Pandharipande-type Equations in Enumerative Geometry

• Pandharipande-type equations are a collection of universal identities that unite recursive, Hamiltonian, and degeneration relations across moduli spaces and curve counting theories.
• They convert geometric identities on moduli spaces into explicit constraints for generating series, clarifying structure in tautological rings and intersection theory.
• These equations underpin key results in Faber–Pandharipande identities, open KdV/Virasoro systems, and PT invariants on threefolds and fourfolds, serving as diagnostic invariants in complex geometric settings.

Searching arXiv for recent and foundational papers on Pandharipande-related equations, relations, and PT theory. In the literature represented here, “Pandharipande-type equations” is not a single standardized designation but an umbrella for several families of identities, recursion relations, and differential equations attached to Pandharipande and collaborators or to Pandharipande–Thomas curve counting. They occur in Gromov–Witten theory, the tautological rings of moduli spaces of curves, open intersection theory, and the stable-pair theories of threefolds and fourfolds. The common feature is not a single formalism but a recurrent mechanism: geometric identities on moduli spaces are converted into recursive, Hamiltonian, holomorphic-anomaly, Virasoro, or wall-crossing equations for generating series and invariants (Bouchard et al., 2011, Kramer et al., 2017, Iglesias et al., 2021, Buryak, 2014, Oberdieck et al., 2023).

1. Scope of the term in current usage

The bodies of work usually grouped under this heading are heterogeneous. Some are tautological relations on $\overline{\mathcal M}_{g,n}$; some are PDE systems for generating functions; some are holomorphic-anomaly or rationality statements for stable-pair partition functions; and some are localization or degeneration identities.

Family	Representative content	Papers
Faber–Pandharipande	constant-map free energies on $\mathbb C^3$; the cycle $K\times K-(2g-2)K_\Delta$	(Bouchard et al., 2011, Qiu, 2024)
Tautological relations	PPZ relations; Liu–Pandharipande relations; JPT formula	(Kramer et al., 2017, Iglesias et al., 2021, Dunin-Barkowski et al., 2015)
Open intersection equations	open KdV and open Virasoro constraints	(Buryak, 2014)
PT equations on threefolds	holomorphic anomaly, quasi-Jacobi structure, rationality, degeneration, GW/PT transfer	(Oberdieck et al., 2023, Anderson et al., 7 Apr 2026, Lin, 2021, Lin et al., 2023)
PT equations on fourfolds	toric CY$_4$ vertex, canonical half, JK-residue chamber structure	(Liu, 2023, Kimura et al., 16 Aug 2025)

This suggests that the expression is best understood as a research-program label rather than as the name of a single equation. The strongest unifying thread is that enumerative geometry supplies universal identities whose outputs are explicit functional constraints on generating series.

2. Faber–Pandharipande identities: free energies and Chow cycles

One major branch concerns the constant-map sector of Gromov–Witten theory. For the framed mirror curve of $\mathbb C^3$,

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$

the Eynard–Orantin topological recursion produces free energies

$$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$

and these free energies reproduce exactly the Faber–Pandharipande constant-map formula

$$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$

The proof passes through

$$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$

using the residue definition of $F_g$, the Chen–Zhou open-sector formula for $\mathbb C^3$0, the dilaton equation, and Mumford’s relation, thereby completing the remodeling conjecture for $\mathbb C^3$1 (Bouchard et al., 2011).

A separate but equally important Faber–Pandharipande object is the $\mathbb C^3$2-cycle on $\mathbb C^3$3

$\mathbb C^3$4

where $\mathbb C^3$5 is a connected smooth projective curve of genus $\mathbb C^3$6, $\mathbb C^3$7 is a canonical divisor, and $\mathbb C^3$8 denotes $\mathbb C^3$9 placed on the diagonal. This cycle has degree $K\times K-(2g-2)K_\Delta$0 and lies in the kernel of the Albanese map. For generic curves of genus $K\times K-(2g-2)K_\Delta$1, Green–Griffiths proved that it is nontorsion over $K\times K-(2g-2)K_\Delta$2, and Yin extended generic nontriviality over arbitrary fields of characteristic $K\times K-(2g-2)K_\Delta$3. For Shimura curves, however, Qiu proves that

$K\times K-(2g-2)K_\Delta$4

via the algebra of divisorial correspondences, motivic decomposition, Rosati-compatible projectors, and Hecke correspondences (Qiu, 2024).

The significance of these two appearances is structural. In the first, the Faber–Pandharipande formula is the constant-map closed-string free energy extracted from recursion on the mirror pair of pants. In the second, the same names identify a canonical Chow-theoretic $K\times K-(2g-2)K_\Delta$5-cycle whose behavior is arithmetic-sensitive: generically nontrivial, but vanishing on Shimura curves. The shared theme is that apparently universal expressions become diagnostic invariants of geometry.

3. Tautological relations and their conversion into equations

A second branch is formed by tautological identities on moduli spaces of curves and their translation into recursive equations for intersection numbers and integrable hierarchies. The Pandharipande–Pixton–Zvonkine relations begin from the $K\times K-(2g-2)K_\Delta$6-spin CohFT and the Givental–Teleman formula. In the formulation used to study the open moduli space $K\times K-(2g-2)K_\Delta$7, the relations are written as

$K\times K-(2g-2)K_\Delta$8

The crucial specialization is

$K\times K-(2g-2)K_\Delta$9

for which the $_4$0-matrix polynomials become explicit: $_4$1 Because many $_4$2 vanish at half-integer values, the resulting relations simplify drastically and yield a triangular linear system on tautological monomials. This produces the bounds

$_4$3

together with explicit lower-degree estimates (Kramer et al., 2017).

The Liu–Pandharipande relations are another universal source of equations. In the form used for the second Dubrovin–Zhang bracket, the basic tautological identity is

$_4$4

After pushforwards, these yield the LP1–LP3 corollaries used to control correlators $_4$5 and the transformed second Poisson bracket

$_4$6

The paper proves that the coefficients are rational with prescribed singularities and, crucially,

$_4$7

so all negative-degree terms vanish (Iglesias et al., 2021).

A third interaction-theoretic example is the Johnson–Pandharipande–Tseng formula for orbifold Hurwitz numbers. The key chain is

$_4$8

For the spectral curve

$_4$9

the Eynard–Orantin recursion yields the same intersection-theoretic expression as the special case of the Johnson–Pandharipande–Tseng formula for the cyclic group $\mathbb C^3$0, giving a new proof of JPT from combinatorics and topological recursion rather than the reverse implication (Dunin-Barkowski et al., 2015).

Taken together, these results illustrate a characteristic Pandharipande-type mechanism: tautological relations among $\mathbb C^3$1-classes and boundary strata become triangular systems, recursion relations, or structural vanishing theorems for generating functions.

4. Open intersection theory: open KdV and open Virasoro

A distinct but closely related family of equations arises in the Pandharipande–Solomon–Tessler theory of Riemann surfaces with boundary. The relevant moduli space is

$\mathbb C^3$2

where $\mathbb C^3$3 is the doubled genus, $\mathbb C^3$4 the number of boundary marked points, and $\mathbb C^3$5 the number of interior marked points, with

$\mathbb C^3$6

The open potential is

$\mathbb C^3$7

The open string equation is

$\mathbb C^3$8

with initial condition

$\mathbb C^3$9

The conjectural open KdV system is

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$0

while the open Virasoro constraints are

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$1

The core theorem is that the open KdV system has a unique formal power series solution satisfying the initial condition, and that this solution automatically satisfies the open Virasoro equations. The proof uses the Burgers–KdV hierarchy, the commutation relations of the open Virasoro operators, and a reduction to low modes $C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$2 (Buryak, 2014).

This equivalence is conceptually important. It shows that the open theory admits both an evolutionary formulation, via KdV-type flows, and a constraint formulation, via Virasoro annihilation. In that sense, the open equations are a boundary analogue of the Witten–Kontsevich paradigm rather than a separate ad hoc recursion.

5. Pandharipande–Thomas equations on threefolds

In stable-pair theory, the phrase “Pandharipande-type equations” most often refers to structural equations for partition functions. For an elliptically fibered threefold

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$3

with $C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$4, the normalized $C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$5-relative PT generating series are conjectured to be meromorphic quasi-Jacobi forms and to satisfy two holomorphic anomaly equations. The quasi-Jacobi property states that

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$6

has weight

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$7

and index

$C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$8

The $C=\{x-y^f+y^{f+1}=0\}\subset (\mathbb{C}^*)^2,$9-anomaly equation expresses

$$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$0

in terms of splitting terms from the diagonal of $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$1, insertions twisted by $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$2, terms involving $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$3, and a quadratic term from a correspondence class $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$4. An $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$5-equation for $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$6 is also stated. These conjectures are proved in the equivariant anti-diagonal theory of $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$7, where the generating series reduce to Bloch–Okounkov $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$8-point functions (Oberdieck et al., 2023).

Another PT equation is rationality. For a smooth connected projective complex $$F_g = (-1)^g \frac{1}{2-2g}\operatorname*{Res}_{q=a}\Phi(q)\,W_{g,1}(q), \qquad g\ge 2,$$9-fold $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$0, an effective curve class $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$1 is positive if $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$2, and superpositive if every effective summand is positive. For superpositive classes, the descendent PT series

$$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$3

is proved to be the Laurent expansion of a rational function with poles only at $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$4 and roots of unity. The same paper does not prove the symmetry

$$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$5

A key intermediate statement is piecewise quasi-polynomiality in the Euler characteristic variable $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$6, derived from Joyce wall-crossing and periodicity under tensoring by ample line bundles (Anderson et al., 7 Apr 2026).

Degeneration and correspondence equations form the third large subfamily. Relative orbifold PT theory is constructed for projective Deligne–Mumford stacks by using expanded degenerations and expanded pairs. The resulting moduli stacks are separated and proper Deligne–Mumford stacks of finite type, and the absolute descendant invariants on a smooth fiber are expressed as sums of products of relative invariants on the two components of the central fiber, glued by the diagonal class on the Hilbert stack of the boundary divisor (Lin, 2021). In a different direction, a projective conifold transition

$$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$7

transfers the GW/PT descendent correspondence from the resolution $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$8 to the smoothing $$F_g = (-1)^g\, \frac{|B_{2g}|\,|B_{2g-2}|}{2(2g)(2g-2)(2g-2)!}, \qquad g\ge 2.$$9 for stationary descendent insertions. The proof compares the two semistable degenerations

$$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$0

and uses degeneration formulas in both GW and PT theories (Lin et al., 2023).

These results make clear that PT theory has acquired its own equation-based architecture: anomaly equations, rationality equations, degeneration identities, and correspondence transformations all act on generating functions rather than on individual invariants.

6. Fourfold vertices, chamber structure, and adjacent terminology

For toric Calabi–Yau $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$1-folds, the PT analogue of the vertex formalism is substantially subtler than in dimension three. One explicit proposal identifies torus-fixed PT loci as quiver Grassmannians and introduces a canonical choice of half of the symmetric obstruction theory, so that the Oh–Thomas localized virtual cycle becomes computable in equivariant $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$2-theory. The DT/PT vertex correspondence is then verified by computer in low degrees. The central geometric message is that PT fixed loci in dimension four can be positive-dimensional, singular, and even non-reduced, so the vertex is no longer a purely point-counting combinatorics (Liu, 2023).

A complementary formulation uses Jeffrey–Kirwan residues. In that approach, the DT and PT $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$3-vertices are obtained from the same contour-integral integrand; the only change is the reference vector: $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$4 Thus the PT $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$5-vertex is produced by a chamber change in JK residue rather than by a new integrand. The corresponding DT/PT relation is multiplication by the universal magnificent-four factor $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$6, and the formalism is extended to higher-rank and supergroup-like settings (Kimura et al., 16 Aug 2025).

A separate literature studies “Painlevé-type equations,” and it should not be conflated with the Pandharipande-related equations above. In the unramified four-dimensional isomonodromic classification there are $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$7 types of $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$8-dimensional Painlevé-type equations, of which $$F_g=\frac12\left\langle \lambda_{g-1}^3 \right\rangle,$$9 are partial differential systems and $F_g$0 are ordinary differential systems, all organized by singularity pattern and spectral type (Kawakami et al., 2012). In the additive difference setting, the Padé interpolation method yields time evolution equations, scalar Lax pairs, and determinant formulae of hypergeometric special solutions for additive $F_g$1-Painlevé equations of types $F_g$2, $F_g$3, $F_g$4, and $F_g$5 (Nagao, 2017).

The juxtaposition is informative. On one side are equations arising from enumerative geometry, tautological rings, and stable-pair moduli; on the other is the Hamiltonian-isomonodromic meaning of “type equations.” The current Pandharipande-related corpus belongs to the former domain. Its characteristic output is not a single canonical differential equation but a network of universal identities—Faber–Pandharipande formulas and cycles, PPZ and Liu–Pandharipande relations, open KdV/Virasoro systems, PT holomorphic anomaly and rationality equations, degeneration formulas, correspondence transforms, and fourfold vertex chamber identities—through which enumerative geometry constrains generating functions at every level of complexity.