Abstract This work provides single cell durability tests of membrane electrode assemblies in dynamic operation regularly interrupted by recovery procedures for the removal of reversible voltage losses. Degradation rates at different loads in one single test can be determined from these tests. Hence, it is possible to report degradation rates versus current density instead of a single degradation rate value. A clear discrimination between reversible and irreversible voltage loss rates is provided. The irreversible degradation rate can be described by a linear regression of voltage values after the recovery steps. Using voltage values before refresh is less adequate due to possible impacts of reversible effects. The reversible contribution to the voltage decay is dominated by an exponential decay after restart, eventually turning into a linear one. A linear-exponential function is proposed to fit the reversible voltage degradation. Due to this function, the degradation behavior of an automotive fuel cell can be described correctly during the first hours after restart. The fit parameters decay constant, exponential amplitude and linear slope are evaluated. Eventually, the reasons for the voltage recovery during shutdown are analyzed showing that ionomer effects in the catalyst layer and/or membrane seem to be the key factor in this process.