The coconut crab (Birgus latro (L.)) is the world’s largest terrestrial crab. Limited to tropical areas of the Indian and Pacific oceans, the crab can weigh up to 4 kilograms and exceed a one-meter legspan over its 50+ year lifetime. But Birgus populations have faced significant pressures over the last century: many small island communities have replaced large native forest trees (e.g. Pisonia and Pandanus sp.) with coconut palm monocultures, disrupting the coconut crab’s natural habitat. Our study focuses on the protected Birgus latro population on Palmyra Atoll, a United States territory in the Northern Line Islands of the Pacific Ocean. Palmyra’s mix of native forest trees and invasive coconut monoculture allows for detailed study of the habitat associations and movement patterns of the crab. We conducted crab surveys across three vegetation types (native non-coconut trees, coconut palm monoculture, and mixed-tree forests) to determine the habitat preferences of Palmyra’s crab populations. We also collected GPS telemetry data from 44 hand-tagged crabs over two field seasons. To further study Birgus’ interactions with atoll vegetation types, these tracks are overlaid on a new forest classification map of Palmyra Atoll created using multispectral satellite imagery and a machine learning algorithm.
Results/Conclusions
Ground surveys across vegetation types reveal a sharp decrease in crab presence in both mixed-tree (>80%) and coconut monoculture areas (>90%) relative to native forest. These results indicate that even on islands with no hunting pressure, coconut crab populations (despite their name) appear to suffer after native forest conversion to coconut monoculture. Our novel forest classification algorithm demonstrates exceptional accuracy in detecting coconut palms from satellite imagery (up to 98%), allowing us to overlay telemetry data onto precise habitat maps. This crab track analysis found that while Birgus home ranges were preferentially based in non-coconut native forests, this relationship could be attributed to the habitat distributions on Palmyra rather than selection by crabs. However, crabs appear to move through coconut forests less consistently than native forests, perhaps indicating preferential sheltering in native forests but sporadic foraging in coconut forests. In addition, crab home range sizes are determined to be highly variable and up to one kilometer wide, though these ranges appear dependent on island size. These results will help inform the design of protected areas and management plans for the coconut crab, which has seen significant population declines in certain locations but lacks coherent population assessments across its range.