Natural cyclical stress perturbations, including wet and solid earth tides, coupled to repeatable seismic monitoring approaches, provide a tool for understanding the elastic properties of rocks at depth. Ocean tides in particular perturb the overburden and pore pressure, affecting hydraulic and elastic rock properties. The recent emergence of distributed acoustic sensing (DAS) and seismology utilizing existing dark telecom fiber offers an unprecedented opportunity to instrument the seas and oceans with seismic monitoring networks of large aperture, fine spatial resolution, and broadband sensitivity. We use a submarine dark fiber array to methodologically assess whether changes in seafloor seismic velocities are modulated by ocean tidal loading. Considering oceanic ambient noise in the frequency range between 1 and 5 Hz, we find it possible to retrieve reliable surface-wave estimates using relatively short time windows (∼60 min). Next, utilizing the power of linear arrays to capture ambient noise energy from stationary-phase zones, and ambient noise processing techniques commonly used in DAS applications for subsurface monitoring in urban settings, we introduce a methodological approach to turn each segment of dark fiber into a short-offset monitoring array (2 km), providing redundant, hourly, surface-wave estimates. Using four days of ambient noise data recorded by the Monterey Accelerated Research System (MARS) 20 km section offshore cable, we generate virtual-source gathers for 10 segments spanning the whole MARS cable. We then use virtual-source data from one segment to estimate hourly velocity changes that are then quantitatively compared to tidal data. The results suggest that changes in seafloor seismic velocities are likely modulated by tidal height, even in regions without large tidal excursions. This advance demonstrates a new application of submarine telecom fibers for monitoring the seismic stress response of near-seafloor sediments.