Abstract:

Aims Our objective was to determine the impacts of climate change (mainly temperature and precipitation) on the phenology of tropical evergreen forests at the community and species level.
Methods Two phenophases—the onset of leaf unfolding and the onset of first flowering—had been monitored in 12 evergreen broadleaf tree species in the tropical arboretum in Jianfengling, Hainan Island, since 2003. Observations were made on three mature and healthy individuals for each species once every two days in the same sunny direction in the afternoon by the same person. After the phenological data were collected, they transformed into the number of days from the first day of a calendar year (DOY, i.e. from 1th January). The outliers were excluded used the Grubbs principle. Meteorological data for the corresponding time period were derived from the National Climate Centre. Ultimately, the phenological and meteorological data were combined with the integrate regression method and used to identify the impacts of changing climate on the onset of leaf unfolding and the onset of first flowering in the 12 tree species and to determine their responses in the two phenophases to climate change.
Important findings Both the onset of leaf unfolding and the onset of first flowering in the tree species were affected by climate change; the starting dates of the two phenophases in nearly all the tree species were closely related to changes in monthly temperature and precipitation. The onset of leaf unfolding was affected by the monthly mean temperature of the preceding winter and spring, and the impacts became more apparent with the time approaching the occurrence of specific phenophases. Monthly precipitation in the preceding autumn had more effects on the onset of the first flowering than any other time period, in support of the hypothesis that the effect of precipitation on phenology had a hysteretic nature. Furthermore, the onset of leaf unfolding was more susceptible to climate change than the onset of first flowering in the native tree species, whereas the introduced tree species showed reversed responses. Generally, if the monthly mean temperature increases by 0.1 ℃ and the monthly precipitation increases by 10 mm in the month with the strongest effects, and given that the conditions of other 11 months preceding the occurrence of the specific phenophases remains unchanged, then most of the species would advance or delay the phenophases by about one to three days. Ultimately, the Integrate Regression method provides an effective and operational approach to understanding the dynamic and complex relationships between climate change and the phenology in tropical plants, as demonstrated by the high values of the coefficient of determination (R²≥ 0.943) for the integrate temperature and precipitation regression models for phenological predictions based on climatic factors. These models can well be used for predicting the trend of phenological changes in tropical evergreen broadleaf tree species in response to changing climate.

Key words: integrate regression, plant phenology, the onset of first flowering, the onset of leaf unfolding, tropical evergreen broadleaf trees