Search

Controversy has surrounded the interpretation of the water isotopic composition (delta D or delta(18)O) in tropical and subtropical ice cores in South America. Although recent modeling studies using AGCM have provided useful constraints at interannual time scales, no direct calibration based on modem observations has been achieved. In the context of the recent ice core drilling at Nevado Illimani (16 degrees 39'S-67 degrees 47'W) in Bolivia, we examine the climatic controls on the modem isotopic composition of precipitation in the Zongo Valley, located on the northeast side of the Cordillera Real, at about 55 km from Nevado Illimani. Monthly precipitation samples were collected from September 1999 to August 2004 at various altitudes along this valley. First we examine the local and regional controls on the common delta D signal measured along this valley. We show that (1) local temperature has definitely no control on delta D variations, and (2) local rainout is a poor factor to explain delta D variations. We thus seek regional controls upstream the Valley potentially affecting air masses distillation. Based on backtrajectory calculations and using satellite data (TRMM precipitation, NOAA OLR) and direct observations of precipitation (IAEA/GNIP), we show that moisture transport history and the degree of rainout upstream are more important factors, explaining seasonal delta D variations. Analysis of a 92-yr simulation from the ECHAM-4 model (T30 version) implemented with water stable isotopes confirms our observations at seasonal time scale and emphasize the role of air masses distillation upstream as a prominent factor controlling interannual delta D variations. Lastly, we focus on the isotopic depletion along the valley when air masses are lifted up. Our results suggest that, if the temperature gradient between the base and the top of the Andes was higher by a few degrees during the Last Glacial Maximum (LGM), less than 10% of the glacial to interglacial isotopic variation recorded in the Illimani ice core could be accounted for by this temperature change. It implies that the rest of the variation would originate from wetter conditions along air masses trajectory during LGM. (C) 2005 Elsevier B.V. All rights reserved.