Tropical dry plant formations comprise diverse forests found in areas that experience pronounced seasonal changes in precipitation. Across broad regions, variation in the amount and distribution of precipitation determine the availability of water, and thus differences in species composition and phenology. Within any given region hydrology and topographic position, and geologic substrate may underlie additional differences in forest attributes.  Of these three factors, geologic substrate is the least understood but it has been suggested that its influence on tropical dry forests, and more broadly speaking tropical forests, results from a combination of mechanisms regulating the use of water and nutrients. Phenological studies at population, community, and system levels acknowledge the significance of plan phenology for different ecological processes. Thus, characterizing phenological patterns at multiple levels of biological organization as a function of geologic substrate is important to assess the diversity of behaviors, and ultimately vulnerability of tropical dry forests to climate change. A study conducted in Curaçao between September of 1992 to February of 1995 at three sites underlaing by different geologies provided a unique opportunity to investigate the extent to which plant phenology diverges in regions underlain by different geologies.

The phenological status of leaves, flowers, and fruits within Christoffelpark and neighboring Wacawa (CP) area in northwestern Curaçao was investigated from September of 1992 to February of 1995. The study targeted three geologic formations, namely Curaçao Lava (basalt; CLF; 48% of CP), Knip (chert, silica-rich sediments; KF; 46%), and Limestone (limestone; LF; 6%), and within each formations transects of variable length were visited to identify and tag at least 10 individuals of the most common species and between 1 – 8 individuals of the least common and rare ones. The number of individuals (n_i) and species (n_s) varied among the three formations [CLF (n_i=229 and n_s=28), KF (n_i =369 and n_s=44), and LF (n_i =283 and n_s=35)] and species could occur on more than one formation. Rroughly 881 individuals belonging to 69 species were visited on a monthly basis for a total of 30 months. The species represent a mixture of life forms including trees, shrubs, vines, epiphytes, and succulents. Using the crown density method, individuals were assigned to a linear scale category between 0 - 4 to quantify leaf (expanded, flushing, senescent), flower, and fruit (unripe and ripe) abundance for all species except succulents.  In this approach, a value of 1, 2, and 3 represent one quarter, half, three quarters of the crown density, respectively while a value of 4 indicates complete cover and a value of 0 no cover.  For succulents (cacti and epiphytes sensu lato) only flowers and fruits were counted, and instead of using the abundance scale, the absolute number of these organs was counted. 

The dataset includes missing values. For the analyses reported in the accompanying paper only individuals for which there were ≥27 months (maximum 30 months) of phenological observations were included. For individuals with 27-29 months of data we estimated the missing phenophase values by averaging the previous and following months. Three columns in this dataset (Include_Analysis_System, Include_Analysis_Community, Include_Analysis_Species) specify whether the individual should be included (yes, no) in the three levels of analysis. We use null across all variables to denote cells that do not have values.