Water and thermal management have been identified as technical hurdles to the successful implementation of low temperature, polymer electrolyte membrane (PEM), fuel cell power systems. In low power applications, miniature PEM fuel cells show significant promise as a competitor to lithium-ion batteries. Significant design work is underway to improve the specific power and energy densities of these fuel cells. However, little attention has been given to characterizing transient response in these miniature applications to enable gains in system design, optimization and control. This work develops, calibrates and experimentally validates two different dynamic control-oriented models for open-loop temperature state observation in miniature PEM fuel cells. Of critical importance, these estimators target operation under dry conditions with no reactant pre-treatment. Operational conditions are then identified for which each model architecture is more suitable, specifically targeting minimal model complexity. A sensitivity analysis was completed that indicates necessary sensor measurements with sensor frugality in mind. The dynamic response under changes in load and fuel stoichiometry are well captured over a range of operating conditions.