The effects of seasonal and long-term climatic variability on Neotropical flowering phenology: An ecoinformatic analysis of aerial pollen data

- We develop a predictive regression model of pollen productivity for two lowland Panamanian forests.
- We evaluate how pollen production responds to variability in temperature, precipitation, and solar irradiance.
- The model is capable of detecting pattern in noisy ecological data sets.

The phenological behavior of tropical forests changes in response to seasonal, annual, and long-term variation in temperature, precipitation, and solar irradiance. However, detecting the respective influence of these variables is difficult due to the relatively small range of change that is observed in the tropics. Analysis is further constrained by the limited duration of many phenological datasets. To address these limitations, we developed a predictive ecoinformatic model using multivariate linear regression and slope correlation analysis that can uncover statistically significant biological responses within short, noisy ecological time series. Our approach correlates all possible combinations of climatic and taxonomic variables using a series of random determination trials on shuffled environmental data. Seasonal and annual fluctuations in temperature, precipitation, and sunlight were used to predict the reproductive response of each individual taxon. This predictive model was applied to two seasonally sampled aerial pollen records collected between 1996 and 2006 from two Panamanian forests, Barro Colorado Island and Parque Nacional San Lorenzo. Our results highlight the degree to which pollen output responds to fine-scale variability in climate. Our results lend support to the hypothesis that the pollen output of tropical species is diminished with prolonged periods of heavy rainfall and that pollen output is sensitive to small, seasonal increases in temperature. Our ecoinformatic approach can be expanded to other observational phenological datasets to better understand how communities will respond to climate change and our results demonstrate the ability of aerial pollen data to track long-term changes in flowering phenology.