A Dolbeault-Grothendieck lemma on complex spaces via Koppelman formulas (Q1930892)

The present paper is part of a series of papers in which M. Andersson and H. Samuelsson Kalm develop a theory of weighted Koppelman formulas (with applications) on singular analytic spaces (see also [\textit{M. Andersson} and \textit{H. Samuelsson Kalm}, ``Koppelman formulas and the \(\dbar\)-equation on an analytic space'', \url{arXiv:0801.0710}], [\textit{M. Andersson} and \textit{H. Samuelsson Kalm}, J. Funct. Anal. 261, No. 3, 777--802 (2011; Zbl 1227.32010)]). These formulas are used to treat many aspects of the \(\overline\partial\)-equation on such spaces, like e.g. local solvability and regularity, resolutions of the structure sheaf, extension of holomorphic functions and \(\overline\partial\)-closed forms, global solvability. The present paper focuses on a resolution of the structure sheaf by a \(\overline\partial\)-complex of fine sheaves. To explain that more precisely, let \(X\) be a complex space of pure dimension \(n\). The authors introduce fine sheaves \(\mathcal{A}^X_q\) of \((0,q)\)-currents containing the sheaf of smooth \((0,q)\)-forms \(\mathcal{E}^X_{0,q}\) on \(X\) such that {\parindent=6mm \begin{itemize}\item[(1)] \(\mathcal{A}^X :=\bigoplus_q \mathcal{A}^X_q\) is a module over \(\mathcal{E}^X_{0,*}\), \item[(2)] \(\mathcal{A}^X_q|_{X_{reg}} = \mathcal{E}^X_{0,q}|_{X_{reg}}\), \item[(3)] the complex \[ 0\rightarrow \mathcal{O}^X \hookrightarrow \mathcal{A}^X_0 \overset{\overline\partial}{\longrightarrow} \mathcal{A}^X_1 \overset{\overline\partial}{\longrightarrow} ... \overset{\overline\partial}{\longrightarrow} \mathcal{A}^X_n \rightarrow 0 \] is exact. \end{itemize}} The construction of the complex \(\mathcal{A}^X_*\) is based on intrinsic explicit semi-global Koppelman formulas (introduced by the authors in [loc. cit.]) which give a \(\overline\partial\)-homotopy formula for the \(\mathcal{A}^X\)-sheaves. Laxly formulated, the currents in \(\mathcal{A}^X_*\) appear iteratively as the values of the integral operators in the Koppelman formulas. The present paper contains a short but instructive summary of the construction of these formulas in [loc. cit.]. Such formulas were long searched for after the initial work of \textit{G. M. Henkin} and \textit{P. L. Polyakov} [C. R. Acad. Sci., Paris, Sér. I 308, No. 13, 405--409 (1989; Zbl 0673.32009)] and have some direct nice applications concerning regularity of the \(\overline\partial\)-equation at singularities and extension of holomorphic functions. The integral kernels appearing in the Koppelman formulas are based on the theory of integral representations developed by \textit{M. Andersson} [Math. Ann. 326, No. 1, 1--18 (2003; Zbl 1024.32005); Math. Z. 254, No. 2, 315--332 (2006; Zbl 1104.32002)]. The kernels are locally integrable on the regular part of the variety and are residue currents at the singularities. So, the integral operators map smooth froms to forms with some poles at the singularities of the variety. To get a hand on this, the authors introduce the module \(\mathcal{A}^X\) as the smallest module of sheaves that is closed under multiplication with smooth forms and the action of any of such integral operator appearing in the Koppelman formulas. The authors show that forms in \(\mathcal{A}^X\) are a natural substitute for smooth froms if we go from manifolds to singular spaces. These forms do not only provide a resolution of the structure sheaf, but come with the analoga of extension properties of smooth forms on manifolds and the analoga of solvability results for the \(\overline\partial\)-equation. The residue currents appearing in the Koppelman formulas at the singularities of the varieties can be described by the so-called structure form introduced by the authors. It is a key point that these structure forms are semi-meromorphic, a very nice property in view of the standard extension property SEP. If the variety is Cohen-Macaulay, the structure sheaf is just a holomorphic \((n,0)\)-form and this has nice applications. The understanding of the structure form -- which gives a realization of the Grothendieck dualizing module (see Proposition 3.4) -- seems to be a key element in the description of the sheaf \(\mathcal{A}^X\) (which is not very explicit by now) and the understanding of the behavior of the \(\overline\partial\)-equation at singularities. It is a very interesting analytic object decoding algebraic properties of the variety. Here, Andersson and Samuelsson Kalm have introduced a very interesting concept which deserves further investigation.