Recent epidemiology studies have reported increase in lumbar spine injuries in frontal crashes. Whole human body finite element models (FEHBM) are frequently used to delineate mechanisms of such injuries. However, the accuracy of these models in mimicking the response of human spine relies on the characterization data of the spine model. The current study set out to generate characterization data that can be input to FEHBM lumbar spine, to obtain biofidelic responses from the models. Twenty-five lumbar functional spinal units were tested under compressive loading. A hydraulic testing machine was used to load the superior ends of the specimens. A 75N load was placed on the superior PMMA to remove the laxity in the joint and mimic the physiological load. There were three loading sequences, namely, preconditioning, 0.5 m/s (non-injurious) and 1.0 m/s (failure). Forces and displacements were collected using six-axis load cell and VICON targets. In addition, acoustic signals were collected to identify the times of failures. Finally, response corridors were generated for the two speeds. To demonstrate the corridors, GHBMC FE model was simulated in frontal impact condition with the default and updated lumbar stiffness. Bi-linear trend was observed in the force versus displacement plots. In the 0.5 m/s tests, mean toe- and linear-region stiffnesses were 0.96±0.37 and 2.44±0.92 kN/mm. In 1.0 m/s tests, the toe and linear-region stiffnesses were 1.13±0.56 and 4.6±2.5 kN/mm. Lumbar joints demonstrated 2.5 times higher stiffness in the linear-region when the loading rate was increased by 0.5 m/s.