Robust Multiblock STATICO for the Modeling of Environmental Indicator Structures

Multivariate environmental studies frequently involve the integration of paired datasets exhibiting high heterogeneity, collinearity, and influential observations, which can undermine the stability and interpretability of classical multiblock analytical methods. In this context, this study aims to enhance the robustness of the STATICO framework for modeling common structures among environmental indicator blocks. To this end, a robust extension of the STATICO method is proposed by incorporating robust covariance estimators and adaptive weighting schemes within the original triadic algebraic structure, preserving its fundamental mathematical formulation. Robustification is introduced both at the interstructure stage, through a reformulated RV coefficient, and in the construction of the compromise space. The performance of classical and robust STATICO approaches is evaluated using simulated environmental datasets calibrated to represent Ecuadorian coastal systems. The results indicate that the robust STATICO approach increases the dominance and stability of the global compromise, produces a more balanced redistribution of inter-block similarities, and yields more discriminative representation values in the factorial space. Graphical analyses further reveal improved parsimony and stability of the latent structure compared with the classical formulation. Overall, the proposed robust STATICO method provides a methodologically sound and reliable tool for multiblock analysis of complex environmental data affected by contamination-related heterogeneity.