Almost-everywhere energy and moment variation for building-valued harmonic maps
OpenHarmonicBuilding.weakFirstVariationDataLet be a continuous, nonconstant Korevaar-Schoen energy-minimizing map from a connected Riemann-surface domain to a complete Euclidean building. Fix a point where the canonical small-radius energy is finite and the boundary moment is positive. In the preferred complex coordinate, let be the energy in the radius- ball and the arclength integral of the squared distance from the central value.
There is , and real functions on the radii, such that and are absolutely continuous on every compact subinterval of and, for almost every such radius,
Here and encode the boundary radial flux and radial energy. These data supply the analytic input for the two-dimensional frequency monotonicity formula.
Formalization Note. The energy and boundary moment are the canonical functions already defined in the harmonic-building interface. No differentiability at exceptional radii is asserted.
import Definitions.Def_frame_2026_harmonic_building_interfaces open HarmonicBuilding Set MeasureTheory Filter open scoped Manifold set_option autoImplicit false
theorem HarmonicBuilding.weakFirstVariationData
{N : ℕ} (C : EuclideanCoxeterData N)
{X S : Type*}
[MetricSpace X] [CompleteSpace X] [MeasurableSpace X] [BorelSpace X]
[TopologicalSpace S] [T2Space S] [SecondCountableTopology S]
[ChartedSpace ℂ S] [IsManifold (modelWithCornersSelf ℂ ℂ) ⊤ S]
(B : EuclideanBuildingData N C X)
(D : RiemannSurfaceDomain S) (u : S → X) (x₀ : D.Point)
(hu : IsKSHarmonic D u) (hnc : NonconstantOn u D.carrier)
(hdef : OrderDefinedAt D u x₀) :
let E : ℝ → ℝ := fun r => scaleEnergy (coordinateDomainAt D (x₀ : S))
(coordinateMapAt u (x₀ : S)) (coordinateCenter (x₀ : S)) r
let I : ℝ → ℝ := fun r => (boundaryMoment (coordinateMapAt u (x₀ : S))
(coordinateCenter (x₀ : S)) r).toReal
∃ r₀ : ℝ, 0 < r₀ ∧ ∃ J F : ℝ → ℝ,
(∀ a b : ℝ, 0 < a → a ≤ b → b < r₀ →
AbsolutelyContinuousOnInterval E a b ∧ AbsolutelyContinuousOnInterval I a b) ∧
∀ᵐ r : ℝ, r ∈ Set.Ioo (0 : ℝ) r₀ →
HasDerivAt I (I r / r + 2 * J r) r ∧
HasDerivAt E (2 * F r) r ∧ E r ≤ J r ∧ J r ^ 2 ≤ I r * F r := by sorry