To increase the stability and current density of molecular-catalyst-based electroanodes for water oxidation, immobilization of the catalysts at the electrode surface is a common strategy. A prominent example is the oligomerized Ru(tda) molecular catalyst, which showed outstanding current densities even at neutral pH values. One of the most challenging aspects of immobilized catalysts is to understand the interaction between the catalyst and the surface under operando conditions. Experiments are often performed under model conditions, and computational methods to study reaction steps are typically limited to a few hundred atoms. In this study, we combined three computational methods, density functional theory electronic structure computations, molecular dynamics for large-scale simulations of the catalyst-solid interaction, and empirical valence bond for reaction modeling the catalyst at the interface of a large carbon support and a phosphate water buffer. These techniques allowed us to explore the combined effects of solvent, hydrophobic directionality, and electric field on the attachment and reactivity of a Ru(tda) pentamer at a graphene surface. Our simulations have a perfect agreement with the experimental characterization under model conditions. However, we find that under operando conditions, where the catalyst is oxidized to the active RuV state, with a phosphate-containing electrolyte and an applied electric field, the attachment is completely reversed compared to the model conditions with RuII and organic solvents. This reversed attachment leads to a water-excluded region close to the active RuV═O center. The EVB reaction modeling showed that the reaction could still proceed to form an O-O bond via an oxide relay mechanism, where a dangling carboxylate reacts with the oxo via nucleophilic attack. We find that the activation energies are identical in water solution and at the electrode surface, showing how this mechanism is key to highly active molecular water oxidation catalysts immobilized on surfaces. Since attachment to surfaces could have a strong, and often negative, influence on the reactions, this study provides a guideline on how to model reactions without compromising the complexity of the electrode environment.