Phenology influences many forest functions and can inform forest conservation and management, yet representative phenological data for most common tropical forest tree species remain sparse or absent. Between June 2011 and December 2013, we investigated flowering, fruiting and leafing patterns in the Bwindi Impenetrable National Park, a montane forest located near the equator in Uganda, drawing on 16,410 observations of 530 trees of 54 species located between 2,066 and 2,527 m in elevation. The park’s climate is equatorial with two wet and dry seasons each year. Flowering and fruiting were strongly seasonal while patterns in leafing were less pronounced. Flower occurrence peaked at the beginning of the short dry season followed by a pronounced trough during the beginning and the middle of the short wet season. Fruit occurrence had a pronounced peak during high rainfall months in March through April with most fruits ripening during drier months in May through July. Fruit scarcity was observed for a 4-month period spanning September to December and most flushing of leaves noted at the end of the wet season in November and December. Our binomial generalised linear mixed models (GLMM) indicated that flowering and fruiting were negatively associated with temperature and that leafing activity was positively associated with rainfall and temperature. These findings are consistent with the insolation- and water-limitation hypotheses suggesting that the seasonally varying availability of resources such as light, water and nutrients determines these phenological patterns. Ideally, prolonged, multi-year community-level studies would be supported so as to better characterise the influence of climate and of climate variability.

A total of 530 trees were monitored each month for two and a half years, starting in June 2011 and ending in December 2013. Our team observed each tree at monthly intervals using binoculars and recorded comparable estimated counts of flowers, fruits and leaves (see Table S4). The presence and quantity per tree of flower buds, open flowers, fallen flowers, unripe fruits, ripe fruits, rotten or dry fruits, fallen fruits, new leaves, old leaves, damaged leaves, senescent leaves and fallen leaves were recorded by assigning a score between zero and four. The score of 0 represents absence of a phenophase, 1 represents an abundance of 1–10, 2 represents an abundance of 10–100, 3 represents an abundance of 100–1000 and 4 represents an abundance greater than 1000 counts of a given phenophase. Categories of flowers, fruits and leaves of some species were deemed “difficult to distinguish” and subsequently dropped from the data collection protocol (see Table S2 for a more complete description). For instance, damaged leaves were not scored for Erica benguelensis. We replaced trees that died (n = 8) with nearby trees of the same species. Categorical abundance measures of each phenological variable for each individual tree were converted into activity indices, i.e. absence–presence data (absent if zero, present if greater than zero), and used to calculate the proportion of each study species observed bearing flowers, fruits and leaves in each calendar month. The absence–presence data were also used to test for seasonality and the role of rainfall and temperature on the occurrence of phenological events.

The data files can be accessed using Office Word reader and MS Office Excel.