A 1000-year record of winter precipitation from northwestern New Mexico, USA: a reconstruction from tree-rings and its relation to El Niño and the Southern Oscillation

A 1000-year reconstruction of winter (cool season) November-May precipitation for northern New Mexico is developed for AD 985-1970 based on six millenium-long tree-ring records from moisture-sensitive coniferous sites in New Mexico and southern Colorado. Two of the most prolonged and severe droughts occurred during the last 100 years: in the AD 1890s-1904 and the AD 1950s-early 1960s. These were exceeded in length and severity only once in the 1000-year reconstructed record, in the late 1500s, and almost equaled in the early 1200s. The five decadal-scale intervals of greatest drought were, in descending order, AD 1577-1598, AD 1955-64, AD 1895-1904, AD 1217-1226, and AD 1778- 1787. The wettest periods were AD 1835-49, AD 1905-28, AD 1429-1440, AD 1609-23, and AD 1487- 98. The most abrupt shift from severe sustained drought to extremely high precipitation occurred in the 1890s to early 1900s. The intensity of this shift is unprecedented over the 1000-year period of the reconstruction. Relatively dry conditions are indicated for part of the 1400-1600s period around the onset of the Little Ice Age, and the reconstruction indicates average conditions during the Medieval Warm Period. The southwestern US is an area of known teleconnection with the El Nine-Southern Oscillation (ENSO), coinciding with a tendency for wetter conditions to occur during ENSO years (warmer events) and drier conditions during colder events. The chronologies and reconstruction show correspondence with indices of ENSO, including Wright's Southern Oscillation Index (SOI). The years AD 1720, 1484, 1816 and 1941 (interpreted as ENSO-related events in instrumental or historical data) are the four wettest individual years in the reconstruction. Variance spectral analysis of the chronologies and reconstruction reveals significant peaks at approximately 2-7 year, 22 year and 80-100 year. There is potential for modelling the amplitude and frequency of the ENSO signature in the southwestern US over the past 1000 years or more, particularly by integrating tree-ring data from this region with other sources of high-resolution palaeoclimatic information.