But now researchers investigating the use of oxygen isotopes in tropical trees as records of past climate have, almost as a sideline, discovered that the toad's demise was instead probably due to a combination of a dry period caused by the 1986–1987 El Niño and the chytrid fungus Batrachochytrium dendrobatidis.

"Our work emphasizes that natural climate variability can contribute to the extinction of species, and that in this specific case it is not necessary to invoke anthropogenic climate change as a cause, nor is their evidence for this from our proxy climate record," Kevin Anchukaitis of Columbia University, US, told environmentalresearchweb. "I want to emphasize, however, that this does not mean that current and future global warming will not be involved in extinction. Rising temperatures and changes in precipitation patterns will without a doubt contribute to stress on ecological communities that could lead to the extinction of species."

Anchukaitis and colleague Michael Evans of the University of Maryland, US, headed to the Monteverde cloud forest because it provided the conditions and data they needed to study the past history of the El Niño Southern Oscillation in Central America. Their technique, dubbed "tropical isotope dendroclimatology", examines the concentration of stable oxygen isotopes in cores taken from tree trunks.

Since tropical conditions can enable plant growth all year round, trees in the tropics often do not exhibit the annual growth rings that have made their more temperate cousins such useful records of past climate. Instead, the team looked at the concentration of δ 18O across the cross section of the trunk; this tends to exhibit a seasonal cycle as the tree switches between obtaining water from rain and, in the dry season, from clouds, which are enriched in δ 18O.

The researchers measured stable oxygen isotope ratios in 200 micron increments along the radial growth axis of two mature canopy Pouteria trees.

"Monteverde had several features I was looking for in a research site – primarily, though, it was an environment wet enough to allow tree growth year round but with relatively shallow-rooted trees that would still be sensitive to changes in moisture availability," explained Anchukaitis. "Monteverde also had a couple decades of weather data that I could use to calibrate my record – more than most tropical mountain locations!"

The fact that the researcher's reconstruction of moisture and hydroclimate at Monteverde turned out to inform them about the climatic context for the extinction of the golden toad was fortuitous, they say, but not the original goal of the research.

The tree data did not reveal any long-term trend in hydroclimate in the cloud forest over the last century, just significant year-to-year and decade-to-decade variability. It did show that the 1986–1987 El Niño caused the driest interval for one hundred years.

"We hypothesize that the dry conditions during 1986 and 1987 caused the majority of the population of the toads to congregate around a few moist microhabitats, where they would have been readily infected," said Anchukaitis. "We're not the first to suggest this mechanism – Alan Pounds and Marty Crump wrote about this in the 1990s, and the phenomenon has been observed elsewhere."

According to the researchers, the lengthy process of extracting chronological and climate information from tropical trees without rings makes their technique more similar to palaeoclimatology using corals or speleothem [cave deposits], than to traditional dendrochronology [tree-ring analysis] where large sample replication and robust crossdating are possible and fundamental. "Improved instrumentation and protocols, however, are already improving our ability to develop longer and better replicated records," said Anchukaitis. "Stay tuned."

Next the team would like to apply the same approach to other cloud forests in Central America and to cloud- and fog-dependent environments throughout the Americas that are sensitive to both natural climate variability and anthropogenic climate change. "We hope that developing a network of sites like this will allow us to better understand variability in tropical climate at timescales from months to centuries," said Anchukaitis.

The researchers reported their work in PNAS.