2
$\begingroup$

What techniques are out there to calculate the cohomology groups of the structure sheaf $\mathcal{O}_X$ of a smooth quasi-projective variety $X$?

For example can we conclude something from the dimension of the complement $Z = \bar{X} \X$, where $\bar{X}$ is smooth and projective. I know I could use Hodge theory to calculate the cohomology groups $H^i(\bar{X},\mathcal{O}_{\bar X})$.

Is there some way to relate the groups $H^i(\bar{X},\mathcal{O}_{\bar X})$ and $H^i(X,\mathcal{O}_X)$,.

I should mention here that I do know the cohomology groups $H^i(\bar{X},\mathcal{O}_{\bar X})$ and the space $Z$ quite well.

$\endgroup$

2 Answers 2

6
$\begingroup$

Not knowing anything else about your situation, the approach that comes to mind is to use the long exact sequence: $$\cdots \to H^i_Z(\overline{X},\mathcal O_{\overline{X}}) \to H^i(\overline{X}, \mathcal O_{\overline{X}}) \to H^i(X,\mathcal O_X) \to \cdots $$ (described somewhere in a Hartshorne exercise). This requires knowing the local cohomology $H^i_Z(\overline{X},\mathcal O_{\overline{X}})$, which you can hopefully get a handle on, if you know enough about $Z$ and its embedding in $\overline{X}$.

$\endgroup$
2
  • $\begingroup$ Is there a standard reference to read about Local Cohomology? $\endgroup$ Commented Feb 14, 2011 at 20:10
  • 1
    $\begingroup$ The standard reference to my knowledge is Local Cohomology, LNM #41, by A. Grothendieck, sometimes found attributed to R. Hartshorne. I.e. the seminar was by Grothendieck and the notes written by Hartshorne. There are also some more recent books by others, to be found on Amazon. $\endgroup$
    – roy smith
    Commented Feb 15, 2011 at 16:00
6
$\begingroup$

Your problem basically boils down to computing the groups $H^i_Z(X,\mathcal O)$, the local cohomology with support in $Z$. Once these groups are known the local cohomology exact sequence takes the form

$\cdots \to H^i_Z(\bar{X},\mathcal O) \to H^i(\bar{X},\mathcal O) \to H^i(X, \mathcal O) \to H^{i+1}_Z(X,\mathcal O) \to \cdots$

If you know $\bar{X}$ and $Z$ well there are various methods to compute $H^i_Z(X,\mathcal O)$, e.g., it is by definition the group $\lim_\leftarrow Ext^i(\mathcal{O}_{\bar{X}}/I_Z^n,\mathcal{O}_{\bar{X}})$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .