Large-scale spatial variation in palm fruit abundance across a tropical moist forest estimated from high-resolution aerial photographs

Abstract

Fruit abundance is a critical factor in ecological studies of tropical forest animals and plants, but difficult to measure at large spatial scales. We tried to estimate spatial variation in fruit abundance on a relatively large spatial scale using low altitude, high-resolution aerial photography. We measured fruit production for all 555 individuals of the arborescent palm Astrocaryum standleyanum across 25 ha of mapped tropical moist forest on Barro Colorado Island, Panama, by visually counting fruits from the ground. Simultaneously, we used high-resolution aerial photographs to map sun-exposed crowns of the palm across the same area, which were then linked to ground-mapped stems. First, we verified that the fruit crop size of individual trees was positively associated with both crown presence on aerial photos and crown area visible on aerial photos. Then, we determined how well spatial variation in Astrocaryum fruit density across the study area was predicted by spatial densities of photo-detected crowns and crown area compared to spatial densities of ground-mapped stems and stem diameters. We found a positive association of fruit crop size with crown visibility on aerial photographs. Although representing just one third of all individuals in the study area, photo-detected crowns represented 57% of all fruits produced. The spatial pattern of photo-detected crowns was strongly correlated with the spatial pattern of fruit abundance based on direct fruit counts, and correctly showed the areas with the highest and lowest fruit abundances. The spatial density of photo-detected crowns predicted spatial variation in fruit abundance equally well as did the spatial density of ground-mapped stems. Photo-detected crown area did not yield a better prediction. Our study indicates that remote sensing of crowns can be a reliable and cost-effective method for estimating spatial variation in fruit abundance across large areas for highly distinctive canopy species. Our study is also among the few to provide empirical evidence for a positive relationship between crown exposure of forest trees and fruit production.