The nearterm fate of the enormous quantities of carbon stored by tropical rain forests will have substantial impacts on the global carbon cycle and on future climate change. Three drivers of change in carbon stocks of tropical forests are increasing temperature and drought intensity (negative impacts) and increasing atmospheric [CO2] (“CO2 fertilization”, positive impact). Long-term growth records from a Costa Rican old-growth tropical wet forest show that 90% of the variation in annual stand-level wood production (all individuals ³10 cm stem diameter) is accounted for by yearly dry season intensity and nighttime temperature. Here we extend these analyses to the species level to determine the effects of individual species on ecosystem performance, and to compare species’ growth responses to annual dry season intensity, nighttime temperature and [CO2]. The stand data are for all species with an average density of ³2 individuals/ha (N=43 species) from 18 0.5 ha plots censused annually from 1997 – 2007. We also analyzed data for 10 focal species censused annually at the same site for period up to 24 years (1983-2007). In the stand-level analyses we classified species as canopy or subcanopy depending on adult size.
Results/Conclusions For canopy species (N=24), most species’ diameter growth was negatively correlated with dry season intensity and/or nighttime temperature (83% for both variables, 22% [dry season] and 39% [temperature] of correlations significant). In contrast, correlations with annual [CO2] were not different from random (52% positive). Subcanopy species (N=19) differed in that correlations with dry season, nighttime temperature and [CO2] were approximately equally in the predicted direction (ca. 73%, ca. 22% significant). For the focal species (1-4 cm and 4-10 cm diameter), we used partial correlation to separate the effects of nighttime temperature and [CO2]. Negative correlations with nighttime temperature were common (84% of 19 species x size class combinations, 42% significant), while [CO2] correlations were not different from random (53% positive). We discuss how sample size, sampling frequency, and correlations among climatic variables can affect this type of analysis. Our preliminary conclusions are that dry season intensity and nighttime temperature are currently the principal climatic drivers of wood production in this forest, both for canopy species and for individuals <10 cm diameter. There is also some evidence for CO2 fertilization for individuals in the 10-30 cm diameter size range.