This paper addresses a procedure to extract error estimates for the physical and dynamical components of a forecast model. This is a two-step process where contributions to the forecast tendencies from individual terms of the model equations are first determined using an elaborate bookkeeping of the forecast. The second step regresses these estimates of tendencies from individual terms of the model equations against the observed total tendencies. This process is executed separately for the entire horizontal and vertical transform grid points of a global model. The summary of results based on the corrections to the physics and dynamics provided by the regression coefficients highlights the component errors of the model arising from its formulation. This study provides information on geographical and vertical distribution of forecast errors contributed by features such as nonlinear advective dynamics, rest of dynamics, deep cumulus convection, large-scale condensation physics, radiative processes and the rest of physics. Several future possibilities from this work are also discussed in this paper.