Weighted wave envelope estimates for the parabola

Jongchon Kim, Hyerim Ko · 2025

Read Weighted wave envelope estimates for the parabola by Jongchon Kim, Hyerim Ko (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-04503.txt.

Introduction

Jongchon Kim and his co-author extend a classic square function estimate to weighted settings. The Córdoba-Fefferman estimate concerns the parabola, a basic curve in Fourier analysis. Weights change the picture, since mass can pile up unevenly across scales.

Their weighted bound grows out of a wave envelope estimate for the parabola. Terms involve families of tubes at many scales plus parameters describing the weight's spread. As applications, they obtain weighted estimates for a class of Fourier multipliers. The same bounds cover solutions of the free Schrödinger equation, linking a geometric estimate to wave evolution.

Catalog record

CREATORJongchon Kim, Hyerim Ko
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-04503

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Weighted wave envelope estimates for the parabola Jongchon Kim, Hyerim Ko Abstract . In this paper, we extend Fefferman’s classical square function estimate for the parabola to a weighted setting. Our weighted square function estimate is derived from a weighted wave envelope estimate for the parabola. The bounds are formulated in terms of families of multiscale tubes together with weight parameters that quantify the distribution of the weight. As an application, we obtain some weighted L p L^{p} -estimates for a class of Fourier multiplier operators and for solutions to free Schrödinger equation. 2020 Mathematics Subject Classification 42B15, 42B25 1. Introduction The paper is concerned with weighted square function estimates for the parabola and some of its applications. Let \mathcal{P} denote the truncated parabola \mathcal{P}=\{(t,t^{2})\in\mathbb{R}^{2}:|t|\leq 1\} and N_{R^{-1}}\mathcal{P}=\{(t,t^{2}+\eta)\in\mathbb{R}^{2}:|t|\leq 1,\;|\eta|\leq R^{-1}\} denote its R^{-1} -neighborhood for a large R\geq 1 . We consider the canonical covering of N_{R^{-1}}\mathcal{P} by finitely overlapping parallelograms \theta of dimensions R^{-1/2}\times R^{-1} . Given a function f whose Fourier transform is supported on N_{R^{-1}}\mathcal{P} , we decompose f=\sum_{\theta}f_{\theta} , where \widehat{f}_{\theta} is supported on \theta . This can be done, for example, by using a smooth partition of unity subordinate to a covering of the interval [-1,1] by finitely overlapping intervals of length \sim R^{-1/2} (see e.g. the proof of Theorem 2.1 ). By Plancherel’s theorem, these functions \{f_{\theta}\} are orthogonal on L^{2}(\mathbb{R}^{2}) : \|f\|_{L^{2}}^{2}\leq C\sum_{\theta}\|f_{\theta}\|_{L^{2}}^{2} . Moreover, the family exhibits certain L^{p} orthogonality due to the curvature properties of the parabola for some p larger than 2 . For instance, the classical square function estimate for the parabola (see [ 9 ] ) states that (1) \|f\|_{L^{4}(\mathbb{R}^{2})}\leq C\big\|\big(\sum_{\theta}|f_{\theta}|^{2}\big)^{1/2}\big\|_{L^{4}(\mathbb{R}^{2})}. This inequality relies on the geometric observation by Fefferman [ 14 ] that the algebraic differences \theta-\theta^{\prime} overlap only finitely often as \theta\neq\theta^{\prime} vary. See also [ 17 , 30 , 20 ] for extensions to non-degenerate curves in higher dimensions. Square function estimates of the form ( 1 ) have several important applications in harmonic analysis. The sharp square function estimate ( 1 ) is known to imply sharp results for the Kakeya maximal function, the Bochner-Riesz multipliers, the Fourier restriction operator, and local smoothing estimates for the Schrödinger equation; see [ 6 , 41 ] and references therein. In higher dimensions, it is conjectured that ( 1 ) holds with L^{4}(\mathbb{R}^{2}) replaced by L^{\frac{2d}{d-1}}(\mathbb{R}^{d}) , which remains wide open. Let H:\mathbb{R}^{2}\to[0,\infty) be a bounded function on \mathbb{R}^{2} . The main goal of this paper is to establish weighted square function estimates of the form \|f\|_{L^{p}(Hdx)}\leq C_{p,H}(R)\big\|\big(\sum_{\theta}|f_{\theta}|^{2}\big)^{1/2}\big\|_{L^{p}(\mathbb{R}^{2})} for 2\leq p\leq 4 and to explore some of its consequences. To describe the constant C_{p,H}(R) , we need to introduce a family of tubes originating from a multiscale analysis. Let s be a dyadic number in the range R^{-1/2}\leq s\leq 1 . At each scale s , we cover N_{s^{2}}\mathcal{P} by canonical blocks \{\tau\} of dimension s\times s^{2} and use |\tau|=s to denote the scale. For the smallest scale s=R^{-1/2} , these blocks are just \{\theta\} . We fix a dyadic s\in[R^{-1/2},1] . For each \tau with |\tau|=s , we consider a linear transform L_{\tau} determined by the parabolic rescaling (see ( 38 )) for which L_{\tau}([-\frac{1}{2},\frac{1}{2}]^{2}) is a parallelepiped dual to \tau of dimensions s^{-1}\times s^{-2} and orthogonal to \tau . Let \mathbb{T}_{\tau} denote the tiling of \mathbb{R}^{2} by translates of the dual parallelepiped: (2) \mathbb{T}_{\tau}=\{L_{\tau}(z+q):z\in\mathbb{Z}^{2}\},\;\;q=[-1/2,1/2]^{2}. Next, we consider the tiling of \mathbb{R}^{2} by the dilated family of tubes \mathbb{U}_{\tau}=\{Rs^{2}\cdot T:T\in\mathbb{T}_{\tau}\}. Each U\in\mathbb{U}_{\tau} is thus a parallelepiped of dimensions Rs\times R . Let \mathbb{U} denote the union of \mathbb{U}_{\tau} for all \tau ranging over all dyadic scales R^{-1/2}\leq s\leq 1 . For a given U\in\mathbb{U} , we let \tau(U) denote the \tau such that U\in\mathbb{U}_{\tau} . \tau s s^{2} T s^{-1} s^{-2} U Rs R Figure 1 . U\in\mathbb{U}_{\tau} , T\in\mathbb{T}_{\tau} such that T\subset U for |\tau|=s . Given U\in\mathbb{U} , we define (3) \displaystyle\kappa_{p,H}(U)=\max_{\begin{subarray}{c}T\in\mathbb{T}_{\tau(U)}:\\ T\subset U\end{subarray}}\left(\frac{H(T)}{|T|}\right)^{\frac{1}{4}}\left(\frac{H(U)}{|U|}\right)^{\frac{1}{p}-\frac{1}{4}}, where we write H(E):=\int_{E}H for a measurable set E\subset\mathbb{R}^{2} . We are now ready to state our weighted square function estimates. Theorem 1.1 . Let 2\leq p\leq 4 and H:\mathbb{R}^{2}\to[0,1] be a function. For any function f whose Fourier transform is supported on N_{R^{-1}}(\mathcal{P}) , we have, (4) \displaystyle\|f\|_{L^{p}(Hdx)}\lessapprox\big(\max_{U\in\mathbb{U}}\kappa_{p,H}(U)+R^{-100}\big)\big\|\big(\sum_{\theta}|f_{\theta}|^{2}\big)^{1/2}\big\|_{L^{p}(\mathbb{R}^{2})}. Here we mean by A\lessapprox B an inequality of the form A\leq C_{p}(\log R)^{O(1)}B . The term R^{-100} on the right-hand side of ( 4 ) is negligible for interesting weights H . For instance, we have \max_{U\in\mathbb{U}}\kappa_{p,H}(U)\gg R^{-100} whenever H is the characteristic function of a union of unit balls. We also use the notation A\lesssim B to denote A\leq CB with an absolute constant C>0 , possibly depending on parameters such as p and \alpha . Remark 1.2 . Theorem 1.1 (and Theorem 1.4 to be stated) remains valid when the parabola \mathcal{P} is replaced by a small perturbation of \mathcal{P} for which the bilinear restriction estimate (see Theorem 4.4 ) is valid. In particular, it holds for any function whose Fourier transform is supported in a small neighborhood of the unit circle under a corresponding modification in the collections \mathbb{T}_{\tau} and \mathbb{U}_{\tau} outlined before the statement of Theorem 2.1 below. In addition, ( 4 ) holds for all non-negative H\in L^{\infty}(\mathbb{R}^{2}) with R^{-100} replaced by R^{-100}\|H\|_{L^{\infty}}^{1/p} by homogeneity and the fact that \kappa_{p,cH}(U)=c^{1/p}\kappa_{p,H}(U) for any U\in\mathbb{U} and constant c>0 . For H\equiv 1 , we have \kappa_{p,H}(U)=1 for any U\in\mathbb{U} . Thus, when p=4 , Theorem 1.1 essentially recovers the classical square function estimate ( 1 ). For 2\leq p<4 , the H\equiv 1 case of Theorem 1.1 recovers square function estimates due to Gan [ 15 ] , where more general small cap square function estimates are established. Our weighted square function estimates are inspired by weighted decoupling inequalities for the paraboloids, which have been extensively studied in recent years and applied to problems such as the Falconer distance set conjecture and Bochner-Riesz means; see, e.g., [ 18 , 11 , 16 , 26 ] and references therein. We compute the constant \max_{U\in\mathbb{U}}\kappa_{p,H}(U) for \alpha -dimensional weights. Example 1 ( \alpha -dimensional weights) . Let 0\leq\alpha\leq 2 . Suppose that H:\mathbb{R}^{2}\to[0,1] is \alpha -dimensional in the sense that \displaystyle\langle H\rangle_{\alpha}:=\sup_{(z,\rho)\in\mathbb{R}^{2}\times[1,\infty]}\rho^{-\alpha}H(B_{\rho}(z))\lesssim 1. Here B_{\rho}(z) denotes the ball of radius \rho centered at z (and we simply write B_{\rho} when centered at the origin). Then (5) \max_{U\in\mathbb{U}}\kappa_{p,H}(U)\;\lesssim\;R^{-(2-\alpha)\left(\frac{1}{p}-\frac{1}{4}\right)}. To see this, let T\in\mathbb{T}_{\tau} and U\in\mathbb{U}_{\tau} for some |\tau|=s . Since T and U are covered by O(s^{-1}) balls of radius s^{-1} and Rs , respectively, we have \displaystyle H(T) \displaystyle\lesssim s^{-1}s^{-\alpha}, \displaystyle H(U) \displaystyle\lesssim s^{-1}(Rs)^{\alpha}. On the other hand, |T|\sim s^{-3} and |U|\sim R^{2}s . Therefore, \max_{U\in\mathbb{U}}\kappa_{p,H}(U)\;\lesssim\;\max_{R^{-1/2}\leq s\leq 1}(s^{2-\alpha})^{\frac{1}{4}}\big((Rs)^{-(2-\alpha)}\big)^{\frac{1}{p}-\frac{1}{4}}, and the maximum is attained at the scale s=1 , which yields ( 5 ). Example 2 (Unit ball) . We examine the sharpness of Theorem 1.1 for the weight H=1_{B_{1}} . Since H is \alpha -dimensional for every \alpha\in[0,2] , Theorem 1.1 together with ( 5 ) shows that \displaystyle\|f\|_{L^{p}(B_{1})}\lessapprox R^{-2(\frac{1}{p}-\frac{1}{4})}\big\|\big(\sum_{\theta}|f_{\theta}|^{2}\big)^{1/2}\big\|_{L^{p}(\mathbb{R}^{2})}. This estimate is essentially sharp for all 2\leq p\leq 4 . Indeed, let \widehat{f_{\theta}} be an L^{1} -normalized smooth bump function supported on \theta . In this case, |\sum_{\theta}f_{\theta}(x)|\gtrsim\#\{\theta\} for x\in B_{c} for a sufficiently small c>0 and |f_{\theta}| decays rapidly away from the tube \theta^{*} dual to \theta

Read the complete text