A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
The flux of F through oriented surface S is ∬_S F · n dS = ∬_S F · (r_u × r_v) du dv, measuring flow through the surface. When F is the curl of another vector field, Stokes' theorem relates this to a line integral around the boundary.
From your work with scalar surface integrals, you know how to integrate a function f over a surface S: parameterize the surface as r(u, v), compute the cross product r_u × r_v to get the area element dS = |r_u × r_v| du dv, and integrate ∬_D f(r(u,v)) |r_u × r_v| du dv. This measures "how much of f accumulates on S." The flux of a vector field through a surface asks a different question entirely: not how much of a scalar quantity sits on S, but how much of a vector field F passes through S. This requires the surface to be oriented — equipped with a consistent choice of "which side is the outside," specified by a unit normal vector n at each point.
The central idea is that only the component of F perpendicular to the surface contributes to flow through it. If F points directly along the normal, it passes through the surface at full strength. If F is tangent to the surface (perpendicular to n), it slides along without crossing. The contribution at each point is F · n — the dot product projects F onto the normal direction. The total flux is the surface integral ∬_S F · n dS. Since n = (r_u × r_v)/|r_u × r_v| and dS = |r_u × r_v| du dv, the magnitude |r_u × r_v| cancels, leaving the clean formula: ∬_S F · n dS = ∬_D F(r(u,v)) · (r_u × r_v) du dv. The cross product encodes both the surface area element and the normal direction in one object.
Orientation is not a formality — it changes the sign of the flux. The cross product r_u × r_v points in one of two possible normal directions depending on the parameterization; reversing the parameterization's order (swapping u and v) reverses the cross product and negates the integral. For a closed surface (like a sphere), the convention is that the outward normal points away from the enclosed region. For a surface with boundary (like a hemisphere or a disk), the orientation of the boundary curve and the surface normal are linked by the right-hand rule. Getting orientation right means checking, after computing r_u × r_v, that the result points in the intended direction — and if not, negating the entire integral.
The physical interpretation is the reason all of this machinery matters. For a fluid with velocity field F, the flux of F through a surface S measures the volume of fluid crossing S per unit time — positive if the flow is in the direction of n, negative if it opposes n. For an electric field, it is electric flux. This quantity is the foundation of the Divergence Theorem: the total flux of F outward through a closed surface equals the integral of ∇·F over the enclosed volume. If ∇·F = 0 (an incompressible fluid or a divergence-free field), exactly as much flows in as flows out through any closed surface, and the net flux is zero. The flux integral is therefore not just a computation — it is the precise formalization of "how much of a vector field crosses a boundary," which is the central object in all three fundamental theorems of vector calculus you are about to encounter.