When the selected parametric model for the covariance structure is far from the true one, the corresponding covariance estimator could have considerable bias. To balance the variability and bias of the covariance estimator, we employ a nonparametric method. In addition, as different mean...