Sea ice is a complex, multiphasic porous medium where the evolution of microstructure comprising an ice matrix, liquid brine, and dissolved salts governs heat and solute transport, ultimately regulating polar biogeochemical activity. This work presents a comprehensive numerical modeling framework that integrates multi-phase, multi-component, multi-scale, and multi-physics phenomena, a "multi-x" approach to simulate sea-ice freezing and associated brine-network dynamics. At the macroscale, sea ice is modeled as a biphasic porous medium using the Extended Theory of Porous Media (eTPM), ensuring thermodynamically consistent mass, momentum, and energy balances. To resolve microscale solidification, we employ a phase-field model within a Landau–Ginzburg setting. This resolves dendritic ice growth and brine-channel formation at the micrometer scale, capturing the coupled evolution of salinity and order parameters. Equilibrium pore geometries are upscaled via an offline coupling strategy to inform macroscale eTPM closure relations, significantly improving the prediction of freezing-driven transport pathways. The physical foundation is integrated with a biogeochemical (BGC) module modeling photosynthesis, nutrient uptake, and carbon assimilation. This coupling illustrates how seasonal microstructure evolution regulates habitat connectivity and algal photoadaptation. To overcome the computational demands of high-fidelity simulations, we introduce a reduced-order surrogate model based on Principal Component Analysis (PCA) and Singular Value Decomposition (SVD). Trained on parametric ensembles, this surrogate enables rapid scenario screening and sensitivity analysis across varying environmental conditions. This integrated pipeline provides a robust tool for investigating the critical interplay between sea ice physics and polar ecosystems.