Let M⊂R3 be a surface, let x(u,v)=⎝⎛x1(u,v)x2(u,v)x3(u,v)⎠⎞ be parametrization on U and x(s,t)=⎝⎛x1(s,t)x2(s,t)x3(s,t)⎠⎞ be parametrization on V.
Let xu=⎝⎛(x1)u(x2)u(x3)u⎠⎞, where (xi)u means derivative ∂u∂xi(u,v), and xv=⎝⎛(x1)v(x2)v(x3)v⎠⎞.
(1) Let (u(s,t)v(s,t)) be change of coordinate, and J:=(∂s∂u∂s∂v∂t∂u∂t∂v) be its Jacobi matrix. Verify that (xuxv)=J(xsxt).
Let E=⟨xu,xv⟩ be square length of xu, G=⟨xv,xv⟩ be square length of xv, and F=⟨xu,xv⟩. Let g=(EFFG).
(2) Let E~=⟨xs,xs⟩, G~=⟨xt,xt⟩ and F~=⟨xs,xt⟩ be the corresponding data on V, and g~:=(E~F~F~G~). Show that g~=JTgJ.
(3) Consider the two form det(g)du∧dv. Assume detJ>0. Show that detgdu∧dv=detg~ds∧dt.
If one can always choose parametrizations such that detJ>0 for every change of coordinate, we call such Moriented. We always assume a surface is oriented.
The above exercise indicates that one can define a positive measure dV on M, which is given on any parametrization U by detgdu∧dv. This is well defined on M because on any coordinate it is of the same form by (3).
Let M be a compact surface, let (Ui)i=1N be its coordinate charts. On each Ui we parametrize M by x(ui,vi). Let dV be the area measure on M. Now we define the integral ∫MfdV for f∈C∞(M).
Choose partition of unity (ρi)i=1N subordinate to (Ui)i=1N. We know that ρif is supported on Ui. Then we define
The statements are given on surface, the same argument will work on compact smooth manifolds with integration defined. If you are really uncomfortable with integration on surfaces, you can replace all “M” by an open set U in Rn.
Let M be a compact smooth surface. Let K∈C∞(M×M). Define TK:C∞(M)→C∞(M) given by TKf(x)=∫MK(x,y)f(y)dV(y).
Let U,V be open sets in R2. φ:V→U be a diffeomorphism (bijective, smooth, det(Jacobi matrix)=0). Let f:U→R be a smooth function. Then f∘φ is a smooth function on V, we denote this by φ∗f. Since φ is a smooth diffeo, we can do samething for φ−1, so that φ∗ and (φ−1)∗ give bijection C∞(U) and C∞(V).
(1) Check that (φ−1)∗Djαφ=((φ−1)∗Djφ))α.
(2)
Let P be a differential operator with degree m on V, we know that (φ−1)∗Pφ∗ is a differential operator on U. The idea is, for f∈C∞(U), we first “pull back” f to a smooth function on V, apply P, then “pull back” back to U. Show that ((φ)−1)∗Pφ∗ is also degree m.
Hint: Write Pf=∑∣α∣≤maα(s,t)Dαf(s,t). Write φ:V→U by (uv)=φ((st))=(u(s,t)v(s,t)). Then (1) can be verified directly, and (2) reduce to check the case P=Dk. By the chain rule, we have Dsφ∗f(s,t)=Ds(f(u(s,t),v(s,t)))=∂s∂uDuf(u(s,t),v(s,t))+∂s∂vDvf(u(s,t),v(s,t))=∂s∂uφ∗Duf+∂s∂vφ∗Dvf.
Let f,g∈Cc∞(U), complex valued. Let ∫Ufgˉdm be the inner product. Let P=∑∣α∣≤maα(x)Dα, define its adjoint P∗ by requiring (Pf,g)=(f,P∗g). Calculate P∗ using integration by parts. The P∗ is called the formal adjoint of P. Note that here it only applies on test functions.