Pietro Colombo
Research title: Multi-fidelity data fusion for environmental data
Research interests
Research interests
My main research interest is forecasting modeling for environmental data with humidity, temperature, and wind data applications.
Projects include Gaussian processes, heteroscedasticity and statistical downscaling techniques.
My PhD is about Data Fusion, since in recent years there has been a rapid shift in the quantity and quality of data generated with a view to bettering our understanding of the natural environment. This upsurge in data availability has been driven in large part by advances in technologies such as drones, in situ sensors and satellite remote sensing instruments. However, although the vast quantities of data now available should present a richer picture of underlying processes in the environment, alongside it comes numerous challenges with regards to statistical modelling. One key challenge is the fusion of data which have originated from different sources and have been collected at mis-matched spatial and temporal resolutions. Furthermore, the complexity is compounded by the need to incorporate different layers of uncertainty that are associated with these data at the different scales (e.g. spatial, temporal, instrument, atmospheric correction, and retrieval). It is the aim of this PhD project to develop statistical methodology to address such challenges.