Bivalves are prized for the ecosystem services they provide. The removal of particles from the water column through filter feeding and resulting water quality benefits, known as the biofiltration services, of bivalves have been studied for over a century. This has created a wealth of knowledge around the mechanistic drivers of bivalve feeding activity. Recently, Chesapeake Bay ecosystem-wide models have begun incorporating Eastern Oyster (Crassostrea virginica) biofiltration. Acute feeding variability is critically important when estimating oyster biofiltration services at ecosystem scale. Typically, natural seston clearance rate studies last a limited timeframe, with a focus on specific environmental events such as an increase in temperature, drop in salinity, or a tidal cycle.To capture the highly variable filter feeding rate of bivalves, such as the Eastern Oyster, studies have used highly controlled laboratory conditions, with single environmental variable modification. These studies often use indirect methods for estimating clearance rates that commonly lack high-resolution capability. Furthermore, these studies are labor intensive and time consuming, and as a result, few studies have monitored bivalve feeding activities over long periods to understand variation in activity or how these rates may change with seasonal shifts in conditions. These limitations have led to a shortage of knowledge around how clearance rates of oysters vary in response to ambient conditions over both short-term (hourly) and long-term (seasonal) time scales. This study leverages advances in semi-autonomous aquatic observing to track high- resolution, long-term feeding responses of bivalves to subtle variations in environmental conditions. Oyster ex situ clearance rates in the Choptank River (Maryland, USA) were estimated under flow-through conditions, and logged in real-time using fluorometers among replicate oysters over 5-day experiments for 9 months. The measured clearance rates from this system were compared to a mechanistic clearance rate model used by the Chesapeake Bay Program, which is used to estimate the role of oysters in controlling water quality in the Bay. Environmental data were evaluated to build a statistical and random forest model to predict how oyster clearance rates respond to prevailing environmental conditions. This monitoring system and resulting models enable a deeper understanding of feeding variability and how natural seston and environmental variability directly influence oyster physiology.