CMStatistics 2016: Start Registration
View Submission - CFE
Title: Spatial variability of the relationship between carbon dioxide and temperature Authors:  Francesca Di Iorio - University of Naples Federico II (Italy)
Umberto Triacca - University of L Aquila (Italy) [presenting]
Abstract: Several recent studies have established that there is a clear relationship between the amount of carbon dioxide (CO2) in the atmosphere and global temperature. We propose a new statistical approach to analyze the spatial variability of this relationship based on a distance measure between pairs of vector autoregressive (VAR) models. In particular, we estimate a bivariate VAR model (in the variables CO2 and temperature) for eight latitude bands and we compute the proposed distance between any pair of models obtaining a matrix of distances. Our analysis by multidimensional scaling of this matrix shows that, despite CO2 being a globally well-mixed gas, the `fingerprints' of CO2 forcing are regionally very different. In the latitude belts surrounding the equator (0 - 24S and 0 - 24N) the estimated VAR models are very close. Another cluster is given by the VAR models concerning the lower mid-latitudes of both hemispheres (24S - 44S and 24N -44N) and the 44S - 64S latitude belt. However, the main result of our analysis is that the VAR models for the polar regions seem to be `isolated points'. In particular, it is interesting to underline that the largest distance is that between the VAR models for Arctic and Antarctic. There is a clear different response of the Arctic and Antarctic to anthropogenic forcing. Also the phenomenon referred to as Arctic warming amplification seems to be confirmed by our analysis.