I apologize if this is straightforward, but I can't seem to find a reference: What is $H^1_{et}(\mathrm{Spec}(\mathbb{Z}), \mathbf{G}_m)$? It shouldn't be as simple as noting that the etale fundamental group of Spec Z is trivial, since Spec Z does admit nontrivial etale covers, just not finite ones.
Asked
Viewed
186 times

11$\begingroup$ It's much simpler: $H^1_{et}(X,\mathbf{G}_m)$ is always the Picard group $\mathrm{Pic}(X)$ for any scheme $X$. If $X = \mathrm{Spec}(R)$ for a PID R (e.g., $R = \mathbf{Z}$), then this group is trivial by algebra. $\endgroup$ – Meric Feb 2 at 20:55

5$\begingroup$ @Meric This is an answer (not a comment). $\endgroup$ – Martin Brandenburg Feb 3 at 0:24

$\begingroup$ What is a nontrivial etale cover of Spec(Z)? Asking for a friend. $\endgroup$ – Yosemite Stan Feb 5 at 4:26

$\begingroup$ The map $\mathrm{Spec}\, \mathbb{Z}[\sqrt{1}, \frac{1}{2}] \times \mathbb{Z}[\frac{1 + \sqrt{5}}{2}, \frac{1}{5}] \to \mathrm{Spec} \, \mathbb{Z}$ works. Extensions of number rings are ramified along the vanishing of their discriminants, so if we delete that vanishing, we have an etale map. Doing this with a pair of number rings with relatively prime discriminants gives us a surjective etale map. (My algebraic number theory is rusty, so I'm using the table on en.wikipedia.org/wiki/… to build my specific example.) $\endgroup$ – SebastianJB Feb 6 at 13:47

1$\begingroup$ You could also just look at Zariski covers, like $\mathbf{Z} \to \mathbf{Z}[1/2] \times \mathbf{Z}[1/3]$. These covers are enough to compute (via Cech cohomology) the relevant cohomology group anyways. $\endgroup$ – Meric Feb 8 at 18:31