Codimension-1 Collapse: Geometric & Dynamic Insights

• Codimension-1 collapse is a process where a family of objects systematically degenerates to lose exactly one dimension while preserving controlled curvature and volume bounds.
• In spectral theory, the collapse leads to divergent non-invariant eigenmodes and convergent S1-invariant modes, demonstrating precise geometric and analytic transitions.
• Applications span from constructing Ricci-flat gravitational instantons and topological reconstructions to analyzing dynamical bifurcations with explicit metrics and stability indicators.

A codimension-1 collapse refers to the systematic degeneration of a family of geometric, topological, or dynamical objects such that their limiting configuration loses exactly one dimension, often under curvature or structural constraints. In geometry, this notion arises within the theory of collapsed Riemannian manifolds under bounded curvature, singular degenerations of Ricci-flat metrics, spectral theory on degenerating bundles, and in topological simplification of complexes. In applied mathematics, “codimension-1 collapse” also denotes dynamical, often stochastic, reductions at critical bifurcation points where a single parameter drives system instability.

1. Geometric Characterization of Codimension-1 Collapse

A central context for codimension-1 collapse is the study of sequences of closed $n$-dimensional Riemannian manifolds $(M_i, g_i)$ with uniformly bounded curvature $|\sec^{M_i}| \leq 1$ and bounded diameter $\operatorname{diam}(M_i)\le D$, converging (in the Gromov–Hausdorff sense) to a compact metric space $Y$ with Hausdorff dimension $n-1$. The defining feature is that at each point $x \in M_i$, the quantity

$$\frac{\operatorname{vol}(B^{M_i}_r(x))}{\operatorname{inj}^{M_i}(x)}$$

admits a uniform lower bound for some fixed $r>0$. This precisely characterizes codimension-1 collapse: such a bound holds if and only if the limiting dimension drops by at most one (i.e., $\dim_{\textrm{Haus}}(Y)\geq n-1$) (Roos, 2017).

A prominent structure theorem states that in the codimension-1 regime, outside a singular subset, $(M_i, g_i)$0 admits, for large $(M_i, g_i)$1, the structure of a principal or orbifold $(M_i, g_i)$2-bundle over the $(M_i, g_i)$3-dimensional orbifold $(M_i, g_i)$4, with collapsed fiber length tending to zero. The corresponding fiber and total space injectivity radii become asymptotically equivalent as the collapse deepens: $(M_i, g_i)$5 with $(M_i, g_i)$6. The volume and curvature of any Gromov–Hausdorff limit space emerging from such sequences retain uniform lower (volume) and upper (curvature) bounds (Roos, 2017).

2. Spectral Theory Under Codimension-1 Collapse

In spin geometry, codimension-1 collapse imposes dramatic effects on the Dirac spectrum. Given a sequence of spin manifolds $(M_i, g_i)$7 with symmetrically bounded $(M_i, g_i)$8 potentials $(M_i, g_i)$9 and a collapse along $|\sec^{M_i}| \leq 1$0-fibers to an orbifold $|\sec^{M_i}| \leq 1$1, the Dirac operator $|\sec^{M_i}| \leq 1$2 exhibits spectral bifurcation:

• For non-invariant fiber modes ($|\sec^{M_i}| \leq 1$3 in the Fourier decomposition), all eigenvalues diverge:

$|\sec^{M_i}| \leq 1$4

• The $|\sec^{M_i}| \leq 1$5-invariant sector ($|\sec^{M_i}| \leq 1$6) converges to the spectrum of an explicitly determined first-order differential operator $|\sec^{M_i}| \leq 1$7 on the base $|\sec^{M_i}| \leq 1$8, perturbed by

$|\sec^{M_i}| \leq 1$9

where $\operatorname{diam}(M_i)\le D$0 captures the limiting curvature of the connection and $\operatorname{diam}(M_i)\le D$1 is the $\operatorname{diam}(M_i)\le D$2-invariant part of the limiting $\operatorname{diam}(M_i)\le D$3-potential (Roos, 2017).

If $\operatorname{diam}(M_i)\le D$4 is even-dimensional orientable, $\operatorname{diam}(M_i)\le D$5 is the standard Dirac operator; in the odd-dimensional case, the limiting operator splits as $\operatorname{diam}(M_i)\le D$6 plus a similar perturbation. The $\operatorname{diam}(M_i)\le D$7-structure of potentials descends with no loss of regularity in the limit. This underpins the spectral stability and explicit degeneration in Dirac-type operators under codimension-1 geometric collapse.

3. Codimension-1 Collapse in Gravitational Instantons

Singular degeneration via codimension-1 collapse plays a core role in the construction of gravitational instantons with non-maximal volume growth. For four-dimensional Ricci-flat (hyperkähler) spaces such as those of ALF, ALG, and ALH-type, families of metrics $\operatorname{diam}(M_i)\le D$8 can be built in which the circle-fiber length is $\operatorname{diam}(M_i)\le D$9 and the entire space Gromov–Hausdorff-converges to a 3-orbifold quotient $Y$0 as $Y$1 (Salm, 2024).

• The metric takes the Gibbons–Hawking ansatz: $Y$2.
• Away from finitely many singular points, local volumes shrink proportionally to $Y$3, reflecting collapse only along the circle direction.
• At exceptional loci (fixed points under symmetries or isolated sources), the geometry bubbles off standard noncollapsed models (Atiyah–Hitchin and Taub–NUT spaces), ensuring completeness and precise control over curvature and topology.
• Resulting instantons interpolate between collapsed 3-orbifolds and the original 4-manifolds, and their homological invariants and moduli space structure are computable in terms of the combinatorics of gluing and the base orbifold (Salm, 2024).

This framework systematically elucidates degenerations at the “boundary” of instanton moduli spaces.

4. Codimension-1 Collapse in Simplicial and Topological Reconstruction

In computational geometry, codimension-1 collapse underpins algorithms for reconstructing smooth embedded submanifolds from finite point samples via $Y$4-complex simplification (Attali et al., 2024). Given a sample $Y$5 of a codimension-1 manifold $Y$6, the $Y$7-complex may contain surplus $Y$8-simplices (“slivers”) obstructing manifold triangulation.

A specific collapse, termed "vertical collapse" relative to $Y$9, removes facets whose projection along the normal direction does not locally cover $n-1$0. The Naive Squash (and its practical variant) algorithm iteratively collapses free codimension-1 faces and their incident $n-1$1-simplices, subject to angle and convexity constraints, until only a pure $n-1$2-dimensional complex homeomorphic to $n-1$3 remains. Quantitative sampling and angle conditions (e.g., $n-1$4) guarantee generic correctness and outperform previous Delaunay-type reconstruction bounds (Attali et al., 2024).

Relatedly, under these conditions, even the restricted Delaunay complex becomes a manifold triangulation for generic samples, with codimension-1 collapse providing the mechanism for abstracting the ambient bulk to the underlying manifold.

5. Dynamical Systems and Saddle-Node Bifurcation

Outside geometry, codimension-1 collapse appears in the context of dynamical bifurcations, particularly in power system stability analysis. When a high-dimensional system approaches a codimension-1 saddle-node bifurcation, the Jacobian matrix of its linearized dynamics acquires a single vanishing real eigenvalue $n-1$5 while others remain regular. Near such an instability,

$n-1$6

where $n-1$7 is the right and $n-1$8 the left null eigenvector.

Critical slowing down and an increased likelihood of transition to system collapse ensue, with mean clearing time (collapse probability) given asymptotically by Kramers’ law,

$n-1$9

where $x \in M_i$0 encodes stochastic load fluctuation, $x \in M_i$1 damping, and $x \in M_i$2 captures nonlinearity (Podolsky et al., 2012). The real-time indicator

$x \in M_i$3

serves as a robust control metric for quantifying proximity to large-scale system collapse, with control strategies designed to maximize $x \in M_i$4 and thereby stabilize the system.

6. Applications and Interdisciplinary Relevance

Codimension-1 collapse is fundamental in:

• The analysis of collapsed Riemannian geometries and moduli space boundaries, particularly in Ricci-flat and Einstein metrics (Salm, 2024).
• Spectral geometry, where the fate of eigenvalues under geometric collapse constrains analytic invariants and index theory (Roos, 2017).
• Topological data analysis and manifold learning, serving as a complexity-reduction tool transforming high-dimensional combinatorial complexes into faithful triangulations of unknown lower-dimensional structures (Attali et al., 2024).
• Power engineering and nonlinear dynamical systems, where it codifies the onset of global instability and blackout phenomena under stochastic perturbation (Podolsky et al., 2012).

A recurring theme is that codimension-1 collapse provides a sharp boundary between topological, geometric, or dynamical regimes, with explicit quantitative criteria governing convergence, regularity, and spectral or stability transitions.

7. Summary Table: Key Contexts of Codimension-1 Collapse

Domain	Collapse Structure	Limiting Object
Riemannian Geometry	$x \in M_i$5 fiber bundle with shrinking fiber	$x \in M_i$6-orbifold
Spectral Theory	Dirac operator eigenmode filtration	Operator on base manifold
Ricci-flat Instantons	Gibbons–Hawking circle reduction	3-orbifold, with glued singularities
Topological Data	α-complex/Simplicial collapse	$x \in M_i$7-dim. complex homeomorphic to $x \in M_i$8
Dynamical Systems	1D center manifold at bifurcation	Critical scalar reduction

Codimension-1 collapse thus constitutes a unifying principle across pure and applied mathematics, capturing degeneration phenomena where exactly one dimension is lost, often under precise geometric, analytic, or combinatorial control.