StEPhI (Statistics for Environmental Phenomena and their Interactions) is a 3-year research project funded by the Italian Ministry of Education, University and Research within the “Futuro in Ricerca” (Future in Research) program.
The project involves 4 main research units from the University of Bergamo, Bologna, Pescara and Torino and a total of 18 researchers.
As a consequence of the increasing cultural, social, political and scientific attention paid to environmental problems, there has been a growing availability of data concerning hazardous phenomena for the environment and human health. Environmental phenomena are usually complex and influenced by many factors that interact with each others and that are linked to climate and anthropogenic pressure. Environmental data, with spatial and/or temporal dimensions, come from heterogeneous sources: monitoring networks (air, water), remote sensing (land use, atmosphere), numerical models (meteorological data, emission inventory), geographic information systems and it includes economic, demographic and health care data as well. Hence, statistical research should be developed considering two fundamental features. Firstly, it needs to take into account the complexity of environmental phenomena observed in time and space and characterized by several variability sources. Secondly, in order to take advantage of all the available information, new methods have to be developed for combining data coming from multiple sources and with different spatial and/or temporal resolutions. The advantage of the statistical approach consists in taking into account possible measurement errors and in providing a probabilistic evaluation of the estimate and prediction uncertainty. The StEPhI project evolves in three specific applicative topics: POLLUTION, METEOROLOGY and HEALTH with attention to their interactions. The main goal shared by the project applicative topics is the development of statistical methods for the study – with explanatory and predictive purposes – of environmental complex phenomena with spatial and temporal components. Such goal can be pursued by means of hierarchical spatio-temporal models which allow managing complex processes through their decomposition in simpler components while simultaneously considering all the uncertainty sources.
In the framework of this general goal, five methodological thematics to be developed are proposed in different applicative environmental topics taking into account specific characteristics of the phenomena and of the observed data.
The project is based on the following working packages
- WP1 – Assimilation and calibration of satellite and numerical model data;
- WP2 – Prediction with change of support;
- WP3 – Geostatistical models for functional data;
- WP4 – Markovian models for areal data;
- WP5 – Probabilistic ensemble forecasting.