Parabolic Tail Condition in Probability and PDE

• Parabolic tail condition is a framework that defines how parabolic scaling influences tail behavior in both stochastic processes and nonlocal PDEs.
• In probability, it yields a stretched-exponential decay (e.g., x^(3/2)) in Brownian motion with a quadratic penalty, while in PDEs it quantifies nonlocal exterior forces.
• The diverse formulations highlight its significance in establishing optimal regularity conditions and asymmetry in tail behavior across different mathematical models.

The expression parabolic tail condition does not denote a single standardized concept across the literature. In probability and stochastic growth, it often refers to a tail law induced by a quadratic or parabolic structure, such as Brownian motion penalized by $t^2$ or the KPZ equation started from a parabola. In nonlocal parabolic PDE, by contrast, it denotes a time-integrated exterior tail functional that measures the influence of values outside a spatial ball on local regularity. Some papers also discuss nearby geometric or coefficient conditions—such as the tusk condition or parabolic Carleson conditions—that are parabolic and multiscale but are explicitly not tail conditions in the same sense (Groeneboom et al., 2010, Kamenev et al., 2016, Liao, 2024, Kar, 11 Feb 2026).

1. Terminological scope and basic dichotomy

Two main usages recur.

First, in probabilistic asymptotics, a parabolic tail condition is a statement about the survival tail of a random maximum formed by subtracting a parabola. The canonical example is

$N=\max_{t\ge 0}\{W(t)-t^2\},$

for standard Brownian motion $W$, where the quadratic penalty produces a stretched-exponential tail of order $e^{-c x^{3/2}}$ rather than a Gaussian $e^{-c x^2}$ or exponential $e^{-c x}$ decay (Groeneboom et al., 2010).

Second, in nonlocal parabolic regularity theory, the same phrase refers to an exterior quantity integrated over a backward time interval, typically of the form

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$

or its Heisenberg-group analogue. This tail is not a probability tail: it is a nonlocal forcing term that must be controlled to obtain continuity, Hölder regularity, or local boundedness (Liao, 2024, Kar, 11 Feb 2026, Kar et al., 9 Apr 2025).

This suggests that the unifying content of the phrase is not a shared formula but a shared parabolic scaling principle. In one class of problems, the parabola $t^2$ reshapes large deviations; in the other, intrinsic parabolic cylinders and backward time averages reshape how nonlocal exterior data enter local estimates.

2. Brownian motion minus a parabola and the canonical $x^{3/2}$ law

The most explicit probabilistic realization appears in the study of the maximum of Brownian motion minus a quadratic drift,

$N=\max_{t\ge 0}\{W(t)-t^2\},$

and its two-sided analogue

$N=\max_{t\ge 0}\{W(t)-t^2\},$0

The principal asymptotic statement is

$N=\max_{t\ge 0}\{W(t)-t^2\},$1

with matching density asymptotic

$N=\max_{t\ge 0}\{W(t)-t^2\},$2

For two-sided Brownian motion, the corresponding tail and density acquire a factor $N=\max_{t\ge 0}\{W(t)-t^2\},$3: $N=\max_{t\ge 0}\{W(t)-t^2\},$4 The rate is therefore

$N=\max_{t\ge 0}\{W(t)-t^2\},$5

and the characteristic exponent is $N=\max_{t\ge 0}\{W(t)-t^2\},$6, which is the defining signature of the parabolic tail in this setting (Groeneboom et al., 2010).

The paper also gives a full asymptotic expansion for the one-sided survival function,

$N=\max_{t\ge 0}\{W(t)-t^2\},$7

with

$N=\max_{t\ge 0}\{W(t)-t^2\},$8

and an analogous density expansion with coefficients

$N=\max_{t\ge 0}\{W(t)-t^2\},$9

Thus the term “parabolic tail condition” here means not merely $W$0, but an Airy-controlled asymptotic series refining the leading stretched-exponential law (Groeneboom et al., 2010).

A further structural point is the quadratic-coefficient scaling. For

$W$1

the paper states

$W$2

so the $W$3 exponent is preserved while the rate is multiplied by $W$4. In particular,

$W$5

The derivation proceeds from exact Airy-function representations for the density, such as

$W$6

followed by contour deformation and saddle-point analysis. The paper explicitly relates this representation to relation (5.10) in Janson, Louchard, and Martin-Löf (2010), with the normalization difference arising because that paper studies $W$7 rather than $W$8 (Groeneboom et al., 2010).

3. KPZ started from a parabola: asymmetric large deviations

A different but closely related meaning arises for the $W$9-dimensional KPZ equation

$e^{-c x^{3/2}}$0

with parabolic initial condition

$e^{-c x^{3/2}}$1

The observable is the one-point height distribution at the origin,

$e^{-c x^{3/2}}$2

with $e^{-c x^{3/2}}$3 taken in the proper moving frame

$e^{-c x^{3/2}}$4

In the weak-noise, short-time regime, the distribution obeys a large-deviation form

$e^{-c x^{3/2}}$5

but the two tails behave differently (Kamenev et al., 2016).

For $e^{-c x^{3/2}}$6,

$e^{-c x^{3/2}}$7

For $e^{-c x^{3/2}}$8,

$e^{-c x^{3/2}}$9

The paper identifies this asymmetry as the central parabolic-tail phenomenon for KPZ started from a parabola: the positive/right tail has exponent $e^{-c x^2}$0 and depends on the curvature through

$e^{-c x^2}$1

whereas the negative/left tail has exponent $e^{-c x^2}$2 and prefactor

$e^{-c x^2}$3

that is independent of $e^{-c x^2}$4 and $e^{-c x^2}$5 (Kamenev et al., 2016).

The dependence of the positive tail is encoded in a scaling function $e^{-c x^2}$6 satisfying

$e^{-c x^2}$7

and

$e^{-c x^2}$8

Hence $e^{-c x^2}$9 corresponds to the sharp-wedge or droplet limit, while $e^{-c x}$0 corresponds to the flat limit. The positive-tail prefactor therefore interpolates monotonically between those two exactly solved regimes (Kamenev et al., 2016).

Mechanistically, the two tails come from different optimal fluctuations in weak-noise theory. The positive tail is governed by an inviscid hydrodynamic system for the slope field $e^{-c x}$1,

$e^{-c x}$2

and yields the universal $e^{-c x}$3 exponent after the scaling

$e^{-c x}$4

The negative tail is instead dominated by a localized boundary-layer solution,

$e^{-c x}$5

which explains why its action, and hence its asymptotic coefficient, is universal across all $e^{-c x}$6 (Kamenev et al., 2016).

4. Nonlocal parabolic equations: tail functionals as exterior forcing

In nonlocal parabolic PDE, a parabolic tail condition is a genuinely different object. It measures the contribution of values of a solution outside a spatial ball over a backward time interval, in a manner compatible with the kernel and intrinsic scaling of the equation.

For the nonlocal $e^{-c x}$7-Laplacian type equation

$e^{-c x}$8

with measurable symmetric kernel comparable to $e^{-c x}$9, the fundamental tail is

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$0

The paper distinguishes a minimal tail regime,

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$1

which is enough for continuity with a quantified modulus, from a stronger regime

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$2

which upgrades continuity to Hölder continuity (Liao, 2024).

The paper’s analytic innovation is a time-dependent truncation level

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$3

whose derivative cancels the exterior contribution in the energy inequality. As a result, the tail enters De Giorgi iteration in an either-or form: either the tail is too large, or one obtains measure-to-pointwise improvement, expansion of positivity, and oscillation reduction (Liao, 2024).

On the Heisenberg group, the same architecture persists, but with the homogeneous distance $$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$4 and homogeneous dimension $$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$5. One normalized definition is

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$6

The final Hölder theorem assumes that for some $$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$7,

$$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$8

and calls this an optimal tail condition, because earlier parabolic nonlocal results used $$\int_{t_0-S}^{t_0}\int_{\mathbb R^N\setminus K_R(x_0)} \frac{|u(x,t)|^{p-1}}{|x-x_0|^{N+sp}}\,dx\,dt,$$9 control in time, whereas here $t^2$0 suffices (Kar, 11 Feb 2026).

A further Heisenberg-group result proves local boundedness for quasilinear equations under the time-integrability condition

$t^2$1

It defines the time-slice tail

$t^2$2

and absorbs the far-field contribution through the accumulated level

$t^2$3

The paper describes this as optimal because it replaces stronger $t^2$4-type tail assumptions by the natural $t^2$5-in-time condition dictated by the equation (Kar et al., 9 Apr 2025).

Across these nonlocal papers, the parabolic tail condition is therefore not an asymptotic statement but a regularity hypothesis and bookkeeping device. It quantifies the nonlocal exterior forcing that survives localization and determines whether intrinsic De Giorgi iteration can be closed.

5. Nearby notions that are explicitly not parabolic tail conditions

Several papers discuss objects that are parabolic and multiscale but explicitly distinguish them from tail conditions.

For the normalized $t^2$6-parabolic equation

$t^2$7

the main exterior regularity criterion is the tusk condition, a parabolic analogue of the exterior cone condition. At $t^2$8, a tusk is

$t^2$9

and if $x^{3/2}$0, then $x^{3/2}$1 is regular. The paper also notes a variant consisting of a union of geometrically spaced ellipses

$x^{3/2}$2

with $x^{3/2}$3, $x^{3/2}$4, $x^{3/2}$5, $x^{3/2}$6. It remarks that this is the closest analogue to a multiscale “tail of holes” condition, but also states unambiguously that the paper does not formulate a condition explicitly called a parabolic tail condition (Björn et al., 2017).

For divergence-form operators

$x^{3/2}$7

another nearby notion is the small parabolic Carleson condition on the coefficients,

$x^{3/2}$8

or the oscillation version

$x^{3/2}$9

The paper explicitly calls this a natural Carleson condition and a parabolic analog of the DKP condition, and proves that for every $N=\max_{t\ge 0}\{W(t)-t^2\},$0, the $N=\max_{t\ge 0}\{W(t)-t^2\},$1 Neumann problem is solvable provided both the Carleson norm and the Lipschitz constant are sufficiently small. It also states that this is not an explicit far-field tail condition: it is a multiscale near-boundary accumulation condition on coefficient roughness (Dindoš et al., 8 Jun 2026).

These examples are important terminologically. They show that not every parabolic multiscale condition is a tail condition, even when it has a cumulative-across-scales structure.

6. Structural themes and comparative interpretation

Despite the diversity of meanings, several structural themes recur.

A first theme is parabolic scaling. In Brownian motion minus a parabola, the competition between Brownian fluctuations of size $N=\max_{t\ge 0}\{W(t)-t^2\},$2 and the deterministic penalty $N=\max_{t\ge 0}\{W(t)-t^2\},$3 yields the rate $N=\max_{t\ge 0}\{W(t)-t^2\},$4. In nonlocal PDE, the relevant cylinders have time scale $N=\max_{t\ge 0}\{W(t)-t^2\},$5, and the tail is averaged over a backward interval whose length is matched to the intrinsic diffusion scale (Groeneboom et al., 2010, Liao, 2024).

A second theme is local–global competition. In the Brownian and KPZ settings, a local observable—the maximum or one-point height—depends on a global optimization in time or on a global optimal fluctuation. In nonlocal parabolic PDE, local boundedness or continuity inside a ball depends on values outside the ball, and the tail functional records exactly that dependence (Kamenev et al., 2016, Kar et al., 9 Apr 2025).

A third theme is asymmetry created by parabolic structure. For Brownian motion minus $N=\max_{t\ge 0}\{W(t)-t^2\},$6, the quadratic penalty replaces Gaussian or exponential decay by a stretched-exponential law with exponent $N=\max_{t\ge 0}\{W(t)-t^2\},$7. For KPZ started from a parabola, the left and right tails become sharply different: the negative tail is universal and has exponent $N=\max_{t\ge 0}\{W(t)-t^2\},$8, while the positive tail has exponent $N=\max_{t\ge 0}\{W(t)-t^2\},$9 and retains memory of the initial curvature through $N=\max_{t\ge 0}\{W(t)-t^2\},$00 (Groeneboom et al., 2010, Kamenev et al., 2016).

A fourth theme is optimality of time integrability. In the nonlocal regularity literature, “optimal tail condition” refers not to the spatial weight but to the weakest time integrability presently sufficient to close the argument. In Euclidean and Heisenberg settings, the shift from $N=\max_{t\ge 0}\{W(t)-t^2\},$01 to $N=\max_{t\ge 0}\{W(t)-t^2\},$02 or $N=\max_{t\ge 0}\{W(t)-t^2\},$03 is presented as the main refinement (Kar, 11 Feb 2026, Kar et al., 9 Apr 2025).

Taken together, these results indicate that parabolic tail condition is best understood as a family of technically precise conditions tied to parabolic scaling, not as a single universal definition. In stochastic asymptotics it denotes a specific stretched-exponential law generated by a quadratic penalty; in nonlocal parabolic PDE it denotes a backward-in-time exterior influence functional controlling regularity; and in adjacent areas, related geometric or Carleson conditions may play analogous multiscale roles without themselves being tail conditions in the strict sense.