I am trying to find a reference for the following well-known result on the functoriality of the Leray spectral sequence:

Let $\pi:X\to Y$ and $\pi':X'\to Y'$ be morphisms of schemes and denote by $E_2^{p,q}(\mathcal{F})=\mathrm{H}^q(Y,R^q\pi_*\mathcal{F})\implies \mathrm{H}^{p+q}(X,\mathcal{F})$ and $(E')_2^{p,q}(\mathcal{G})\implies \mathrm{H}^{p+q}(X',\mathcal{G})$ their associated Leray spectral sequences (in étale cohomology for some étale sheafs $\mathcal{F},\mathcal{G}$). Assume that there are $f:X'\to X$ and $g:Y'\to Y$ such that the obvious square commutes. Then there is a morphism of spectral sequences $E_r^{p,q}(\mathcal{F})\to (E')^{p,q}_r(g^{-1}\mathcal{F})$ induced by the natural pullback maps on cohomology $E_2^{p,q}(\mathcal{F})\to (E')_2^{p,q}(g^{-1}\mathcal{F})$ such that the map on the infinity-pages agrees with the canonical pullback morphism $\mathrm{H}^n(X,\mathcal{F})\to \mathrm{H}^n(X',f^{-1}\mathcal{F})$.

It would be amazing if someone knows a reliable source for this frequently used result. Thanks in advance!


1 Answer 1


I apologize for the self promotion, but page 570 of my article The Leray spectral sequence is motivic has a very brief discussion of the functoriality of Leray.

Added In a bit more detail, here are the key points:

  • To every object $(C, F)$ in the (bounded below, biregularly filtered) filtered derived category, one has a spectral sequence with $E_1^{pq}= H^{p+q}(Gr^p_FC)$. This construction is functorial. See Deligne, Théorie de Hodge II, sect. 1; III, sect. 7.
  • With your notation, the spectral sequence associated to $(\mathbb{R} \pi_* \mathcal{F},\tau)$ becomes the Leray spectral sequence after reindexing to transform $E_1$ to $E_2$ [loc. cit.]. Here $\tau$ is the canonical filtration, given by truncations.
  • Let $\mathcal{F}'= f^{-1}\mathcal{F}$. Then we have a morphism in the filtered derived category $$ (\mathbb{R} \pi_* \mathcal{F},\tau)\to \mathbb{R}g_*(\mathbb{R} \pi'_* \mathcal{F}',\tau)$$ This will induce a morphism of Leray spectral sequences.
  • $\begingroup$ Thank you for the answer. I can see that you give a construction for the Leray spectral sequence and write under (3) that functoriality follows from the construction. It is still a bit unclear to me how this implies that one has the morphism of spectral sequences and that it works well with the convergence to the cohomology of $X$ and $X'$. Sorry I am quite new to the on the subject. $\endgroup$ Commented Feb 21, 2023 at 17:18
  • 1
    $\begingroup$ Amazing! This is something I can work with. Thank you for taking your time to answer. $\endgroup$ Commented Feb 22, 2023 at 14:15
  • 1
    $\begingroup$ You’re welcome. $\endgroup$ Commented Feb 22, 2023 at 14:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.