A phenology model for tropical species that flower multiple times each year

Abstract

AbstractPhenology models developed for temperate and boreal plants predict a single, population‐level first flowering date for each year. These models cannot accommodate species that flower multiple times each year in humid tropical forests nor flowering data with census‐interval rather than daily temporal resolution. Here, we present a new model framework able to predict the timing of multiple annual flowering events from census data. We extend a recent model, which predicted tropical flowering probabilities as discrete events occurring on census dates, by integrating predicted flowering probabilities over all dates between censuses. We evaluate our model against 29 years of daily climate and weekly flowering records for Hybanthus prunifolius (Violaceae) and Handroanthus guayacan (Bignoniaceae) from Barro Colorado Island, Panama. Previous experiments demonstrate that both species flower shortly after a heavy, dry‐season rain interrupts an extended dry period. Our model captures this sequence of rainfall events. Best‐fit model parameters are consistent with previous experimental results. This match suggests the new model framework will provide novel insights for other humid tropical forest species.