CARDAMOM-JPL is a Bayesian model-data fusion framework developed to support NASA Earth science via multi-sensor integration within a mechanistic and physics-based land model framework. Though CARDAMOM-JPL has demonstrated applications in quantifying warming-induced emissions and feedbacks in the contemporary era-- including wetland CO2/CH4 feedbacks, fire and biomass forecasts, and evaluation of global drivers of land-sink variability--  its approach to estimating initial conditions from contemporary observations has placed it in different research space than bottom-up TBMs, which typically leverage a spin-up protocol. As a result, CARDAMOM-JPL and bottom-up TBMs both explore sensitivities of complex environmental process routines, but embark on fundamentally different modeling experiments to do so (different timescales, initial condition assumptions, baseline conditions, etc), obscuring clear interpretation of how the observational record embedded in CARDAMOM impacts estimation of future trajectories. Funded participation in WIE-MIP enables the development of novel spin up protocols and other compatibility adaptations to partially bridge the gap and support closer comparisons between data-aware frameworks like CARDAMOM and traditional bottom-up land surface models. We emphasize that CARDAMOM remains a complementary approach to bottom-up modeling due to its sparse mechanistic structure and coarse time step—both required due to the high computational cost of parameter inversion-- and cannot capture all relevant feedbacks. Nevertheless, we anticipate unprecedented insights on how remote-sensing-based parameter calibration, anchored to contemporary land surface dynamics, influences inferences of ecosystem sensitivity and centennial-scale evolution across represented routines.