Soils contain the largest pool of carbon in terrestrial ecosystems. Soil carbon dynamics and associated nutrient dynamics play significant roles in regulating global carbon cycle and atmospheric CO2 concentrations. Our capability to predict future climate change depends to a large extent on a well-constrained representation of soil carbon dynamics in ESMs. Here we evaluate two decomposition schemes - converging trophic cascade (CTC) and Century - in CLM4.5/ACME V0 using data from the long-term intersite decomposition experiment team (LIDET), radiocarbon (14C) observations, and Harmonized World Soil Database (HWSD). For the evaluation against LIDET, We exercise the full CLM4.5/ ACME V0 land model, including seasonal variability in nitrogen limitation and environmental scalars (temperature, moisture, O2), in order to represent LIDET experiment in a realistic way. We show that the proper design of model experiments is crucial to model evaluation using data from field experiments such as LIDET. We also use ¹⁴C profile data at 10 sites to evaluate the performance of CTC and CENTURY decomposition scheme. We find that the ¹⁴C profiles at these sites are most sensitive to the depth dependent decomposition parameters, consistent with previous studies.