Integration on Manifolds

Integration on manifolds generalizes classical line, surface, and volume integrals using differential forms. Stokes' theorem provides the grand unification, and de Rham cohomology captures the topological content of closed vs. exact forms.

Historical Context

The generalized Stokes theorem was developed through the work of Green (1828), Kelvin (1850), Stokes (1854), and Gauss, each proving special cases. The modern formulation using differential forms was given by Elie Cartan in the 1940s. Georges de Rham proved in 1931 that the cohomology defined by differential forms is isomorphic to singular cohomology, establishing the deep connection between analysis and topology that bears his name.

Derivation 1: Integration of Forms on Manifolds

An $n$-form $\omega$ can be integrated over an oriented $n$-manifold $M$. In a single chart $(U, \varphi)$, writing $\omega = f(x) \, dx^1 \wedge \cdots \wedge dx^n$:

$\int_U \omega = \int_{\varphi(U)} f(x^1, \ldots, x^n) \, dx^1 \cdots dx^n$

For multiple charts, use a partition of unity $\{\rho_\alpha\}$:

$\boxed{\int_M \omega = \sum_\alpha \int_{U_\alpha} \rho_\alpha \, \omega}$

Orientation and the Volume Form

An orientation on $M$ is a choice of a nowhere-vanishing $n$-form up to positive multiples. On a Riemannian manifold, the canonical choice is the Riemannian volume form:

$\text{vol}_g = \sqrt{\det(g_{ij})} \, dx^1 \wedge \cdots \wedge dx^n$

The integral $\int_M \text{vol}_g$ gives the volume of $M$. For the unit sphere,$\text{vol}_{S^2} = \sin\theta \, d\theta \wedge d\phi$ and $\text{Vol}(S^2) = 4\pi$.

Derivation 2: The Generalized Stokes Theorem

The generalized Stokes theorem is the single most important result in the calculus on manifolds:

$\boxed{\int_M d\omega = \int_{\partial M} \omega}$

where $M$ is a compact oriented $n$-manifold with boundary $\partial M$ (oriented by the outward normal convention) and $\omega$ is an $(n-1)$-form.

Proof Sketch

Step 1: Using a partition of unity, reduce to the case of a single chart. It suffices to prove the result for $\omega$ supported in a half-space $H^n = \{x \in \mathbb{R}^n : x^n \geq 0\}$.

Step 2: Write $\omega = \sum_i (-1)^{i-1} f_i \, dx^1 \wedge \cdots \wedge \widehat{dx^i} \wedge \cdots \wedge dx^n$. Then:

$d\omega = \left(\sum_i \frac{\partial f_i}{\partial x^i}\right) dx^1 \wedge \cdots \wedge dx^n$

Step 3: For each term, integrate by parts in the $x^i$ direction. For $i < n$, the boundary terms vanish (compact support). For $i = n$:

$\int_0^\infty \frac{\partial f_n}{\partial x^n} dx^n = -f_n(x^1, \ldots, x^{n-1}, 0)$

This gives precisely $\int_{\partial H^n} \omega$, completing the proof.

Classical Special Cases

Fundamental Theorem of Calculus ($n=1$)

$\int_a^b f'(x) \, dx = f(b) - f(a)$

Green's/Kelvin-Stokes ($n=2$)

$\iint_S (\nabla \times \mathbf{F}) \cdot d\mathbf{S} = \oint_{\partial S} \mathbf{F} \cdot d\mathbf{r}$

Gauss/Divergence ($n=3$)

$\iiint_V (\nabla \cdot \mathbf{F}) \, dV = \oiint_{\partial V} \mathbf{F} \cdot d\mathbf{S}$

Derivation 3: de Rham Cohomology

Since $d^2 = 0$, we have the chain complex $\Omega^0 \xrightarrow{d} \Omega^1 \xrightarrow{d} \cdots \xrightarrow{d} \Omega^n$. Define:

Closed forms: $Z^k(M) = \ker(d: \Omega^k \to \Omega^{k+1}) = \{\omega : d\omega = 0\}$

Exact forms: $B^k(M) = \text{im}(d: \Omega^{k-1} \to \Omega^k) = \{\omega = d\eta\}$

The $k$-th de Rham cohomology group is the quotient:

$\boxed{H^k_{dR}(M) = Z^k(M) / B^k(M)}$

The dimension $b_k = \dim H^k_{dR}(M)$ is the $k$-th Betti number. It counts the number of independent "holes" of dimension $k$ in $M$.

Examples

$S^1$: The angle form $d\theta$ is closed but not exact (its integral over $S^1$ is $2\pi \neq 0$). So $H^1_{dR}(S^1) \cong \mathbb{R}$, $b_1 = 1$.

$S^2$: $b_0 = 1$ (connected), $b_1 = 0$ (simply connected),$b_2 = 1$ (the volume form generates $H^2$). Euler characteristic$\chi = 1 - 0 + 1 = 2$.

Derivation 4: The Poincaré Lemma

The Poincare lemma states that on a contractible open set, every closed form is exact:

$H^k_{dR}(\mathbb{R}^n) = \begin{cases} \mathbb{R} & k = 0 \\ 0 & k > 0 \end{cases}$

Proof via Homotopy Operator

The key is to construct a homotopy operator $K: \Omega^k(\mathbb{R}^n) \to \Omega^{k-1}(\mathbb{R}^n)$ satisfying:

$Kd + dK = \text{id} - \iota_0^*$

where $\iota_0^*$ is the pullback to the origin. Explicitly, for $\omega = \omega_{i_1\ldots i_k} dx^{i_1} \wedge \cdots \wedge dx^{i_k}$:

$(K\omega)(x) = k \int_0^1 t^{k-1} \omega_{i_1 \ldots i_k}(tx) \, x^{i_1} dt \wedge dx^{i_2} \wedge \cdots \wedge dx^{i_k}$

If $d\omega = 0$ and $k \geq 1$, then $\omega = dK\omega + Kd\omega = d(K\omega)$, so $\omega$ is exact. This is a constructive proof—it gives an explicit antiderivative.

Derivation 5: de Rham's Theorem

De Rham's theorem establishes a profound bridge between analysis (differential forms) and topology (singular cohomology):

$\boxed{H^k_{dR}(M) \cong H^k(M; \mathbb{R})}$

The isomorphism is given by integration. Given a closed $k$-form $\omega$ and a singular $k$-cycle $c$, the pairing:

$\langle [\omega], [c] \rangle = \int_c \omega$

depends only on the cohomology class $[\omega]$ and the homology class $[c]$ (by Stokes' theorem). This gives a non-degenerate pairing, proving the isomorphism.

The Hodge Decomposition

On a compact oriented Riemannian manifold, every $k$-form decomposes uniquely as:

$\omega = d\alpha + \delta\beta + h$

where $h$ is harmonic ($\Delta h = 0$). Each cohomology class has a unique harmonic representative, giving $H^k_{dR}(M) \cong \mathcal{H}^k(M)$ (harmonic $k$-forms).

Applications to Physics

General Relativity

The Einstein-Hilbert action $S = \int_M R \, \text{vol}_g$ integrates the scalar curvature over the spacetime manifold. Boundary terms in the action (Gibbons-Hawking-York term) arise from Stokes' theorem applied to second derivatives of the metric.

Gauge Theory

Magnetic charge quantization follows from de Rham cohomology: the magnetic field 2-form $F$ on$\mathbb{R}^3 \setminus \{0\} \simeq S^2$ represents a class in $H^2(S^2) \cong \mathbb{R}$. The Dirac quantization condition $eg = n/2$ comes from integrality of this class.

Condensed Matter

The quantized Hall conductance $\sigma_{xy} = \frac{e^2}{h} \cdot n$ where $n = \frac{1}{2\pi}\int_{BZ} F$ is the first Chern number—an integral of the Berry curvature 2-form over the Brillouin zone torus, guaranteed to be an integer by de Rham's theorem.

Interactive Simulation

This simulation numerically verifies Stokes' theorem, computes the Gauss-Bonnet integral for the 2-sphere, displays Betti numbers for standard manifolds, and plots Euler characteristics for orientable surfaces of varying genus.

Integration on Manifolds: Stokes Theorem & de Rham Cohomology

Python

script.py194 lines

#!/usr/bin/env python3
"""
integration_manifolds.py - Stokes Theorem and de Rham Cohomology
Numerical verification of Stokes theorem and computation of Betti numbers
"""
import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import base64
from io import BytesIO

# ============================================================
# 1. Stokes theorem verification on a disk
# ============================================================
# omega = -y dx + x dy (angular 1-form)
# d(omega) = 2 dx ^ dy
# Stokes: integral of omega over boundary circle = integral of d(omega) over disk

# Boundary integral: circle of radius R
R = 2.0
N_boundary = 1000
t_param = np.linspace(0, 2*np.pi, N_boundary+1)[:-1]
dt = 2*np.pi / N_boundary

# On the circle: x = R cos(t), y = R sin(t)
# dx = -R sin(t) dt, dy = R cos(t) dt
# omega = -y dx + x dy = R sin(t) * R sin(t) dt + R cos(t) * R cos(t) dt = R^2 dt
boundary_integral = 0.0
for i in range(N_boundary):
    x_i = R * np.cos(t_param[i])
    y_i = R * np.sin(t_param[i])
    dx_i = -R * np.sin(t_param[i]) * dt
    dy_i = R * np.cos(t_param[i]) * dt
    boundary_integral += (-y_i * dx_i + x_i * dy_i)

# Area integral: integral of 2 dx dy over disk of radius R
# = 2 * pi * R^2
area_integral = 2.0 * np.pi * R**2

print("=" * 72)
print("  STOKES THEOREM AND DE RHAM COHOMOLOGY")
print("=" * 72)
print()
print("  STOKES THEOREM VERIFICATION")
print("  " + "-" * 40)
print("  1-form: omega = -y dx + x dy")
print("  d(omega) = 2 dx ^ dy")
print("  Domain: disk of radius R = {:.1f}".format(R))
print()
print("  Boundary integral (numerical):  {:.6f}".format(boundary_integral))
print("  Area integral (2 * pi * R^2):   {:.6f}".format(area_integral))
print("  Exact value:                    {:.6f}".format(2 * np.pi * R**2))
print("  Relative error: {:.2e}".format(abs(boundary_integral - area_integral)/area_integral))
print()

# ============================================================
# 2. Gauss-Bonnet via integration of curvature on S^2
# ============================================================
# K = 1/R^2 for unit sphere
# integral K dA = integral (1) sin(theta) dtheta dphi = 4 pi
# = 2 pi * chi(S^2) = 2 pi * 2 = 4 pi

N_theta = 500
N_phi = 500
dtheta = np.pi / N_theta
dphi = 2 * np.pi / N_phi

curvature_integral = 0.0
theta_arr = np.linspace(dtheta/2, np.pi - dtheta/2, N_theta)
for th in theta_arr:
    curvature_integral += np.sin(th) * dtheta * dphi * N_phi

print("  GAUSS-BONNET ON S^2")
print("  " + "-" * 40)
print("  Gaussian curvature K = 1 (unit sphere)")
print("  integral K dA (numerical): {:.6f}".format(curvature_integral))
print("  Exact value (4 pi):        {:.6f}".format(4 * np.pi))
print("  2 pi chi(S^2) = 4 pi:     {:.6f}".format(4 * np.pi))
print("  Relative error: {:.2e}".format(abs(curvature_integral - 4*np.pi)/(4*np.pi)))
print()

# ============================================================
# 3. de Rham cohomology: Betti numbers
# ============================================================
print("  DE RHAM COHOMOLOGY")
print("  " + "-" * 40)
print()
print("  Betti numbers b_k = dim H^k_dR(M):")
print()

manifolds = [
    ("R^n", [1, 0, 0, 0]),
    ("S^1", [1, 1, 0, 0]),
    ("S^2", [1, 0, 1, 0]),
    ("S^n (n>=1)", [1, 0, "...", 1]),
    ("T^2 (torus)", [1, 2, 1, 0]),
    ("T^n", ["C(n,0)", "C(n,1)", "C(n,2)", "..."]),
    ("RP^2", [1, 0, 0, 0]),
]

header = "  {:20s} {:>6s} {:>6s} {:>6s} {:>6s}".format("Manifold", "b_0", "b_1", "b_2", "b_3")
print(header)
print("  " + "-" * len(header.strip()))
for name, betti in manifolds:
    print("  {:20s} {:>6s} {:>6s} {:>6s} {:>6s}".format(
        name, str(betti[0]), str(betti[1]), str(betti[2]), str(betti[3])
    ))
print()
print("  Euler characteristic: chi(M) = sum (-1)^k b_k")
print("  chi(S^2) = 1 - 0 + 1 = 2")
print("  chi(T^2) = 1 - 2 + 1 = 0")
print()

# ============================================================
# Visualization
# ============================================================
fig, axes = plt.subplots(2, 2, figsize=(14, 11))
fig.patch.set_facecolor('#0a0a0a')

for ax in axes.flat:
    ax.set_facecolor('#111111')
    for spine in ax.spines.values():
        spine.set_color('#2dd4bf')
    ax.tick_params(colors='#99f6e4')
    ax.grid(True, color='#134e4a', alpha=0.4)

# Plot 1: Stokes theorem - boundary vs area
ax1 = axes[0, 0]
radii = np.linspace(0.1, 4.0, 50)
boundary_vals = []
area_vals = []
for r in radii:
    boundary_vals.append(2 * np.pi * r**2)  # exact boundary integral
    area_vals.append(2 * np.pi * r**2)       # area integral
ax1.plot(radii, boundary_vals, 'o', color='#2dd4bf', markersize=4, label='Boundary integral', alpha=0.7)
ax1.plot(radii, area_vals, '-', color='#f59e0b', linewidth=2.5, label='Area integral')
ax1.set_xlabel('Radius R', color='#5eead4')
ax1.set_ylabel('Integral value', color='#5eead4')
ax1.set_title('Stokes Theorem Verification', color='#99f6e4', fontsize=12, fontweight='bold')
ax1.legend(facecolor='#111111', edgecolor='#134e4a', labelcolor='#99f6e4')

# Plot 2: Curvature on S^2
ax2 = axes[0, 1]
theta_plot = np.linspace(0, np.pi, 300)
integrand = np.sin(theta_plot)
cumulative = np.zeros_like(theta_plot)
for i in range(1, len(theta_plot)):
    cumulative[i] = cumulative[i-1] + np.sin(theta_plot[i]) * (theta_plot[i] - theta_plot[i-1])
cumulative *= 2 * np.pi  # integrate over phi

ax2.plot(theta_plot, 2*np.pi*integrand, color='#f472b6', linewidth=2, label='Integrand: 2pi sin(theta)')
ax2.plot(theta_plot, cumulative, color='#34d399', linewidth=2.5, label='Cumulative integral')
ax2.axhline(y=4*np.pi, color='#fbbf24', linestyle='--', linewidth=1, label='4 pi (Gauss-Bonnet)')
ax2.set_xlabel('theta', color='#5eead4')
ax2.set_ylabel('Value', color='#5eead4')
ax2.set_title('Curvature Integration on S^2', color='#99f6e4', fontsize=12, fontweight='bold')
ax2.legend(facecolor='#111111', edgecolor='#134e4a', labelcolor='#99f6e4', fontsize=9)

# Plot 3: Betti numbers for torus
ax3 = axes[1, 0]
dims = [0, 1, 2]
betti_sphere = [1, 0, 1]
betti_torus = [1, 2, 1]
bar_width = 0.35
x_pos = np.array(dims)
ax3.bar(x_pos - bar_width/2, betti_sphere, bar_width, color='#2dd4bf', label='S^2', alpha=0.8)
ax3.bar(x_pos + bar_width/2, betti_torus, bar_width, color='#f59e0b', label='T^2', alpha=0.8)
ax3.set_xlabel('Degree k', color='#5eead4')
ax3.set_ylabel('Betti number b_k', color='#5eead4')
ax3.set_title('Betti Numbers: S^2 vs T^2', color='#99f6e4', fontsize=12, fontweight='bold')
ax3.set_xticks(dims)
ax3.legend(facecolor='#111111', edgecolor='#134e4a', labelcolor='#99f6e4')

# Plot 4: Euler characteristic for various surfaces
ax4 = axes[1, 1]
genus = np.arange(0, 8)
euler_char = 2 - 2*genus
ax4.bar(genus, euler_char, color='#a78bfa', alpha=0.8, edgecolor='#7c3aed')
ax4.set_xlabel('Genus g', color='#5eead4')
ax4.set_ylabel('Euler characteristic chi', color='#5eead4')
ax4.set_title('chi = 2 - 2g for Orientable Surfaces', color='#99f6e4', fontsize=12, fontweight='bold')
ax4.axhline(y=0, color='#475569', linestyle='--', linewidth=1)

plt.tight_layout(pad=2.0)
buf = BytesIO()
plt.savefig(buf, format='png', dpi=150, bbox_inches='tight', facecolor='#0a0a0a')
buf.seek(0)
img_b64 = base64.b64encode(buf.read()).decode()
print('<img src="data:image/png;base64,' + img_b64 + '" alt="Integration on Manifolds Visualization" />')
buf.close()
print()
print("Visualization complete.")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

← Differential Forms Next: Riemannian Geometry →