Phenology & Autumn senescence

In temperate regions, when summer ends, the golden season of autumn begins. Before the cold winter sets in, temperate trees change into their most colourful version. It is the time to enjoy a nice walk in the forest, breath in the smell of earth and mushrooms, while enjoying the rustling sound of fallen leaves beneath the feet. This show, which nature provides every year, is part of the seasonal cycle, also called phenological cycle, of deciduous trees. Leaves change colour and drop, in a process called “leaf senescence”, necessary to protect temperate trees from the cold temperature of the coming winter. 

However, climate warming is affecting the timing of the seasonal cycle. Spring leaf-out tends to happen earlier. Researchers have also predicted that warming will cause autumn senescence to happen later, but is that really what happens? What are the factors regulating the timing of autumn? In order to understand how the timing of autumn leaf senescence may shift under future climate conditions, our lab had a closer look at the underlying environmental triggers.

In this study, we used not just one, but three different approaches to test and cross-validate our results: on-the-ground observations, laboratory experiments, and phenological modelling.

The citizen science program, Pan European Phenology Project, provided us with a database of phenological observations of spring leaf out and autumn leaf senescence. In total, we analysed over 434,000 phenological observations of six deciduous tree species across Central Europe, from 1948 to 2015. We assessed the effects of several environmental factors - including spring leaf out, spring and summer photosynthesis, temperature, and precipitation - to evaluate the mechanisms governing the timing of the autumn senescence.

To support the results obtained from the field observations, we performed several manipulative experiments in the lab. The controlled lab conditions allowed us to evaluate independently the role of sun-exposition, CO2 level, and temperature on the timing of autumn leaf senescence.

To improve confidence in future phenology and carbon cycling forecasts, we developed a new autumn phenology model, explicitly accounting for both seasonal carbon uptake and autumn environmental cues, by incorporating mechanistic representations of these drivers.

The tree observations in Europe over the last six decades show an unexpectedly large effect of spring and summer productivity on autumn senescence, counteracting the expected delays in senescence due to autumn warming. In fact, careful analysis of the data shows that, on average, an increase of 10% in photosynthetic activity during spring and summer is linked to an eight-day advance of autumn senescence. Our analyses thus suggest seasonal photosynthesis and autumn temperature as the main drivers of senescence, whereas CO2-levels, summer temperatures and precipitation affect photosynthesis and therefore, only have an indirect effect on autumn senescence.

In experiments, we independently assessed the roles of sun-exposure, CO2 concentrations, and temperature on leaf senescence. The results confirmed that higher photosynthetic activity, induced by elevated CO2, temperature, or light levels, advanced autumn senescence. Furthermore, this experiment allowed us to highlight that the correlation between photosynthesis and leaf senescence operates at the plant level rather than at leaf level. In other words, the higher the photosynthetic activity of leaves during spring and summer, the sooner the tree hits its productivitylimitation,and is more likely to undergo autumnal senescence.

Traditionally, dropping temperatures and shorter day-length in autumn were accepted as the main drivers of autumn senescence. Thus, current phenological models mainly account for these parameters to predict when autumn occurs. As we now have a better grasp of the interactive triggers of autumn senescence, it seemed crucial to develop a new model including the pivotal role of photosynthesis on leaf senescence. The newly developed model, called PIA model, turned out to perform well at predicting the dates of autumn onset. The high predictive power and accuracy of the PIA model thus confirm seasonal photosynthesis as a key player in the timing of autumn.

Using a business-as-usual climate-warming scenario (RCP 8.5), previous models suggested that, as a result of autumn warming, senescence would happen 2 to 3 weeks later by the end of the 21st century. In contrast, our new PIA model, accounting for the role of photosynthesis, predicts a slight advance in autumn leaf senescence of 3 to 6 days over the rest of the century. These updated predictions reverse our expectations of autumn timing under future climate conditions, underscoring the need to represent the detected mechanisms in global vegetation models.  

The present study, based on field observations, modelling and lab experiments, shows that seasonal carbon uptake exerts an important constraint on autumn senescence. Similar to when we have a good meal, it does not matter how nice the food being put in front of you is, there will always be a limit to how much you can eat. ​Therefore, if a tree experiences more photosynthesis during a season, it will shed its leaves earlier. This discovery can be explained by carbon sink limitation in plants, predicting that seasonal carbon uptake is internally limited by nutrient availability and developmental constraints. Thus, if spring and summer photosynthetic rates are higher, the carbon-sink of plants should be saturated earlier, triggering the onset of leaf senescence. While this phenomenon has already been observed at the molecular level in herbaceous plants, our study is the first evidence that sink limitation is affecting the phenological cycle of trees.

The development of a new powerful model, implementing the important role of photosynthesis on leaf senescence, led us to reassess the timing of autumn under future climate conditions. Previous models predicted that the temperature increase would allow trees to be active for longer and therefore increase the yearly carbon uptake of temperate forests. In contrast, our study predicts that warmer conditions will slightly advance autumn senescence, lowering our expectations of the extent to which longer growing seasons will increase seasonal carbon uptake in forests.

Trees play a critical role in mitigating climate change. Temperate and boreal forests alone absorb over 10 gigatons of carbon annually. Understanding their performance under future climate conditions is therefore crucial. To improve our confidence in future carbon cycle and climate projections, we need to be able to forecast the productivity cycles of trees, and the strong constraint of growing-season productivity on autumn senescence we show in this study represents an important step in this direction. Future research will need to generate a thorough spatial understanding of the extent of sink limitation. Only then will we be able to forecast plant phenology and forest productivity over space and time.