1
$\begingroup$

Let $\varphi:X\to Y$ be a surjective morphism of schemes which are connected and of finite type.

Let $A$ be an abelian group, $\mathscr{F}$ be the constant sheaf on $X$ with fibers $A$ and $\mathscr{G}$ the constant sheaf on $Y$ with fibers $A$.

Is it true, then, that we have an isomorphism $$ \varphi^* \mathscr{G} \cong \mathscr{F} \quad ?$$

Or does this hold only under some more restrictive hypothesis?

In case it is true, could you give me a reference for a proof of it or give me some hints to prove it by myself?

$\endgroup$
4
  • 10
    $\begingroup$ First of all, one should write $f^{-1}$, the inverse image sheaf. $f^*$ is suitable for sheaves of $\mathcal{O}_Y$-modules. Next, it has nothing to do with schemes: it works in the category of topological spaces. As for the question, I suggest you read carefully any introductory text on sheaf theory. The proof is obvious once you understand all the definitions. $\endgroup$ Dec 6, 2013 at 10:23
  • 2
    $\begingroup$ Thanks for the suggestion Anton. I already read an introductory text on sheaves, but I'm still not very experienced about the theory. Give me some more time to practice! ;-) $\endgroup$
    – Abramo
    Dec 6, 2013 at 10:33
  • $\begingroup$ Now I see it, thinking in terms of the inverse image this is really a triviality. I was making my life difficult using the pullback when it was not needed. $\endgroup$
    – Abramo
    Dec 6, 2013 at 11:00
  • 1
    $\begingroup$ The constant sheaf on $X$ with fibre $A$ is the pullback of sheaf $A$ on a point $pt$ by the unique map $\pi_X:X\to pt$. In particular, for any map $f:Y\to X$, its composition with $\pi_X$ is $\pi_Y$. In particular, $f^*A_X=f^*\pi_X^*A=\pi_Y^*A=A_Y$. This argument works just as well in the category of topological spaces as in schemes. $\endgroup$
    – Pulcinella
    Mar 14, 2021 at 16:03

2 Answers 2

5
$\begingroup$

Yes, it is true. Maybe the easiest way to see it is to think of your sheaves as "étalé space" (see Wikipedia): $\mathcal{F}$ is the sheaf of (local) sections of the projection $X\times A\rightarrow X$, and similarly for $\mathcal{G}$. Now the pull back $\varphi ^*\mathcal{G}$ corresponds to the pull back on $X$ of $Y\times A$, which is obviously $X\times A$.

$\endgroup$
5
$\begingroup$

If $\mathscr{F} = g^{-1}\underline{A}$ is the constant sheaf on $Y$, $f^{-1}\mathscr{F} = (g \circ f)^{-1}\underline{A}$ is constant on $X$, where $f: X \to Y$.

$g$ is the map to the final object of the category $X$ and $Y$ live in.

$\endgroup$
3
  • $\begingroup$ Could you make explicit what is the map $g$ you are using? Thanks! $\endgroup$
    – Abramo
    Dec 6, 2013 at 12:25
  • 1
    $\begingroup$ Unless the "final object of the category" is a point, you are invoking the fact that the pullback of a constant sheaf is constant in full generality, which is exactly what the question asks. $\endgroup$
    – Ryan Reich
    Dec 6, 2013 at 15:25
  • 1
    $\begingroup$ I would say it also holds if the final object is an irreducible topological space (so the constant presheaf is flasque and a sheaf), e.g. $\mathrm{Spec}\mathbf{Z}$. $\endgroup$
    – user19475
    Dec 6, 2013 at 15:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.