In this work, some potential identification/observation problems that arise in the anode line of a typical Polymer Electrolyte Membrane (PEM) fuel cell based system are highlighted and then solved. They involve estimating the hydrogen input flow at the stack anode and the water transport across the membrane. It is argued that estimating the membrane water transport is of special interest since it gives relevant information on the actual performance and state of the fuel cell. Both problems are solved by constructing robust observers and parameter identification algorithms based on the Generalized Super-Twisting Algorithm.