The Stacks project

59.79 Cohomology with support in a closed subscheme

Let $X$ be a scheme and let $Z \subset X$ be a closed subscheme. Let $\mathcal{F}$ be an abelian sheaf on $X_{\acute{e}tale}$. We let

\[ \Gamma _ Z(X, \mathcal{F}) = \{ s \in \mathcal{F}(X) \mid \text{Supp}(s) \subset Z\} \]

be the sections with support in $Z$ (Definition 59.31.3). This is a left exact functor which is not exact in general. Hence we obtain a derived functor

\[ R\Gamma _ Z(X, -) : D(X_{\acute{e}tale}) \longrightarrow D(\textit{Ab}) \]

and cohomology groups with support in $Z$ defined by $H^ q_ Z(X, \mathcal{F}) = R^ q\Gamma _ Z(X, \mathcal{F})$.

Let $\mathcal{I}$ be an injective abelian sheaf on $X_{\acute{e}tale}$. Let $U = X \setminus Z$. Then the restriction map $\mathcal{I}(X) \to \mathcal{I}(U)$ is surjective (Cohomology on Sites, Lemma 21.12.6) with kernel $\Gamma _ Z(X, \mathcal{I})$. It immediately follows that for $K \in D(X_{\acute{e}tale})$ there is a distinguished triangle

\[ R\Gamma _ Z(X, K) \to R\Gamma (X, K) \to R\Gamma (U, K) \to R\Gamma _ Z(X, K)[1] \]

in $D(\textit{Ab})$. As a consequence we obtain a long exact cohomology sequence

\[ \ldots \to H^ i_ Z(X, K) \to H^ i(X, K) \to H^ i(U, K) \to H^{i + 1}_ Z(X, K) \to \ldots \]

for any $K$ in $D(X_{\acute{e}tale})$.

For an abelian sheaf $\mathcal{F}$ on $X_{\acute{e}tale}$ we can consider the subsheaf of sections with support in $Z$, denoted $\mathcal{H}_ Z(\mathcal{F})$, defined by the rule

\[ \mathcal{H}_ Z(\mathcal{F})(U) = \{ s \in \mathcal{F}(U) \mid \text{Supp}(s) \subset U \times _ X Z\} \]

Here we use the support of a section from Definition 59.31.3. Using the equivalence of Proposition 59.46.4 we may view $\mathcal{H}_ Z(\mathcal{F})$ as an abelian sheaf on $Z_{\acute{e}tale}$. Thus we obtain a functor

\[ \textit{Ab}(X_{\acute{e}tale}) \longrightarrow \textit{Ab}(Z_{\acute{e}tale}),\quad \mathcal{F} \longmapsto \mathcal{H}_ Z(\mathcal{F}) \]

which is left exact, but in general not exact.

Lemma 59.79.1. Let $i : Z \to X$ be a closed immersion of schemes. Let $\mathcal{I}$ be an injective abelian sheaf on $X_{\acute{e}tale}$. Then $\mathcal{H}_ Z(\mathcal{I})$ is an injective abelian sheaf on $Z_{\acute{e}tale}$.

Proof. Observe that for any abelian sheaf $\mathcal{G}$ on $Z_{\acute{e}tale}$ we have

\[ \mathop{\mathrm{Hom}}\nolimits _ Z(\mathcal{G}, \mathcal{H}_ Z(\mathcal{F})) = \mathop{\mathrm{Hom}}\nolimits _ X(i_*\mathcal{G}, \mathcal{F}) \]

because after all any section of $i_*\mathcal{G}$ has support in $Z$. Since $i_*$ is exact (Section 59.46) and as $\mathcal{I}$ is injective on $X_{\acute{e}tale}$ we conclude that $\mathcal{H}_ Z(\mathcal{I})$ is injective on $Z_{\acute{e}tale}$. $\square$

Denote

\[ R\mathcal{H}_ Z : D(X_{\acute{e}tale}) \longrightarrow D(Z_{\acute{e}tale}) \]

the derived functor. We set $\mathcal{H}^ q_ Z(\mathcal{F}) = R^ q\mathcal{H}_ Z(\mathcal{F})$ so that $\mathcal{H}^0_ Z(\mathcal{F}) = \mathcal{H}_ Z(\mathcal{F})$. By the lemma above we have a Grothendieck spectral sequence

\[ E_2^{p, q} = H^ p(Z, \mathcal{H}^ q_ Z(\mathcal{F})) \Rightarrow H^{p + q}_ Z(X, \mathcal{F}) \]

Lemma 59.79.2. Let $i : Z \to X$ be a closed immersion of schemes. Let $\mathcal{G}$ be an injective abelian sheaf on $Z_{\acute{e}tale}$. Then $\mathcal{H}^ p_ Z(i_*\mathcal{G}) = 0$ for $p > 0$.

Proof. This is true because the functor $i_*$ is exact and transforms injective abelian sheaves into injective abelian sheaves (Cohomology on Sites, Lemma 21.14.2). $\square$

Lemma 59.79.3. Let $i : Z \to X$ be a closed immersion of schemes. Let $j : U \to X$ be the inclusion of the complement of $Z$. Let $\mathcal{F}$ be an abelian sheaf on $X_{\acute{e}tale}$. There is a distinguished triangle

\[ i_*R\mathcal{H}_ Z(\mathcal{F}) \to \mathcal{F} \to Rj_*(\mathcal{F}|_ U) \to i_*R\mathcal{H}_ Z(\mathcal{F})[1] \]

in $D(X_{\acute{e}tale})$. This produces an exact sequence

\[ 0 \to i_*\mathcal{H}_ Z(\mathcal{F}) \to \mathcal{F} \to j_*(\mathcal{F}|_ U) \to i_*\mathcal{H}^1_ Z(\mathcal{F}) \to 0 \]

and isomorphisms $R^ pj_*(\mathcal{F}|_ U) \cong i_*\mathcal{H}^{p + 1}_ Z(\mathcal{F})$ for $p \geq 1$.

Proof. To get the distinguished triangle, choose an injective resolution $\mathcal{F} \to \mathcal{I}^\bullet $. Then we obtain a short exact sequence of complexes

\[ 0 \to i_*\mathcal{H}_ Z(\mathcal{I}^\bullet ) \to \mathcal{I}^\bullet \to j_*(\mathcal{I}^\bullet |_ U) \to 0 \]

by the discussion above. Thus the distinguished triangle by Derived Categories, Section 13.12. $\square$

Let $X$ be a scheme and let $Z \subset X$ be a closed subscheme. We denote $D_ Z(X_{\acute{e}tale})$ the strictly full saturated triangulated subcategory of $D(X_{\acute{e}tale})$ consisting of complexes whose cohomology sheaves are supported on $Z$. Note that $D_ Z(X_{\acute{e}tale})$ only depends on the underlying closed subset of $X$.

Lemma 59.79.4. Let $i : Z \to X$ be a closed immersion of schemes. The map $Ri_{small, *} = i_{small, *} : D(Z_{\acute{e}tale}) \to D(X_{\acute{e}tale})$ induces an equivalence $D(Z_{\acute{e}tale}) \to D_ Z(X_{\acute{e}tale})$ with quasi-inverse

\[ i_{small}^{-1}|_{D_ Z(X_{\acute{e}tale})} = R\mathcal{H}_ Z|_{D_ Z(X_{\acute{e}tale})} \]

Proof. Recall that $i_{small}^{-1}$ and $i_{small, *}$ is an adjoint pair of exact functors such that $i_{small}^{-1}i_{small, *}$ is isomorphic to the identify functor on abelian sheaves. See Proposition 59.46.4 and Lemma 59.36.2. Thus $i_{small, *} : D(Z_{\acute{e}tale}) \to D_ Z(X_{\acute{e}tale})$ is fully faithful and $i_{small}^{-1}$ determines a left inverse. On the other hand, suppose that $K$ is an object of $D_ Z(X_{\acute{e}tale})$ and consider the adjunction map $K \to i_{small, *}i_{small}^{-1}K$. Using exactness of $i_{small, *}$ and $i_{small}^{-1}$ this induces the adjunction maps $H^ n(K) \to i_{small, *}i_{small}^{-1}H^ n(K)$ on cohomology sheaves. Since these cohomology sheaves are supported on $Z$ we see these adjunction maps are isomorphisms and we conclude that $D(Z_{\acute{e}tale}) \to D_ Z(X_{\acute{e}tale})$ is an equivalence.

To finish the proof we have to show that $R\mathcal{H}_ Z(K) = i_{small}^{-1}K$ if $K$ is an object of $D_ Z(X_{\acute{e}tale})$. To do this we can use that $K = i_{small, *}i_{small}^{-1}K$ as we've just proved this is the case. Then we can choose a K-injective representative $\mathcal{I}^\bullet $ for $i_{small}^{-1}K$. Since $i_{small, *}$ is the right adjoint to the exact functor $i_{small}^{-1}$, the complex $i_{small, *}\mathcal{I}^\bullet $ is K-injective (Derived Categories, Lemma 13.31.9). We see that $R\mathcal{H}_ Z(K)$ is computed by $\mathcal{H}_ Z(i_{small, *}\mathcal{I}^\bullet ) = \mathcal{I}^\bullet $ as desired. $\square$

Lemma 59.79.5. Let $X$ be a scheme. Let $Z \subset X$ be a closed subscheme. Let $\mathcal{F}$ be a quasi-coherent $\mathcal{O}_ X$-module and denote $\mathcal{F}^ a$ the associated quasi-coherent sheaf on the small étale site of $X$ (Proposition 59.17.1). Then

  1. $H^ q_ Z(X, \mathcal{F})$ agrees with $H^ q_ Z(X_{\acute{e}tale}, \mathcal{F}^ a)$,

  2. if the complement of $Z$ is retrocompact in $X$, then $i_*\mathcal{H}^ q_ Z(\mathcal{F}^ a)$ is a quasi-coherent sheaf of $\mathcal{O}_ X$-modules equal to $(i_*\mathcal{H}^ q_ Z(\mathcal{F}))^ a$.

Proof. Let $j : U \to X$ be the inclusion of the complement of $Z$. The statement (1) on cohomology groups follows from the long exact sequences for cohomology with supports and the agreements $H^ q(X_{\acute{e}tale}, \mathcal{F}^ a) = H^ q(X, \mathcal{F})$ and $H^ q(U_{\acute{e}tale}, \mathcal{F}^ a) = H^ q(U, \mathcal{F})$, see Theorem 59.22.4. If $j : U \to X$ is a quasi-compact morphism, i.e., if $U \subset X$ is retrocompact, then $R^ qj_*$ transforms quasi-coherent sheaves into quasi-coherent sheaves (Cohomology of Schemes, Lemma 30.4.5) and commutes with taking associated sheaf on étale sites (Descent, Lemma 35.9.5). We conclude by applying Lemma 59.79.3. $\square$


Comments (2)

Comment #2150 by Katha on

In Lemma 50.72.5(2) it should be instead of .

Comment #2186 by on

Thanks. If you want to be listed as a contributor, then leave a first and last name in your next comment. Fixed here.


Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.




In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 09XP. Beware of the difference between the letter 'O' and the digit '0'.