Deseasonalized geophysical time series are often used in time series models (Hipel and McLeod 1994). In this article an optimal method for selecting the deseasonalization transformation is suggested and an R package implementation (McLeod and Gweon 2012) is discussed. Our deseasonalization method may be used with the recently developed periodic autoregression model for daily river flow suggested by Tesfaye, Anderson, and Meerschaert (2011) and for the hierarchical Bayes modeling for multi-site daily tempera- ture series discussed by Craigmile and Guttorp (2011).