What DBH Tells You That Tree Height Doesn't
The DBH_cm field is where this template diverges from a basic presence/absence survey. Trunk diameter at breast height is the proxy measurement for tree biomass, estimated canopy volume, and productive age. Two chinaberry trees of identical height can have radically different DBH values — one a young fast-growth individual pushing skyward at the expense of girth, one an older established tree with a thick bole and dense spreading crown. Their Fruit_Amount categories may overlap. Their productivity dynamics are different.
In an urban foraging study, the relationship between DBH and Fruit_Amount across the dataset tells you whether high-DBH trees consistently support higher fruit loads within a species. If the correlation is strong, DBH becomes a predictive field — trees below a certain diameter threshold do not constitute meaningful foraging resources regardless of their fruit_amount category at the time of survey. That threshold, extracted from the data, is a finding.
The Habitat multichoice — Garden, Park, Urban — in this version uses multichoice rather than single-choice, acknowledging that urban tree placement rarely conforms to clean category boundaries. A row of chinaberries along a park boundary adjacent to a private garden straddles both. The multichoice format records that accurately.
The Ripeness-Depletion Combination
Fruit_Ripeness and Top_Depleted work together. Fruit_Ripeness is a checkbox field capturing Ripe, Unripe, and Overripe simultaneously — a tree may carry all three stages at once as the phenological window opens from the upper canopy downward. Top_Depleted adds the behavioral layer: has the upper canopy been cleared while lower clusters remain intact?
A tree with Fruit_Ripeness = Ripe only, Top_Depleted = Yes tells a clear story. The ripe fruit in the upper canopy was accessed and consumed. A tree with Fruit_Ripeness = Overripe only, Top_Depleted = No suggests the resource peaked and was not exploited — either the species avoids overripe fruit at this site, or the tree was not visited during the ripe window. Two different management implications for corridor planning.
Leftover_Species confirms what was consumed at the ground level. Fresh husks on the ground, recent seeds, still-moist loquat skins — these are signs of consumption within the last 48 hours. Old, desiccated residue from two species simultaneously means the site has been used across multiple sessions. In a plantation setting, where trees are densely packed and ground evidence accumulates quickly, the Fresh/Old distinction in the multichoice options is the temporal anchor that separates yesterday from last week.
The ClusterID and Tree_ID at a Hundred Records
At one hundred records across twenty clusters, the two integer identifiers — ClusterID and Tree_ID — become the backbone of the analysis. Filter by ClusterID to reconstruct any cluster's full resource profile: all trees logged, their DBH, height, habitat, depletion state. Sort by Tree_Type within a cluster to compare perch tree architecture against food tree architecture. The Perch type records in the dataset are the control group — trees selected for non-feeding use within the same cluster as actively depleted food trees.
What separates them? If Perch trees in a given cluster consistently show lower Fruit_Amount categories than Food trees at the same DBH range, the species is selecting for high-yield trees rather than perching opportunistically near feeding sites. If perch trees show equivalent Fruit_Amount but higher Lowest_Branch_m values than food trees, the access threshold is what drives selection.
The Comment field captures what the structured fields cannot. The nest activity two trees over from the target. The group size. The arrival direction. The specific branch that every bird used. That contextual information, tied to a specific Tree_ID and ClusterID, rounds out the record into something a reviewer can actually use.