Singular integral operators on tent spaces over spaces of homogeneous type: Fefferman-Stein box maximal functions and pointwise Carleson type estimates (Q1626638)

Tent spaces were introduced in [\textit{R. R. Coifman} et al., J. Funct. Anal. 62, 304--335 (1985; Zbl 0569.42016)] to provide a unified functional analytic framework in which to express ideas, and prove fundamental results in harmonic analysis, related to Hardy spaces, BMO spaces, and Carleson measures. They then turned out to be natural solution spaces for non-smooth elliptic boundary value problems, where the notion of well-posedness involves \(L^{p}\) boundedness of a nontangential maximal function or square function, i.e. a tent space norm estimate (see e.g. Definition 3.6 in [\textit{C. E. Kenig} and \textit{J. Pipher}, Invent. Math. 113, No. 3, 447--509 (1993; Zbl 0807.35030)]). More recently, appropriate tent spaces have also been used as solution spaces for parabolic Cauchy problems in [\textit{P. Auscher} et al., Ann. Sc. Norm. Super. Pisa, Cl. Sci. (5) 19, No. 2, 387--471 (2019; Zbl 1447.35168)], and even wave equations [\textit{A. Hassell} et al., ``Off singularity bounds and Hardy spaces for Fourier integral operators'', Trans. Amer. Math. Soc. (to appear)]. In these PDE applications, one relies on a Calderón-Zygmund extrapolation theory for singular integral integral operators with operator valued kernels to extend \(L^2\) estimates to tent spaces estimates. Such a theory was first systematically developed in [\textit{P. Auscher} et al., J. Evol. Equ. 12, No. 4, 741--765 (2012; Zbl 1279.47072)] for tent spaces over \(\mathbb{R}^{d}\). In the paper under review, the author extends the key result from [Auscher et al., loc. cit.] to doubling metric measure spaces \(X\): i.e. he proves that an operator of the form \[ Tf(t,.) = \int \limits _{0} ^{t} K(t,s)f(s,.)ds, \] such that \(\{K(t,s) \;;\; t>s\}\subset B(L^{2}(X))\) satisfies appropriate off-diagonal bounds, is bounded on the tent space \(T^{p,2}(X)\) for \(p\) above a critical exponent, as soon as it is \(L^{2}(X\times \mathbb{R}_{+})\) bounded. Such an extension has great potential applicability, to extend PDE results from full space to domains, for instance. The proof is based on the following nice idea. Most of the arguments in [Auscher et al., loc. cit.] extend verbatim to doubling metric measure spaces. However, one crucial step typically uses a sharp change of aperture formula for Lusin area functionals and/or an atomic decomposition. These steps can easily be reproduced in some, but not all, doubling metric measure spaces. The author avoids such an extension by controlling the Carleson functional of \(Tf\) by a version of the Fefferman-Stein box maximal function of \(f\). This gives the result because the Lusin area functional, traditionally used to define tent spaces, is controlled by the Carleson functional (a fact going back to [Coifman et al., loc. cit.]), and because the Fefferman-Stein box maximal function can be controlled by the Lusin area functional using a Carleson measure argument that goes back to [\textit{S. Chanillo} and \textit{R. L. Wheeden}, Proc. Am. Math. Soc. 88, 509--512 (1983; Zbl 0522.42016)]. The paper concludes with some interesting remarks about the challenges and approaches to extending all of [Auscher et al., loc. cit.] to doubling metric measure spaces.