This preprint investigates whether information-geometric quantities derived from multivariate abundance data can approximate instability signals predicted by fluctuation–dissipation theory (FDT) in stochastic dynamical systems. We introduce the CRTI (Covariance–Response Temperature Index) estimator T_est = Tr(F⁻¹)/Tr(F), constructed from the empirical Fisher Information Matrix, and evaluate its performance against the theoretically grounded observable T_true = Tr(Σ)/Tr(Σ⁻¹) derived from stochastic generalized Lotka–Volterra (gLV) dynamics. A reduced ensemble simulation (10 replicates, S=15 species, 30 timesteps) yields mean Spearman correlation ρ(T_true, T_est) = 0.983, demonstrating that the information-geometric estimator closely tracks the true fluctuation–dissipation signal even in small-sample regimes. Simulation code is included as supplementary material. early warning signals · Fisher information · fluctuation-dissipation theory · critical transitions · generalized Lotka-Volterra · information geometry · complex systems · regime shifts · critical slowing down · compositional data · microbiome · ecological networks Resource Type Publication → PreprintLicense Creative Commons Attribution 4.0 International (CC BY 4.0)Language English