Rigidity of the harmonic map heat flow from the sphere to compact Kähler manifolds (Q2380815)

The main purpose of this paper is to derive a Lojasiewicz-type estimate for maps \(\varphi\) from \(\mathbb{S}^2\) to a compact Kähler manifold \(N\) with nonnegative holomorphic bisectional curvature generalizing the one obtained by \textit{P. M. Topping} when \(N=\mathbb{S}^2\) [J. Differ. Geom. 45, No. 3, 593--610 (1997; Zbl 0955.58013)]. While Topping used a similar estimate due to Leon Simon, the authors use the following Sobolev inequality, generalized to compact Riemann surfaces \(\Sigma\): There is a constant \(C\) depending only on \((\Sigma, g)\) such that \[ \int_{\Sigma} f^4\,dv_g \leq C\int_{\Sigma}f^2\,dv_g\int_{\Sigma}(f^2+ |\nabla f|^2)\,dv_g \] holds for any smooth function \(f\) on \(\Sigma\). Then, they derive a key Lojasiewicz-type estimate for the \(\partial\)-energy (resp., \(\overline{\partial}\)-energy) in terms of the tension of \(\varphi\) for maps with sufficiently small \(\partial\)-energy (resp., \(\overline{\partial}\)-energy). To derive such an estimate, the authors also use the harmonic map heat flow \(\frac{\partial}{\partial t}\varphi_t=\tau(\varphi_t)\), with initial condition \(\varphi_0=\varphi\) (and the fact that \(E_{\partial}(v)-E_{\overline\partial}(v)\) is homotopy invariant), and the assumption on the bisectional curvature of \(N\). The authors give some applications of these estimates. Embedding \(N\) into an Euclidean space, the harmonic flow \(\varphi_t\) corresponds to the evolution equation \(\frac{\partial}{\partial t}v=\Delta_{\mathbb{S}^2}v+ A(v)(\nabla v, \nabla v)\) where \(A\) is the second fundamental form of \(N\). From a well-known result of Struwe, a global weak solution \(v\) exists that is regular away from finitely many points \(\widehat{x}^k\), \(k=1,\dots, m\) (the blow-up points). Other authors also describe how bubbles may occur in the flow. There is a sequence of times going to infinity such that the flow converges weakly to the body map \(\widehat{\omega}_0\) in \(W^{1,2}(\mathbb{S}^2;\mathbb{R}^K)\) and strongly in \(W^{1,2}_{\text{loc}}(\mathbb{S}^2\backslash \{\widehat{x}^1, \dots,\widehat{x}^m\};\mathbb{R}^K)\), and the energy of the flow converges to the sum of the energies of the body map \(\widehat{\omega}_0\) and of the bubble maps \(\widehat{\omega}_i\), \(i=1,2\dots\), that are nonconstant harmonic maps. The key estimate leads then to the uniformity properties of the harmonic map flow, as Topping pointed out in case \(N\) is \(\mathbb{S}^2\): Supposing that at a infinite time, the bubble maps and the body map are all holomorphic or anti-holomorphic, then all the flow \(v(\cdot,t)\) converges to \(\widehat{\omega}^0\) weakly in \(W^{1,2}(\mathbb{S}^2;\mathbb{R}^K)\), strongly in \(L^p(\mathbb{S}^2;\mathbb{R}^K)\), for any \(p\geq 1\), and uniformly in \(C^l_{\text{loc}}(\mathbb{S}^2\backslash \{\widehat{x}^1, \dots, \widehat{x}^m\})\) for any \(l\), and for any \(r>0\) and \(k\), the total energy of \(v(\cdot,t)\) on the geodesic ball \(B_r(\widehat{x}^k)\) converges to a limit \(F_{k,r}\), uniformly. Furthermore, if there is another solution \(w\) of the flow, with no bubble maps at finite time, and \(v\) and \(w\) start close in \(W^{1,2}(\mathbb{S}^2)\), then they will stay close in \(L^2(\mathbb{S}^2)\) for all time, and close in \(W^{1,2}(\Omega)\) on compact domains \(\Omega\) away from the blow-up points.