Compliance at joint level in human locomotion has played a central role in many studies. In particular joints in the lower body were mainly considered in order to create mechanisms able to reproduce human-like gaits such as prostheses, exoskeletons or walking robots. In a previous study, we have used an 11 DOF 2D human model to carry out the analysis of compliance in the leg joints during level ground walking, by introducing torsional springs with variable stiffness in the leg joints and dampers in the ankle joints. In this paper we have significantly extended this study to different walking scenarios, such as slope and stair walking, in addition to level ground walking. In the dynamic model, two degrees of freedom have been added to the trunk in order to better reproduce the flexibility of the human trunk, resulting in a 13 DOF 2D human model. In addition, biarticular springs as coupling between hip and knee joints have been introduced. Optimal control is applied to identify the stiffness profiles, the rest positions of the springs and the value of the damper at the ankle that best reproduce measured human joint trajectories in these walking scenarios.