The Southern Ocean plays a critical role in the global carbon cycle and climate system, surfacing old, carbon-rich water via upwelling, and absorbing anthropogenic carbon from the atmosphere.  However, Southern Ocean carbon cycling is understudied in comparison to its importance, due to the challenges of collecting carbon-related measurements from a remote location and under rough environmental conditions.  Autonomous Biogeochemical-Argo floats are helping to address this gap in knowledge by collecting measurements of ocean acidity and other physical and biogeochemical parameters as they drift around the Southern Ocean with the currents.  However, this new Lagrangian approach to ocean biogeochemical observation is not represented in the current suite of Earth system models used to project future carbon-climate feedbacks in this critical region.  In this presentation, I will describe our DOE RGMA-funded efforts to simulate online biogeochemical floats in E3SMv2, to assess model bias using Lagrangian float observations, and to use our model testbed to assess observational sampling and interpolation techniques for improved projections of the Southern Ocean carbon cycle.