This paper deals with the relationship between the CO2 emissions and the global temperatures across the various pandemic episodes that have been taken place in the last 100 years. To carry out the analysis, first we conducted unit root tests finding evidence of nonstationary I(1) behavior, which means that a shift in time causes a change in the shape of distribution. However, due to the low statistical power of unit root tests, we also used a methodology based on long memory and fractional integration. Our results indicate that the emissions display very heterogeneous behavior in relation to the degree of persistence across pandemics. The temperatures are more homogeneous, finding values for the orders of integration of the series smaller than 1 in all cases, thus showing mean reverting behavior.