Monge Gauge Parametrization

• Monge gauge parametrization is a method representing surfaces as height functions over a reference domain, with metrics explicitly derived from the function's derivatives.
• Generalizations such as spherical and local Monge gauges extend the approach to curved and complex surfaces, facilitating detailed tensor analysis and geometric operator formulation.
• Applications span computational physics, quantum models, and differential geometry by enabling efficient tensor approximations and preserving tangency on embedded manifolds.

The Monge gauge parametrization is a foundational technique in differential geometry for representing two-dimensional surfaces as graphs of scalar functions—height functions—over a reference manifold. Its origins lie in the classical “Monge form,” where a surface is realized as a height field above a supporting plane, but extensive generalizations encompass local and global parameterizations over both flat and curved base surfaces. This approach underpins the formulation of geometric operators on embedded manifolds, the approximation of tensor fields for computational physics, the study of constrained quantum systems, and the analytical classification of geometric structures.

1. Classical Monge Gauge: Parametrization, Metric, and Geometry

The classical Monge gauge describes a surface as the graph of a function $z = f(x, y)$ over a planar domain $D \subset \mathbb{R}^2$. For surfaces such as the upper hemisphere of a sphere of radius $r$ in $\mathbb{R}^3$, the Monge representation is

$$z = f(x, y) = \sqrt{r^2 - x^2 - y^2}, \qquad (x, y) \in D: x^2 + y^2 \leq r^2.$$

With this parametrization, the induced metric is given by

$$g_{ij} = \delta_{ij} + f_{,i} f_{,j}, \quad i,j \in \{x,y\},$$

which for the sphere results in

$$g_{xx} = 1 + \frac{x^2}{D}, \quad g_{yy} = 1 + \frac{y^2}{D}, \quad g_{xy} = \frac{xy}{D}, \quad D = r^2 - x^2 - y^2.$$

The determinant and inverse of the metric are

$$\det g = \frac{r^2}{D}, \quad g^{ij} = \frac{1}{r^2} \begin{pmatrix} r^2 - x^2 & -x y \ -x y & r^2 - y^2 \end{pmatrix}.$$

The unit normal vector field is

$$n(x, y) = \left( \frac{x}{r},\, \frac{y}{r},\, \frac{\sqrt{D}}{r} \right).$$

This parametrization facilitates the explicit calculation of intrinsic and extrinsic geometric quantities, such as the surface Laplacian and mean curvature, in terms of derivatives of $f(x, y)$ (Xun et al., 2012).

2. Generalizations: Spherical and Local Monge Gauges

Generalizations of the Monge gauge allow the height function to be defined above non-planar reference surfaces. For example, a “spherical Monge gauge” represents a surface via a radial height function $h(\theta, \varphi)$ over a sphere of radius $R$ (Mazharimousavi et al., 2017): $$\Sigma:\ \quad F(r, \theta, \varphi) = r - h(\theta, \varphi) = 0,$$ with

$$x = h(\theta, \varphi) \sin\theta \cos\varphi,\quad y = h(\theta, \varphi) \sin\theta \sin\varphi,\quad z = h(\theta, \varphi) \cos\theta.$$

For arbitrary surfaces, the Local Monge Parametrization (LMP) constructs local charts around each node $X_0$ of a mesh, expressing the neighborhood as a graph over a supporting plane $\Pi$: $X(u, v) = X_0 + u\,e_1 + v\,e_2 + h(u, v)\, n,$ where $n$ is the unit normal at $X_0$ (Torres-Sánchez et al., 2019). This approach enables patchwise representations suitable for high-fidelity computations on complex topologies without global coordinate charts or harmonic decompositions.

3. Geometric Operators in the Monge Gauge

In the Monge gauge, the transition from bulk differential operators to their surface-restricted analogs is systematic. For the classical Monge case, the surface gradient and Laplacian are given by

$$\nabla_S = r^x \partial_x + r^y \partial_y, \quad \Delta_S = \frac{1}{\sqrt{g}} \partial_i\left( \sqrt{g} g^{ij} \partial_j \right),$$

where $r_i = \partial_i (x, y, f(x, y))$, and $r^i = g^{ij} r_j$ (Xun et al., 2012).

For quantum mechanics on surfaces, the geometric momentum operator in the Monge gauge takes the form

$$\mathbf{p}_g = -i\hbar \left( \nabla_S + M n \right),$$

with $M$ the mean curvature. On the sphere, explicit expressions for $p_x, p_y, p_z$ are derived in terms of coordinates and derivatives, ensuring self-adjointness and invariance under smooth coordinate changes.

For height functions $h(\theta, \varphi)$ above a sphere, the induced metric, second fundamental form, and curvatures are all explicitly available in terms of the derivatives of $h$ (Mazharimousavi et al., 2017). Flat-space limits recover the classical formulas.

4. Monge Gauge in Global, Local, and Computational Contexts

The Monge parametrization’s local version is exploited for computational purposes in finite element and NURBS discretizations. The LMP method constructs Monge charts at every node, with height functions determined by least-squares polynomial fits to sampled points from the global parametrization. This enables efficient and accurate pull-back and push-forward of tensor fields between the Monge basis and the underlying surface element coordinates. Change-of-basis matrices $T_{[E,I]}^A{}_a$ and their inverses are computed for coordinate transitions (Torres-Sánchez et al., 2019).

Discrete fields—vector fields, 1-forms, and higher-rank tensors—are represented minimally, with the number of degrees of freedom per node matching the intrinsic tangent or tensorial dimension. This circumvents the excess DOF required by global potential methods, and avoids “extraneous constraints or potentials.” The method is robust across arbitrarily complex surfaces, including tori and higher genus geometries.

The table below outlines Monge gauge approaches for different contexts:

Context	Reference Chart	Parametrization
Classical Monge	Planar $(x, y)$	$z = f(x, y)$
Spherical Monge	Spherical $(\theta,\varphi)$	$r = h(\theta, \varphi)$
Local Monge (LMP)	Local plane at $X_0$	$X(u, v) = X_0 + u e_1 + v e_2 + h(u, v) n$

5. Applications: Tensor Approximation and Physical Models

The Monge gauge and its local generalizations enable direct numerical solution of tensor-valued PDEs on arbitrary surfaces (Torres-Sánchez et al., 2019). Key applications include:

• Approximation of tangential fields: LMP achieves optimal ($\sim$3rd–4th order) convergence in $L^2$ for vector and tensor fields on spheres and tori, surpassing Hodge-based approaches, particularly on non-simply-connected topologies.
• Marangoni flow modeling: Surface-tension-driven vector PDEs discretized with the Monge framework yield robust, optimal convergence with minimal DOF.
• Nematic ordering: The evolution and minimization of symmetric traceless 2-tensor fields $Q_{ab}$ on surfaces, governed by Rayleighian variational principles, are tractable with local Monge representation.
• Preservation of tangency and optimality: The Monge gauge ensures that vector/tensor fields remain confined to the tangent bundle globally, without the need for Lagrange multipliers or penalty methods.

6. Advanced Generalizations and Gauge Structures

Monge parametrizations admit further structural generalizations. For rolling distributions, explicit Monge normal forms can be established for Pfaffian systems describing nonholonomic constraints, such as the rolling of two hyperboloids. Here, five-dimensional Pfaffian systems are converted to Monge normal forms involving explicit functions $F(x, y, p, q)$, with residual projective-type gauge symmetries preserving the functional form up to scale (Randall, 2021). The manuscript provides algebraic manipulations to bring coordinate and one-form data into Monge normal form and catalogues the associated gauge redundancies.

The Monge gauge is not unique: transformations such as $$(x, y, p, q) \mapsto (\lambda^2 x, \lambda y, \lambda^{-1} p, q)$$ and involutive exchanges generate a gauge group that leaves the distribution invariant. In the context of geometric analysis, this structure underpins the classification and equivalence of complex distributions.

7. Noncanonical Commutators, Constraints, and Quantum Geometry

On embedded surfaces, classical canonical commutation relations are deformed by the presence of holonomic constraints. For a surface defined by the constraint $\phi(x) = x^2 + y^2 + z^2 - r^2 = 0$, the Dirac bracket construction yields

$$\{ x_i, p_j \}_D = \delta_{ij} - \frac{x_i x_j}{r^2},$$

which upon quantization gives the non-canonical commutator

$$[\, x_i, p_j \,] = i\hbar \left( \delta_{ij} - \frac{x_i x_j}{r^2} \right),$$

revealing that geometric momentum operators are not canonical, but remain geometric invariants under coordinate transformations. These operators are manifestly self-adjoint in the surface Hilbert space $L^2(D, \sqrt{g} dx dy)$ and preserve invariance under reparametrizations (Xun et al., 2012).

---

The Monge gauge parametrization, in its various forms and extensions, remains central to the analytical and numerical treatment of surface geometry, tangential field discretization, and physical models across geometry, mechanics, and quantum theory. Foundational approaches such as those detailed in (Xun et al., 2012, Torres-Sánchez et al., 2019, Mazharimousavi et al., 2017), and (Randall, 2021) continue to provide explicit constructions and gauge-theoretic insights that underpin both theoretical analysis and computational methodologies.