we prove that partial sums of linear multiparameter stochastic processes can be represented as partial sums of independent innovations plus components that are uniformly of smaller order. This representation is exploited to establish functional central limit theorems and strong approximations for random fiels