Quotients of groups of birational transformations of cubic del Pezzo fibrations (Q2209258)

The Cremona group \(\operatorname{Cr}_n(\mathbb{C})\) of rank \(n>1\) is the group of birational transformations of the \(n\)-dimensional complex projective space. It was a long-standing question whether \(\operatorname{Cr}_n(\mathbb{C})\) is simple group. Several years ago Serge Cantat and Stéphane Lamy made a breakthrough in [\textit{S. Cantat} et al., Acta Math. 210, No. 1, 31--94 (2013; Zbl 1278.14017)] by proving that \(\operatorname{Cr}_2(\mathbb{C})\) is not simple. Recently Jérémy Blanc, Stéphane Lamy and Susanna Zimmermann proved that \(\operatorname{Cr}_n(\mathbb{C})\) is not simple for any \(n>2\). The strategies for \(n=2\) and \(n>2\) are very different. The paper under review can be viewed as a complement to the above mentionned article to appear of Blanc-Lamy-Zimmermann and is completely based on the framework developped in that article. Roughly speaking the idea is to construct a homomorphism of groups from \(\operatorname{Cr}_n(\mathbb{C})\) to a free product of \(\mathbb{Z}/2\mathbb{Z}\), obtained by counting certain special Sarkisov links in the factorization of a birational map. In the original paper of Blanc-Lamy-Zimmermann the special Sarkisov links are chosen to be certain links between conic bundles while in this paper the authors work in dimension three and chose them to be certain links involving del Pezzo fibrations and classical Bertini involutions on del Pezzo surfaces. A very interesting by-product of the proof is that \(\operatorname{Cr}_3(\mathbb{C})\) is not generated by its algebraic subgroups.