Temperature signals in tree-ring oxygen isotope series from the northern slope of the Himalaya

Abstract Oxygen isotope ratios ( δ 18 O) are the most commonly used parameters recorded in paleoclimate archives since they link different natural archives via the water cycle. Tree-ring δ 18 O ( δ 18 O TR ) has been widely used for hydroclimate reconstructions in the Himalaya. However, few of them record temperature signals, which are dominant in Himalaya ice-core δ 18 O. We hypothesize that the “precipitation amount effect” due to the South Asian Summer Monsoon (SASM) may overprint temperature signals in δ 18 O TR series. The purpose of this study is to investigate whether temperature signals could be found in the δ 18 O TR in locations where the influence of SASM is weak. We developed a 105-yr δ 18 O TR chronology from the northern slope of the western Himalaya which greatly blocks the SASM. Our δ 18 O TR clearly shows stronger correlations with temperature (dominant winter and weak summer) than summer precipitation signals. It also agrees well with summer soil moisture δ 18 O simulated by the global isotope model LMDZ4 ( r = 0.72, 1979–2010). In LMDZ4, low winter temperature was found to increase winter snowfall and subsequent snow melt, and thus to increase the contribution of winter snowfall to soil moisture in summer at the expense of summer precipitation. Since winter snowfall is more depleted than summer precipitation, this leads to lower summer soil moisture δ 18 O. The temperature signals found in our δ 18 O TR series are consistent with those found in the Dasuopu ice-core δ 18 O. This implies that δ 18 O TR series from the southwest Tibetan Plateau (TP), with a weak monsoon, hold great potential to capture temperature signals. Climate interpretations of δ 18 O proxies in the Himalaya largely depend on the influence of seasonal water from the dominant atmosphere circulation systems of the westerlies or monsoon. The δ 18 O proxies from the monsoon-affected region have a higher potential for the reconstruction of boreal summer hydroclimate, whereas δ 18 O proxies from westerly-affected sites have a higher potential for temperature reconstructions.