In precision agriculture, most studies focus on spatial crop variability whereas temporal variability and its role in decision-making is equally important. The classical methods for temporal analysis have limitations, potentially resulting in information loss. A novel method based on a Bayesian functional Linear regression with Sparse Steps functions (BLiSS method) is evaluated in this paper to investigate continuous influence analysis when working with time series data. The example of the influence of temperature on the number of clusters per vine during the year before harvest was considered as an example application. The evaluation of the BLiSS results was done by comparing identified critical time periods with traditional viticulture knowledge in the literature. It showed the relevance of the BLiSS method, highlighting already known results and identifying new critical time periods for yield elaboration.