The importance of tropical tree-ring chronologies for global change research

Abstract

Tropical forests and woodlands are key components of the global carbon and water cycles. Yet, how climate change affects these biogeochemical cycles is poorly understood because of scarce long-term observations of tropical tree growth. The recent rise in tropical tree-ring studies may help to fill this gap, but a large-scale quantitative analysis of their potential in global change research is missing.

We compiled a list of all tropical tree species known to form annual tree rings and built a network encompassing 492 tropical ring-width chronologies to evaluate the potential to generate insights on climate sensitivity of woody productivity and to build centuries-long reconstructions of climate variability. We assess chronology quality, length, and climatic representativeness and explore how these change along climatic gradients. Finally, we applied species-distribution modeling to identify regions with potential for tree-ring studies in ecological and climatic studies.

The number of tropical chronologies has rapidly increased, with ∼400 added over the past two decades. Yet, tree-ring studies are biased towards high-elevation locations, with gaps in warmer and wetter climates, on the African continent, and for angiosperm species. The longest chronologies with strongest climate signals (i.e., synchronous growth variations among trees) are from cool regions. In wet regions, climate signals and precipitation sensitivity decrease. Most tropical regions harbor 5–15 (and up to 80) species with proven potential to generate chronologies. The potential for long climate reconstructions is particularly high in drier high elevation sites. Our findings support strategies to effectively expand tree-ring research in the tropics, by targeting specific species and regions. Tropical dendrochronology can importantly contribute to global change research by generating historical context of climate extremes, quantifying climate sensitivity of woody productivity and benchmarking vegetation models.