In many childrens’ picture books, we read that “Spring is coming, trees wake up, stretch themselves, and sprout pale green leaves.” But unlike people, trees cannot simply look at a calendar. So, how can they know the cycle of the seasons? Scientists have shown that plants are equipped with sensors that feel the environment. When the days are lengthening and the temperature warms up after winter chilling, trees sense that it is time to green up and reactivate growth.

Under accelerating global change, large changes are happening in the timing and duration of the growing season in extratropical ecosystems. Recent studies have demonstrated that trees tend to sprout and resume growth earlier as climate warms. But important questions remain of whether and how much an earlier spring results in higher carbon sequestration from the atmosphere, enhanced drought stress or more frequent late-frost damage to developing buds and leaves. Despite a growing body of literature on earlier leaf emergence in spring, inference of its consequences for the forest carbon sink and tree growth have remained rather speculative. With their short lifespan lasting from spring to autumn in deciduous species, leaves only represent a short-term carbon storage pool in trees. Besides, the beneficial effects of spring warmth on growing season productivity can be dramatically offset by increasing carbon losses due to summer droughts or autumn warming. We thought a missing puzzle here are extensive observations of the growth of tree stems, because a carbon surplus from increased uptake by the leaves may not mean that trees necessarily grow more. A wood-oriented view on phenological impacts is thus essential for predicting changes in productivity, because wood is the primary long-term carbon storage pool in forests (Figure 1). Therefore, we asked the fundamental but still unresolved questions of whether and how an advance in spring onset influences tree growth.

Figure 1 Illustration of the influences of phenology on growth.

To address this question, we would ideally assess wood formation at sub-seasonal time steps, but such data are sparse, and often cover short time frames of less than a decade. As such, we addressed a broad spatiotemporal scale using a massive collection of annual tree-ring width data from the International Tree-Ring Data Bank, the world's largest public archive of tree ring data containing the effort of numerous and generous dendrochronologists. Past research has shown that the onset of wood formation is closely related to the fulfilment of critical temperatures and photoperiod, in either form of sum or threshold. Therefore, we analyzed on a hemispheric scale, how annual tree-ring width relates to thermal thresholds for growth start.

Perhaps you would ask, whether the impact of a small shift in the onset of the vegetative season can actually be detected in annual ring width. Highly resolved local investigations of wood formation have demonstrated a tight link between wood phenology and annual growth. The onset of wood formation was shown to be the main factor that directly or indirectly triggers all subsequent phases of xylem maturation and determines the period of xylem growth. Small advances in the period of cell division can thus lead to substantial increases in xylem cell production, which eventually results in wider growth rings. Therefore, tree growth in cold climates would be enhanced throughout the entire growing season by an earlier onset of cambial activity or by higher growth rates at the peak of the growing season. Evidence for this arise from both field and manipulative experiments. We thus argued that much can be learned still from annually resolved growth information (Figure 2) – especially when it is collected from across a vast array of environmental conditions.

Figure 2 Illustration of how warming impacts on wood phenology affect tree growth.

Although we learned from the wood formation processes that an earlier growth onset would result in wider tree rings, we still needed to verify this expectation at a large spatial scale. This was quite a challenge because annual ring width is usually being correlated with monthly or seasonally climatic data, and we were not sure whether their correlation with a specific date (i.e., the onset of the vegetative season), calculated from daily temperature data would pass rigorous significance tests. After extensive processing of daily climatic data and conducting different correlation analyses, we found distinct spatial patterns of tree growth response to an earlier spring onset. Areas where tree growth benefit from an advanced spring are generally located at the higher latitudes (above 60°N), in central Europe, as well as in eastern and western coastal North America. These cool and humid regions are not strongly limited by water availability during the growing season. The regions with negative effects of advanced springs on growth were mainly located on the Colorado Plateau and the Tibetan Plateau, which correspond to cold and dry conditions, where forests are typically limited by a number of factors including low temperatures, drought events, and poor soil fertility. From these observations, we conclude that trees will grow faster under advanced springs, unless they are thirsty.

After analyzing the spatial patterns in climate correlations, we further asked not only where, but why a changing timing of spring affects tree growth. We tested a series of hypotheses: (1) an advanced spring will extend the vegetative season, so that trees have more time to grow; (2) an advanced spring will result in higher heat accumulation, so that trees can grow faster; (3) an advanced spring will alter the soil moisture conditions and thereby affect tree growth. Based on these hypotheses, we proposed a path model and decomposed the effect of advanced spring on growth. We found distinct latitudinal responses. In boreal forests of northern Asia and Europe, advanced spring enhanced tree growth primarily due to the alleviation of cold stress. In the temperate forests of central Europe and the eastern US coast, as well as in forests of the Mediterranean region and along the western US coast, advanced spring also enhanced growth, but primarily due to the extension of the growing season. In semi-arid forests of the Colorado Plateau and dry subalpine forests of the Tibetan Plateau, advanced spring did not benefit growth, as a longer growing season induces both atmospheric and soil drought there, and will also increase the risk of tree exposure to spring frost.

Overall, we found that the impact of advanced spring could be detected in tree rings at regional to hemispheric scales. Our study will stimulate deeper explorations of the impacts of climatic trends and variability on forest growth worldwide. Such information is essential for integrating information regarding the responses of forests to climate change, and for predicting future vegetation productivity and performance.

The paper in Nature Ecology & Evolution can be found here.