Dirichlet Parabolicity: Theory & Applications

• Dirichlet parabolicity is a framework ensuring that bounded harmonic functions are uniquely determined by prescribed boundary or exterior data.
• It connects potential theory, p-parabolicity, and nonlocal parabolic problems using tools like vanishing capacity, absence of positive Green functions, and maximum principles.
• This concept underpins geometric analysis and quantitative solvability criteria for classical and modern boundary value problems.

Dirichlet parabolicity is a family of closely related notions at the intersection of potential theory, geometric analysis, and parabolic partial differential equations. In one classical sense, for a manifold with boundary it expresses the uniqueness of bounded harmonic functions from their boundary values; in $p$-potential theory it is encoded by vanishing capacity, the absence of a positive Green function, and comparison principles on exterior domains; in nonlocal and time-dependent PDE it refers to well-posedness of Dirichlet problems, sometimes with exterior data rather than boundary traces; and in modern boundary regularity theory it is tied to quantitative solvability of parabolic Dirichlet problems through $A_\infty$ properties of parabolic or caloric measure, reverse Hölder estimates for Poisson kernels, and geometric conditions such as parabolic uniform rectifiability (Pessoa et al., 2016, Aiolfi et al., 2021, Felsinger et al., 2013, Auscher et al., 2016, Bortz et al., 2023).

1. Potential-theoretic meaning on manifolds and domains

A classical boundary-based formulation appears for manifolds with nonempty boundary: a manifold $M$ is parabolic if and only if every bounded harmonic function on $M$ is determined by its boundary values; equivalently, if $f_1,f_2$ are bounded harmonic and $f_1=f_2$ on $\partial M$, then $f_1=f_2$ on $M$ (Prete et al., 2023). In the same direction, a smooth Riemannian manifold with boundary is called Dirichlet parabolic, or $D$-parabolic, if every bounded $A_\infty$0 satisfying

$A_\infty$1

vanishes identically (Pessoa et al., 2016).

This notion is equivalent to several global maximum principles. One formulation states that $A_\infty$2 is $A_\infty$3-parabolic if and only if, for every bounded harmonic $A_\infty$4,

$A_\infty$5

A subharmonic version asserts that for every domain $A_\infty$6 and every bounded $A_\infty$7 with $A_\infty$8 on $A_\infty$9,

$M$0

The same paper gives an exhaustion characterization via harmonic functions $M$1 on relatively compact domains and a Khas’minskii-type test using a function $M$2 that diverges at infinity and is weakly superharmonic outside a compact set (Pessoa et al., 2016).

The potential-theoretic formulation is also linked to Green kernels. For manifolds without boundary, the papers on $M$3-parabolicity define parabolicity by the nonexistence of a positive Green function; in the Dirichlet setting with boundary, the Dirichlet Green kernel $M$4 controls the associated Dirichlet $M$5-Liouville property. A point emphasized in the boundary case is that $M$6-parabolicity and $M$7-$M$8-Liouville are distinct notions: $M$9-parabolicity concerns uniqueness of bounded harmonic functions vanishing on $M$0, whereas $M$1-$M$2-Liouville is characterized by non-integrability of $M$3, equivalently by divergence of the global Dirichlet mean exit time $M$4 (Pessoa et al., 2016).

2. $M$5-parabolicity, capacity, and boundary behavior at infinity

For complete Riemannian manifolds, the nonlinear extension is $M$6-parabolicity. The $M$7-Laplacian is

$M$8

and a complete manifold is $M$9-parabolic if it does not admit a positive Green function for $f_1,f_2$0. The paper on equivalences among parabolicity, comparison principle, and capacity proves that the following are equivalent: $f_1,f_2$1 is $f_1,f_2$2-parabolic; the $f_1,f_2$3-capacity of every compact subset is zero; the $f_1,f_2$4-capacity of some precompact open set is zero; and every bounded from below supersolution of $f_1,f_2$5 is constant (Aiolfi et al., 2021).

That work also identifies Dirichlet-type uniqueness on exterior domains as an equivalent formulation. For an exterior domain $f_1,f_2$6, $f_1,f_2$7 satisfies the comparison principle for the $f_1,f_2$8-Laplace operator if, whenever $f_1,f_2$9 are bounded sub- and supersolutions and $f_1=f_2$0, then $f_1=f_2$1 in $f_1=f_2$2. The paper proves that $f_1=f_2$3 is $f_1=f_2$4-parabolic if and only if this comparison principle holds for exterior $f_1=f_2$5-harmonic problems, and extends the conclusion to more general divergence-form operators $f_1=f_2$6 under $f_1=f_2$7-growth and monotonicity conditions (Aiolfi et al., 2021).

A complementary metric-space formulation appears in the sphericalization approach to unbounded uniform domains. There the Dirichlet problem for $f_1=f_2$8-harmonic functions is transferred to a bounded sphericalized space, and the point at infinity becomes a boundary point. In that framework, Dirichlet parabolicity is the regime in which the boundary point $f_1=f_2$9 created by sphericalization has zero $\partial M$0-capacity. If $\partial M$1 has positive $\partial M$2-capacity, one enters the $\partial M$3-hyperbolic regime, where uniqueness can fail unless a value at infinity is prescribed; the paper then shows that for unbounded uniform domains with unbounded boundary, uniqueness still holds because zero-trace perturbations automatically have zero trace at infinity (Korte et al., 17 Feb 2026).

Curvature and volume growth supply geometric criteria. Under

$\partial M$4

outside a compact set, the cited paper proves $\partial M$5-parabolicity for every $\partial M$6. Under

$\partial M$7

it proves $\partial M$8-parabolicity for every

$\partial M$9

For $f_1=f_2$0, these statements imply classical Dirichlet parabolicity, hence recurrence, nonexistence of a positive Green function, and triviality of the bounded harmonic Dirichlet problem at infinity (Priebe et al., 2023).

3. Nonlocal parabolic Dirichlet problems and exterior data

In nonlocal analysis, the phrase has a more operational meaning. The paper on nonlocal operators studies

$f_1=f_2$1

with measurable, possibly nonsymmetric kernel $f_1=f_2$2, and formulates both elliptic and parabolic Dirichlet problems. The fundamental structural change is that data are prescribed on the complement $f_1=f_2$3, not on $f_1=f_2$4, because $f_1=f_2$5 depends on values of $f_1=f_2$6 for all $f_1=f_2$7 (Felsinger et al., 2013).

For the parabolic problem one prescribes

$f_1=f_2$8

The functional framework uses the energy space $f_1=f_2$9, the Gelfand triple

$M$0

and

$M$1

For kernels $M$2 with measurable $M$3, standard abstract evolution theory yields a unique solution $M$4 and the energy estimate

$M$5

The same paper explicitly states that it does not introduce a formal “Dirichlet parabolicity” definition in the classical potential-theoretic sense; rather, it establishes parabolic Dirichlet well-posedness for nonlocal operators through existence, uniqueness, and stability in Hilbert spaces (Felsinger et al., 2013).

A related degenerate fractional setting appears for

$M$6

on a bounded domain with Dirichlet boundary data on $M$7. There the regional fractional Laplacian is chosen because constants lie in its kernel and a Green formula with a fractional normal derivative is available. The resulting weak entropy solutions satisfy an $M$8-type contraction within the perturbative class, again realizing Dirichlet parabolicity primarily as well-posedness with boundary data in a nonlocal parabolic equation (Huaroto et al., 2022).

4. Quantitative solvability for local parabolic operators

For divergence-form parabolic operators, Dirichlet parabolicity often means quantitative solvability of boundary value problems. In the parabolic upper half-space

$M$9

the operator

$D$0

with real, bounded, measurable, uniformly elliptic, not necessarily symmetric coefficients has parabolic measure $D$1. The main theorem in the cited work states that $D$2, where $D$3 on $D$4; equivalently, there exists $D$5 such that $D$6, or, equivalently again, the Dirichlet problem $D$7 with $D$8 is solvable and

$D$9

The proof reduces $A_\infty$00 to a Carleson measure estimate in sawtooth regions and uses parabolic functional calculus, non-tangential maximal estimates, and a key lemma controlling

$A_\infty$01

by $A_\infty$02 (Auscher et al., 2016).

A different but related line treats continuous and Hölder Dirichlet problems on rough space-time domains. For

$A_\infty$03

with bounded measurable uniformly elliptic coefficients, the paper introduces the time-backwards capacity density condition (TBCDC) and the time-backwards Hausdorff content condition (TBHCC). Under these conditions it proves existence of parabolic measure even for unbounded domains, Bourgain-type non-degeneracy,

$A_\infty$04

boundary Hölder decay, and well-posedness of the continuous and Hölder Dirichlet problems. One of its explicit points is that the operator dependence of the parabolic Wiener criterion is unavoidable, a feature tied to the one-sided time geometry of parabolic equations (Hidalgo-Palencia et al., 6 Oct 2025).

For hypoelliptic evolution equations, the Perron–Wiener construction and a cone-type regularity criterion play the analogous role. There the Dirichlet problem

$A_\infty$05

is solved in a Doob $A_\infty$06-harmonic space framework, and an exterior intrinsic parabolic cone guarantees boundary regularity. This extends the Effros–Kazdan parabolic-cone criterion from the heat operator to a broad class of hypoelliptic operators (Kogoj, 2016).

5. Caloric measure, $A_\infty$07 solvability, and parabolic uniform rectifiability

In recent boundary regularity theory, Dirichlet parabolicity is increasingly encoded through quantitative absolute continuity of caloric measure. For a parabolic Lipschitz graph domain

$A_\infty$08

the heat equation $A_\infty$09 has caloric measure $A_\infty$10. The central equivalence proved for such domains is that the following are equivalent: $A_\infty$11 is parabolic $A_\infty$12 with respect to surface measure $A_\infty$13; the graph function satisfies

$A_\infty$14

the boundary is parabolic uniformly rectifiable; and the $A_\infty$15 Dirichlet problem is solvable for some finite $A_\infty$16 (Bortz et al., 2023).

The necessity direction is proved by analyzing level sets of the Green function as approximate boundary graphs and establishing a Littlewood–Paley square function estimate of the form

$A_\infty$17

This yields $A_\infty$18, hence parabolic uniform rectifiability (Bortz et al., 2023).

Variable-coefficient analogues extend this picture. For

$A_\infty$19

in a graph domain, with $A_\infty$20 satisfying an $A_\infty$21 Carleson oscillation condition, the paper shows that $A_\infty$22-solvability of the Dirichlet problem for $A_\infty$23 and $A_\infty$24 implies

$A_\infty$25

equivalently that the boundary is parabolic uniformly rectifiable; in the symmetric case, solvability for $A_\infty$26 alone suffices (Bortz et al., 2 Mar 2025).

At the level of minimal geometric assumptions, the one-phase free-boundary result in (Bortz et al., 24 Oct 2025) proves that if $A_\infty$27 is time-symmetrically parabolic ADR, $A_\infty$28 has interior corkscrews, and caloric measure satisfies local weak-$A_\infty$29, then $A_\infty$30 is parabolic uniformly rectifiable. The paper explicitly presents this as identifying parabolic uniform rectifiability as the correct geometric framework for $A_\infty$31 solvability of the Dirichlet problem for the heat equation (Bortz et al., 24 Oct 2025).

6. Geometric incarnations on surfaces, submanifolds, and singular spaces

A geometric meaning of Dirichlet parabolicity appears for invariant surfaces. If $A_\infty$32 is a complete immersed surface invariant under a complete Killing vector field $A_\infty$33, and $A_\infty$34 is a complete curve orthogonal to $A_\infty$35, then $A_\infty$36 is parabolic if and only if either $A_\infty$37 is compact, or the integral curves of $A_\infty$38 are compact and

$A_\infty$39

or the integral curves are noncompact and the analogous two-sided divergence condition holds at $A_\infty$40. The paper also states explicitly that, for a manifold with nonempty boundary, parabolicity is equivalent to uniqueness of bounded harmonic functions from boundary data (Prete et al., 2023).

For minimal surfaces with boundary, Dirichlet parabolicity is expressed probabilistically: Brownian motion hits the boundary almost surely, bounded harmonic functions are determined by their boundary values, and no positive Green’s function exists. The stochastic comparison method in the cited paper proves parabolicity for stochastically complete minimal surfaces contained in regions of the form

$A_\infty$41

and quadratic area growth under the slightly smaller profile

$A_\infty$42

The resulting parabolicity is exactly Dirichlet uniqueness for bounded harmonic functions on the surface with boundary (Neel, 2010).

Weighted analogues replace $A_\infty$43 by the drifted Laplacian $A_\infty$44. For submanifolds in weighted model manifolds, $A_\infty$45-parabolicity is characterized by vanishing $A_\infty$46-capacity and by the criterion

$A_\infty$47

for radial weights $A_\infty$48. Capacity comparison for extrinsic balls then yields parabolicity or hyperbolicity of properly immersed submanifolds with controlled $A_\infty$49-mean curvature (Hurtado et al., 2018).

Finally, singular complex geometry supplies a local version at singular sets. The regular locus $A_\infty$50 of a complex variety is locally parabolic at $A_\infty$51, and if $A_\infty$52 is compact then $A_\infty$53 is parabolic. The construction uses explicit exhaustion functions of log–log type with $A_\infty$54-gradient control near the singular set, producing cutoffs of arbitrarily small Dirichlet energy. A direct consequence is that bounded $A_\infty$55-forms belong to the minimal domain $A_\infty$56, since the singular set carries no extra Dirichlet boundary contribution in the $A_\infty$57 theory (Ruppenthal, 2014).

Across these settings, Dirichlet parabolicity consistently marks the collapse of nontrivial boundary or infinity data into uniqueness, recurrence, or vanishing capacity. What changes from one field to another is the analytic mechanism: Green kernels and maximum principles in potential theory, Gelfand triples and coercive forms in nonlocal evolution equations, parabolic measure and $A_\infty$58 estimates in rough-boundary PDE, and geometric integral tests in the study of surfaces, submanifolds, and singular spaces.