A Klein TQFT: the local real Gromov-Witten theory of curves (Q2232721)

This paper takes a step toward proving a Real version of the Gupakumar-Vafa conjecture. The latter predicts that the rational-valued (real) Gromov- Witten invariants of a (Real) Calabi-Yau threefold can be expressed, through an explicit function, in terms of certain integer-valued invariants. This explicit function is calculated from local Gromov-Witten invariants of an open Calabi-Yau manifold, which is the total space of a rank two complex vector bundle on a Riemann surface. The conjecture was recently proved in the classical case by \textit{E.-N. Ionel} and \textit{T. Parker} [Ann. Math. (2) 187, No. 1, 1--64 (2018; Zbl 1459.53078)]. On the other hand, (also recently) real GW invariants of Calabi-Yau threefolds were constructed by \textit{P. Georgieva} and \textit{A. Zinger} [Ann. Math. (2) 188, No. 3, 685--752 (2018; Zbl 1432.53128)] in arbitrary genus. The authors aim to combine these results and prove a real version of the Gupakumar-Vafa conjecture, which was explained more than a decade ago by \textit{J. Walcher} [Commun. Number Theory Phys. 3, No. 1, 111--172 (2009; Zbl 1168.81023)]. This paper proves a formula for the local Real Gromov-Witten invariants that realizes \textit{J. Walcher}'s prediction [Commun. Number Theory Phys. 3, No. 1, 111--172 (2009; Zbl 1168.81023), (5.41)]. The authors package these local invariants into a TQFT formalism to streamline their calculations, extending earlier results of \textit{J. Bryan} and \textit{R. Pandharipande} [J. Am. Math. Soc. 21, No. 1, 101--136 (2008; Zbl 1126.14062)]. They also consider relative invariants and use a degeneration technique [\textit{P. Georgieva} and \textit{E. Ionel}, ``Splitting formulas for the local real Gromov-Witten invariants'', Preprint, \url{arXiv:2005.05928}] to decompose a (higher genus) Riemann surface into a union of lower genus ones. One of the main issues in working with real moduli spaces is the orientation problem. The authors indicate that (they have shown that) there is a canonical orientation for the local moduli spaces. This leads to local invariants that only depend on the degree and Euler characteristic.