Uniqueness and nonuniqueness of limits of Teichmüller harmonic map flow (Q2156030)

Let \(M\) be a closed oriented surface, \(N\) a compact Riemannian manifold. The Teichmüller harmonic map flow is a sort of gradient flow for the Dirichlet energy \(E(u,g):=\int_M|du|^2_g\,dv_g\) viewed as a function of a mapping \(u:M\to N\) and a Riemannian metric \(g\) on \(M\). The idea is that the flow should converge to a minimal immersion. Weak solutions are known to blow up at a finite number of times, at most, where bubbling occurs, changing the topology of the domain surface. If the metric of the domain does not degenerate, there is a useful limit for the flow at infinite times. This is a known fact only for a sequence of times \(t_i\to\infty\), and only up to pullback by a sequence of diffeomorphisms. This paper addresses the question whether the limit can actually depend on the choice of the time sequence \((t_i)\), and whether the diffeomorphisms are really needed. It turns out that the answer is yes for both questions. There are examples exhibiting a sort of ``winding behaviour'' of the domain metric as \(t\to\infty\), resulting in limits depending on the sequence of times selected. Moreover, this can be arranged in such a way that the pullback by diffeomorphisms is needed to have a limit for any sequence. None of this is prompted by degeneration, since the examples can be constructed in such a way that the energy density is bounded uniformly from above, and the injectivity radius from below. The target \(N\) in these examples is not analytic. In contrast to this, the authors prove that no such nonuniqueness of limit configurations can appear if \(N\) is analytic. Both the nonuniqueness and the uniqueness results are constructed following a parallel with the case of the harmonic map heat flow, where similar theorems are known.