Constructions of harmonic maps between Hadamard manifolds
Constructions of harmonic maps between Hadamard manifolds
Constructions of harmonic maps between Hadamard manifolds
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
3.2. HARMONIC MAPS BETWEEN DAMEK-RICCI SPACES 63<br />
3.2.1 Harmonic map equation<br />
We compute explicitly the tension field <strong>of</strong> a <strong>harmonic</strong> map <strong>between</strong> Damek-Ricci spaces. Let<br />
(S, g S ) be a Damek-Ricci space with left invariant metric g S , and set s = a ⊕ n = a ⊕ v ⊕ z.<br />
Take the unit vector h ∈ a such that ad h(v) =v, ad h(z) =2z for any v ∈ v,z ∈ z and<br />
an orthonormal basis {h, v 1 ,... ,v n ,z 1 ,... ,z m } <strong>of</strong> s, {v i } (resp. {z j }) being an orthonormal<br />
basis <strong>of</strong> v (resp. z). We define a map ϕ : R × N → S by<br />
ϕ(t, n) :=n(exp(th)),<br />
where N is the simply connected Lie group associated with n and exp is the exponential<br />
map on a. Then the induced metric ϕ ∗ g S is given by<br />
ϕ ∗ g S = dt 2 + e −2t g Ú<br />
+ e −4t g Þ<br />
,<br />
where g Ú<br />
+ g Þ<br />
is a left invariant metric on N. Setting y = e t , one can verify that (S, g S )is<br />
isometric to M := R + × N with the Riemannian metric<br />
g M := y −2 dy 2 + y −2 g Ú<br />
+ y −4 g Þ<br />
,<br />
where R + = {r ∈ R | r>0}. A straightforward calculation then yields that the Levi-Civita<br />
connection ∇ on (S, g S ) is given by<br />
⎧<br />
∇ η η = −y −1 η, ∇ η v i = −y −1 v i , ∇ η z k = −2y −1 z k ,<br />
(3.2.1)<br />
⎪⎨<br />
⎪⎩<br />
∇ vi v j = y −1 δ ij η + 1 2<br />
∇ vi z k = 1 n∑<br />
2 y−2<br />
j=1<br />
m∑<br />
k=1<br />
a k<br />
ijz k ,<br />
where a k<br />
ij are the structure constants, [v i ,v j ]=<br />
regarded as left invariant vector fields on S.<br />
a k<br />
j iv j , ∇ zk z l =2y −3 δ kl η,<br />
m∑<br />
k=1<br />
a k<br />
ijz k ,η= ∂/∂y and, v i and z k<br />
Now we compute the tension field. Let (S, g S ) and (S ′ ,g S ′) be Damek-Ricci spaces and<br />
u ∈ C 2 (S, S ′ ). We represent (S, g S ) (resp. (S ′ ,g S ′)) as<br />
(R + × N,y −2 dy 2 + y −2 g Ú<br />
+ y −4 g Þ<br />
) (resp. (R + × N ′ , ȳ −2 dȳ 2 +ȳ −2 g Ú ′ +ȳ −4 g Þ<br />
′)),<br />
are