Skip to content

Commit

Permalink
Improve handling LC/X versus X
Browse files Browse the repository at this point in the history
  • Loading branch information
aisejohan committed Feb 14, 2017
1 parent d23abe3 commit cf5b3ef
Showing 1 changed file with 188 additions and 54 deletions.
242 changes: 188 additions & 54 deletions sites-cohomology.tex
Original file line number Diff line number Diff line change
Expand Up @@ -4947,56 +4947,63 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}

\begin{lemma}
\label{lemma-describe-pullback-pi}
Let $X$ be an object of $\textit{LC}_{qc}$. Let $\mathcal{F}$ be a
sheaf on $X_{Zar}$. Then the sheaf $\pi^{-1}\mathcal{F}$ on
$\textit{LC}_{Zar}/X$ is given by the rule
Let $X$ be an object of $\textit{LC}$. Let $\mathcal{F}$ be a
sheaf on $X$. The rule
$$
\pi^{-1}\mathcal{F}(Y) = \Gamma(Y_{Zar}, f^{-1}\mathcal{F})
\textit{LC} \longrightarrow \textit{Sets},\quad
(f : Y \to X) \longmapsto \Gamma(Y, f^{-1}\mathcal{F})
$$
for $f : Y \to X$ in $\textit{LC}_{qc}$. Moreover $\pi^{-1}\mathcal{F}$
is a sheaf for the qc topology, i.e., the sheaf
$\epsilon^{-1}\pi^{-1}\mathcal{F}$ on $\textit{LC}_{qc}$
is given by the same formula.
is a sheaf in the qc topology, i.e., this rule defines a sheaf on
$\textit{LC}_{qc}$ and a fortiori on $\textit{LC}_{Zar}$.
If we view $\mathcal{F}$ as a sheaf on $X_{Zar}$, then the sheaf
above is equal to $\epsilon^{-1}\pi^{-1}\mathcal{F}$ on
$\textit{LC}_{qc}$ and $\pi^{-1}\mathcal{F}$ on $\textit{LC}_{Zar}$.
\end{lemma}

\begin{proof}
Of course the pullback $f^{-1}$ on the right hand side indicates
usual pullback of sheaves on topological spaces
(Sites, Example \ref{sites-example-continuous-map}).
The equality of the lemma follows directly from the definitions.
Of course the pullback $f^{-1}$ in the formula denotes usual
pullback of sheaves on topological spaces. Let us denote
$\mathcal{F}^a$ the presheaf on the category $\textit{LC}$
defined by the rule given in the lemma. It is immediate
from the defitions that $\mathcal{F}^a$ is a sheaf for the Zar
topology.

\medskip\noindent
Let $\mathcal{V} = \{g_i : Y_i \to Y\}_{i \in I}$ be a covering of
$\textit{LC}_{qc}/X$. It suffices to show that
$\pi^{-1}\mathcal{F}(Y) \to H^0(\mathcal{V}, \pi^{-1}\mathcal{F})$
Let $Y \to X$ be a morphism in $\textit{LC}$.
$\mathcal{V} = \{g_i : Y_i \to Y\}_{i \in I}$ be a qc covering.
To prove $\mathcal{F}^a$ is a sheaf for the qc topology it
suffices to show that
$\mathcal{F}^a(Y) \to H^0(\mathcal{V}, \mathcal{F}^a)$
is an isomorphism, see Sites, Section \ref{sites-section-sheafification}.
We first point out that the map is injective as a qc covering
is surjective and we can detect equality of sections at stalks
(use Sheaves, Lemmas \ref{sheaves-lemma-sheaf-subset-stalks} and
\ref{sheaves-lemma-stalk-pullback-presheaf}). Thus we see that
$\pi^{-1}\mathcal{F}$ is a separated presheaf on $\textit{LC}_{qc}$
\ref{sheaves-lemma-stalk-pullback-presheaf}). Thus
$\mathcal{F}^a$ is a separated presheaf on $\textit{LC}_{qc}$
hence it suffices to show that any element
$(s_i) \in H^0(\mathcal{V}, \pi^{-1}\mathcal{F})$
maps to an element in the image of $\pi^{-1}\mathcal{F}(Y)$
$(s_i) \in H^0(\mathcal{V}, \mathcal{F}^a)$
maps to an element in the image of $\mathcal{F}^a(Y)$
after replacing $\mathcal{V}$ by a refinement
(Sites, Theorem \ref{sites-theorem-plus}).

\medskip\noindent
Observe that $\pi^{-1}\mathcal{F}|_{Y_{i, Zar}}$ is the pullback
of $f^{-1}\mathcal{F} = \pi^{-1}\mathcal{F}|_{Y_{Zar}}$ under
Identifying sheaves on $Y_{i, Zar}$ and sheaves on $Y_i$ we find that
$\mathcal{F}^a|_{Y_{i, Zar}}$ is the pullback of $f^{-1}\mathcal{F}$ under
the continuous map $g_i : Y_i \to Y$. Thus we can choose an open covering
$Y_i = \bigcup V_{ij}$ such that for each $j$ there is an open
$W_{ij} \subset Y$ and a section $t_{ij} \in \pi^{-1}\mathcal{F}(W_{ij})$
such that $s|_{U_{ij}}$ is the pullback of $t_{ij}$. In other words,
$W_{ij} \subset Y$ and a section $t_{ij} \in \mathcal{F}^a(W_{ij})$
such that $V_{ij}$ maps into $W_{ij}$ and such that
$s|_{V_{ij}}$ is the pullback of $t_{ij}$. In other words,
after refining the covering $\{Y_i \to Y\}$ we may assume there
are opens $W_i \subset Y$ such that $Y_i \to Y$ factors through $W_i$
and sections $t_i$ of $\pi^{-1}\mathcal{F}$ over $W_i$ which restrict
and sections $t_i$ of $\mathcal{F}^a$ over $W_i$ which restrict
to the given sections $s_i$. Moreover, if $y \in Y$ is in the image
of both $Y_i \to Y$ and $Y_j \to Y$, then the images $t_{i, y}$
and $t_{j, y}$ in the stalk $f^{-1}\mathcal{F}_y$ agree
(because $s_i$ and $s_j$ agree over $Y_i \times_Y Y_j$).
Thus for $y \in Y$ there is a well defined element $t_y$ of
$f^{-1}\mathcal{F}_y$ agreeing with $t_{i, y}$ whenever $y \in Y_i$.
$f^{-1}\mathcal{F}_y$ agreeing with $t_{i, y}$ whenever $y$
is in the image of $Y_i \to Y$.
We will show that the element $(t_y)$ comes from a global section
of $f^{-1}\mathcal{F}$ over $Y$ which will finish the proof of the lemma.

Expand All @@ -5006,34 +5013,58 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}
Pick $i_1, \ldots, i_n \in I$ and
quasi-compact subsets $E_j \subset Y_{i_j}$ such that
$\bigcup g_{i_j}(E_j)$ is a neighbourhood of $y_0$.
Then we can find an open neighbourhood $V \subset Y$ of $y_0$
contained in $W_{i_1} \cap \ldots \cap W_{i_n}$ such that
the sections $t_{i_j}|_V$, $j = 1, \ldots, n$ agree.
Hence we see that $(t_y)_{y \in V}$ comes from this section
and the proof is finished.
Let $V \subset Y$ be an open neighbourhood of $y_0$ contained
in $\bigcup g_{i_j}(E_j)$ and contained in $W_{i_1} \cap \ldots \cap W_{i_n}$.
Since $t_{i_1, y_0} = \ldots = t_{i_n, y_0}$, after shrinking $V$
we may assume the sections $t_{i_j}|_V$, $j = 1, \ldots, n$ of
$f^{-1}\mathcal{F}$ agree. As $V \subset \bigcup g_{i_j}(E_j)$
we see that $(t_y)_{y \in V}$ comes from this section.

\medskip\noindent
We still have to show that $\mathcal{F}^a$ is equal to
$\epsilon^{-1}\pi^{-1}\mathcal{F}$ on $\textit{LC}_{qc}$,
resp.\ $\pi^{-1}\mathcal{F}$ on $\textit{LC}_{Zar}$.
In both cases the pullback is defined by taking the presheaf
$$
(f : Y \to X)
\longmapsto
\colim_{f(Y) \subset U \subset X} \mathcal{F}(U)
$$
and then sheafifying. Sheafifying in the Zar topology
exactly produces our sheaf $\mathcal{F}^a$ and the fact
that $\mathcal{F}^a$ is a qc sheaf, shows that it works as well
in the qc topology.
\end{proof}

\begin{lemma}
\label{lemma-compare-cohomology-LC}
Let $X$ be an object of $\textit{LC}_{qc}$. Let $\mathcal{F}$ be an
abelian sheaf on $X_{Zar}$. Then we have
Let $X$ be an object of $\textit{LC}$. Let $\mathcal{F}$ be an
abelian sheaf on $X$. Then
$$
H^q(X_{Zar}, \mathcal{F}) =
H^q(\textit{LC}_{qc}/X, \epsilon^{-1}\pi^{-1}\mathcal{F})
H^q(X, \mathcal{F}) =
H^q(\textit{LC}_{Zar}/X, \mathcal{F}^a) =
H^q(\textit{LC}_{qc}/X, \mathcal{F}^a)
$$
where $\mathcal{F}^a = \epsilon^{-1}\pi^{-1}\mathcal{F} = \pi^{-1}\mathcal{F}$
is as in Lemma \ref{lemma-describe-pullback-pi}.
In particular, if $A$ is an abelian group, then we have
$H^q(X, \underline{A}) = H^q(\textit{LC}_{qc}/X, \underline{A})$.
\end{lemma}

\begin{proof}
The statement is more precisely that the canonical map
The equality $H^q(X, \mathcal{F}) = H^q(\textit{LC}_{Zar}/X, \mathcal{F}^a)$
is a general fact coming from the trivial observation that
coverings of $X$ in $\textit{LC}_{Zar}$ are the same thing as open
coverings of $X$. The reader who wishes to see a detailed proof
should apply Lemma \ref{lemma-cohomology-bigger-site} to the functor
$X_{Zar} \to \textit{LC}_{Zar}$. In the rest of the proof we
show that
$$
H^q(X_{Zar}, \mathcal{F}) \longrightarrow
H^q(\textit{LC}_{qc}/X, \epsilon^{-1}\pi^{-1}\mathcal{F})
H^q(X, \mathcal{F}) = H^q(\textit{LC}_{qc}/X, \mathcal{F}^a)
$$
is an isomorphism for all $q$.
The result holds for $q = 0$ by Lemma \ref{lemma-describe-pullback-pi}.
We argue by induction on $q$. Pick $q_0 > 0$. We will assume the result
We argue by induction on $q$.
The base case $q = 0$ is Lemma \ref{lemma-describe-pullback-pi}.
Pick $q_0 > 0$. We will assume the result
holds for $q < q_0$ and prove it for $q_0$.

\medskip\noindent
Expand All @@ -5047,25 +5078,25 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}
E_2^{p, q} = \check{H}^p(\mathcal{U}, \underline{H}^q(\mathcal{F}))
\longrightarrow
E_2^{p, q} =
\check{H}^p(\mathcal{U}, \underline{H}^q(\epsilon^{-1}\pi^{-1}\mathcal{F}))
\check{H}^p(\mathcal{U}, \underline{H}^q(\mathcal{F}^a))
$$
between the spectral sequences of
Cohomology, Lemma \ref{cohomology-lemma-cech-spectral-sequence} and
Lemma \ref{lemma-cech-spectral-sequence}.
Since the maps
Lemma \ref{lemma-cech-spectral-sequence}. The maps
$\underline{H}^q(\mathcal{F})(U_{i_0 \ldots i_p}) \to
\underline{H}^q(\epsilon^{-1}\pi^{-1}\mathcal{F}))(U_{i_0 \ldots i_p})$
are isomorphisms for $q < q_0$ we see that
\underline{H}^q(\mathcal{F}^a)(U_{i_0 \ldots i_p})$
are isomorphisms for $q < q_0$ by the induction hypothesis
applied to $\mathcal{F}|_{U_{i_0 \ldots i_p}}$. Thus
$$
\text{Ker}(H^{q_0}(X, \mathcal{F}) \to \prod H^{q_0}(U_i, \mathcal{F}))
$$
maps isomorphically to the corresponding subgroup of
$H^{q_0}(\textit{LC}_{qc}/X, \epsilon^{-1}\pi^{-1}\mathcal{F})$.
$H^{q_0}(\textit{LC}_{qc}/X, \mathcal{F}^a)$.
In this way we conclude that our map is injective for $q_0$.

\medskip\noindent
Surjective.
Let $\xi \in H^{q_0}(\textit{LC}_{qc}/X, \epsilon^{-1}\pi^{-1}\mathcal{F})$.
Let $\xi \in H^{q_0}(\textit{LC}_{qc}/X, \mathcal{F}^a)$.
If for every $x \in X$ we can find a neighbourhood $x \in U \subset X$
such that $\xi|_U = 0$, then we can use the {\v C}ech complex argument
of the previous paragraph to conclude that $\xi$ is in the image of
Expand All @@ -5077,16 +5108,17 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}
We may replace $X$ by $\bigcup f_{i_j}(E_j)$ and set $Y = \coprod E_{i_j}$.
Then $Y \to X$ is a surjective continuous map of Hausdorff and
quasi-compact topological spaces,
$\xi \in H^{q_0}(\textit{LC}_{qc}/X, \epsilon^{-1}\pi^{-1}\mathcal{F})$,
$\xi \in H^{q_0}(\textit{LC}_{qc}/X, \mathcal{F}^a)$,
and $\xi|_Y = 0$. Set $Y_p = Y \times_X \ldots \times_X Y$ ($p + 1$-factors)
and denote $\mathcal{F}_p$ the pullback of $\mathcal{F}$ to $Y_p$.
Then the spectral sequence
Consider the spectral sequence
$$
E_1^{p, q} =
\check{C}^p(\{Y \to X\}, \underline{H}^q(\epsilon^{-1}\pi^{-1}\mathcal{F}))
\check{C}^p(\{Y \to X\}, \underline{H}^q(\mathcal{F}^a))
$$
of Lemma \ref{lemma-cech-spectral-sequence} has
rows for $q < q_0$ which are (by induction) the complexes
of Lemma \ref{lemma-cech-spectral-sequence}. By induction hypothesis
applied to the pullback of $\mathcal{F}$ to $Y_p$ the $q$th
row of $E_1^{p, q}$ for $q < q_0$ is the complex
$$
H^q(Y_0, \mathcal{F}_0) \to
H^q(Y_1, \mathcal{F}_1) \to
Expand All @@ -5105,9 +5137,9 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}
H^q(Y_1 \times_X U, \mathcal{F}_1) \to
H^q(Y_2 \times_X U, \mathcal{F}_2) \to \ldots\right)
$$
are exact (some details omitted). By the proper base change
are exact (small detail omitted). By the proper base change
theorem in topology
(for example Cohomology, Lemma \ref{cohomology-lemma-proper-base-change})
(Cohomology, Lemma \ref{cohomology-lemma-proper-base-change})
the colimit is equal to
$$
H^q(Y_x, \underline{\mathcal{F}_x}) \to
Expand Down Expand Up @@ -5139,6 +5171,108 @@ \section{Cohomology on Hausdorff and locally quasi-compact spaces}
has the required exactness properties this finishes the proof of the lemma.
\end{proof}

\begin{lemma}
\label{lemma-cohomological-descent-LC}
Let $X$ be an object of $\textit{LC}$. For $K \in D^+(X)$ the maps
$$
K \to R\pi_*\pi^{-1}K
\quad\text{and}\quad
K \to R(\pi \circ \epsilon)_*(\pi \circ \epsilon)^{-1}K
$$
are isomorphisms where
$\pi : \Sh(\textit{LC}_{Zar}/X) \to \Sh(X)$ and
$\pi \circ \epsilon : \Sh(\textit{LC}_{qc}/X) \to \Sh(X)$
are the morphisms of topoi described above.
\end{lemma}

\begin{proof}
This is a consequence of Lemma \ref{lemma-compare-cohomology-LC}
and simple properties of cohomology; we suggest the reader skip the proof.
Suppose that $\mathcal{F}$ is an abelian sheaf on $X$.
Let $\mathcal{F}^a = \epsilon^{-1}\pi^{-1}\mathcal{F} = \pi^{-1}\mathcal{F}$
be as in Lemma \ref{lemma-describe-pullback-pi}. Recall that
$$
R^q\pi_*\mathcal{F}^a
\quad\text{resp. }\quad
R^q(\pi \circ \epsilon)_*\mathcal{F}^a
$$
is the sheaf associated to the presheaf
$$
U \longmapsto H^q(\textit{LC}_{Zar}/U)
\quad\text{resp. }\quad
H^q(\textit{LC}_{qc}/U)
$$
See Lemma \ref{lemma-higher-direct-images}. By
Lemma \ref{lemma-compare-cohomology-LC} this is equal to
$U \mapsto H^q(U, \mathcal{F})$, i.e., the presheaf
$\underline{H}^q(\mathcal{F}$. Since this presheaf sheafifies
to zero for $q > 0$ and to $\mathcal{F}$ for $q = 0$ we find
that $R^q\pi_*\mathcal{F}^a = 0$,
resp.\ $R^q(\pi \circ \epsilon)_*\mathcal{F}^a = 0$
for $q > 0$ and that $\pi_*\mathcal{F}^a = \mathcal{F}$,
resp.\ $(\pi \circ \epsilon)_*\mathcal{F}^a = \mathcal{F}$.
Represent $K$ by a bounded below complex $\mathcal{F}^\bullet$.
By Leray's acyclicity lemma
(Derived Categories, Lemma \ref{derived-lemma-leray-acyclicity})
applied to $\pi^{-1}\mathcal{F}^\bullet$,
resp.\ $(\pi \circ \epsilon)^{-1}\mathcal{F}^\bullet$
and $\pi_*$, resp.\ $(\pi \circ \epsilon)_*$ we conclude.
\end{proof}

\begin{lemma}
\label{lemma-push-pull-LC}
Let $f : X \to Y$ be a morphism of $\textit{LC}$.
Then there is a commutative diagram
$$
\xymatrix{
\Sh(\textit{LC}_{qc}/X) \ar[r]_j \ar[d]_a &
\Sh(\textit{LC}_{qc}/Y) \ar[d]^b \\
\Sh(X) \ar[r]^f &
\Sh(Y)
}
$$
of topoi. If $f$ is proper, then $b^{-1} \circ f_* = j_* \circ a^{-1}$.
\end{lemma}

\begin{proof}
The morphism of topoi $j$ is the one from
Sites, Lemma \ref{sites-lemma-relocalize}
which in our case comes from the continuous functor
$Z/Y \mapsto Z \times_Y X/X$, see
Sites, Lemma \ref{sites-lemma-relocalize-given-fibre-products}.
The diagram commutes simply because the corresponding
continuous functors compose correctly
(see Sites, Lemma \ref{sites-lemma-composition-morphisms-sites}).
Let $\mathcal{F}$ be a sheaf on $X$. Let $g : Z \to Y$ be an object of
$\textit{LC}/Y$. Consider the fibre product
$$
\xymatrix{
Z' \ar[r]_{f'} \ar[d]_{g'} & Z \ar[d]^g \\
X \ar[r]^f & Y
}
$$
Then we have
$$
(j_*a^{-1}\mathcal{F})(Z/Y) =
(a^{-1}\mathcal{F})(Z'/X) =
\Gamma(Z', (g')^{-1}\mathcal{F}) =
\Gamma(Z, f'_*(g')^{-1}\mathcal{F})
$$
the second equality by Lemma \ref{lemma-describe-pullback-pi}.
On the other hand
$$
(b^{-1}f_*\mathcal{F})(Z/Y) = \Gamma(Z, g^{-1}f_*\mathcal{F})
$$
again by Lemma \ref{lemma-describe-pullback-pi}.
Hence by proper base change for sheaves of sets
(Cohomology, Lemma \ref{cohomology-lemma-proper-base-change-sheaves-of-sets})
we conclude the two sets are canonically isomorphic.
The isomorphism is compatible with restriction mappings
and defines an isomorphism $b^{-1}f_*\mathcal{F} = j_*a^{-1}\mathcal{F}$.
Thus an isomorphism of functors
$b^{-1} \circ f_* = j_* \circ a^{-1}$.
\end{proof}

\begin{lemma}
\label{lemma-proper-surjective-is-qc-covering}
Let $f : X \to Y$ be a morphism of $\textit{LC}$.
Expand Down

0 comments on commit cf5b3ef

Please sign in to comment.