NOAA/WDS Paleoclimatology - Evans 2007 Model for Isotopic Composition of Tropical Tree Cellulose

A forward model of the oxygen isotopic composition (δ18O) of wood cellulose is parameterized for time series prediction in tropical environments and driven with meteorological data observed at La Selva Biological Research Station, Costa Rica, for 1985–2001. Monthly-resolution model results correlate modestly (r = 0.34, p < 0.05) with observed isotopic data, with higher correlation (r = 0.45, p < 0.01) over the earliest 10 years of comparison and nonsignificant correlation over the most recent 6 years of record. Analysis of model output for La Selva suggests that isotopic variations are strongly controlled by rainfall amount. The model simulates an analogous but stronger than observed negative isotopic anomaly associated with positive July–September rainfall anomalies during El Niño–Southern Oscillation (ENSO) warm phase event years. Simulated tree isotope data for the global tropics suggest that a network of well-replicated data series from selected locations may resolve the large-scale precipitation anomaly pattern associated with ENSO.