Seasonal and interannual patterns of reproduction in tropical forest in Puerto Rico: hurricanes, drought, and mast reproduction.

Abstract

Climate variation is becoming more evident in tropical regions. Phenological activities in tropical regions are susceptible to interannual variations in climate change and weather. In resource-limited environments in tropical areas, many plants species flower and fruit when the environmental conditions are favorable to allow the successful seedling establishment process. Light-demanding species correlate with a high peak of energy to time their reproductive activities in the forest, while other species have more opportunistic strategies for flowering and fruiting activities. Other species may use day length, high peak of irradiance, rainfall or temperature as environmental cues to reproduce flower and fruit. A diversity of phenological patterns (annual, sub-annual, and supra-annual) has emerged in tropical environments. Tropical phenology is challenging to define because plant species use numerous reproductive strategies. This variation in phenological patterns may have an impact on the food chain in the tropical forests and the consumers.<br /> <br /> Some tropical regions are experiencing extreme natural disturbances such as drought and hurricanes, which may influence plant species&rsquo; reproductive strategies. Species may differ in phenological responses to drought and hurricane events. Studies are limited in tropical forests, which link environmental drivers and disturbances to phenological responses at community and species levels. I also looked for reproductive strategies for lifeforms and plant functional types (PFTs). PFTs tend to share the same ecological strategies that may use similar environmental drivers as cues for flowering and fruiting. There is a scientific interest in conducting studies in different tropical forest types to address the knowledge gap in phenological responses to those factors. There is a trend in conducting synthesis research in reproductive phenology to better understand climate variation on forest reproduction, and to detect potential phenological patterns emerging from those factors. The overall goal of this dissertation was to assess the phenological responses of various types of plant species and lifeforms to environmental drivers, drought, and hurricanes in a subtropical forest in Puerto Rico.<br /> <br /> SECOND CHAPTER: I analyzed 30+ years (1992-2022) phenological records of flowers and fruit or seeds collected in the 16-ha Luquillo Forest Dynamic Plot (LFDP) in a subtropical forest, eastern Puerto Rico. Phenological sampling was conducted biweekly for approximately 370 months using 120 stationary traps placed inside the 16-ha plot. Flowering and fruiting were analyzed using vector algebra for the mean date and concentration for each species. I analyzed the timing of community-wide flowering and fruiting, testing the hypothesis that reproduction would occur around the summer solstice. The community was further grouped into plant functional types (PFTs) based on morphology (palm, liana, and vines and dicotyledonous trees) and demographic trade-offs of dicot trees (Fast, slow, long-lived pioneer [LLP], and Short-lived Breeder [SLB]and intermediate) to assess the effect of PFT on reproductive timing. Community-wide flowering and fruiting occurred around the summer solstice, aligned with the peak of solar irradiance in the forest. Furthermore, plant functional types (PFTs) flowered and fruited at various times away from the winter solstice. Reproduction in short-lived breeders (SLB) coincided within the summer solstice and concentration of flowers and seeds were higher in the intermediate species Reproductive events in the 16-ha were associated with high levels of irradiance, but flowering and fruiting concentration did not differ across plant function types.<br /> <br /> THIRD CHAPTER: In the tropical forest, plant species may use similar or different environmental drivers to trigger flower production. This chapter summarizes flowering timing responses to various climatic variables. I evaluated the influence of solar irradiance, temperature, rainfall, and relative humidity on flower production in the community. I used environmental data from the National Solar Radiation Database and El Verde Field Station data along with the 30+ years of flower records to perform the following analyses and test a few hypotheses. I performed Ordinary Least Square (OLS) regression models, including the environmental variables as predictors to assess flower production response. Furthermore, I conducted phylogenetic analyses to reveal evolutionary history within the community. The findings revealed environmental variables strongly influence community-wide flowering in the forest. Especially, solar irradiance and maximum temperature (TMAX) significantly influenced flower production at both community and species levels (p < 0.01). Flowering timing across lifeforms was positively correlated with the climatic variables. This study suggests environmental variables are the primary drivers of flowers production timing, but phylogenetically structured species-specific responses shape overall community phenology.<br /> <br /> FOURTH CHAPTER: I used the 30+ years of interannual seed production collected from the 16-ha plot to identify species with masting behavior in the forest. Many plant species often produce huge numbers of seeds (masting) in some years following low production in the subsequent year. I summarized seed production annually for each species to detect trends and patterns within the community. I hypothesized that masting is driven by the resource-matching hypothesis, weather, and climate factors. I identified species with masting behavior when the coefficient of variation is greater than one (COV >1). I used Lag autocorrelation method to test the resource-matching hypothesis and the deviation from the long-term mean to the disturbance (hurricanes and droughts) hypothesis. I quantified the effects of hurricanes and droughts on the seed production records using Autoregressive Integrated Moving Average (ARIMA) analysis. Results show seed production varies across species. Findings show roughly 80 % of the species exhibited masting behaviors which indicated by the coefficient of variation (CV>1). I found no species exhibited lag effect, or no evidence of high seed production followed by a decline in the following year (resource-matching hypothesis). The results indicate that disturbances are not major forces driven mast reproduction in the forest. Hurricanes Georges and María had negative effects on seed production. This suggests hurricanes may have a long-term effect on the forest. This study suggests that other environmental factors may play important roles in masting events in plant communities in the forest.