---
title: Tame Stacky Node
url: https://www.emergentmind.com/topics/tame-stacky-node
type: topic
---

# Tame Stacky Node

Searching arXiv for the cited papers to ground the article in current records.
A tame stacky node is the local stack structure at a node of a tame Deligne–Mumford stacky curve such that, after passing to a strict henselization or étale neighborhood, the node is modeled by a quotient of the ordinary nodal curve \(\operatorname{Spec}\mathbbm k[x,y]/(xy)\) by a finite stabilizer group. In the split-node setting, the local model is \([\operatorname{Spec} R/G]\), where \(R\) is the strict henselization of the local ring of \(\mathbbm k[x,y]/(xy)\) at the origin and \(G\) is either cyclic or branch-swapping [2509.20629]. In logarithmic formulations, the same phenomenon is encoded by a simple morphism of log structures and appears as the cyclic orbifold chart \(\left[\operatorname{Spec}(k[u,v]/(uv))/\mu_r\right]\) with weights \((1,-1)\) [2412.03408]. The tame condition means that the relevant stabilizer orders are prime to the residue characteristics, so the local geometry is governed by cyclic inertia, root stacks, rigidifications, and representable ramification rather than by wild higher-ramification data [2602.19771].

## 1. Local notion and ambient hypotheses

In the narrow sense developed for nodal stacky curves, a nodal stacky curve is a proper, geometrically integral, one-dimensional Deligne–Mumford stack \(\mathcal C\) over \(\mathbbm k\) with trivial generic stabilizer, such that at each closed geometric point the completed strict henselian local ring is either \(\bar{\mathbbm k}[[x]]\) or \(\bar{\mathbbm k}[[x,y]]/(xy)\). The analysis of tame stacky nodes in this setting assumes that all nodes are split, meaning that the node itself and its preimages in the normalization are \(\mathbbm k\)-points [2509.20629].

A broader logarithmic framework treats node stackiness as one part of a generalized log twisted curve
\[
\mathbf C=(C/S,\{s_i\}_{i=1}^n,\ell:M_S\hookrightarrow M_S',N),
\]
where the simple inclusion \(\ell:M_S\hookrightarrow M_S'\) produces stackiness at nodes and the admissible sheaf
\[
N\subset \bigoplus_i s_{i*}\mathbf Q_{\ge 0}
\]
produces stackiness at markings. The associated stack is
\[
\mathcal C:=\mathcal C^{}\times_C \mathcal C^N,
\]
with coarse space \(C\) and abelian stabilizers supported at nodes and markings [2412.03408].

The tameness hypothesis is decisive in each of these formulations. For tame Deligne–Mumford stacks and tame stacky curves, geometric stabilizer orders are prime to the characteristic of the base field, which is precisely what permits cyclic quotient models, root-stack descriptions, and tractable Picard- and Brauer-theoretic calculations [2507.08780]. In the modular-curve setting, this same tame behavior is built into the cyclotomic Deligne–Mumford formalism: the relevant local phenomena are finite cyclic inertia at stacky points together with ramification indices in representable maps, and the paper is explicit that these local stacky phenomena are not wild [2602.19771].

## 2. Classification of local models

The classification theorem for tame stacky nodes in the split-node case is completely explicit. After strict henselization,
\[
\mathcal C^{sh}\cong [\operatorname{Spec} R/G],
\]
and there are exactly two kinds of local stabilizer structure [2509.20629].

| Type | Stabilizer | Local description |
|---|---|---|
| Twisted node | \(G=\mu_n\) | \((x,y)\mapsto (\zeta_n x,\zeta_n^a y)\), \(a\in (\mathbb Z/n)^\times\) |
| Doubly-twisted node | \(1\to \mu_n\to G\to \mathbb Z/2\to 1\) | \(\mathbb Z/2\) swaps branches, with \(a^2\equiv 1\pmod n\) |

A twisted node is thus a cyclic quotient
\[
[\operatorname{Spec} R/\mu_n],
\qquad
(x,y)\longmapsto (\zeta_n x,\zeta_n^a y),
\]
with \(a\in (\mathbb Z/n)^\times\). Many authors reserve “twisted node” for the balanced case
\[
a\equiv -1 \pmod n,
\]
but in this classification \(a\) is allowed to be any unit modulo \(n\) [2509.20629]. The paper introduces the numerical parameters
\[
d_-=\gcd(a-1,n),\qquad d_+=\gcd(a+1,n),\qquad n_+=\frac{n}{d_+},
\]
which control both local inertia behavior and the Picard contribution of the node.

A doubly-twisted node has local structure
\[
[\operatorname{Spec} R/G],
\qquad
1\to \mu_n\to G\to \mathbb Z/2\to 1,
\]
where the quotient \(\mathbb Z/2\) exchanges the two branches and the cyclic subgroup still acts by
\[
(x,y)\mapsto (\zeta_n x,\zeta_n^a y).
\]
The branch-swapping involution forces
\[
a^2\equiv 1\pmod n.
\]
The local group is not determined by \(a\) alone: if \(\sigma\in G\) lifts the nontrivial element of \(\mathbb Z/2\), then
\[
\sigma^2=t^m
\]
for a generator \(t\) of \(\mu_n\), with \(m\in \mathbb Z/n\). The parameter \(m\) is well-defined modulo \((1+a)\), lies in the kernel of multiplication by \(1-a\), determines the extension class
\[
[G]\in H^2(\mathbb Z/2,\mu_n),
\]
and the extension is split iff
\[
d_+\mid m.
\]
The notation for this group is \(G_{n,a,m}\), and every admissible triple \((n,a,m)\) occurs on some nodal stacky curve [2509.20629].

An arithmetic dichotomy governs these local models:
\[
d_-d_+=n
\qquad\text{or}\qquad
d_-d_+=2n.
\]
This distinction later controls when the node contributes nontrivially to the Brauer group [2509.20629].

## 3. Logarithmic, root-stack, and ramification descriptions

In logarithmic language, a nodal coarse curve is locally given by the smoothing equation
\[
xy=t_e,
\]
with canonical log chart
\[
\mathbf N^2 \to \mathcal O_S[x,y]/(xy-t_e).
\]
The associated tame stacky node has the orbifold chart
\[
\left[\operatorname{Spec}(k[u,v]/(uv))/\mu_r\right],
\]
with action
\[
\zeta\cdot u=\zeta\,u,\qquad \zeta\cdot v=\zeta^{-1}v.
\]
This is the standard cyclic tame node model in the generalized log twisted-curve construction [2412.03408].

The same paper makes the deformation theory concrete under contraction. If a node with stabilizer \(\mu_{e_i}\) is contracted to one with smaller stabilizer \(\mu_{e_i'}\), where \(e_i' \mid e_i\), the local map is
\[
R[u^\dagger,v^\dagger]/(u^\dagger v^\dagger-(t')^{e_i/e_i'})
\to
R[u,v]/(uv-t'),
\]
with
\[
u^\dagger\mapsto u^{e_i/e_i'},
\qquad
v^\dagger\mapsto v^{e_i/e_i'}.
\]
This gives an explicit local rule for how tame stacky node structure changes under contraction [2412.03408].

Root stacks furnish the corresponding smooth-point and branch-point models. For tame stacky curves with trivial generic stabilizer, the local form near a stacky point is
\[
\left[\operatorname{Spec}\big(O_{X,x}[z]/(z^n-f)\big)/\mu_n\right],
\]
and globally such curves are identified with iterated root stacks along stacky points [2306.08227]. This does not itself give a nodal classification, but it isolates the same cyclic Kummer mechanism that underlies tame node charts.

In the modular-curve setting, rigidification and root-stack structure are carried out explicitly. For the Deligne–Rapoport stack \(\mathscr X_0(N)\), the core structural statement is
\[
\mathscr X_0(N)\cong \sqrt[2]{\Lambda_N^{\rig}/\mathscr X_0(N)^{\rig}},
\]
obtained from
\[
\mathscr X_0(1)\cong (4,6),
\qquad
\mathscr X_0(1)^{\rig}\cong (2,3),
\qquad
\mathscr X_0(1)\cong \sqrt[2]{\Lambda^{\rig}/\mathscr X_0(1)^{\rig}}
\]
by base change along \(J_N:\mathscr X_0(N)\to \mathscr X_0(1)\) [2602.19771]. The local model after rigidification is a root stack over a smooth formal disc,
\[
\sqrt[m]{\operatorname{Spec}(F[[T]])},
\]
with closed-point automorphism group \(\mu_m\); the paper also writes
\[
\sqrt[m]{\operatorname{Spec}(F[[T]])}
\cong
[\operatorname{Spec}(F[[T]][Y,Y^{-1}]) /_{(1,m)}]
\]
and, for the \(\mu_2\)-gerbe structure,
\[
[\operatorname{Spec}(F[[T]][Y,Y^{-1}]) /_{(2,m)}].
\]
This is a branch-point model rather than a nodal one, but it uses the same tame cyclic and root-stack technology [2602.19771].

## 4. Normalization, coarsening, Picard groups, and Brauer groups

The coarsening and normalization of a tame stacky node separate the singular quotient from the residual cyclic stabilizer structure. If \(\pi:\mathcal C\to C\) is the coarse moduli map and \(\eta:\tilde{\mathcal C}\to\mathcal C\) is the partial normalization, then in the doubly-twisted case the étale local structure of the normalization is
\[
\left[\operatorname{Spec} \mathbbm k[z]/\mu_n\right].
\]
Thus the normalization is smooth with cyclic stabilizer \(\mu_n\), and its composition with \(\mathcal C\to C\) is the coarse moduli map of \(\tilde{\mathcal C}\) [2509.20629].

The local node type is visible in the Picard group. For a twisted node,
\[
0\to \mathbbm k^\times \to \operatorname{Pic}\mathcal C \to \operatorname{Pic}\tilde{\mathcal C}\to \mathbb Z/n\to 0,
\]
and ultimately
\[
\operatorname{Pic} \mathcal C
=
\mathbbm k^{\times}\oplus\operatorname{Pic}\tilde C\left<\frac1{n_+}\hat p\right>\oplus\mathbb Z/d_+.
\]
In the balanced case \(a\equiv -1\pmod n\),
\[
\operatorname{Pic}\mathcal C=\mathbbm k^\times\oplus \operatorname{Pic}\tilde C\oplus \mathbb Z/n.
\]
For a doubly-twisted node,
\[
\operatorname{Pic}\mathcal C=\mu_2\oplus \operatorname{Pic} C\left<\frac1{d_-}p\right>.
\]
The \(\mu_2\) factor is the branch-swap contribution, and the root order is governed by \(d_-=\gcd(a-1,n)\) [2509.20629].

The Brauer-theoretic behavior is sharper. A twisted node contributes nothing because
\[
H^2(\mu_n,\mathbbm k^\times)=0.
\]
A doubly-twisted node can contribute a \(\mathbb Z/2\) only in the split case and only when
\[
d_-d_+=2n.
\]
Otherwise the local Brauer contribution is zero [2509.20629].

A more global cohomological formulation treats any singular stacky curve pointwise through stabilizer cohomology. For a separated tame Deligne–Mumford stack \(\mathcal X\to X\) with coarse moduli map \(f\), the local stalk formula is
\[
(\mathbf R^n f_*\mathbb G_m)_x=\mathcal H^n(G,K^\times),\qquad n\ge 2,
\]
and for a stacky curve
\[
\mathcal H^k(\mathcal C,\mathbb G_m)=\bigoplus_{i=1}^n \mathcal H^k(G_i,\mathbb k^\times),\qquad k\ge 2.
\]
The paper notes explicitly that the \(p_i\) contributing in this way must be singular points, since smooth stacky points in a tame stacky curve have only cyclic stabilizers [2507.08780]. This provides a cohomological criterion for distinguishing ordinary tame stacky points from genuinely singular stacky nodes.

## 5. Arithmetic realization on modular curves

The modular stack \(\mathscr X_0(N)\) supplies an arithmetic setting in which tame local stacky structure controls global height geometry. For
\[
N\in\mathfrak N_0=\{1,2,3,4,5,6,7,8,9,10,12,13,16,18,25\},
\]
these are precisely the values for which the coarse moduli space of \(\mathscr X_0(N)\) is isomorphic to \(\mathbb P^1\), and the stacky Batyrev–Manin conjecture is proved for the naive height when \(F=\mathbb Q\) [2602.19771].

For these stacks, the generic stabilizer is always \(\mu_2\), coming from the involution \([-1]\) on elliptic curves. The only points with larger stabilizer lie above the special \(j\)-values \(0\) and \(1728\), where the automorphism groups become \(\mu_6\) and \(\mu_4\), respectively. Their counts are encoded by
\[
\varepsilon_3(N):=\begin{cases}
0,&\text{ if } 9|N\\
\prod_{p|N}\bigg(1+ \bigg(\dfrac{-3}{p}\bigg)\bigg),&\text{otherwise}
\end{cases}
\]
and
\[
\varepsilon_2(N):=\begin{cases}
0&\text{ if } 4|N\\
\prod_{p|N}\bigg(1+\bigg(\dfrac{-1}{p}\bigg)\bigg)&\text{otherwise.}
\end{cases}
\]
The ordinary canonical degree is then
\[
\deg(K_{\mathscr X_0(N)})=-1+\frac{\varepsilon_3(N)}{3}+\frac{\varepsilon_2(N)}{4},
\]
which is exactly the contribution of the tame stacky points with stabilizers \(\mu_6\) and \(\mu_4\) [2602.19771].

The orbifold structure is recorded by
\[
K_{\mathscr X_0(N),\orb}=
\big(K_{\mathscr X_0(N)},(\operatorname{age}(y)-1)_{y\in\pi_0^*\mathcal J_0\mathscr X_0(N)}\big).
\]
There is always the untwisted sector \(0\), and there is the ubiquitous sector \(1/2\) of age \(0\) coming from the generic \(\mu_2\)-stabilizer. Over points with automorphism group \(\mu_6\) or \(\mu_4\), there are additional twisted sectors:
\[
\text{for }T_6\in\mathcal K_6,\quad (T_6,\zeta)\text{ has age }\{4\zeta\},
\]
and
\[
\text{for }T_4\in\mathcal K_4,\quad (T_4,\zeta)\text{ has age }\{2\zeta\}.
\]
These sectors enter the orbifold Néron–Severi space \(\operatorname{NS}_{\orb}\) and determine the orbifold corrections to the canonical class and height class [2602.19771].

The local ramification formulas are equally explicit. For a representable map of root stacks over DVRs,
\[
\sqrt[m]{\operatorname{Spec}(R')}\to\sqrt[n]{\operatorname{Spec}(R)},
\]
one has
\[
m=\frac{n}{\gcd(n,e)},
\qquad
\mu_m\to\mu_n,\quad x\mapsto x^{e/\gcd(n,e)}.
\]
Specialized formulas for the \(\mu_6\) and \(\mu_4\) cases are
\[
m=\frac{6}{\gcd(6,e)} \quad\text{or}\quad m=\frac{6}{\gcd(6,3+e)},
\]
with
\[
\mu_m\to\mu_6,\qquad x\mapsto x^e \quad\text{or}\quad x\mapsto x^{e+3},
\]
and
\[
m=\frac{4}{\gcd(e,4)} \quad\text{or}\quad m=\frac{4}{\gcd(e+2,4)},
\]
with
\[
\mu_m\to\mu_4,\qquad x\mapsto x^e \quad\text{or}\quad x\mapsto x^{e+2}.
\]
These formulas determine which sector a stacky point contributes to and thereby feed into the orbifold \(a\)- and \(b\)-invariants for the naive height [2602.19771].

## 6. Tame versus wild behavior and related frameworks

The tame classification depends on the fact that cyclic stabilizer order suffices to describe the local quotient geometry. In characteristic \(p>0\), this fails for cyclic stabilizers of order \(p\). The local root-stack model
\[
\sqrt[r]{(L,s)/X}
\]
must then be replaced by an Artin–Schreier root stack
\[
\mathcal R_m((L,s,f)/X),
\]
and the local quotient is governed by an Artin–Schreier equation
\[
y^p-y=F(x)
\]
together with the ramification jump \(m\) [1910.03146]. The paper’s central point is that, unlike the tame case, the order \(p\) of the stabilizer does not determine the local structure: one must record higher ramification data. A plausible implication is that the clean two-type classification of tame stacky nodes cannot extend verbatim to the wild cyclic case.

This contrast clarifies the role of root stacks in the tame theory. For tame stacky curves with trivial generic stabilizer, the local quotient form
\[
\left[\operatorname{Spec}\big(O_{X,x}[z]/(z^n-f)\big)/\mu_n\right]
\]
underlies Picard computations and rigidification arguments [2306.08227]. That paper does not give a separate theory of nodal stacky curves, but it isolates the same cyclic Kummer mechanism that reappears at tame nodes. This suggests that tame stacky nodes should be viewed as the nodal analogue of tame stacky points, with the ordinary node \(\operatorname{Spec}\mathbbm k[x,y]/(xy)\) replacing the smooth local parameter.

At the level of birational and toroidal geometry, smooth tame stacks with diagonalisable stabilisers admit a functorial destackification algorithm by sequences of ordinary blow-ups and root stacks, ending with a smooth coarse space and a residual gerbe over a root stack of a simple normal crossings divisor [1409.5713]. This is not a classification of nodal stacky curves, but it places tame local quotient singularities into a broader framework in which stackiness is systematically reduced to root-stack data. In that sense, the tame stacky node sits at the intersection of three parallel formalisms: quotient classification, logarithmic twisted curves, and root-stack rigidification.

Source: https://www.emergentmind.com/topics/tame-stacky-node