Explicit Dehn filling and Heegaard splittings (Q376091)

In 1997, Moriah and Rubinstein showed that if a bound is imposed on the genus of the Heegaard surfaces and finitely many Dehn filling slopes are excluded, then every irreducible Heegaard surface in the filled manifold is isotopic to one of a finite collection of surfaces in the original manifold. Later, results of Rieck and Sedgwick suggested that if one excluded a finite number of Dehn filling slopes, and a finite number of destabilization lines in the Dehn surgery space, any bounded-genus Heegaard surface in a Dehn filling will be a Heegaard surface in the original manifold. However, there has not been an effective characterization of which slopes and destabilization lines must be excluded. In the paper under review, the authors show the following: { Theorem 1.1.} Let \(X\) be a cusped, orientable hyperbolic \(3\)-manifold. Choose disjoint horospherical neighborhoods \(C_1, \cdots, C_k\) about some subset of the cusps, and let \(s_i\) be a Dehn filling slope on each torus \(\partial C_i\). Let \(\Sigma\) be a Heegaard surface of genus \(g \geq1\) for the Dehn filled manifold \(M = X(s_1, \cdots, s_k)\). Then we have the following. 1) If the length \(l(s_i)\) satisfies \(l(s_i) > 2\pi(2g-1)\) for every \(i\), then each core curve \(\gamma_i\) for the filling solid torus is isotopic into \(\Sigma\) (although these cores may not be simultaneously isotopic into \(\Sigma\)). 2) If, in addition, the shortest longitude \(\lambda_i\) for each \(s_i\) satisfies \(l(\lambda_i) > 6(2g-3)\), the surface \(\Sigma\) can be isotoped into \(M\setminus(\gamma_1 \cup\cdots\cup \gamma_k)\cong X\), and forms a Heegaard surface for \(X\). In the theorem, part (1) gives an explicit list of possible bad slopes, and part (2) gives a list of possible destabilization lines. The results give an effective version of the theorems of Moriah-Rubinstein and Rieck-Sedgwick.