Carlos Pires initiated the teaching on geosciences (meteorology) at the Faculty of Sciences, University of Lisbon in 1989 and the research at CGUL (former IDL) in the domain of atmospheric energetic. In the period 1992-96 he concluded the PhD on meteorology at the Université Pierre et Marie-Curie, during which he published works on data assimilation, nonlinear atmospheric dynamics, predictability and non-linear statistical climatology aiming to build seasonal forecast models.
Later, in 1998, he joined IDL as a senior research in the area of climate variability and long-range predictability. There he has co-participated in an ICAT project for the Electrical Portuguese Company, aiming the development of monthly statistical forecast models of the potential flow in the Portuguese hydro-electrical dams. He has participated in several projects financed by the Portuguese Scientific Foundation, being the PI of three of them, especially in the area of large-scale predictability and droughts. Since 2000, he has developed studies on nonlinear statistics, information theory and non-Gaussianity applied to data assimilation, statistical climatology and nonlinear dynamics, which has led several international collaborations (France, Belgium, Austria) and publications in peer-reviewed journals as main author and also some invited talks on international meetings (AGU, EGU).