Geostatistica e Kriging: Teoria e applicazioni per uno Studio di Inquinamento dell'Aria in Italia

The Master Thesis work I have created is about air pollution, a threat for our health and our planet. The interest in air pollution analysis in the last decades has increased making it an active research area. Mostly of the research articles aims at health problems and medical issues. The vast majority of them are applied to diseases or viruses, mostly respiratory ones, and they investigate correlation, relation or peculiar pattern among the two issues. For example, plenty of articles were published about how much air pollution affects COVID 19 in the latter years. Statistical applications on air pollution lay in the Spatial Statistic field where the data have a spatial component, the location. The growing interest about air pollution, due to its health affection and the necessity of monitoring it by rules and directives is making easier the analysis of air pollution. On the same page, it is also down to the development of more efficient technology in measuring air pollution and to the fact that the cost of installing measuring stations is getting smaller and smaller every year. Every year, though, the number of measuring stations increases. Nowadays it is indeed common to find on the market private monitoring stations to install inside houses, or find them on cars detecting the pollution level around it. But despite this, the research area remains strongly linked to the medical one. Which is why, the data collected are mostly located on town centers or near them, where people live or work, spending time exposed to pollution. But here is the problem, if we aimed to understand how air pollution affects plants or ecosystem we would be more interested in data collected far from towns, called rural, rather than industrial or residential data, called urban or suburban. I have further developed this concept in this Thesis where, headed for environmental consequences I applied some basic spatial statistic theories on real database. The different nature of data is deeply investigated in my Thesis, both with the question: “Can we use all the data we have for this environmental logical direction?”. The amount of rural data, for example, is small. That is why I would like to use all the data, including the ones shortly related to environmental area. It is one of the problem we could face if we think about environmental affection of air pollution and it could be one of the causes of a less active research area compared to the respiratory disease one. Frequently, indeed, air pollution for rural area studies are not general, needing their own database to collect and they are developed for a local small region. Alongside the type of area matter, the driving force of the Thesis was the intent of performing Multivariate Analysis and Principal Component Analysis on different pollutants trying to synthesize the information reducing the number of variables used. This is why I kept working on more than one pollutant. The problem of this approach for Italian data was the size gap among databases of diverse pollutants. This means that we have measured values for a pollutant in some locations, and values of an other pollutants in different locations (it is possible of course that there are common ones, but not all of them are). Thus, a station does not provide each pollutants value so some of them need to be estimated. The solution I have chosen is the kriging method.