Model-driven development of covariances for spatiotemporal environmental health assessment.

Known conceptual and technical limitations of mainstream environmental health data analysis have directed research to new avenues. The goal is to deal more efficiently with the inherent uncertainty and composite space-time heterogeneity of key attributes, account for multi-sourced knowledge bases (health models, survey data, empirical relationships etc.), and generate more accurate predictions across space-time. Based on a versatile, knowledge synthesis methodological framework, we introduce new space-time covariance functions built by integrating epidemic propagation models and we apply them in the analysis of existing flu datasets. Within the knowledge synthesis framework, the Bayesian maximum entropy theory is our method of choice for the spatiotemporal prediction of the ratio of new infectives (RNI) for a case study of flu in France. The space-time analysis is based on observations during a period of 15 weeks in 1998-1999. We present general features of the proposed covariance functions, and use these functions to explore the composite space-time RNI dependency. We then implement the findings to generate sufficiently detailed and informative maps of the RNI patterns across space and time. The predicted distributions of RNI suggest substantive relationships in accordance with the typical physiographic and climatologic features of the country.

Multi-perspective analysis and spatiotemporal mapping of air pollution monitoring data.

Space-time data analysis and assimilation techniques in atmospheric sciences typically consider input from monitoring measurements. The input is often processed in a manner that acknowledges characteristics of the measurements (e.g., underlying patterns, fluctuation features) under conditions of uncertainty; it also leads to the derivation of secondary information that serves study-oriented goals, and provides input to space-time prediction techniques. We present a novel approach that blends a rigorous space-time prediction model (Bayesian maximum entropy, BME) with a cognitively informed visualization of high-dimensional data (spatialization). The combined BME and spatialization approach (BME-S) is used to study monthly averaged NO2 and mean annual SO4 measurements in California over the 15-year period 1988-2002. Using the original scattered measurements of these two pollutants BME generates spatiotemporal predictions on a regular grid across the state. Subsequently, the prediction network undergoes the spatialization transformation into a lower-dimensional geometric representation, aimed at revealing patterns and relationships that exist within the input data. The proposed BME-S provides a powerful spatiotemporal framework to study a variety of air pollution data sources.

Spatiotemporal infectious disease modeling: a BME-SIR approach.

This paper is concerned with the modeling of infectious disease spread in a composite space-time domain under conditions of uncertainty. We focus on stochastic modeling that accounts for basic mechanisms of disease distribution and multi-sourced in situ uncertainties. Starting from the general formulation of population migration dynamics and the specification of transmission and recovery rates, the model studies the functional formulation of the evolution of the fractions of susceptible-infected-recovered individuals. The suggested approach is capable of: a) modeling population dynamics within and across localities, b) integrating the disease representation (i.e. susceptible-infected-recovered individuals) with observation time series at different geographical locations and other sources of information (e.g. hard and soft data, empirical relationships, secondary information), and c) generating predictions of disease spread and associated parameters in real time, while considering model and observation uncertainties. Key aspects of the proposed approach are illustrated by means of simulations (i.e. synthetic studies), and a real-world application using hand-foot-mouth disease (HFMD) data from China.