9
$\begingroup$

Let $\mathcal{O}_{X}$ be a sheaf of rings on a topological space $X$, and let $U\subseteq X$ be an open set. It is common to give the sheaf axioms as exactness of a sequence. For any open covering $\left\{U_{i}\right\}$ of $U$, we have the exact sequence:

$ 0 \to \mathcal{O}_{X}(U) \to \prod_{i}\mathcal{O}_{X}(U_{i}) \to \prod_{i,j} \mathcal{O}_{X}(U_{i}\cap U_{j})$

The last map's projection to each factor sends a sequence $(s_{i}) \mapsto s_{i}\vert_{U_{i}\cap U_{j}} - s_{j}\vert_{U_{i}\cap U_{j}}$.

My question: Are there conditions in which this last map is a surjection? Do these kinds of sheaves have a name?

Thanks!

$\endgroup$
7
$\begingroup$

Surjectivity at that point is a very tough restriction as we need to consider all possible choices of $\{U_i\}$. Suppose $U_i=U_j=U_k$. Then for three arbitrary elements $s_{i,j}, s_{i,k}, s_{j,k}$ of $\mathcal O_X(U_i\cap U_j)=\mathcal O_X(U_i\cap U_k)=\mathcal O_X(U_j\cap U_k)$, you'd need to exhibit $s_i,s_j,s_k\in \mathcal O_X(U_i)=\mathcal O_X(U_j)=\mathcal O_X(U_k)$ such that $s_{i,j}=s_i-s_j$, $s_{i,k}=s_i-s_k$, $s_{j,k}=s_j-s_k$. Hence we obtain the condition $s_{i,j}+s_{j,k}=s_{i,k}$, which is not in general true. The only way to make this work is to have $\mathcal O_X(U)=0$ for all $U$.

So, yes, these sheaves have a name: trivial sheaf.

$\endgroup$

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.