TALOS: The Magnetic Record of Mars
MARS HOLDS A RECORD.
EOLISA TALOS 1.0.0
TALOS brings magnetic evidence, geographic controls, and thermal physics into one reproducible investigation of Mars.
A planetary question, precisely defined
Mars no longer generates a global magnetic field from its core. Yet ancient crust retains a strong, uneven magnetic record. Several of the planet’s largest impact basins coincide with weak orbital signatures. Their contrast is important because the surviving field offers evidence about rocks and conditions that reach deep into planetary history.
Reading that evidence requires care. Heating, excavation, magnetic mineralogy, source depth, and cancellation between opposing directions can all influence the field measured from orbit. TALOS, developed by Eolisa Space, investigates which conclusions survive those ambiguities. The work connects a local magnetic measurement, a test of geographic prediction, and controlled physical experiments.
Establish the magnetic contrast
The observational analysis evaluates published magnetic-field models by Langlais and colleagues and by Morschhauser and colleagues. It integrates total vector magnitude over spherical cells, with weights that respect their different surface areas. Basin geometry follows the source catalogue of Frey: Hellas, Utopia, Argyre, and Isidis form the primary comparison.
Each basin interior is paired with an adjacent annulus of equal continuous spherical area. At 200 km above the primary model’s reference radius, the four-basin joint geometric-mean interior-to-ring field ratio is 0.410. The corresponding ratio in the comparison model is 0.424. These values describe weaker modelled orbital fields inside the adopted footprints.
The study then moves the complete basin constellation through 1,999 rigid rotations, preserving its internal geometry. A separate rotation control retains the basins’ latitudes. Both place the observed alignment in an unusual part of their reference distributions. Individual basin evidence is uneven, and neither control assigns an impact cause. The two field models also share some underlying spacecraft observations.
Test what prediction has learned
A model can predict a field pattern from distances to basins while learning something more general: location. TALOS makes that ambiguity explicit. When landmark vectors span three dimensions, the cosines of angular distances contain enough information to reconstruct position on a sphere.
This identity supplies a geographic control. Prediction experiments compare basin distances with direct coordinates, surface composition and topography, and a low-degree harmonic baseline. Buffered validation excludes nearby training locations around held-out regions. Basin-distance performance weakens as the geographic gap increases, and the tested combinations provide no stable gain over the geographic alternatives. Predictive success therefore requires interpretation in the context of the task and its spatial structure.
Follow the thermal consequences
The physical investigation asks what follows from stated assumptions about heat extraction at the core–mantle boundary. TALOS compares 2,048 pairs of thermal histories that share background parameters and differ in prescribed impact forcing. Within this unfitted design, the median additional duration below a selected thermal threshold is 81 million years. That is a scenario summary, not an inferred age of Martian dynamo loss.
A separate experiment deposits the same retained energy at two depths in an idealized spherical mantle. Heat close to the core boundary and heat close to the surface produce very different boundary responses. The solver tracks stored and transferred energy through a conservative finite-volume formulation, with independent convergence checks. This experiment isolates deposition depth; it does not calculate an impact shock, mantle convection, or magnetic induction.
A research record open to inspection
TALOS combines the source data, catalogue provenance, executable code, derived grids, numerical results, scientific figures, and manuscript in one research package. Its executed suite records 91 passing tests. The checks address scientific behavior: spherical geometry, geographic separation, limiting cases, energy conservation, and analytic diffusion. Numerical verification supports the computational record while leaving geological uncertainty visible.
The project credits Ahmad Jabakenji’s 2025 concept note as its intellectual starting point and retains attribution to the researchers and archives behind its source data. Its contribution lies in the integrated application of established methods to a precise problem of inference. A distinctive framework becomes scientifically useful when its assumptions and consequences can be inspected, repeated, and challenged.
Watch the investigation
The English TALOS documentary connects the planetary question to the full technical study. Animated field reconstructions show how orbital altitude changes the view. Geometric experiments reveal the information inside distances. Calculated thermal evolution follows heat through the model, with clear distinctions between observations, evaluated fields, and conditional simulations.
The research identifies concrete targets for the next investigation: dated remanence, signed-field coherence, magnetic mineralogy, and constraints on source depth. Each can help separate histories that a weak orbital signature alone cannot resolve. TALOS establishes the computational foundation and the questions those observations need to answer.
