---
title: Time-Periodic Gevrey-Sobolev-Kato Spaces
url: https://www.emergentmind.com/topics/time-periodic-gevrey-sobolev-kato-spaces
type: topic
---

# Time-Periodic Gevrey-Sobolev-Kato Spaces

Time-periodic Gevrey-Sobolev-Kato spaces are anisotropic periodic ultradifferentiable spaces on products such as \(T^n\times X\) and \(\mathbb T^n\times \mathbb R^d\), where the time variable is periodic, the time regularity is Gevrey, and the spatial regularity is measured in weighted Sobolev scales that the relevant papers describe as Sobolev-Kato spaces. Their distinctive feature is that they are built to admit spectral characterizations through suitable elliptic operators, so that periodic evolution equations can be reduced to infinite families of scalar ODEs indexed by spatial eigenmodes. The directly named construction appears on asymptotically Euclidean manifolds in 2025, while a closely related time-periodic weighted Sobolev version on \(\mathbb T^n\times\mathbb R^d\) was developed earlier the same year; both are best understood against a longer background in operator-domain Gevrey theory, periodic Navier–Stokes analysis, and Kato-Sobolev frameworks [2509.03745, 2501.03414].

## 1. Direct definitions and basic models

The manifold model uses a \(d\)-dimensional asymptotically Euclidean manifold \(X\), equivalently a scattering manifold, and periodic time \(t\in T^n\), \(T=\mathbb R/(2\pi\mathbb Z)\). The natural spatial regularity scale is \(H^{r,\rho}(X)\), described there as weighted Sobolev spaces, also called Sobolev-Kato spaces. For fixed \(\sigma>1\), \(r,\rho\in\mathbb R\), and \(C>0\), the basic Banach block is
\[
\mathcal G^\sigma\mathcal H_{r,\rho}^C(\mathbb T^n\times X)
\]
with defining norm
\[
\|u\|_{\sigma,C,r,\rho}
=
\sup_{\gamma\in\mathbb Z_+^n}
\left\{
C^{-|\gamma|}(\gamma!)^{-\sigma}
\sup_{t\in\mathbb T^n}
\|\partial_t^\gamma u(t)\|_{H^{r,\rho}(X)}
\right\}.
\]
The associated inductive limits are
\[
\mathcal G^\sigma\mathcal H_{r,\rho}(T^n\times X)
=
\bigcup_{C>0}\mathcal G^\sigma\mathcal H_{r,\rho}^C(\mathbb T^n\times X),
\qquad
\mathcal G\mathcal H_{r,\rho}(T^n\times X)
=
\bigcup_{\sigma>1}\mathcal G^\sigma\mathcal H_{r,\rho}(T^n\times X),
\]
and the global test-function space is
\[
F(\mathbb T^n\times X)=\bigcap_{r,\rho\in\mathbb R}\mathcal G\mathcal H_{r,\rho}(\mathbb T^n\times X),
\]
with dual
\[
F'(\mathbb T^n\times X)=\bigcup_{r,\rho\in\mathbb R}\mathcal G\mathcal H'_{r,\rho}(\mathbb T^n\times X).
\]
This construction is explicitly introduced as a family of time-periodic Gevrey-Sobolev-Kato spaces [2509.03745].

The Euclidean SG version is formulated on \(\mathbb T^n\times\mathbb R^d\). Its spatial scale is
\[
H^{r,p}(\mathbb R^d)
=
\left\{
u\in\mathcal S'(\mathbb R^d):
\|u\|_{r,p}
:=
\|\langle x\rangle^r\langle D\rangle^p u\|_{L^2(\mathbb R^d)}<\infty
\right\},
\]
with \(\langle x\rangle=(1+|x|^2)^{1/2}\) and \(\langle D\rangle^p=\operatorname{Op}(\langle\xi\rangle^p)\). The paper explicitly calls these weighted Sobolev spaces also Sobolev-Kato spaces. For \(\sigma>1\), \(r,p\in\mathbb R\), and \(C>0\), it defines
\[
G^{\sigma,C}H^{r,p}(\mathbb T^n\times\mathbb R^d)
\]
by
\[
\|u\|_{\sigma,C,r,p}
=
\sup_{\gamma\in\mathbb Z_+^n}
\left\{
C^{-|\gamma|}(\gamma!)^{-\sigma}
\sup_{t\in\mathbb T^n}
\|\partial_t^\gamma u(t)\|_{H^{r,p}(\mathbb R^d)}
\right\}
<\infty.
\]
The corresponding inductive limit is
\[
\mathcal G H^{r,p}(\mathbb T^n\times\mathbb R^d)
=
\bigcup_{\sigma>1}\bigcup_{C>0} G^{\sigma,C}H^{r,p}(\mathbb T^n\times\mathbb R^d),
\]
and the projective-limit test space is
\[
\mathcal F(\mathbb T^n\times\mathbb R^d)
=
\bigcap_{r,p\in\mathbb R}\mathcal G H^{r,p}(\mathbb T^n\times\mathbb R^d).
\]
Here again the time regularity is Gevrey and the space regularity is Sobolev-Kato [2501.03414].

These definitions are anisotropic by construction. Time is treated through factorial Gevrey control of \(\partial_t^\gamma\), while space is treated through graph-type Sobolev norms associated either with scattering operators on \(X\) or SG-elliptic operators on \(\mathbb R^d\). A common misconception is to identify these spaces with ordinary torus-periodic Gevrey classes in all variables; the direct constructions are more specific, because they separate the periodic time variable from the spatial operator scale [2509.03745].

## 2. Spectral characterizations

A central reason these spaces are useful is that they admit exact coefficientwise descriptions. On asymptotically Euclidean manifolds, the spatial operator is a positive self-adjoint elliptic scattering operator
\[
P\in \Psi_{scc}^{m,\mu}(X)
\]
with eigenbasis \(\{\phi_j\}_{j\in\mathbb N^\ast}\) and eigenvalues \(\{\lambda_j\}\). With \(\widetilde P=P+P_0\), where \(P_0\) is the orthogonal projection onto \(\ker P\), one writes
\[
u=\sum_{j\in\mathbb N^\ast}u_j\phi_j,
\qquad
u_j=\langle u,\overline{\phi_j}\rangle.
\]
The spatial Sobolev-Kato scale is characterized by
\[
\|u\|_{H^{rm,r\mu}(X)}^2
\asymp
\sum_{j\in\mathbb N^\ast}|u_j|^2|\widetilde\lambda_j|^{2r},
\]
and \(f\in F(T^n\times X)\) if and only if, for every \(M\in\mathbb N\), there exist \(\sigma_M>1\) and \(C_M>0\) such that
\[
\sup_{t\in T^n}\sum_{j\in\mathbb N^\ast}\widetilde\lambda_j^{2M}|\partial_t^\gamma f_j(t)|^2
\le
C_M^{2(|\gamma|+1)}(\gamma!)^{2\sigma_M}
\qquad
\text{for every }\gamma\in\mathbb Z_+^n.
\]
When \(P\) is positive, self-adjoint, classical, and elliptic, this is equivalent to the pointwise decay estimate
\[
\sup_{t\in T^n}|\partial_t^\gamma f_j(t)|
\le
C^{|\gamma|+1}(\gamma!)^\sigma |\widetilde\lambda_j|^{-M},
\qquad
j\in\mathbb N^\ast,\ \gamma\in\mathbb Z_+^n.
\]
The dual space \(F'\) is described by coefficient distributions \(\theta_j\in(\mathcal G^\sigma(T^n))'\) with polynomial spectral growth, and every \(\theta\in F'\) has an expansion
\[
\theta=\sum_{j\in\mathbb N^\ast}\theta_j\otimes\phi_j.
\]
These statements are the sharp characterization results around which the manifold theory is organized [2509.03745].

The Euclidean SG theory is structurally parallel. One fixes a normal elliptic SG-operator
\[
P\in \operatorname{Op}(S^{m,\mu}(\mathbb R^d)),
\qquad m,\mu>0,
\]
defines \(\widetilde P=P+P_0\), and uses its orthonormal eigenbasis \(\{\phi_j\}\) with eigenvalues \(\{\lambda_j\}\). For \(u=\sum_j u_j\phi_j\),
\[
u\in H^{rm,r\mu}(\mathbb R^d)
\Longleftrightarrow
\sum_{j\in\mathbb N}|u_j|^2|\lambda_j|^{2r}<\infty,
\]
with the exact identity
\[
\|u\|_{H^{rm,r\mu}(\mathbb R^d)}^2
=
\sum_{j\in\mathbb N}|u_j|^2|\lambda_j|^{2r}.
\]
The time-periodic test space \(\mathcal F(\mathbb T^n\times\mathbb R^d)\) is then characterized by the coefficient condition: for every \(M\in\mathbb N\), there exist \(\sigma_M>1\) and \(C_M>0\) such that
\[
\sup_{t\in\mathbb T^n}\sum_{j\in\mathbb N} |\lambda_j|^{2M}|\partial_t^\gamma f_j(t)|^2
\le
C_M^{2(|\gamma|+1)}(\gamma!)^{2\sigma_M},
\qquad
\forall \gamma\in\mathbb Z_+^n.
\]
When \(P\) is self-adjoint, positive, and elliptic, this is equivalent to
\[
\sup_{t\in\mathbb T^n} |\partial_t^\gamma f_j(t)|
\le
C^{|\gamma|+1}(\gamma!)^\sigma |\lambda_j|^{-M}.
\]
Thus both direct theories reduce regularity to Gevrey control of the scalar periodic mode functions \(f_j(t)\) plus rapid decay in the spectral parameter [2501.03414].

This spectral representation is not a secondary reformulation. It is the main mechanism by which periodic PDEs become tractable: a time-periodic evolution equation on \(T^n\times X\) or \(\mathbb T^n\times\mathbb R^d\) is converted into a countable family of scalar periodic ODEs, and the function space is tailored so that uniform estimates on those ODEs are equivalent to membership in the ambient regularity class. That design principle is explicit in the direct papers [2509.03745, 2501.03414].

## 3. Gevrey, Sobolev, and Kato lineages

The term “Gevrey-Sobolev-Kato” combines three ingredients that arose historically in partially separate literatures. The Gevrey component is the periodic-time estimate
\[
C^{-|\gamma|}(\gamma!)^{-\sigma}\sup_t \|\partial_t^\gamma u(t)\|<\infty,
\]
the Sobolev component is the use of spatial weighted Sobolev scales such as \(H^{r,\rho}(X)\) or \(H^{r,p}(\mathbb R^d)\), and the Kato component refers to the operator-generated weighted Sobolev scale natural on scattering or SG geometries. In the manifold paper, the authors state that the Kato label refers to the weighted Sobolev scale \(H^{r,\rho}(X)\) natural on scattering manifolds, and in the Euclidean SG paper the spaces \(H^{r,p}\) are explicitly described as Sobolev-Kato spaces [2509.03745, 2501.03414].

An important antecedent is Arsu’s “On Kato-Sobolev type spaces,” which studies the Wiener amalgam family
\[
\mathcal B_k^p = W(\mathcal B_k,L^p),
\qquad
\mathcal B_k=B_{2,k},
\]
as a generalization of uniformly local Sobolev/Kato spaces. Those spaces are neither periodic in time nor Gevrey, and the admissible weights are of polynomial growth, but the paper supplies a foundational blueprint for localization, multiplication, Wiener–Lévy theorems, and pseudodifferential operator bounds in Kato-Sobolev settings. It is therefore background rather than a direct source for the time-periodic Gevrey constructions [1209.6465].

A second lineage comes from abstract Gevrey regularity for dissipative equations. In “Gevrey regularity for a class of dissipative equations with analytic nonlinearity,” the mild formulation
\[
u(t)=e^{-tA^\kappa}u_0+\int_0^t e^{-(t-s)A^\kappa}G(u(s))\,ds
\]
is developed in scale-critical \(L^p\)-Sobolev spaces, with Gevrey norms of the form
\[
\|v\|_{Gv(s,\beta,p)}
=
\|e^{cs^{1/\kappa}A_1}\Lambda^\beta v\|_{L^p}
\]
or, in the inhomogeneous setting,
\[
\|v\|_{Gv(s,\beta,p)}
=
\|e^{cs^{1/\kappa}A_1}(I+\Lambda)^\beta v\|_{L^p}.
\]
The paper does not study time-periodic solutions, time-periodic forcing, or periodicity in the time variable, but it explicitly states that its machinery is directly relevant to attempts to build a theory of time-periodic Gevrey-Sobolev-Kato spaces. Its most useful contributions are the Kato/Fujita–Kato mild-solution framework, critical scaling analysis, Gevrey-weighted semigroup smoothing, and analytic-majorant estimates for nonlinearities [1403.1603].

A third lineage is operator-domain Gevrey theory on periodic boxes. Hoang–Martinez formulate the spaces
\[
G_{\alpha,\sigma}
=
D\!\left(A^\alpha e^{\sigma A^{1/2}}\right),
\qquad
|u|_{\alpha,\sigma}
=
\left|A^\alpha e^{\sigma A^{1/2}}u\right|,
\]
for the three-dimensional incompressible Navier–Stokes equations on a periodic box, and prove that the Foias–Saut asymptotic expansion holds in all Gevrey classes for every Leray–Hopf weak solution. This work is fundamental for the spatially periodic Gevrey-Sobolev operator framework, but the “periodic” aspect there is spatial periodicity, not time periodicity [1511.03523].

The resulting picture is precise. Time-periodic Gevrey-Sobolev-Kato spaces do not arise from a single tradition; they synthesize periodic Gevrey regularity in time, operator-generated Sobolev-Kato scales in space, and modewise analysis of periodic evolution equations. The direct 2025 formulations codify that synthesis explicitly [2509.03745, 2501.03414].

## 4. Time-periodic evolution operators and global hypoellipticity

The most developed applications of these spaces are to global hypoellipticity and solvability for periodic evolution equations. On asymptotically Euclidean manifolds, the constant-coefficient model is
\[
L=D_t+\omega P,
\qquad
\omega=\alpha+i\beta,
\]
and global hypoellipticity on \(T\times X\) means
\[
u\in F'(T\times X),\ Lu\in F(T\times X)\Rightarrow u\in F(T\times X).
\]
After expansion in the eigenbasis of \(P\), the equation becomes
\[
D_tu_j(t)+\omega\lambda_j u_j(t)=f_j(t),
\qquad
j\in\mathbb N^\ast.
\]
The paper proves that \(L\) is globally hypoelliptic if and only if either \(\beta\ne0\), or \(\beta=0\) and \(\alpha\) satisfies Condition \((\mathcal H)\), namely the existence of \(\epsilon,C>0\) such that
\[
|\tau-\alpha\lambda_j|\ge C j^{-\epsilon},
\qquad
\forall (j,\tau)\in\mathbb N^\ast\times\mathbb Z.
\]
For the variable-coefficient model
\[
\mathcal L=D_t+c(t)P,
\qquad
c(t)=a(t)+ib(t)\in \mathcal G^s(T),
\]
the sign of \(b(t)\) controls the outcome: if \(b<0\) on \([0,2\pi]\), or \(b>0\) on \([0,2\pi]\), then \(\mathcal L\) is globally hypoelliptic; if \(b\) is not identically zero on any subinterval and \(b\) changes sign, then \(\mathcal L\) is not globally hypoelliptic; and the case \(b\equiv0\) is left open [2509.03745].

The Euclidean SG theory yields an analogous criterion on \(\mathbb T\times\mathbb R^d\). For
\[
L=D_t+\omega P,
\qquad
\omega=\alpha+i\beta,
\]
the paper proves global hypoellipticity on \(\mathbb T\times\mathbb R^d\) if and only if either \(\beta\ne0\), or \(\beta=0\) and \(\alpha\) satisfies Condition (A):
\[
|\tau-\alpha\lambda_j|\ge C j^{-\varepsilon},
\qquad
\forall (j,\tau)\in \mathbb N\times \mathbb Z.
\]
The proof is modewise. Writing
\[
u(t)=\sum_j u_j(t)\phi_j,
\qquad
f(t)=\sum_j f_j(t)\phi_j,
\]
one reduces to
\[
D_tu_j(t)+\omega\lambda_j u_j(t)=f_j(t),
\]
whose periodic solutions are expressed through denominators of the form
\[
1-e^{-2\pi i\omega\lambda_j}.
\]
When \(\beta\ne0\), those denominators are harmless. When \(\beta=0\), small divisors are controlled exactly by Condition (A) [2501.03414].

A closely related solvability theory in time-periodic Gelfand-Shilov spaces was developed by Fernando de Ávila Silva and Marco Cappiello for
\[
L=D_t+c(t)P(x,D_x)
\]
on \(\mathbb T\times\mathbb R^n\), with \(P\) self-adjoint and globally elliptic. Their space is not described as a Sobolev-Kato scale, since the spatial regularity is Gelfand-Shilov rather than weighted Sobolev, but the mechanism is the same: spectral diagonalization reduces the PDE to
\[
\partial_t u_j(t)+i\lambda_j c(t)u_j(t)=if_j(t).
\]
The resulting solvability criterion is sharper in the ultradifferentiable scale. If \(b\) does not change sign, solvability is equivalent to a subexponential Diophantine condition
\[
|\tau-c_0\lambda_j|> C_\varepsilon \exp\!\big(-\varepsilon j^{1/(2n\mu)}\big),
\]
while if \(b\) changes sign, solvability is equivalent to the finiteness of the nonresonant complement \(Z^c\) together with connectedness of all superlevel sets
\[
\Omega_r=\left\{t\in\mathbb T:\int_0^t b(s)\,ds>r\right\}.
\]
This paper is not a Sobolev-Kato theory, but it confirms that the decisive mechanisms in time-periodic evolution problems are mode reduction, resonant sets, and small-divisor or damping estimates matched to the target regularity scale [2402.10006].

## 5. Spatially periodic and dissipative Gevrey-Sobolev frameworks

Much of the literature most relevant to time-periodic Gevrey-Sobolev-Kato spaces is not itself time periodic. Hoang–Martinez study the three-dimensional incompressible Navier–Stokes equations on a periodic box and prove that the Foias–Saut asymptotic expansion holds in all Gevrey classes. Their norms
\[
|u|_{\alpha,\sigma}
=
\left|A^\alpha e^{\sigma A^{1/2}}u\right|
\]
and spaces
\[
G_{\alpha,\sigma}=D(A^\alpha e^{\sigma A^{1/2}})
\]
are purely spatial operator-domain Gevrey-Sobolev spaces. The “periodic” variable is the spatial torus, and the asymptotic expansion
\[
u(t)\sim \sum_{n=1}^\infty q_n(t)e^{-nt}
\quad\text{in every }G_{\alpha,\sigma}
\]
is a large-time result, not a time-periodic one. Even so, the paper is a major source for bilinear Gevrey estimates, eventual analyticity, and spectral mode decomposition on periodic domains [1511.03523].

Biswas, Hudson, and Tian study persistence times for strong solutions of the three-dimensional Navier–Stokes equations in time-varying analytic Gevrey classes based on Sobolev spaces \(H^s\), \(s>\frac12\). Their basic norm is
\[
|u|_{s,\alpha}
=
\left(\sum_k |k|^{2s} e^{2\alpha |k|} |\hat u_k|^2\right)^{1/2}
=
|A^{s/2}e^{\alpha A^{1/2}}u|_{L^2},
\]
and the analyticity radius is allowed to vary linearly in time,
\[
\alpha(t)=\beta_0+\beta t
\quad\text{or}\quad
\alpha(t)=\beta t.
\]
They show that persistence times in these stronger Gevrey norms match the best known persistence times in Sobolev classes and are optimal for \(\frac12<s<\frac52\). This is not a periodic-in-time theory, but it demonstrates that weighted Gevrey-Sobolev energy estimates can track a time-dependent analyticity radius without losing the sharp Sobolev lifespan scale [1912.11192].

A different periodic-box line is the Fourier-side method for a priori Gevrey-Sobolev bounds on equations of hydrodynamic type. There the norm is
\[
|v|_{\sigma,q}^2=\sum_{n\neq 0} |\widehat v_n|^2 e^{2\sigma |n|}|n|^{2q},
\]
and a nonlinear transformation of Fourier coefficients produces a feedback between the analyticity index and the transformed Sobolev norm. The paper gives finite-time bounds for Euler and inviscid Burgers, global-in-time bounds for Voigt-type regularisations of Euler and Navier–Stokes, and in the critical Voigt-type Navier–Stokes case a global uniform analytic bound. Again, the periodicity is spatial and the framework is not Kato-type in the semigroup sense, but it supplies techniques for time-uniform analytic control [1001.4237].

The abstract whole-space theory of Gevrey regularity for dissipative equations with analytic nonlinearities provides the nearest direct bridge to Kato semigroup methods. Its fixed-point space
\[
E=\Big\{u\in C([0,T);\dot H_p^{\beta_0}) :
\|u\|_E=\max\{\|u\|_{E_1},\|u\|_{E_2}\}<\infty\Big\},
\]
with
\[
\|u\|_{E_1}=\sup_{s\in(0,T)}\|u(s)\|_{Gv(s,\beta_0,p)},
\qquad
\|u\|_{E_2}=\sup_{s\in(0,T)} s^{\delta/\kappa}\|u(s)\|_{Gv(s,\gamma,p)},
\]
is explicitly described as analogous to Kato spaces with time-weighted derivative norms, except that Gevrey weights are built in. The paper states directly that what is missing for a genuinely time-periodic theory are forcing terms, periodic solution operators, temporal Floquet or Poincaré formulations, and norms on a time-periodic orbit rather than on an initial-value evolution [1403.1603].

These antecedents matter because they show that the essential analytic machinery already existed before the directly named time-periodic Gevrey-Sobolev-Kato spaces were introduced. What the 2025 papers add is a periodic-time functional architecture in which that machinery can be combined with spectral ODE reduction [2509.03745, 2501.03414].

## 6. Approximation theory, periodicity, and delimitations

The word “periodic” is used in two different senses across the literature, and confusion between them is common. Several important Gevrey-Sobolev papers are periodic only in space. The Navier–Stokes asymptotic expansion of Hoang–Martinez, the hydrodynamic a priori bounds with Gevrey-Sobolev norms, and the approximation-theoretic Fourier-weighted Gevrey spaces on \(\mathbb T^d\) are all torus-periodic in spatial variables, not time-periodic in the sense relevant to periodic evolution equations [1511.03523, 1001.4237, 1505.00631].

The approximation-theoretic paper is nevertheless conceptually relevant because it gives a clean periodic Fourier model for Sobolev and Gevrey scales. On \(\mathbb T^d\), it studies
\[
H^{\mathbf w}(\mathbb T^d)
=
\left\{
f\in L_2(\mathbb T^d):
\sum_{k\in\mathbb Z^d} w(k)^2 |c_k(f)|^2<\infty
\right\},
\]
the isotropic Sobolev family
\[
H^{s,p}(\mathbb T^d),
\qquad
w_{s,p}(k)=
\begin{cases}
(1+\|k\|_p^p)^{s/p}, & 0<p<\infty,\\
\max(1,|k_1|,\dots,|k_d|)^s, & p=\infty,
\end{cases}
\]
and the Gevrey-type family
\[
G^{\alpha,\beta,p}(\mathbb T^d)
=
\left\{
f\in C^\infty(\mathbb T^d):
\sum_{k\in\mathbb Z^d} \exp\!\bigl(2\beta \|k\|_p^\alpha\bigr)|c_k(f)|^2<\infty
\right\}.
\]
Its main results concern approximation numbers and entropy numbers, not PDE, and it contains no Kato framework. Still, it explicitly notes that one may interpret one periodic variable as time and work on a joint torus \(\mathbb T_t\times\mathbb T_x^d\), even though the paper itself does not introduce anisotropic time-space weights of parabolic type [1505.00631].

This suggests a useful distinction. A plausible implication is that the directly named time-periodic Gevrey-Sobolev-Kato spaces sit between two earlier paradigms: isotropic torus-periodic Fourier-weighted Gevrey spaces, where all variables are treated symmetrically, and evolution-adapted Gevrey or Kato spaces, where time and space have different analytic roles. The 2025 constructions choose the second paradigm explicitly: time is periodic and Gevrey, space is noncompact and controlled by a weighted Sobolev-Kato operator scale [2509.03745, 2501.03414].

The principal open and missing points are also sharply delineated in the sources. In the asymptotically Euclidean manifold theory, the variable-coefficient case with \(b\equiv0\) is left open [2509.03745]. In the abstract dissipative Gevrey theory, there is no forcing term, no periodic-response operator, and no time-periodic norm; the paper itself states that these elements are missing for a genuine time-periodic theory [1403.1603]. In the periodic Navier–Stokes Gevrey expansion and related periodic-box works, the “periodic” structure is spatial rather than temporal [1511.03523, 1001.4237]. Those delimitations are not peripheral. They define the conceptual boundary of the subject: a time-periodic Gevrey-Sobolev-Kato space is not merely a periodic Gevrey space, nor merely a Kato-Sobolev scale, but a space built so that periodic time dependence, spatial operator regularity, and spectral mode analysis coexist in a single functional framework.

Source: https://www.emergentmind.com/topics/time-periodic-gevrey-sobolev-kato-spaces