The Stacks project

25.5 Čech cohomology and hypercoverings

Let $\mathcal{C}$ be a site. Consider a presheaf of abelian groups $\mathcal{F}$ on the site $\mathcal{C}$. It defines a functor

\begin{eqnarray*} \mathcal{F} : \text{SR}(\mathcal{C})^{opp} & \longrightarrow & \textit{Ab} \\ \{ U_ i\} _{i \in I} & \longmapsto & \prod \nolimits _{i \in I} \mathcal{F}(U_ i) \end{eqnarray*}

Thus a simplicial object $K$ of $\text{SR}(\mathcal{C})$ is turned into a cosimplicial object $\mathcal{F}(K)$ of $\textit{Ab}$. The cochain complex $s(\mathcal{F}(K))$ associated to $\mathcal{F}(K)$ (Simplicial, Section 14.25) is called the Čech complex of $\mathcal{F}$ with respect to the simplicial object $K$. We set

\[ \check{H}^ i(K, \mathcal{F}) = H^ i(s(\mathcal{F}(K))). \]

and we call it the $i$th Čech cohomology group of $\mathcal{F}$ with respect to $K$. In this section we prove analogues of some of the results for Čech cohomology of open coverings proved in Cohomology, Sections 20.9, 20.10 and 20.11.

Lemma 25.5.1. Let $\mathcal{C}$ be a site with fibre products. Let $X$ be an object of $\mathcal{C}$. Let $K$ be a hypercovering of $X$. Let $\mathcal{F}$ be a sheaf of abelian groups on $\mathcal{C}$. Then $\check{H}^0(K, \mathcal{F}) = \mathcal{F}(X)$.

Proof. We have

\[ \check{H}^0(K, \mathcal{F}) = \mathop{\mathrm{Ker}}(\mathcal{F}(K_0) \longrightarrow \mathcal{F}(K_1)) \]

Write $K_0 = \{ U_ i \to X\} $. It is a covering in the site $\mathcal{C}$. As well, we have that $K_1 \to K_0 \times K_0$ is a covering in $\text{SR}(\mathcal{C}, X)$. Hence we may write $K_1 = \amalg _{i_0, i_1 \in I} \{ V_{i_0i_1j} \to X\} $ so that the morphism $K_1 \to K_0 \times K_0$ is given by coverings $\{ V_{i_0i_1j} \to U_{i_0} \times _ X U_{i_1}\} $ of the site $\mathcal{C}$. Thus we can further identify

\[ \check{H}^0(K, \mathcal{F}) = \mathop{\mathrm{Ker}}( \prod \nolimits _ i \mathcal{F}(U_ i) \longrightarrow \prod \nolimits _{i_0i_1 j} \mathcal{F}(V_{i_0i_1j}) ) \]

with obvious map. The sheaf property of $\mathcal{F}$ implies that $\check{H}^0(K, \mathcal{F}) = H^0(X, \mathcal{F})$. $\square$

In fact this property characterizes the abelian sheaves among all abelian presheaves on $\mathcal{C}$ of course. The analogue of Cohomology, Lemma 25.5.2 in this case is the following.

Lemma 25.5.2. Let $\mathcal{C}$ be a site with fibre products. Let $X$ be an object of $\mathcal{C}$. Let $K$ be a hypercovering of $X$. Let $\mathcal{I}$ be an injective sheaf of abelian groups on $\mathcal{C}$. Then

\[ \check{H}^ p(K, \mathcal{I}) = \left\{ \begin{matrix} \mathcal{I}(X) & \text{if} & p = 0 \\ 0 & \text{if} & p > 0 \end{matrix} \right. \]

Proof. Observe that for any object $Z = \{ U_ i \to X\} $ of $\text{SR}(\mathcal{C}, X)$ and any abelian sheaf $\mathcal{F}$ on $\mathcal{C}$ we have

\begin{eqnarray*} \mathcal{F}(Z) & = & \prod \mathcal{F}(U_ i) \\ & = & \prod \mathop{\mathrm{Mor}}\nolimits _{\textit{PSh}(\mathcal{C})}(h_{U_ i}, \mathcal{F})\\ & = & \mathop{\mathrm{Mor}}\nolimits _{\textit{PSh}(\mathcal{C})}(F(Z), \mathcal{F})\\ & = & \mathop{\mathrm{Mor}}\nolimits _{\textit{PAb}(\mathcal{C})}(\mathbf{Z}_{F(Z)}, \mathcal{F}) \\ & = & \mathop{\mathrm{Mor}}\nolimits _{\textit{Ab}(\mathcal{C})}(\mathbf{Z}_{F(Z)}^\# , \mathcal{F}) \end{eqnarray*}

Thus we see, for any simplicial object $K$ of $\text{SR}(\mathcal{C}, X)$ that we have

25.5.2.1
\begin{equation} \label{hypercovering-equation-identify-cech} s(\mathcal{F}(K)) = \mathop{\mathrm{Hom}}\nolimits _{\textit{Ab}(\mathcal{C})}(s(\mathbf{Z}_{F(K)}^\# ), \mathcal{F}) \end{equation}

see Definition 25.4.1 for notation. The complex of sheaves $s(\mathbf{Z}_{F(K)}^\# )$ is quasi-isomorphic to $\mathbf{Z}_ X^\# $ if $K$ is a hypercovering, see Lemma 25.4.5. We conclude that if $\mathcal{I}$ is an injective abelian sheaf, and $K$ a hypercovering, then the complex $s(\mathcal{I}(K))$ is acyclic except possibly in degree $0$. In other words, we have

\[ \check{H}^ i(K, \mathcal{I}) = 0 \]

for $i > 0$. Combined with Lemma 25.5.1 the lemma is proved. $\square$

Next we come to the analogue of Cohomology on Sites, Lemma 21.10.6. Let $\mathcal{C}$ be a site. Let $\mathcal{F}$ be a sheaf of abelian groups on $\mathcal{C}$. Recall that $\underline{H}^ i(\mathcal{F})$ indicates the presheaf of abelian groups on $\mathcal{C}$ which is defined by the rule $\underline{H}^ i(\mathcal{F}) : U \longmapsto H^ i(U, \mathcal{F})$. We extend this to $\text{SR}(\mathcal{C})$ as in the introduction to this section.

Lemma 25.5.3. Let $\mathcal{C}$ be a site with fibre products. Let $X$ be an object of $\mathcal{C}$. Let $K$ be a hypercovering of $X$. Let $\mathcal{F}$ be a sheaf of abelian groups on $\mathcal{C}$. There is a map

\[ s(\mathcal{F}(K)) \longrightarrow R\Gamma (X, \mathcal{F}) \]

in $D^{+}(\textit{Ab})$ functorial in $\mathcal{F}$, which induces natural transformations

\[ \check{H}^ i(K, -) \longrightarrow H^ i(X, -) \]

as functors $\textit{Ab}(\mathcal{C}) \to \textit{Ab}$. Moreover, there is a spectral sequence $(E_ r, d_ r)_{r \geq 0}$ with

\[ E_2^{p, q} = \check{H}^ p(K, \underline{H}^ q(\mathcal{F})) \]

converging to $H^{p + q}(X, \mathcal{F})$. This spectral sequence is functorial in $\mathcal{F}$ and in the hypercovering $K$.

Proof. We could prove this by the same method as employed in the corresponding lemma in the chapter on cohomology. Instead let us prove this by a double complex argument.

Choose an injective resolution $\mathcal{F} \to \mathcal{I}^\bullet $ in the category of abelian sheaves on $\mathcal{C}$. Consider the double complex $A^{\bullet , \bullet }$ with terms

\[ A^{p, q} = \mathcal{I}^ q(K_ p) \]

where the differential $d_1^{p, q} : A^{p, q} \to A^{p + 1, q}$ is the one coming from the differential on the complex $s(\mathcal{I}^ q(K))$ associated to the cosimplicial abelian group $\mathcal{I}^ p(K)$ and the differential $d_2^{p, q} : A^{p, q} \to A^{p, q + 1}$ is the one coming from the differential $\mathcal{I}^ q \to \mathcal{I}^{q + 1}$. Denote $\text{Tot}(A^{\bullet , \bullet })$ the total complex associated to the double complex $A^{\bullet , \bullet }$, see Homology, Section 12.18. We will use the two spectral sequences $({}'E_ r, {}'d_ r)$ and $({}''E_ r, {}''d_ r)$ associated to this double complex, see Homology, Section 12.25.

By Lemma 25.5.2 the complexes $s(\mathcal{I}^ q(K))$ are acyclic in positive degrees and have $H^0$ equal to $\mathcal{I}^ q(X)$. Hence by Homology, Lemma 12.25.4 the natural map

\[ \mathcal{I}^\bullet (X) \longrightarrow \text{Tot}(A^{\bullet , \bullet }) \]

is a quasi-isomorphism of complexes of abelian groups. In particular we conclude that $H^ n(\text{Tot}(A^{\bullet , \bullet })) = H^ n(X, \mathcal{F})$.

The map $s(\mathcal{F}(K)) \longrightarrow R\Gamma (X, \mathcal{F})$ of the lemma is the composition of the map $s(\mathcal{F}(K)) \to \text{Tot}(A^{\bullet , \bullet })$ followed by the inverse of the displayed quasi-isomorphism above. This works because $\mathcal{I}^\bullet (X)$ is a representative of $R\Gamma (X, \mathcal{F})$.

Consider the spectral sequence $({}'E_ r, {}'d_ r)_{r \geq 0}$. By Homology, Lemma 12.25.1 we see that

\[ {}'E_2^{p, q} = H^ p_ I(H^ q_{II}(A^{\bullet , \bullet })) \]

In other words, we first take cohomology with respect to $d_2$ which gives the groups ${}'E_1^{p, q} = \underline{H}^ q(\mathcal{F})(K_ p)$. Hence it is indeed the case (by the description of the differential ${}'d_1$) that ${}'E_2^{p, q} = \check{H}^ p(K, \underline{H}^ q(\mathcal{F}))$. By the above and Homology, Lemma 12.25.3 we see that this converges to $H^ n(X, \mathcal{F})$ as desired.

We omit the proof of the statements regarding the functoriality of the above constructions in the abelian sheaf $\mathcal{F}$ and the hypercovering $K$. $\square$


Comments (0)


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 01GU. Beware of the difference between the letter 'O' and the digit '0'.