Numerous Department of Defense (DoD) installations across the Pacific contain degraded and endangered dry tropical forest ecosystems. Effectively managing and restoring these ecosystems is challenging due to the presence of non-native species, associated wildfires, and potential legacies of past land uses. Unfortunately, little is known about environmental controls on the composition, structure, and function of tropical dry forest vegetation. Gaining an understanding of the range of environmental variability and the impacts on dry forest form and function is critical to managing and restoring these systems effectively. This is particularly important in light of current projections of changes in global and regional climate. An ideal approach to examine how ecosystems develop and respond to changing environmental conditions is to reconstruct the history of vegetation change using paleoecological proxies.

The focus of this study was to examine the viability of developing coupled reconstructions of vegetation change, wildfires, and extreme typhoons over the last few millennia for the island of Guam. The main objective of the reconstructions was to understand the history of ecosystem change on the island of Guam and determine the role of changing typhoon activity in driving changes to ecosystem form and function.  The results from this suite of reconstructions will provide essential data that will inform dry forest restoration and management activities throughout the tropical Pacific. Given changing disturbance, climatic, and land use regimes it is essential to develop this kind of base line data to set appropriate and attainable management and restoration goals.

Technical Approach

The technical approach for this project involved collecting a series of sediment cores from Cocos Lagoon on the south coast of Guam guided by geophysical surveys. This project used portable 10 kHz Stratabox™ and 4 to 24 kHz EdgeTech profilers to map Cocos Lagoon and the target core sites. Nine 3 and 5 meter cores were taken with a portable underwater vibracore system hand driven with extension rods. Cores were transported back to Woods Hole Oceanographic Institution (WHOI) and refrigerated at 4°C.

Pollen and spores were collected from live plants in the Geus River watershed and from existing specimens housed at the University of Guam Herbarium to create taxonomic references specific to Guam. Seventy-seven samples were collected, including 63 species (24% of which are thought to be non-native) representing 33 families of angiosperms and pteridophytes common to the ravine forests and savanna complex in southern Guam. Samples were subjected to 10% potassium hydroxide and acetolysis and then mounted in silicon oil on microscopic slides to create an archival library of pollen reference material. Select types were photographed at 400x magnification.

To test the viability of a wildfire reconstruction, this study quantified macroscopic charcoal particle concentration (pieces/cm3) in several samples from a lagoon core and a delta core. Eleven samples from the lagoon core were analyzed, each of which contained 10 cm3 of sediment sampled over 1 to 4 cm of depth. Nine samples from the delta core were also analyzed, each of which contained 3 to 5 cm3 of sediment sampled over 2 cm of depth. All samples were treated with 10% hydrogen chloride, washed through a 125 μm sieve, and treated with H2O2 to bleach non-carbonized organic material. Each of the 11 samples from the lagoon core also were washed through a 63 μm sieve to capture the 63 to 125 μm size class. All 31 samples were counted on a dissecting microscope at 20x.


More than forty coarse layers were deposited in the deepest portion of Cocos Lagoon over the last 2700 years. Cores from the Geus River delta provide a high-resolution record dating back more than 600 years of both marine and river sourced flooding. Pollen is well preserved in both the lagoon and delta sediments, though more abundant in the later. Fossil charcoal is also well preserved in the delta sediments providing an excellent proxy record of wildfires. The results demonstrate that the sediments deposited in Cocos Lagoon over the past several millennia provide a unique opportunity to reconstruct detailed paleoecological and paleoclimatological data that can help improve our understanding of the complex interplay between climate, typhoon activity, fire, human land use, and dry tropical forest ecosystems. In addition, the combined preliminary records of catastrophic typhoon strikes, terrestrial runoff, vegetation change, and wildfires generated from this proof of concept study clearly demonstrate that valuable detailed paleoclimatological and paleoecological reconstructions are possible from the Cocos Lagoon sediments.


Numerous DoD installations contain dry tropical forest that are susceptible to similar changes in climate, land use practices, and disturbance regimes. As a result, this work will provide valuable insights into how to best manage and restore these ecosystems at many DoD installations.  This research can inform current decision making on a variety of levels. It can provide a detailed understanding of how different vegetation types, brought about through different land use regimes, respond to typhoon events of varying intensity and to fire. It also can classify vegetation response in terms of plant functional types, thus making these data relevant for islands across the Pacific. Perhaps most importantly, this work can answer fundamental questions regarding what is a native dry tropical forest and what is a sustainable dry tropical forest in the future given changing climate and disturbance regimes.