Automatic transversality and orbifolds of punctured holomorphic curves in dimension four (Q2269705)

It is known since Gromov's seminal paper that moduli of pseudoholomorphic curves in dimension four enjoy higher level of regularity. In particular, they can be smooth for non-generic data and around multiply covered curves. This paper is a major development in the field that concerns punctured curves with Morse-Bott asymptotic orbits in four-dimensional symplectic cobordisms. It extends previous results of Hofer-Lizan-Sikorav and Ivashkovich-Shevchishin among others, and is part of the author's ongoing program to study the regularity of curves in foliations. In contrast to Ivashkovich-Shevchishin, who dealt with non-immersed curves by replacing normal bundles with normal sheaves, the author's approach is more analytic relying on Banach space splittings and Fredholm theory. Due to the paper's size applications are not developed here, but two follow-up papers apply the main results to symplectic fillings and non-separating hypersurfaces of contact type in four dimensions. As is commonly known, when the Cauchy-Riemann operator linearized at a pseudoholomorphic curve \(u\) is surjective, a local neighborhood of \(u\) in the moduli can be modeled on its kernel. Such \(u\) are called regular and the regularity is equivalent to the operator, interpreted as a section of a certain bundle, being transversal to the zero section. Transversality results usually require \(u\) to be generic, and, if it is, also somewhere injective the moduli around \(u\) is shown to be smooth. But Hofer-Lizan-Sikorav proved that for closed embedded curves in dimension four the transversality is automatic as long as the virtual dimension of the moduli (the index of the Cauchy-Riemann operator) is sufficiently large. The author studies the moduli \({\mathcal M}^c\) of punctured curves in a symplectic cobordism \(W\) with cylindrical ends, where \(c\) represents a choice of Morse-Bott asymptotic orbit for each puncture. Given any curve \(u\in{\mathcal M}^c\), he defines the normal first Chern number \(C_N(u)\), which reduces to the first Chern number of the normal subbundle of \(u^*TW\) for immersed \(u\); and a zero count \(Z(du)\), which counts critical points of \(u\) entering halves for the boundary points. It turns out that, if \(\dim_{\text{vir}}({\mathcal M}^c)>C_N(u)+Z(du)\), then \(u\) is automatically regular. It follows that \({\mathcal M}^c\) is a smooth orbifold around \(u\) with isotropy group equal to its group of automorphisms even if \(u\) is a multiple cover. This result is however local while the author is interested in the entire connected component \({\mathcal M}^c_u\) of \(u\). To this end, he introduces a notion of nicely embedded curves that formalizes properties of somewhere injective curves that do not intersect their neighbors in one or two dimensional families. If \(u\) is a nicely embedded curve and the almost complex structure is generic, he proves that the entire component \({\mathcal M}^c_u\) consists of regular curves with identical or disjoint images making \({\mathcal M}^c_u\) a smooth orbifold of the expected dimension. Moreover, all curves in \({\mathcal M}^c_u\) are embedded except for a discreet subset of unbranched multiple covers. The theorem is stronger for symplectizations \(\mathbb R\times M\) of contact 3-manifolds \(M\) with stable Hamiltonian structures. In them, multiple covers do not occur at all and \({\mathcal M}^c_u\) is a smooth manifold. It is this theorem that leads to strong new results for contact 3-manifolds with open book decompositions that can be lifted to pseudoholomorphic foliations in their symplectizations.