The Homology of Moduli Spaces on a Riemann Surface as a Representation of the Mapping Class Group (Q4487585)

Let \(\Sigma\) be a compact Riemann surface of genus \(g\geq 2\), \(p\) a fixed point of \(\Sigma\) and \(M_{g,k}\) the moduli space of semistable rank 2 vector bundles on \(\Sigma\) with fixed determinant \({\mathcal O}_\Sigma(kp)\). It is a compact complex variety of real dimension \(6g-6\). In this paper, the author gives an explicit decomposition of the rational (intersection) homology of \(M_{g,k}\) into irreducible representations of the oriented mapping class group \(\text{Map}^+(\Sigma)\) of oriented diffeomorphisms of \(\Sigma\) (which acts on \(M_{g,k}\), too, by diffeomorphisms). In the non-singular case, \(k\) odd, the author uses a result of \textit{P. E. Newstead} [Trans. Am. Math. Soc. 169, 337-345 (1972; Zbl 0256.14008)] which gives a system of generators of the cohomology ring \(H^*(M_{g,1})\) and a lemma of \textit{M. Thaddeus} [J. Differ. Geom. 35, 131-149 (1992; Zbl 0772.53013)] which computes the cohomology class of \(M_{g-1,1}\) embedded in \(M_{g,1}\) such that \(H^*(M_{g-1,1})\) is a subring of \(H^*(M_{g,1})\). In the more complicated singular case, \(k\) even, the author considers the moduli space \(N_{g,0}\) of semistable rank 2 vector bundles on \(\Sigma\) of degree 0, its realization as a quotient by geometric invariant theory, and uses the ``cohomology of quotients'' method of \textit{F. C. Kirwan} [Ann. Math. (2) 122, 41-85 (1985; Zbl 0592.14011)]. It turns out, in both cases, that the intersection homology of \(M_{g,k}\) is a direct sum of (several) copies of the standard representations \(\bigwedge^{g-p}H_1(\Sigma,\mathbb{Q})\) of \(\text{Map}^+ (\Sigma)\). Their multiplicities are independent of the genus \(g\). As a corollary, the author observes that \(\text{IH}^*(M_{g,0};\mathbb{Q})\) is a \(\text{Map}^+ (\Sigma)\)-equivariant quotient of \(H^*(M_{g,1};\mathbb{Q})\). Using the fact that a 0-surgery on a fibred knot \(K\) in \(S^3\) results in a mapping torus \(\Sigma \times[0,1]\) with ends identified together by the monodromy map \(f:\Sigma \to \Sigma\), the author gives some applications to knot theory.