Seasonal degradation of aquatic habitats from hypoxia occurs in numerous freshwater and coastal marine systems and can result in direct mortality or displacement of fish. Yet, fishery landings from these systems are frequently unresponsive to changes in the severity and extent of hypoxia, and population-scale effects have been difficult to measure except in extreme hypoxic conditions with hypoxia-sensitive species. We investigated fine-scale temporal and spatial variability in dissolved oxygen in Lake Erie as it related to fish distribution and catch efficiencies of both active (bottom trawls) and passive (trap nets) fishing gears. Temperature and dissolved oxygen loggers placed near the edge of the hypolimnion exhibited much higher than expected variability. Hypoxic episodes of variable durations were frequently punctuated by periods of normoxia, consistent with high-frequency internal waves. High-resolution interpolations of water quality and hydroacoustic surveys suggest that fish habitat is compressed during hypoxic episodes, resulting in higher fish densities near the edges of hypoxia. At fixed locations with passive commercial fishing gear, catches with the highest values occurred when bottom waters were hypoxic for intermediate proportions of time. Proximity to hypoxia explained significant variation in bottom trawl catches, with higher catch rates near the edge of hypoxia. These results emphasize how hypoxia may elevate catch rates in various types of fishing gears, leading to a lack of association between indices of hypoxia and fishery landings. Increased catch rates of fish at the edges of hypoxia have important implications for stock assessment models that assume catchability is spatially homogeneous.