Wavelet-Scattering Signatures of Fuzzy Dark Matter in Simulated 21 cm Brightness-Temperature Maps

We study how fuzzy dark matter (FDM) affects the multiscale morphology of the redshifted 21 cm brightness-temperature field during Cosmic Dawn and the Epoch of Reionization. Using FDM-modified \texttt{21cmFAST} simulations, we apply the two-dimensional wavelet scattering transform (WST) to simulated 21 cm maps. The first-order coefficients $S_1(j)$ trace localized variance-like information, while the normalized second-order ratios $R(j_1,j_2)=S_2/S_1$ measure non-Gaussian cross-scale coupling. FDM shifts and reshapes both summaries, especially for scale pairs anchored on fine wavelet bands, reflecting delayed Ly$\alpha$ coupling, X-ray heating, and reionization. We compare power-spectrum-only, WST-only, and combined power-spectrum+WST Fisher forecasts using the same SKA1-Low-like mock-observation assumptions, redshift bins, nuisance parameters, thermal-noise model, and $uv$ filtering. The combined power-spectrum+WST data vector gives the tightest marginalized constraints, showing that WST carries information complementary to the power spectrum in the adopted local Fisher setup. We also include an idealized foreground-wedge avoidance test on three-dimensional light-cone chunks, demonstrating how contaminated low-$k_\parallel$ modes are removed before extracting WST summaries. A compact wedge-impact diagnostic shows that the normalized second-order ratio is less strongly reshaped by horizon-wedge filtering than the first-order amplitudes, while not constituting a foreground-aware detectability forecast. These calculations are not an end-to-end SKA foreground-subtraction pipeline, but they establish WST as an interpretable morphological summary that remains useful when compared directly with the power spectrum.