[Paper Review] Nonstandard estimates for a class of 1D dispersive equations and applications to linearized water waves

Jennifer N. Beichman|arXiv (Cornell University)|Sep 29, 2014

Advanced Mathematical Physics Problems9 references3 citations

TL;DR

This paper establishes sharp nonstandard decay estimates for a class of 1D dispersive equations, including the linearized water wave equation, by analyzing oscillatory integral operators and identifying a surprising time-growth factor in solutions when low-frequency initial data are present. The key contribution is proving that this growth factor is optimal and arises from singularities in the initial data's Fourier transform near the origin, resolving a long-standing question about the necessity of small initial height assumptions in long-time existence results for water waves.

ABSTRACT

In this work, we obtain decay bounds for a class of ID dispersive equations that includes the linearized water wave. These decay bounds display a surprising growth factor, which we show is sharp, The proofs rely on careful analysis of certain oscillatory integral operators. In addition, these results have applications to the linearized water wave operator for low regularity data.

Motivation & Objective

• To remove the small initial height assumption in long-time existence results for the 2D water wave equation by analyzing decay properties without relying on the $ L^2 $ norm of the antiderivative of initial data.
• To understand the role of small-frequency waves in the long-time behavior of solutions to the linearized water wave equation.
• To derive sharp $ \dot{H}^s \to L^\infty $ decay estimates for a general class of 1D dispersive operators $ \partial_t u - i|D|^a u = 0 $, $ a \in (0,1) \setminus \{1/2\} $, under minimal regularity assumptions.
• To identify the precise threshold on initial data regularity (via $ \dot{H}^s $ norms) that determines whether solutions exhibit decay or growth in time, particularly for singular initial data near $ \xi = 0 $.

Proposed method

• The analysis relies on a detailed study of oscillatory integral operators associated with the fundamental solution of the dispersive equation, focusing on phase functions with critical points and stationary phase behavior.
• A novel approach avoids the use of the problematic vector field $ \Omega_0 $ from previous works, instead using a reduced set of vector fields inspired by Keel-Smith-Sogge's method for nonlinear wave equations.
• The proof involves analyzing the behavior of the phase function $ \phi(\xi) = \xi t - |\xi|^a \alpha $ and its derivatives to control the decay of oscillatory integrals, particularly near critical frequencies.
• The authors derive sharp bounds on the $ L^\infty $ norm of solutions by estimating oscillatory integrals through stationary phase and non-stationary phase techniques, with special attention to the singularity at $ \xi = 0 $.
• A key technical step is the analysis of the function $ \mathcal{J}(\xi) $, defined via the phase and amplitude, whose derivative and critical points are used to establish lower bounds on the oscillatory integral's size.
• The method generalizes to a class of dispersive operators with $ a \in (0,1) \setminus \{1/2\} $, and the results are applied specifically to the linearized water wave equation with $ a = 1/2 $.

Experimental results

Research questions

• RQ1 What is the sharp decay rate of solutions to the linearized water wave equation when the initial data are only in $ \dot{H}^s $ for $ s > 0 $, without control on the $ L^2 $ norm of the antiderivative?
• RQ2 How does the presence of small-frequency waves—characterized by a singularity in the Fourier transform at $ \xi = 0 $—affect the long-time decay or growth of solutions?
• RQ3 Is the time-growth factor observed in previous estimates for the water wave equation sharp, and what is its origin in the frequency space structure of the initial data?
• RQ4 Can the small initial height assumption in Wu’s almost global existence result be removed by proving improved decay estimates that do not rely on antiderivative control?
• RQ5 What is the precise relationship between the order of singularity at $ \xi = 0 $ in the initial data and the rate of spatial decay of the solution at infinity?

Key findings

• The paper establishes a sharp $ \dot{H}^s \to L^\infty $ decay estimate for solutions to the linearized water wave equation with initial data in $ \dot{H}^s $, $ s > 0 $, revealing a time-growth factor of order $ t^{a/(2(1-a))} $ for $ a \in (0,1) \setminus \{1/2\} $, which is optimal.
• For the linearized water wave equation ($ a = 1/2 $), the solution exhibits a time-growth factor of order $ t^{1/2} $, which is shown to be sharp, arising from the singularity in the initial data at low frequencies.
• The growth factor originates from the behavior of oscillatory integrals near $ \xi = 0 $, where the phase function has a degenerate critical point, and is quantified via the analysis of the function $ \mathcal{J}(\xi) $ and its derivative.
• The analysis proves that the $ \Omega_0 $ vector field used in earlier works is not necessary for deriving decay estimates, and a reduced set of vector fields suffices when combined with careful frequency-localized analysis.
• The paper shows that the decay rate of the solution at spatial infinity is directly tied to the order of the singularity in the initial data’s Fourier transform at $ \xi = 0 $, with more singular initial data leading to slower spatial decay.
• By combining the new $ \dot{H}^s \to L^\infty $ estimate with the classical decay estimate from Wu’s work (valid for frequencies bounded away from zero), the paper provides a refined, unified decay estimate that improves upon both in the full frequency range.