How simple can we make models for CO2 injection, migration, and leakage?