Several traditional methods for discretizing random fields in stochast
ic mechanics applications are considered; they are the midpoint method
, the interpolation method, and the local averaging method. A simple a
nd computationally convenient criterion for estimating the accuracy of
these discretization methods is developed. Also, the Volterra series
representation of nonlinear input/output relationships is utilized to
assess the effect of the random field discretization methods on the re
sponse variability of stochastic mechanics problems. The theoretical d
evelopments are elucidated by a numerical example involving a beam pro
blem.