For a 4-year conservation agriculture experiment:

• If you have field data on biomass inputs, weather, and clay %, use: RothC.
• If you also want crop-soil feedback and have full datasets, consider DayCent or DNDC.
• For quick SOC estimates with minimal data, ICBM is the simplest.

Use RothC (Rothamsted Carbon Model), as it effectively incorporates clay percentage in SOC turnover and is relatively simple to run with limited data.

Can you find some site where the practices have been in place for a (much) longer time for insurance ??