Response of almond flowering and dormancy to Mediterranean temperature conditions in the context of adaptation to climate variations

International audience; Determining the flowering and the dormancy phases in fruit trees is a crucial process because of their substantial role in some agricultural practices and crop managements. However, few of these studies were conducted on almond flowering and dormancy, particularly in mild climate areas. This study aimed to simulate the dormancy phases, closely involved in the determination of flowering time of almond species in response to temperature variations. To reach this objective, Partial Least Squares analysis was used. In this regard, climatic and flowering data were collected from Ain Taoudjate in Morocco during the period from 1974 to 2014. In fact, a significant temperature increase was highlighted, inducing a decrease in amounts of chill during the studied period. Using Partial Least Squares analysis, a good fit of data was obtained, particularly using mean and maximal temperatures. Two relevant periods controlling the flowering process were highlighted in Tuono cultivar, in concordance with the sequential model in prediction of flowering times. The first long chilling period started from October 01st to January 11th. During this period, temperatures correlated positively with flowering dates, inducing consequently a delayed flowering dates because of low and slow accumulation of chilling requirements. However, the second short forcing period extended from January 18th to March 09th, during which flowering dates were negatively correlated with temperatures. In fact, the accumulation of certain threshold of Growing Degree Hours and Chill Portions during the two periods respectively could delay flowering, particularly in the context of chilling privation in the coming future, causing a serious problem for fruit trees. Significantly, temperatures during the chilling period seemed to affect effectively the flowering dates of almond than temperatures during the forcing period. In this investigation, we demonstrated that Partial Least Squares showed a good fit in explanation of the flowering process and can be used for prediction of dormancy phases and flowering process.