Cluster detection using Bayes factors from overparameterized cluster models

In this paper, we consider the use of a partition model to estimate regional disease rates and to detect spatial clusters. Formal inference regarding the number of partitions (or clusters) can be obtained with a reversible jump Markov chain Monte Carlo algorithm. As an alternative, we consider the ability of models with a fixed, but overly large, number of partitions to estimate regional disease rates and to provide informal inferences about the number and locations of clusters using local Bayes factors. We illustrate and compare these two approaches using data on leukemia incidence in upstate New York and data on breast cancer incidence in Wisconsin.     

Broad-band versus narrow-band irradiance for estimating latitude by archival tags

The relative effectiveness of different bands of irradiance to estimate the latitude of archival tags was evaluated. These tags are placed on fishes in order to describe their movements during long distance migrations. Measurements were recorded of broad-band irradiance with and without a cosine collector and narrow-band irradiance of seven narrow bands with 50% attenuation 30 nm on either side of their central wavelength of 400 (violet), 450 (blue), 500 (blue–green), 550 (green), 600 (yellow), 650 (orange), and 700 nm (red). A holographic, cosine collector was used to reduce the vertical transmission of irradiance to the sensor and to increase horizontal transmission of irradiance so the sensor detected more of the diffuse irradiance penetrating the water at dawn and dusk. Daily measurements were made during seven periods of 1–2 days each, beginning 28 June (after 21 June solstice) and ending on 6 October 1999 (after September 23 equinox) of submarine irradiance at 15-s intervals at a fixed depth (10 m) and location (38.31°N; 123.08°W) in Horseshoe Cove, California. Irradiance transmission at this site is intermediate between the clearest offshore waters, where blue irradiance (450 nm) penetrates farther with depth than green irradiance (550 nm) and most oceanic and coastal waters, where green penetrates farther than blue irradiance. Two algorithms were used to estimate latitude, the maximum slope method and the maximum logarithmic difference method. The broad-band, cosine-corrected light, excluding those deployments near the equinox when error is highest, produced an estimate of latitude of 38.30° for both methods and a latitudinal error of ±34.4 km for the former and ±27.2 km for the latter. The mean latitudinal error for non-cosine-collected, broad-band irradiance was ±190.9 km, using the slope algorithm and ±184.8 km using the difference algorithm. The blue band of irradiance, which attenuates least with increasing depth in clear, oceanic water, also produced a comparatively high-latitudinal error of ±163.8 km error for the former algorithm and ±170.4 km for the latter algorithm. Tag designers should consider using cosine-collectors over the irradiance sensors on their archival tags to increase the accuracy of position estimates.     

Scale-dependent variation in coral community similarity across sites, islands, and island groups

Community similarity is the proportion of species richness in a region that is shared on average among communities within that region. The slope of local richness (alpha diversity) regressed on regional richness (gamma diversity) can serve as an index of community similarity across regions with different regional richness. We examined community similarity in corals at three spatial scales (among transects at a site, sites on an island, and islands within an island group) across a 10 000-km longitudinal diversity gradient in the west-central Pacific Ocean. When alpha diversity was regressed on gamma diversity, the slopes, and thus community similarity, increased with scale (0.085, 0.261, and 0.407, respectively) because a greater proportion of gamma diversity was subsumed within alpha diversity as scale increased. Using standard randomization methods, we also examined how community similarity differed between observed and randomized assemblages and how this difference was affected by spatial separation of species within habitat types and specialization of species to three habitat types (reef flats, crests, and slopes). If spatial separation within habitat types and/or habitat specialization (i.e., underdispersion) occurs, fewer species are shared among assemblages than the random expectation. When the locations of individual coral colonies were randomized within and among habitat types, community similarity was 46-47% higher than that for observed assemblages at all three scales. We predicted that spatial separation of coral species within habitat types should increase with scale due to dispersal/extinction dynamics in this insular system, but that specialization of species to different habitat types should not change because habitat differences do not change with scale. However, neither habitat specialization nor spatial separation within habitat types differed among scales. At the two larger scales, each accounted for 22-24% of the difference in community similarity between observed and randomized assemblages. At the smallest scale (transect-site), neither spatial separation within habitat types nor habitat specialization had significant effects on community similarity, probably due to the small size of transect samples. The results suggest that coral species can disperse among islands in an island group as easily as they can among sites on an island over time scales that are relevant to their establishment and persistence on reefs.     

High male reproductive success in a low-density Antarctic fur seal (Arctocephalus gazella) breeding colony

Understanding how population density influences mating systems may lead to important insights into the plasticity of breeding behavior, but few natural systems allow for such studies. Antarctic fur seals (Arctocephalus gazella) provide an interesting model system because they breed in colonies of varying densities. Previous studies have largely focused on a high-density site at Bird Island, South Georgia. Here, 13 highly polymorphic microsatellite loci were used to conduct a genetic analysis of a low-density breeding colony of this species at Livingston Island, approximately 1,600 km south of South Georgia. The majority of adults seen ashore (n?=?54) were sampled together with every pup born (n?=?97) over four consecutive years. Paternities were confidently assigned for 34 out of the 97 pups. Two out of 23 sampled males accounted for the paternity of 28 % of all pups sampled during the study and 82 % of the pups with an assigned father. Moreover, a full likelihood pedigree inference method assigned a further eight paternities to a single unsampled male seal that is inferred to have held a territory during the season before the study began. The most successful males in our study easily surpassed the previous record for the total number of pups sired per male seal for the species. Furthermore, we identified two triads of full siblings implying that their parents remated in three consecutive years. These findings suggest that territorial male fur seals may achieve greater success in monopolizing access to breeding females when population density is relatively low.     

A Modeling System to Assess Land Cover Land Use Change Effects on SAV Habitat in the Mobile Bay Estuary

Estuarine ecosystems are largely influenced by watersheds directly connected to them. In the Mobile Bay, Alabama watersheds we examined the effect of land cover and land use (LCLU) changes on discharge rate, water properties, and submerged aquatic vegetation, including freshwater macrophytes and seagrasses, throughout the estuary. LCLU scenarios from 1948, 1992, 2001, and 2030 were used to influence watershed and hydrodynamic models and evaluate the impact of LCLU change on shallow aquatic ecosystems. Overall, our modeling results found that LCLU changes increased freshwater flows into Mobile Bay altering temperature, salinity, and total suspended sediments (TSS). Increased urban land uses coupled with decreased agricultural/pasture lands reduced TSS in the water column. However, increased urbanization or agricultural/pasture land coupled with decreased forest land resulted in higher TSS concentrations. Higher sediment loads were usually strongly correlated with higher TSS levels, except in areas where a large extent of wetlands retained sediment discharged during rainfall events. The modeling results indicated improved water clarity in the shallow aquatic regions of Mississippi Sound and degraded water clarity in the Wolf Bay estuary. This integrated modeling approach will provide new knowledge and tools for coastal resource managers to manage shallow aquatic habitats that provide critical ecosystem services.     

Variable Irrigation District Action in Water Trading

Irrigation districts (IDs) in the American west are highly diverse in their economic attributes and local water scarcity circumstances. This diversity may affect reallocative action via water transactions as scarcity rises. The institutional background defining and constraining IDs is described here. For a Texas study region the progress of permanent water right transfers involving IDs is documented and examined. An econometric analysis of multiple decades of ID water transfer activities in the Lower Rio Grande Valley finds that IDs with larger initial water right holdings and higher populations in nearby cities are more likely to participate in agricultural‐to‐municipal water transfer activities. The findings suggest that consolidation of smaller water right holding IDs may be an avenue for quickening the pace of reallocation, especially in more populated areas.     

Sensitivity analysis of the asymptotic behavior of a model for the environmental movement of radionuclides

The asymptotic behavior of a linear compartment model for the environmental movement of radionuclides is investigated. Here, the expression asymptotic behavior is used to designate the behavior of q(t) as t → ∞, where q is the solution of a vector differential equation of the form dq/dt = h + Kq. The asymptotic behavior of such equations is described. For the model and conditions under consideration, each element of q converges monotonically to a steady-state value. A hydrologic system is defined and used to illustrate this behavior. An approach to sensitivity analysis employing Latin hypercube sampling, rank transformations and stepwise regression is presented and then applied to this system. A total of 20 independent variables is introduced and the following dependent variables are investigated for the various components of the system: amount of radionuclide present at steady state. concentration of radionuclide at steady state, and time required to reach 90% of steady state. Finally, an application of asymptotic behavior in the analysis of a hypothetical site for the geologic isolation of high-level radioactive waste is described and a brief discussion of differential sensitivity analysis is given.     

Identifying fecal sources in a selected catchment reach using multiple source-tracking tools

Given known limitations of current microbial source-tracking (MST) tools, emphasis on small, simple study areas may enhance interpretations of fecal contamination sources in streams. In this study, three MST tools-Escherichia coli repetitive element polymerase chain reaction (rep-PCR), coliphage typing, and Bacteroidales 16S rDNA host-associated markers-were evaluated in a selected reach of Plum Creek in south-central Nebraska. Water-quality samples were collected from six sites. One reach was selected for MST evaluation based on observed patterns of E. coli contamination. Despite high E. coli concentrations, coliphages were detected only once among water samples, precluding their use as a MST tool in this setting. Rep-PCR classification of E. coli isolates from both water and sediment samples supported the hypothesis that cattle and wildlife were dominant sources of fecal contamination, with minor contributions by horses and humans. Conversely, neither ruminant nor human sources were detected by Bacteroidales markers in most water samples. In bed sediment, ruminant- and human-associated Bacteroidales markers were detected throughout the interval from 0 to 0.3 m, with detections independent of E. coli concentrations in the sediment. Although results by E. coli-based and Bacteroidales-based MST methods led to similar interpretations, detection of Bacteroidales markers in sediment more commonly than in water indicates that different tools to track fecal contamination (in this case, tools based on Bacteroidales DNA and E. coli isolates) may have varying relevance to the more specific goal of tracking the sources of E. coli in watersheds. This is the first report of simultaneous, toolbox approach application of a library-based and marker-based MST analyses to flowing surface water.     

TWO STRATEGIES FOR LEARNING A ROUTE IN A DRIVING SIMULATOR

The study of human navigation has long been dominated by the so-called stage theory, i.e. the notion that there are three distinct types of spatial knowledge (landmark, route, and survey knowledge), that are acquired sequentially during spatial learning and development. Based on the results of a route learning experiment in a driving simulator, an alternative to the stage theory is proposed. The authors suggest that subjects follow either a visually dominated or a spatially dominated strategy to solve a route-learning problem. In the visually dominated strategy, subjects base their wayfinding decisions on visually recognizing decision points along a route; the decision points are not integrated into any kind of survey representation. In the spatially dominated strategy, on the other hand, subjects represent the environment as a survey map right from the start; that is, they do not pass through a landmark or route stage. These strategies may be subserved by different cortical areas recently characterized in neurophysiological studies of animals solving maze problems.     

Morbidity and pollution: Model specification analysis for time-series data on hospital admissions

Time-series analysis of effects of pollutants on emergency hospital admissions indicates important synergistic interactions among pollutants and to a lesser degree nonlinearities in effects of single pollutants. Comparisons of alternative econometric specifications are made to determine the appropriateness of incorporating nonuniform pollution impacts. The data substantially support the existence of synergisms among pollutants with high levels of sulfur dioxide, SO₂ (particulates), increasing the impact of particulates (SO₂) on emergency hospital admissions. Marginal effects of either pollutant are, however, small at current ambient air quality levels. These results indicate that damage estimates were likely to be understated during the 1960's when pollution levels were high, while, at current levels of those pollutants considered here, marginal damages are lower than would be estimated in studies failing to incorporate synergistic and nonlinear impacts.