Modelling daily precipitation is challenging due to the nature of zero occurrences repeatedly. This paper illustrates a truncated Bayesian spatio-temporal modelling to address this issue. It extends the R package spTimer to accommodate the truncated Gaussian spatio-temporal model through user friendly R front-end input and C back-end to facilitate faster computation and data storage. The paper also illustrates truncated models to address the big-n problem (or large data problem) for spatio-temporal data. The proposed method is applicable only, when the data truncation is occurring at the left side of a distribution. The method is illustrated using daily rainfall and ground level ozone concentration data from the United States, and can be applied to any data with a similar truncated distributional pattern.