Clockwork Forest: Day Length and Temperature Steer Subtropical Trees’ Hidden Rhythms

In a small cluster of urban forest fragments in southern Brazil, an international-style team of ecologists has spent two years watching trees breathe, sprout, flower, and fruit with almost forensic patience. The result, published in Neotropical Biology and Conservation, is one of the most detailed community-level portraits of tree phenology ever assembled for a subtropical city, and it delivers a finding that challenges the assumptions long imported from temperate-zone science: the eight dominant tree species of this forest act in synchrony when they build and drop their leaves, but go their separate, idiosyncratic ways when it comes time to reproduce. Even more striking is what drives the synchrony. Not rainfall, the variable most people would bet on in the tropics, but the length of the day and the temperature of the air.

Phenology, the study of recurring life-cycle events such as leaf flushing, leaf shedding, flowering, and fruiting, has historically been built on a temperate foundation. In Europe and North America, winter dormancy imposed by low light and freezing temperatures gives plants an unmistakable annual rhythm, making it relatively easy to read climatic signals in their behavior. Tropical and subtropical forests, by contrast, have long been assumed to lack strong seasonal rhythms, a paradigm that recent advances in tropical dendrochronology have begun to dismantle. Yet the theoretical framework for interpreting plant responses to climate change remains heavily biased toward temperate systems, leaving scientists with a genuine blind spot about how tropical and subtropical trees will fare as the planet warms.

The new study set out to chip away at that blind spot from an unusual vantage point: an urban forest. The research took place in five hectares of Semideciduous Seasonal Forest fragments at the Universidade do Vale do Rio dos Sinos in São Leopoldo, Rio Grande do Sul, divided into twenty randomly distributed plots of one hundred square meters each. The region’s humid subtropical climate brings average January temperatures of 24.5 degrees Celsius, cool Julys averaging 14.3 degrees, and abundant rainfall of roughly 1,324 millimeters per year. Day length swings from just over fourteen hours in December to barely ten hours in June, providing exactly the kind of annual light gradient that can serve as a reliable calendar signal for plants.

From a prior phytosociological survey that catalogued 469 tree individuals across 60 species, the researchers selected the eight most frequent and abundant natives: Allophylus edulis, Casearia sylvestris, Guarea macrophylla, Mimosa bimucronata, Myrsine coriacea, Myrsine umbellata, Schinus glandulosum, and Schinus terebinthifolia. For each species they chose eight individuals, sixty-four trees in total, and monitored them every two weeks from January 2016 through December 2017, forty-eight visits recording the presence or absence of each of the four phenophases. Several of these species, including Mimosa bimucronata and both Schinus species, had never before been studied at this resolution in southern Brazil, making the dataset a genuine first for the region.

To detect seasonality in the resulting binary records, the team turned to circular statistics, converting calendar dates into angles around a 360-degree circle and applying the Rayleigh test to determine whether activity was uniformly distributed or concentrated in particular periods. All species showed statistically non-uniform patterns for every phenophase, with p-values below 0.001, but concentration varied dramatically. Leaf flushing was strongly seasonal in five of the eight species, peaking in late winter and spring, while leaf shedding was remarkably diffuse for most, with only Schinus glandulosum concentrating its leaf drop in early June at the start of winter. Flowering was the sharpest signal of all: every species reached a Rayleigh value above 0.6, with peaks scattered across the calendar from September through December and into the following autumn.

The real innovation came at the community level. Rather than simply describing each species in isolation, the researchers ran a Principal Component Analysis on the activity-index series of all eight species for each phenophase, using Spearman’s correlation to handle the peaked, non-normal data. If the analysis yielded significant shared axes, the community could be said to move together; if not, each species was marching to its own drummer. The verdict was unambiguous. Leaf flushing produced one significant axis explaining 43 percent of the variation, with all species loading positively, meaning the entire community produces new leaves from roughly May or June through November or December. Leaf shedding yielded two significant axes, splitting the community into two groups, one shedding leaves from July to October and the other from March to July, while a second common pattern showed all species avoiding leaf loss during the hottest, brightest months.

Reproduction told a completely different story. For both flowering and fruiting, no PCA axis reached significance, meaning the reproductive schedules of the eight species are statistically independent of one another. Flowering peaks ran the gamut: A. edulis and C. sylvestris flowered in early September, G. macrophylla in December, M. bimucronata in early March, M. coriacea in June, M. umbellata in May, S. glandulosum in early October, and S. terebinthifolia in late March. Notably, no species flowered in January or July, the months of climatic extremes, a pattern the authors link to unfavorable conditions for pollinators and fertilization physiology, effects that urban heat and urban climate extremes may amplify further. Fruiting peaks clustered in early winter and spring for the species with concentrated fruiting, consistent with the zoochorous, animal-dispersed seeds that dominate pioneer Atlantic Forest communities and tie reproductive timing tightly to the behavior of birds and other frugivores.

When the significant community axes were correlated with local climate records from a meteorological station roughly forty kilometers away, precipitation dropped out of the picture entirely. Instead, temperature and day length emerged as the governing signals. Leaf flushing was negatively correlated with both temperature and day length, meaning the community builds new leaves during cooler months and shorter days, preemptively staging vegetative growth before conditions peak rather than gambling on midsummer. Leaf shedding showed an inverse relationship: the shared second axis correlated positively with temperature and day length, confirming that no species drops leaves during the hottest, brightest periods. The authors attribute the diffuseness of leaf fall in most species partly to wind and storm damage common in the region, and suggest that litter-driven changes in soil nutrition at the start of spring may explain why one group of species sheds leaves precisely when days grow long.

The ecological and practical implications ripple outward. All eight species are early-successional natives widely used in agroforestry systems and Atlantic Forest restoration programs, so knowing when they leaf out, flower, and fruit directly informs seed collection schedules, plantation planning, and the harvest of non-timber forest products that support small family economies in Rio Grande do Sul. The community-level approach itself is a methodological contribution, since most phenological studies analyze species in isolation and rarely test whether a shared pattern exists at all. The finding that vegetative activity synchronizes around the light calendar while reproduction disperses across the year offers a template for how subtropical communities may be organized, and a caution against assuming reproductive phenology can be predicted from climatic variables alone when factors like tree age, pollination, competition, and disperser behavior intervene.

Perhaps the most provocative thread is the urban setting itself. Previous work has shown that city trees often grow faster but live shorter lives, and that urban conditions can delay or advance phenological events, yet almost no data exist for subtropical urban forests. Here, the community’s overall climatic responses matched those documented in natural subtropical forests, suggesting day length remains the master signal even amid pruning and urban edge effects. But subtle shifts in timing were observed, and the authors argue that disentangling urbanization from climate will require longer time series, larger spatial coverage, and ultimately genetic and physiological studies of how these species perceive and respond to environmental cues. As climate change reshapes seasonal calendars worldwide, forests like this five-hectare fragment, wedged inside a Brazilian city, may become some of the most informative natural laboratories we have for watching trees renegotiate their relationship with time.

Subject of Research: Climatic drivers of vegetative and reproductive phenology in a subtropical urban forest tree community

Article Title: Climatic signals on phenological patterns among tree species in a subtropical forest community

Article References: Olmedo, G. M., Raguse-Quadros, M., Conrado, G. T., & Oliveira, J. M. (2023). Climatic signals on phenological patterns among tree species in a subtropical forest community. Neotropical Biology and Conservation, 18(3), 191-208. https://doi.org/10.3897/neotropical.18.e103506

Image Credits: AI Generated

DOI: 10.3897/neotropical.18.e103506

Keywords: phenology, subtropical forest, urban forest, leaf flushing, leaf shedding, flowering, fruiting, day length, temperature, Atlantic Forest, circular statistics, climate change