The adjusting of input parameters until model preditions match the observed response. This is performed by setting the majority of the parameters to measured values and adjusting only the few parameters that lack measurement or have the greates uncertainty.