Solar power plays a significant and sometimes primary role in the energy budgets of many terrestrial sensor systems due to its reliability, power density, and simplicity. More importantly, photoelectric generation can be accurately predicted for a given terrestrial location. This knowledge of expected harvested power is critical for both provisioning in design and scheduling of activities in operations. The importance of reliable prediction algorithms is highlighted by the dearth of solar energy harvesting designs for submarine sensor systems. Platforms such as autonomous underwater vehicles and marine wildlife telemetry tags almost exclusively rely on battery power for their entire deployments. Reliance on fixed charge batteries significantly impacts the frequency of measurements, types of measurements, and the transmission of those data to operators. For these reasons, our group has worked to develop a method of predicting photovoltaic energy production for solar cells in marine environments. To validate and improve the model, we have developed a device, referred to as a datalogger, for the characterization of solar energy harvesting, with an initial focus on harvesting when attached to marine wildlife. This device measures and logs for analysis the current-voltage relationship of a silicon solar cell as its host animal moves through the water column. This paper discusses the development of this datalogger, presenting design requirements, design decisions, test results, and preliminary data resulting from a deployment on a Northern elephant seal (Mirounga angustirostris) in the spring of 2017. Results are discussed and future design changes for improving the system are presented.