Advancements in autonomous sensors have enhanced our understanding of carbon cycling in freshwater systems by enabling more numerous and precise measurements of gross primary production (GPP), respiration (R), and net ecosystem production (NEP). Researchers can now easily deploy multiple dissolved oxygen sensors within a single waterbody, improving the spatial representativeness of collected data. We examined high-frequency time series logged by sensors at multiple depths to improve to account for vertical variations in the water column in two shallow Texas ponds (<2.5 m max depth). We hypothesized that photosynthesis would primarily occur in the surface mixed layer due to limited light penetration to the benthos, and that benthic respiration would exceed water column respiration. To test this, we deployed dissolved oxygen miniDOT loggers in two ponds every other month over one year for five-day deployments. We placed sensors at three depths: just below the surface, 0.5 meters below the surface, and just above the bottom sediments. Concurrently, we collected Secchi depth and water samples for analyses of inorganic and organic constituents. Incorporating variation in temperature and water clarity across multiple dates and seasons, results revealed that GPP and R were coupled and had similar patterns with depth. Photosynthesis occurred throughout the water column under conditions of greater water clarity, as indicated by deeper Secchi depths, but was restricted to the top half meter under shallow Secchi depths. Vertical variations in metabolism were apparent between top and middle depths. A surface-only approach to estimating metabolism would have missed these variations. Our findings underscore the importance of accounting for depth-specific processes when estimating whole-system metabolism, especially in shallow systems with variable water clarity.