Report by Andreas Prinzing, University of Rennes. Photos by Sven Finnberg
Spring phenology of temperate forests is widely understood as a temperature-driven process. As a result of global warming, buds are expected to be sprouting around 3 days earlier per decade. However, empirical trends consistently show shifts to be weaker than that. This discrepancy suggests that additional processes counteract the climatic control of phenology. In a study recently published in Nature Ecology & Evolution (https://doi.org/10.1038/s41559-026-03071-9), we were able to give evidence for the first time and at the landscape level that trees with heavy defoliation in one year delay the emergence of buds in the following spring by around three days, which in turn reduces subsequent defoliation intensity by more than half.
Detecting the variability of bud break and leaf feeding on a landscape scale requires observations over longer periods of time with very high temporal resolution. This is logistically not feasible for thousands of individual trees across dozens of forest locations. Here we used newly developed remote sensing methods based on cloud-independent radar data (Sentinel 1) to track leaf emergence and feeding in 27,500 satellite pixels (10×10m²) across 60 oak-dominated forest locations aver 2,400 km² and 5 years in South-Eastern Germany.
In each of 60 forests we found that pixels with higher leaf loss in one year delayed bud break by an average of 3 days in the following year. This delay in bud break is known to impact early feeding caterpillars as they synchronize hatching with budburst to access young, nutritious and little defended oak leaves. A delay of several days means the freshly hatched caterpillars face closed buds and starvation. Indeed, we found 55 percent less defoliation on trees having delayed budburst. The delay was efficient even during a pest outbreak – a mass gypsy-moth reproduction in 2019. This could explain why some oak trees ultimately experienced little leaf damage despite many gypsy moth clutches. As delaying budburst depends on prior herbivory it represents a form of previously overlooked delayed induced-defence over annual growth cycles.
Importantly, the delay in budburst is likely to be more than just exhaustion of trees after herbivory, but a selected adaptive response to spring herbivory. We found trees to delay budburst more in forests where such budburst delay most strongly statistically reduces herbivory.
These results have direct implications for the interpretation of long-term phenological trends. Most predictions assume that bud emergence is primarily controlled abiotically, by temperature and photoperiod. Our results show that biotic interactions can also have a large effect. Due to its complexity, this could not previously be observed using conventional methods. The mechanism provides an important additional explanation for why many trees suffer little herbivory despite being highly apparent, and why observed advance shifts in spring phenology often fall short of expectations. When it comes to climate change, we should therefore take better account not only of the direct effects of temperatures, but also of biotic interactions.