In the midst of climate change, academic travels - one salient aspect of the carbon footprint of research activities - are at the center of a growing concern. Mitigation options often focus on two dimensions : (i) decreasing the frequency of attendance to conferences and (ii) modal shift in transport. Here, we analyze professional travel in academia from a unique database compiling about 100 000 travels from about 150 research labs across a large array of disciplines and localities in France to detail the structure, patterns and heterogeneity of national and international research travels for research purposes. We estimate the mitigation potential of a series of options encompassing but not limited to institutional options. We show that, if short distance traveling (typically below 1000 km) are largely dominant in number, their relative mitigation potential via modal shift is small (i.e., below 15%). On the other hand, long distance traveling, which is often associated with international collaborations or field work hold a much larger mitigation potential but question the very nature of research activities. We propose ambitious sobriety options to robustly decrease travel-induced GHG emissions in academia and discuss their acceptability in the context of the French public research system.