Processing math: 100%

The Stacks project

Lemma 21.40.1. Assumptions and notation as in Situation 21.38.1. For \mathcal{F} in \textit{PAb}(\mathcal{C}) and n \geq 0 consider the abelian sheaf L_ n(\mathcal{F}) on \mathcal{D} which is the sheaf associated to the presheaf

V \longmapsto H_ n(\mathcal{C}_ V, \mathcal{F}|_{\mathcal{C}_ V})

with restriction maps as indicated in the proof. Then L_ n(\mathcal{F}) = L_ n(\mathcal{F}^\# ).

Proof. For a morphism h : V' \to V of \mathcal{D} there is a pullback functor h^* : \mathcal{C}_ V \to \mathcal{C}_{V'} of fibre categories (Categories, Definition 4.33.6). Moreover for U \in \mathop{\mathrm{Ob}}\nolimits (\mathcal{C}_ V) there is a strongly cartesian morphism h^*U \to U covering h. Restriction along these strongly cartesian morphisms defines a transformation of functors

\mathcal{F}|_{\mathcal{C}_ V} \longrightarrow \mathcal{F}|_{\mathcal{C}_{V'}} \circ h^*.

By Example 21.39.3 we obtain the desired restriction map

H_ n(\mathcal{C}_ V, \mathcal{F}|_{\mathcal{C}_ V}) \longrightarrow H_ n(\mathcal{C}_{V'}, \mathcal{F}|_{\mathcal{C}_{V'}})

Let us denote L_{n, p}(\mathcal{F}) this presheaf, so that L_ n(\mathcal{F}) = L_{n, p}(\mathcal{F})^\# . The canonical map \gamma : \mathcal{F} \to \mathcal{F}^+ (Sites, Theorem 7.10.10) defines a canonical map L_{n, p}(\mathcal{F}) \to L_{n, p}(\mathcal{F}^+). We have to prove this map becomes an isomorphism after sheafification.

Let us use the computation of homology given in Example 21.39.2. Denote K_\bullet (\mathcal{F}|_{\mathcal{C}_ V}) the complex associated to the restriction of \mathcal{F} to the fibre category \mathcal{C}_ V. By the remarks above we obtain a presheaf K_\bullet (\mathcal{F}) of complexes

V \longmapsto K_\bullet (\mathcal{F}|_{\mathcal{C}_ V})

whose cohomology presheaves are the presheaves L_{n, p}(\mathcal{F}). Thus it suffices to show that

K_\bullet (\mathcal{F}) \longrightarrow K_\bullet (\mathcal{F}^+)

becomes an isomorphism on sheafification.

Injectivity. Let V be an object of \mathcal{D} and let \xi \in K_ n(\mathcal{F})(V) be an element which maps to zero in K_ n(\mathcal{F}^+)(V). We have to show there exists a covering \{ V_ j \to V\} such that \xi |_{V_ j} is zero in K_ n(\mathcal{F})(V_ j). We write

\xi = \sum (U_{i, n + 1} \to \ldots \to U_{i, 0}, \sigma _ i)

with \sigma _ i \in \mathcal{F}(U_{i, 0}). We arrange it so that each sequence of morphisms U_ n \to \ldots \to U_0 of \mathcal{C}_ V occurs are most once. Since the sums in the definition of the complex K_\bullet are direct sums, the only way this can map to zero in K_\bullet (\mathcal{F}^+)(V) is if all \sigma _ i map to zero in \mathcal{F}^+(U_{i, 0}). By construction of \mathcal{F}^+ there exist coverings \{ U_{i, 0, j} \to U_{i, 0}\} such that \sigma _ i|_{U_{i, 0, j}} is zero. By our construction of the topology on \mathcal{C} we can write U_{i, 0, j} \to U_{i, 0} as the pullback (Categories, Definition 4.33.6) of some morphisms V_{i, j} \to V and moreover each \{ V_{i, j} \to V\} is a covering. Choose a covering \{ V_ j \to V\} dominating each of the coverings \{ V_{i, j} \to V\} . Then it is clear that \xi |_{V_ j} = 0.

Surjectivity. Proof omitted. Hint: Argue as in the proof of injectivity. \square


Comments (0)


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.