Get access

Influence of precipitation seasonality on piñon pine cellulose δD values


and present address: Campus Box 450 – INSTAAR, University of Colorado, Boulder, CO 80309-0450, USA, tel. +1/303-492-5792, fax +1/303-492-6388, e-mail


The influence of seasonal to interannual climate variations on cellulose hydrogen isotopic composition (δD) was assessed by analysing tree rings and needles of piñon pine (Pinus edulis and P. monophylla). Sites spanned a gradient of decreasing summer precipitation, from New Mexico to Arizona to Nevada. Tree rings were divided into earlywood, latewood and whole-year increments, and annual cohorts of needles were collected. The study period (1989–96) included two La Niña events (1989, 1996) and a prolonged El Niño event (1991–95). Winter and spring moisture conditions were strongly related to October–March Southern Oscillation Index (SOI) in New Mexico and Arizona, with above-average precipitation occurring in El Niño years. Wood δD values at these sites were correlated with winter and spring moisture conditions. Needle δD values were correlated with summer moisture conditions in New Mexico and with winter moisture and SOI in Arizona. Low cellulose δD values observed from 1991 to 1993 in both wood and needles occurred during wet El Niño years, whereas high δD values in needles were present during the dry, La Niña years of 1989 and 1996. North-eastern Nevada does not receive precipitation anomalies related to ENSO, and thus cellulose δD values did not reflect the ENSO pattern observed at the other sites. Cellulose δD values were strongly, inversely correlated with relative humidity variations at all sites, as predicted by a mechanistic model. Contrary to predictions from the same model and observations from more mesic areas, time series of cellulose δD values were not directly correlated with interannual or seasonal variations in precipitation δD values or temperature at any of the sites. On a regional basis, however, mean δD values in needles and wood were correlated with mean annual temperature and δD values of precipitation. This suggests that temporal averaging may bias relationships between biological systems and climate.