The rule for whether or not you need to use a Source element in your model is relatively straightforward: If you cannot represent the introduction of mass into your model (i.e., the source term) in an accurate manner using a simple equation, you should probably use a Source element.
That is, if you can't describe the introduction of mass into your system by using an initial or boundary condition (or a discrete change) in a pathway, you should use a Source element. This will often be the case if 1) your source involves barriers which fail over time; 2) your source includes one or more matrix materials which release species mass as they degrade; and/or 3) the source term is modified by reactions (e.g., decay and ingrowth). If any of these processes are active, describing the source term using simple algebraic equations becomes difficult, if not impossible (since in these situations, representing the source term accurately involves the solution of integrals and/or coupled systems of equations). These processes, however, can be explicitly represented using a Source element.
Note: Source elements cannot be activated or deactivated during a simulation (by a conditional Container). If a Source is located within a conditional Container, the Container must be set to inactive or active at the beginning of the simulation and its status cannot change during the simulation.