Celebrating the 25th International Conference of The International Environmetrics Society

Abstract

This special double issue of the journal Environmetrics is devoted to celebrating the 25th international conference of The International Environmetrics Society (TIES). The papers included in the issue represent a selection of invited papers from the 25th Annual TIES meeting that was held in 2015 (November 22–25) at the United Arab Emirates University (UAEU). The decision to create TIES was made during the first Environmetrics conference, which was held in Cairo, Egypt, in April 1989. At this time, Environmetrics was viewed as a nascent scientific discipline focused on the development and use of quantitative methods to understand and solve environmental problems. The word Environmetrics was first introduced in 1971 by Philip Cox in a proposal submitted to the U.S. National Science Foundation. Cox used the word again at the 1972 meeting of the International Biometrics Society (Eastern North American Region: ENAR), to describe a session on statistics in environmental research. The word was independently reborn in discussions within the Committee of National Statistics of the U.S. National Academy of Sciences in 1976, and in testimony in 1977 before the U.S. House of Representatives Subcommittee on the Environment and Atmosphere (El-Shaarawi & Hunter, 2012). Like so many of the big challenges, environmental problems demand interdisciplinary collaboration. To address issues such as climate change, decline of biodiversity, decline of fisheries, and destruction of tropical forests, it is important to build, sharpen, and demonstrate quantitative tools useful across multiple fields of human knowledge. This interdisciplinary approach using diverse examples helps separate the tool (the methods and general principles) from the particular applications and makes its transferability to other problems more evident. This interaction will benefit the progress of science and enhance the range of applications of the quantitative methods in question. As the evidence of human impact on the environment continues to mount, answers are required for the question, “What are the right solutions to our urgent environmental challenges?” The core mission of TIES is to build a means of collaboration between quantitative scientists and environmental stakeholders, thus helping to provide realistic and efficient solutions to manage environmental issues of concern. This is done through its regular and regional meetings, publications such as the journal Environmetrics, and the Encylopedia of Environmetrics. Below, I will briefly discuss the articles included in this issue. Sylvia Esterby surveyed the history of the first 25 TIES conferences, including their impact on the development of the Society and its official journal, Environmetrics. The article highlights the extent of international collaboration, with conferences having taken place on six continents. It also tracks the growth of the field, as the need for more sophisticated quantitative methods to solve global environmental challenges and greater interdisciplinary cooperation has grown. It is important to point out that the first Board of Directors of TIES and the first Editorial Board of Environmetrics were composed of members from several scientific areas, whose work focused on providing realistic and quantitative solutions to environmental problems. Two articles cover environmental health and disease mapping. The first, “Bivariate geostatistical modelling of the relationship between Loa loa prevalence and intensity of infection”, was the 2015 TIES Hunter Lecture that was delivered at the conference by Peter Diggle. Its methodology is designed to meet the challenge faced in the analysis of data generated by multinational programs to control loiasis (known as the African eye worm) in several countries across central and western Africa. The extension made to account for excess of zeros in two bivariate spatial processes will have wider applications because the use of current standard geostatistical methods for prevalence data is inappropriate. The second article “Bayesian inference in time-varying additive hazards models with applications to disease mapping” offers a flexible semiparametric additive hazards model with spatial frailties. The proposed model allows both the frailties and the regression coefficients to be time varying, thus relaxing the proportionality assumption of Cox's hazard model. The estimation framework is Bayesian, powered by carefully tailored posterior sampling strategies via Markov chain Monte Carlo techniques. The authors used a data set on prostate cancer survival from the U.S. state of Louisiana to illustrate the advantages of the proposed model and the estimation procedure. Ecological statistics are covered in the two papers “Riding Down the Bay: Space-Time Clustering of Ecological Trends” and “Applications of random search methods to foraging in ecological environments and other natural phenomena—A review”. The first paper here develops a new data-driven procedure for optimal selection of tuning parameters in dynamic clustering algorithms using the notion of stability probe. The procedure is called downhill riding (DR) because of the dynamics of the clustering stability probe. The procedure is applied to data generated by the Chesapeake Bay Program, which was initiated in 1983 by the federal and several state governments to clean the Bay. The second paper by Jandhyala and Fotopoulos focuses on the application of random search methods to study predator–prey movements in the environment. The research concentrates on movements modeled by Brownian and Levy processes and is related to biodiversity and sustainability. The paper by J. M. Monteroa et al., which studies the connection between environmental quality and housing prices in the city of Madrid, provides an example of environmental economics. It extends spatial linear hedonic models to incorporate spatial autocorrelation, spatial heterogeneity, and nonlinearity. The data set used is massive for it includes the prices and characteristics of a sample size of 10,512 homes selected from the housing population frame of Madrid. The authors conclude that the environment has a significant impact on housing prices; however, when the model includes a drift and/or areal variables, these components absorb a substantial part of the environmental impact. Spatiotemporal modeling was covered in three articles. The first “Testing for local structure in spatio-temporal point pattern data” was presented by Jorge Mateu et al. at the conference. The authors extend local indicators of spatial association to the spatiotemporal context. The extensions are then used to build local tests of clustering to analyze differences in local spatiotemporal structures. The results of a simulation study demonstrated the efficiencies of the proposed testing procedures, and an earthquake data set is used for illustration. The second paper by Yawen Guan, Murali Haran, and David Pollard, “Inferring Ice Thickness from a Glacier Dynamics Model and Multiple Surface Datasets”, discusses the future behavior of the West Antarctic Ice Sheet, which may have a major impact on future climate. It develops a hierarchical Bayesian model that integrates multiple ice sheet surface data sets with a glacier dynamics model. The results of this work infer important parameters describing the glacier dynamics, learning about ice sheet thickness and accounting for errors in the observations and the model. The third article by Yuan Yan and Marc G. Genton examines the effect of non-Gaussianity on Gaussian likelihood inference for the parameters of the Matérn covariance model. Monte Carlo simulations of spatial data from a Tukey g-and-h random field, with the Matérn covariance function, where g controls skewness and h controls tail heaviness, were conducted to evaluate the method. The simulation results show that maximum likelihood estimator under both the increasing and fixed domain asymptotics for spatial data is preferable to the nondistribution-based weighted least squares estimator for data from the Tukey g-and-h random field. The authors conclude that Gaussian kriging based on Matérn covariance estimates with data from the Tukey g-and-h random field provides an overall satisfactory performance. In “Fractional Gaussian noise: Prior specification and model comparison”, Sigrunn Holbek Srbyea and Havard Rue argue that a uniform prior is unreasonable for the Hurst exponent (H) and suggest instead the use of a penalized complexity prior. They conclude that the same prior can be used for the autocorrelation coefficient ϕ of AR(1) and thus allows us to use the Bayes factors to compare ϕ and H in a fractional Gaussian noise model, which is useful in climate regression models where inference for underlying linear or smooth trends depends heavily on the assumed noise model. So far, the history of the Society indicates that TIES is making major progress in advancing its core objectives. The growing academic interest in the field is further evidence of this fact. I wish to sincerely thank the authors for their excellent contributions to this issue and also to extend my gratitude to all the delegates of the TIES25 Conference. In particular, I would to thank the UAEU for hosting the conference and H.E. Sheikh Hamdan bin Mubarak al Nahyan, the Chancellor of the UAEU; Prof. Geralyn McClure Franklin, the Dean of the College of Business and Economics; Prof. Taoufik Zoubeidi, the Chair of the Department of Statistics; and my conference cochair Dr. Ali Gargoum for their support. I would also like to thank the local organizing committee for their dedication and commitment in fulfilling this major responsibility. Finally, I wish to thank Professor Walter Piegorsch for his support during the work on this issue.