Is there any example of a smooth, projective surface $S$ over $\mathbb{C}$, with Picard group $\mathbb{Z}$ and such that $H^1(S, L)$ is not zero for some ample line bundle $L$ ?


2 Answers 2


Edit. My original exact sequences were wrong. I straightened them out now.

Let $X$ be a smooth, projective hyper-Kähler fourfold with $\text{Pic}(X)$ isomorphic to $\mathbb{Z}$, e.g., a sufficiently general deformation of the pair $(\text{Hilb}^2_{M/\mathbb{C}},\mathcal{O}(1))$ for $M$ a polarized K3 surface and $\mathcal{O}(1)$ the natural Plücker invertible sheaf. Let $Y\subset X$ be a smooth ample hypersurface such that $h^r(X,\mathcal{O}_X(\underline{Y}))$ vanishes for $r=1,2$. Let $Z\subset X$ be a very general hypersurface that is "sufficiently" ample. In particular, assume that $h^r(X,\mathcal{O}_X(\underline{Z}))$ vanishes for $r=1,2$ and $h^2(X,\mathcal{O}_X(\underline{Z}-\underline{Y}))$ vanishes. Let $i:S\hookrightarrow X$ be $Y\cap Z$, and denote by $\mathcal{I}$ the ideal sheaf of $S$ in $X$. Let $L$ be $\mathcal{O}_X(\underline{Y})|_S$.

There are short exact sequences, $$ 0 \to \mathcal{I}(\underline{Y}) \to \mathcal{O}_X(\underline{Y}) \to i_* L \to 0,$$ $$ 0 \to \mathcal{O}_X(-\underline{Z}) \to \mathcal{O}_X\oplus \mathcal{O}_X(\underline{Y}-\underline{Z}) \to \mathcal{I}(\underline{Y}) \to 0.$$ By hypothesis, $h^r(X,\mathcal{O}_X(\underline{Y})) = 0$ for $r= 1,2$. Thus, via the long exact sequence of cohomology, the following connecting map is an isomorphism, $$\delta_\Sigma^1:H^1(S,L)\xrightarrow{\cong} H^2(X,\mathcal{I}(\underline{Y})).$$ By Serre duality, $h^{2+r}(X,\mathcal{O}_X(-\underline{Z}))$ equals $h^{2-r}(X,\mathcal{O}_X(\underline{Z}))$. Thus, for $r=0,1$, both of these vanish. Finally, by Serre duality, $h^2(X,\mathcal{O}_X(\underline{Y}-\underline{Z})))$ equals $h^2(X,\mathcal{O}_X(\underline{Z}-\underline{Y}))$, so both of these vanish. Thus, the long exact sequence of the second short exact sequence gives an isomorphism, $$ H^2(X,\mathcal{O}_X)\to H^2(X,\mathcal{I}(\underline{Y})), $$ so also $H^1(S,L)$ is isomorphic to $H^2(X,\mathcal{O}_X)$. Of course for the hyper-Kähler fourfold $X$, $H^2(X,\mathcal{O}_X)$ is a $1$-dimensional vector space.

By the Lefschetz hyperplane theorem, the restriction map $\text{Pic}(X)\to \text{Pic}(Y)$ is an isomorphism (and the same holds for $Z$). By the generalized Noether-Lefschetz theorem (as in SGA7), for $Z$ sufficiently ample and very general, the restriction map $\text{Pic}(Y)\to \text{Pic}(S)$ is also an isomorphism. Thus $\text{Pic}(S)$ is isomorphic to $\mathbb{Z}$.

Edit. I was misremembering what is proved in SGA 7_2. The generalized Noether-Lefschetz theorem was not proved there. The generalized Noether-Lefschetz was proved by Kirti Joshi.

MR1299006 (96f:14005)
Joshi, Kirti(6-TIFR-SM)
A Noether-Lefschetz theorem and applications.
J. Algebraic Geom. 4 (1995), no. 1, 105–135.

  • $\begingroup$ Dear Jason, many many thanks for this :) $\endgroup$ Feb 12, 2017 at 21:11

For example, for an irregular surface of general type of Picard number one, $H^1(K_X)\neq0$. See Kovacs' answer here Smooth projective varieties of Picard number one.

Edit: As pointed out by @abx, in this case the picard group could not be $\mathbb{Z}$, as the surface is irregular.

Maybe the way of construct such example is to look at certain surface with $K_X=2L$ and try to show $H^1(L)\neq 0$.

  • 2
    $\begingroup$ For such a surface the Picard group is far from being $\mathbb{Z}$ — it contains a nontrivial abelian variety. $\endgroup$
    – abx
    Feb 3, 2017 at 7:06
  • $\begingroup$ @abx, you are right... $\endgroup$
    – Chen Jiang
    Feb 3, 2017 at 8:08
  • $\begingroup$ @ChenJiang. I made the same mistake that you made (and then I deleted my solution). I read "Picard group" as "Neron-Severi group". $\endgroup$ Feb 3, 2017 at 12:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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