Statistical Issues in the Creation of Environmental Indicators
Daniela Cocchi
Department of Statistics, University of Bologna, Bologna, Italy

Environmental indicators are currently requested by citizens and politicians: they ought to be constructed for permitting comparisons in time and space. General purpose environmental indicators are difficult to build and motivate but, as regards the linkage between anthropization and environment, statistics is particularly useful since it constitutes the most effective scientific tool for linking social, economical and health issues. Relatively large sets of indicators serving a multiplicity of purposes are very advisable.

The search for environmental indicators is motivated by the idea of monitoring and controlling the environmental status. Human activities concur in modifying the environment when many resources are not renewable while some changes are permanent and not to be considered as random fluctuations.

The help of environmental disciplines is fundamental for defining the basic statistical properties of the proposed syntheses. Indicators aiming at describing pollution in the three main media, soil, water and air, are perhaps the most popular. In particular, air pollution indicators, motivated by the popularity of some environmental laws, have been developed with more details. Biodiversity indicators constitute another important subset of indicators.

At present, the redundancy of data in some environmental field, like air monitoring sites for traditional pollutants, can be remarked. Statistical sampling is, on the contrary, the necessary premise to good quality data collection, with special attention to the issues of optimization and to the research for satisfying results with limited information.

The main statistical issues in the field of environmental indicators cover the questions of measurement, data collection, use of probabilistic models for enhancing the portability of local findings to more general contexts. These characteristics together contribute to see variability and uncertainty as a richness rather than as a default.

Statistical properties of indicators follow from the point just mentioned, and provide the guidelines for inter-subjective understandable data syntheses.

Keywords: Air pollution indicators; Biodiversity indicators; Environmental statistical sampling; Quality of environmental data collection

Biography: Full Professor of Statistics, University of Bologna.

PhD in Statistics at the Université Catholique de Louvain.

TIES president 2009-2011. President of the Italian Statistical Society (SIS) 2004-2008.

Associate Editor of “Metron” and “Environmetrics”.

Research fields: methods for finite population sampling and statistical models for environmental problems. Main results in environmental statistics develop hierarchical models and methods for air quality indicators. Main results in finite populations concern Bayesian solutions for hierarchical superpopulation models, small area inference and design based population inference for spatially organized data.