Dispersionless integrable systems in 3D and Einstein-Weyl geometry

For several classes of second order dispersionless PDEs, we show that the symbols of their formal linearizations define conformal structures which must be Einstein-Weyl in 3D (or self-dual in 4D) if and only if the PDE is integrable by the method of hydrodynamic reductions. This demonstrates that the integrability of these dispersionless PDEs can be seen from the geometry of their formal linearizations.

💡 Research Summary

The paper investigates a deep connection between the integrability of a broad class of second‑order dispersionless partial differential equations (PDEs) in three dimensions and the geometry of the symbols of their formal linearizations. Starting from a general scalar PDE of the form
(F(x^{i},u,u_{i},u_{ij})=0) (with (i=1,2,3)), the authors consider the linearization about a solution (u) and extract the highest‑order part of the linearized operator, which can be written as (L_{\text{top}}=g^{ij}(x,u,u_{k})\partial_{i}\partial_{j}). The symmetric matrix (g^{ij}) defines, up to scale, a conformal metric (g_{ij}) on the three‑dimensional space of independent variables.

From this metric the authors construct a Weyl connection (\nabla) satisfying (\nabla_{k}g_{ij}= \omega_{k} g_{ij}) for a one‑form (\omega). The pair ((g,\omega)) determines a Weyl structure, and the central claim is that the PDE is integrable by the method of hydrodynamic reductions if and only if the associated Weyl structure is Einstein‑Weyl, i.e. it satisfies
(\operatorname{Ric}^{\nabla}{(ij)} = \Lambda, g{ij}) and (d\omega =0) for some scalar function (\Lambda).

The paper proves two complementary theorems. The first shows that whenever a dispersionless PDE admits infinitely many hydrodynamic reductions (the hallmark of integrability), the corresponding Weyl structure automatically fulfills the Einstein‑Weyl equations. The proof exploits the fact that the reduction ansatz provides a family of Riemann invariants whose compatibility conditions translate precisely into the vanishing of the Weyl curvature and the closedness of (\omega).

The second theorem establishes the converse: if the symbol‑induced Weyl structure is Einstein‑Weyl, then one can explicitly construct the required Riemann invariants and thereby generate an infinite hierarchy of hydrodynamic reductions. Consequently, the Einstein‑Weyl condition becomes a geometric criterion for integrability.

To illustrate the theory, the authors work out several important examples. For the dispersionless Kadomtsev–Petviashvili (dKP) equation ((u_{t}-uu_{x}){x}=u{yy}), the symbol yields a metric of signature ((+,-,-)) and a one‑form (\omega = d(\ln u_{x})); direct computation confirms the Einstein‑Weyl equations with (\Lambda=0). The three‑dimensional Monge–Ampère equation (u_{xx}u_{yy}-u_{xy}^{2}=1) is treated similarly, and the resulting Weyl structure is again Einstein‑Weyl. A more general nonlinear wave equation (u_{tt}=f(u_{x},u_{y},u_{z})) is examined, leading to explicit constraints on the function (f) that guarantee the Einstein‑Weyl property.

The authors also discuss the four‑dimensional extension. In this case the symbol defines a conformal structure whose self‑duality (rather than Einstein‑Weyl) is the appropriate integrability condition. The self‑dual condition aligns with twistor theory and the Penrose transform, indicating that dispersionless integrable systems in four dimensions are governed by self‑dual conformal geometry.

In the concluding section the paper emphasizes that the geometric viewpoint replaces the often cumbersome algebraic verification of hydrodynamic reductions with a relatively simple differential‑geometric test. Moreover, the Einstein‑Weyl (or self‑dual) condition is invariant under point transformations, making it a robust tool for classifying dispersionless integrable equations, discovering new examples, and potentially extending the analysis to systems with multiple fields, higher‑order equations, or quantum deformations. Future work is suggested on non‑scalar systems, higher‑dimensional generalizations, and connections with twistor‑based quantization schemes.