Photoelectrochemical cells (PECs) are integrated devices that offer an attractive approach to converting solar energy into storable chemical energy, such as hydrogen. In PEC-driven water electrolysis, the product species, hydrogen and oxygen, are first evolved as gas bubbles at the surface of the electrode. However, these gas bubbles can lead to significant efficiency losses, due to increasing electrode surface coverage by gas bubbles that block active material and reactants from reaching the photoelectrode surface. It is known that when gas bubbles are attached to the surface of a photoelectrode, they can induce a variety of loss mechanisms, however, these bubble losses are difficult to quantifiably measure and still not completely understood. Therefore, it is desirable to study the gas evolution dynamics on these photoelectrodes so that systems can be designed to minimize efficiency losses. In this study, in situ scanning photocurrent microscopy (SPCM) is used, for the first time, to investigate the local photocurrent losses associated with isolated hydrogen bubbles attached to the surface of a photoelectrode. By measuring the change in photocurrent response of an area of interest, we can quantify and determine local optical bubble-induced photocurrent losses on a photoelectrode at the sub-bubble level. Silicon-based photocathodes based on a metal-insulator-semiconductor (MIS) architecture containing a continuous 3 nm metal layer, are used for this work. SPCM is used to systematically investigate the influence of varying bubble size on the performance of these photoelectrodes under varied experimental conditions. The quantitative SPCM measurements are combined with optical modeling based on Snell’s Law. Comparisons between the experimentally determined losses and those predicted with the optical model are used to elucidate relationships between local photocurrent losses and bubble size. Significant increases in optical losses for larger bubbles (diameter > 150 μm) are seen, due to regions of total internal reflection that are not present for the small bubbles studied. Finally, the knowledge gained from the SPCM single-bubble measurements are applied to a large photoelectrode surface under uniform AM1.5 illumination. Using these measurements, we can model the current-time profile associated with multiple bubble evolution off of a fully-illuminated photoelectrode surface. Overall, this study sets the stage for quantitatively modeling bubble-related losses for various photoelectrode geometries and operating conditions. We have also provided valuable insight into the optical losses associated with having a single surface bound gas bubbles on a photoelectrode surface.