition of large rivers in Idaho (U.S.A.)

SUMMARY 1. Diatoms were collected from 49 sites located on 23 rivers and identified to species. We defined 26 attributes of the diatom assemblage on the basis of tolerance and intolerance, autecological guild, community structure, morphological guild and individual condition. 2. We grouped sites into three geographical regions on the basis of altitude, land cover and use. Within each region, we tested for an association between diatom attributes and human disturbance. 3. We selected nine attributes for inclusion in a multimetric index, the river diatom index (RDI): percentage of valves belonging to species sensitive to disturbance, percentage of valves belonging to species very tolerant of disturbance, eutrophic species richness, percentage of valves belonging to nitrogen heterotrophic species, percentage of valves belonging to polysaprobic species, alkaliphilic species richness, percentage of valves belonging to species that require high oxygen, percentage of valves belonging to very motile genera and percentage of deformed valves. 4. The RDI was significantly correlated with measures of human disturbance made at the site (conductivity, percentage of fine sediments and number of human activities) and at the catchment level (percentage of urbanisation and agriculture in the upstream catchment). 5. The percentage of the total variance of RDI related to transect location was very small (1%) compared with among‐site differences (73%) and time of sampling (26%). The RDI could reliably detect approximately three categories of biological condition based on annual monitoring and potentially more if sampling were restricted to the same month each year.     

Longleaf pine cone–radial growth relationships in the southeastern U.S.A.

Longleaf pine (Pinus palustris Mill.) cones have been counted annually by the United States Forest Service (USFS) at eleven locations throughout the species’ range since 1958. These data have been useful for understanding spatiotemporal patterns in longleaf pine cone production, and are beneficial in timing regeneration efforts. Variations in annual mast (i.e. seed crop) are known to influence ring widths in numerous tree species, yet this relationship is poorly understood for longleaf pine. This research examines the relationship between longleaf pine cone data and tree-ring growth from trees sampled in the multi-decadal USFS cone-crop study. We examined cone–radial growth relationships using individual tree-ring data and proprietary cone data for each tree from six sites in four locations in the southeastern USA. We found that longleaf pine cones were correlated with basal area increment growth (BAI) over the three-year cone-development cycle. Low BAI years were more frequently associated with above-average cone crop and BAI during years that coincided with the largest cone-crop class (bumper, > 100 cones per tree) were statistically less than any other cone class. We prepared linear models that predicted radial growth using PDSI and cones as predictors, and found that including cones in the models did not improve adjusted R² values. We conclude that while cone production is inversely related to radial growth, the combination of infrequent bumper years and the concentration of cone production by a few trees per stand, creates an environment where radial-growth chronologies assembled from longleaf pine for dendroclimatic purposes are unlikely to be significantly influenced by reproductive strain.     

Colonial to modern skeletal change in the U.S.A.

Eighty-two people dating from 1675 to 1879 compared with 182 modern middle-class White and Black skeletons test the myths of radical changes produced by improved diet, less disease, and nineteenth century immigration. Longevity increases and health and growth improvement is clearest in reduced juvenile deaths (census data) and deepening of true pelvis. Stature increase is minimal (though seventeenth century Londoners and modern West Africans are shorter than Colonial to Modern Americans); teeth deteriorate and for cultural reasons fractures increase. Clavicles and forearms elongate. From Old to New World Colonial samples there is a noticeable skull change (and a greater Old World to Modern contrast) but White Colonial to Modern shows strong continuity surprisingly, the key changes being increasing head height, and retraction of face with increasing nose projection, and longer mastoids, resulting from selection and mixture. Blacks change more, possibly from Indian and White mixture. Variabilities are above average. Change is much less than expected, and apparently involves heterosis and selection as well as the obvious health advance and mixtures.     

Using Fault Tree Analysis to Assess Bioterrorist Risks to the U.S. Food Supply

Concerns about terrorist use of biological agents have increased at state and federal levels and in the popular press. Vulnerability of U.S. agricultural assets to bioterrorism has received less attention. A bioterrorist could use the U.S. food supply system to deliver a bioagent to a human target or use a bioagent to destroy it's inputs or infrastructure. Establishing effective agriculture security programs requires condensing generalizations about such bioterrorist threats into specific attack scenarios. This paper suggests a two-stage process to achieve this: problem formulation to define and rank scenarios, then quantitative risk assessment of selected, more highly ranked, scenarios. The quantitative component is illustrated for two specific scenarios: the actual 1984 Rajneeshee incident in Oregon and a hypothetical release of oral anthrax into the food supply system. The fault tree analysis of these scenarios suggests that the food supply system may be a less than optimal bioagent delivery mechanism, particularly if the objective is a mass casualty attack, as a number of events must occur simultaneously with reasonable probability if a bioagent is likely to be deployed effectively. In addition, an environmentally stable bioagent would be required if only the food system were used to deliver it to human consumers, but early detection of an intentional release common bioagent may be difficult absent specific intelligence or “unusual” epidemiological circumstances.     

A genome-wide association study of malting quality across eight U.S. barley breeding programs

Key message

We report malt quality QTLs relevant to breeding with greater precision than previous mapping studies. The distribution of favorable alleles suggests strategies for marker-assisted breeding and germplasm exchange.

Abstract

This study leverages the breeding data of 1,862 barley breeding lines evaluated in 97 field trials for genome-wide association study of malting quality traits in barley. The mapping panel consisted of six-row and two-row advanced breeding lines from eight breeding populations established at six public breeding programs across the United States. A total of 4,976 grain samples were subjected to micro-malting analysis and mapping of nine quality traits was conducted with 3,072 SNP markers distributed throughout the genome. Association mapping was performed for individual breeding populations and for combined six-row and two-row populations. Only 16 % of the QTL we report here had been detected in prior bi-parental mapping studies. Comparison of the analyses of the combined two-row and six-row panels identified only two QTL regions that were common to both. In total, 108 and 107 significant marker-trait associations were identified in all six-row and all two-row breeding programs, respectively. A total of 102 and 65 marker-trait associations were specific to individual six-row and two-row breeding programs, respectively indicating that most marker-trait associations were breeding population specific. Combining datasets from different breeding program resulted in both the loss of some QTL that were apparent in the analyses of individual programs and the discovery of new QTL not identified in individual programs. This suggests that simply increasing sample size by pooling samples with different breeding history does not necessarily increase the power to detect associations. The genetic architecture of malting quality and the distribution of favorable alleles suggest strategies for marker-assisted selection and germplasm exchange.

     

Expansions and contractions of the genetic map relative to the physical map of yeast chromosome III.

To examine the relationship between genetic and physical chromosome maps, we constructed a diploid strain of the yeast Saccharomyces cerevisiae heterozygous for 12 restriction site mutations within a 23-kilobase (5-centimorgan) interval of chromosome III. Crossovers were not uniformly distributed along the chromosome, one interval containing significantly more and one interval significantly fewer crossovers than expected. One-third of these crossovers occurred within 6 kilobases of the centromere. Approximately half of the exchanges were associated with gene conversion events. The minimum length of gene conversion tracts varied from 4 base pairs to more than 12 kilobases, and these tracts were nonuniformly distributed along the chromosome. We conclude that the chromosomal sequence or structure has a dramatic effect on meiotic recombination.     

Extrapolation of the relative risk of radiogenic neoplasms across mouse strains and to man

We have examined two interrelated questions: is the susceptibility for radiogenic cancer related to the natural incidence, and are the responses of cancer induction by radiation described better by an absolute or a relative risk model. Also, we have examined whether it is possible to extrapolate relative risk estimates across species, from mice to humans. The answers to these questions were obtained from determinations of risk estimates for nine neoplasms in female and male C3Hf/Bd and C57BL/6 Bd mice and from data obtained from previous experiments with female BALB/c Bd and RFM mice. The mice were exposed to 137Cs gamma rays at 0.4 Gy/min to doses of 0, 0.5, 1.0, or 2.0 Gy. When tumors that were considered the cause of death were examined, both the control and induced mortality rates for the various tumors varied considerably among sexes and strains. The results suggest that in general susceptibility is determined by the control incidence. The relative risk model was significantly superior in five of the tumor types: lung, breast, liver, ovary, and adrenal. Both models appeared to fit myeloid leukemia and Harderian gland tumors, and neither provided good fits for thymic lymphoma and reticulum cell sarcoma. When risk estimates of radiation-induced tumors in humans and mice were compared, it was found that the relative risk estimates for lung, breast, and leukemia were not significantly different between humans and mice. In the case of liver tumors, mice had a higher risk than humans. These results indicate that the relative risk model is the appropriate approach for risk estimation for a number of tumors. The apparent concordance of relative risk estimates between humans and mice for the small number of cancers examined encourages us to undertake further studies.     

Litterfall in a secondary forest with special reference to the relationships between leaf-fall rate,basal area and relative growth rate on a species basis

Total basal area (BA), mean relative growth rate of individual basal area (RGR _bam), annual leaf-fall rate (L), and number of trees (N) on a species basis were measured in a secondary forest mixed with evergreen and deciduous trees in the warm-temperate zone of Japan. Every species had a characteristic tree size and composed the tree and the shrub layers. The layer for each species, and the tree size of the species were represented by the species mean basal area per tree (BA/N orba _m). A power form equation was obtained for the relation between the species mean leaf-fall rate per tree (L/N orl _m) andba _m.RGR _bam was correlated with leaf-fall rate per unit basal area (L/BA) rather than withba _m. This suggested thatRGR _bam is associated less with mean tree size (ba _m) and dominance as total basal area (BA), and more with annual leaf production per unit basal area, provided thatL is equal to the annual leaf production per species.Rhus succedanea, the most dominant species in the forest plot, was typical in that it showed lower values ofRGR _bam andL/BA than other tree species. This would suggest a decline ofR. succedanea in a secondary forest.